Re: [Emc-users] For Peter C. Wallace

[Gene Heskett]

On Wednesday 15 November 2017 10:02:23 Peter C. Wallace wrote:

> On Wed, 15 Nov 2017, Gene Heskett wrote:
> > Date: Wed, 15 Nov 2017 09:14:49 -0500
> > From: Gene Heskett <ghesk...@shentel.net>
> > Reply-To: "Enhanced Machine Controller (EMC)"
> > <emc-users@lists.sourceforge.net>
> > To: emc-users@lists.sourceforge.net
> > Subject: Re: [Emc-users] For Peter C. Wallace
> >
> > On Sunday 12 November 2017 05:12:54 andy pugh wrote:
> >> On 12 November 2017 at 01:49, Gene Heskett <ghesk...@shentel.net>
> >
> > wrote:
> >>> If I write a hal thing that changes the value pushed into the
> >>> encoder's scale input, how much of a time lag might there be to
> >>> make it effective?
> >>
> >> The scaling does not happen in the Mesa layer, it happens in the
> >> HAL driver.
> >>
> >> But, the encoder scale is a parameter, not a pin, so you can't
> >> change it in HAL.
> >>
> >> What I would do would be to pass the hm2_..._position and
> >> hm2..._...velocity through a pair of "scale" HAL functions with the
> >> scaling factor for both chosen by a "mux2" controlled by the gear
> >> selection before connecting to the motion... pins.
> >>
> >> You might be able to use the "gearchange" component for this (or,
> >> specifically, two more instances of the component) as this is
> >> basically what it does, ie switch between two scales.
> >
> > That I can do, but it does bring up a related question. Actually 2
> > or 3.
> >
> > 1. How long, probably in cycles, does it take motion to recover when
> > it finds a huge change in either its position, or velocity inputs
> > which the scale switching causes?
> >
> > 2. at what point does the accumulated position overflow? or is
> > position reset to zero on the index pulse? Seems to me like it ought
> > to be, but I don't recall anyone saying.
> >
> > I intend to temporarily set up a counter to be frozen and reset in
> > one pulse, useing probably an edge module to tap the index pulse to
> > transfer the counter to a sample-hold, then reset the counter when
> > the edge recovers from a very short output pulse, with the
> > sample-hold becoming the value for the 2 scale modules for that
> > gear. By running the motor slow, I ought to be able to define the
> > unknown, in this case the actual gear ratio when in each of its 2
> > gears. Or maybe even use it as the scale.n.gain input since its a
> > pin, however I'd expect a count or more to be skipped at the higher
> > motor speeds and that would confuse the scale values derived. Not a
> > Good Thing.
> >
> > Best to carve it into the ini file once determined.
> >
> > Cheers, Gene Heskett
> > --
> > "There are four boxes to be used in defense of liberty:
> > soap, ballot, jury, and ammo. Please use in that order."
> > -Ed Howdershelt (Author)
> > Genes Web page <http://geneslinuxbox.net:6309/gene>
>
> I dont think any of that is needed, as long as you do not change the
> gear ratio in the middle of a spindle synchronized move.

Which would wreck the part so thats why I alway cut air with the first 
pass. Cautious George.

But having that ratio right for each gear is important, something I can't 
know w/o taking it apart and counting teeth, a major teardown that I 
will precede by ordering the best quality bearings to put in as 
I reassemble it, in place of the roller skate crap in there now. Unless I 
can count from index to index to get that count. Setting that up in hal 
code is 30x easier than the mechanical teardown.

> Spindle synchronized moves always start at index = spindle position 0,
> and dealing with the spindle position "steps" is all handled in motion

True, but the size of those steps determines the working tpi. Must be 
correct else motion may put a step function into Z at the index when it 
comes around again.  That could be interesting in a bad way if it 
requires a couple backlash takeups in the middle of a 30 ipm 
synchronized move.

But you knew all that. I'm just thinking out loud. :)

Thank you, Peter.

> Peter Wallace
> Mesa Electronics

Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page <http://geneslinuxbox.net:6309/gene>

--
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] For Peter C. Wallace

[Gene Heskett]

On Wednesday 15 November 2017 09:59:35 andy pugh wrote:

> On 15 November 2017 at 14:14, Gene Heskett  
wrote:
> > 1. How long, probably in cycles, does it take motion to recover when
> > it finds a huge change in either its position, or velocity inputs
> > which the scale switching causes?
>
> Unless it is actually using them at the time (being in the middle of a
> feed-per-rev or spindle-synched move) it won't even be looking, so no
> time at all.
>
Thats great!

> > 2. at what point does the accumulated position overflow?
>
> Never. (for practical purposes).
> A 1024 count encoder at 10,000 rpm will overflow in 3 million years.

At which time none of the 3 of us (me, motor or encoder) will still exist 
to give a s :) So this should all work ok if I don't screw up the 
hal file too badly. Its currently loaded with comments explaining what 
it does, but even with those excised its still 600 lines.

While I am waiting on the encoder, perhaps I can get the A axis working 
since I've moved the stuff that was on those pins to the 5i25 p2 
connector.

Probably by building yet another box for a psu and a 2m542 driver.  But 
I'll jury rig the 2m542 off the existing supply for x first to see if 
it goes into foldback with a third motor on it. No room left in the big 
box for a third 2m542 though without seriously impinging on my ability 
to get to the bob. I should have made it 4" bigger in all 3 dimensions 
but its final size was determined by the 1/8" sheet stock I had on hand 
at the time, most of which was the surviving, nearly flat stock out of 
the sides of a big wrecked ups step-van. I've since acquired some of the 
paneling from an old Townsend tv transmitter and have another piece or 2 
wedged into a stashing place beside the garage door.

But first, feed Dee again and go get the papers.

Thanks Andy.

Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page 

--
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] For Peter C. Wallace

[Peter C. Wallace]

On Wed, 15 Nov 2017, Gene Heskett wrote:


Date: Wed, 15 Nov 2017 09:14:49 -0500
From: Gene Heskett <ghesk...@shentel.net>
Reply-To: "Enhanced Machine Controller (EMC)"
<emc-users@lists.sourceforge.net>
To: emc-users@lists.sourceforge.net
Subject: Re: [Emc-users] For Peter C. Wallace

On Sunday 12 November 2017 05:12:54 andy pugh wrote:


On 12 November 2017 at 01:49, Gene Heskett <ghesk...@shentel.net>

wrote:

If I write a hal thing that changes the value pushed into the
encoder's scale input, how much of a time lag might there be to make
it effective?


The scaling does not happen in the Mesa layer, it happens in the HAL
driver.

But, the encoder scale is a parameter, not a pin, so you can't change
it in HAL.

What I would do would be to pass the hm2_..._position and
hm2..._...velocity through a pair of "scale" HAL functions with the
scaling factor for both chosen by a "mux2" controlled by the gear
selection before connecting to the motion... pins.

You might be able to use the "gearchange" component for this (or,
specifically, two more instances of the component) as this is
basically what it does, ie switch between two scales.


That I can do, but it does bring up a related question. Actually 2 or 3.

1. How long, probably in cycles, does it take motion to recover when it
finds a huge change in either its position, or velocity inputs which the
scale switching causes?

2. at what point does the accumulated position overflow? or is position
reset to zero on the index pulse? Seems to me like it ought to be, but I
don't recall anyone saying.

I intend to temporarily set up a counter to be frozen and reset in one
pulse, useing probably an edge module to tap the index pulse to transfer
the counter to a sample-hold, then reset the counter when the edge
recovers from a very short output pulse, with the sample-hold becoming
the value for the 2 scale modules for that gear. By running the motor
slow, I ought to be able to define the unknown, in this case the actual
gear ratio when in each of its 2 gears. Or maybe even use it as the
scale.n.gain input since its a pin, however I'd expect a count or more
to be skipped at the higher motor speeds and that would confuse the
scale values derived. Not a Good Thing.

Best to carve it into the ini file once determined.

Cheers, Gene Heskett
--
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page <http://geneslinuxbox.net:6309/gene>



I dont think any of that is needed, as long as you do not change the gear 
ratio in the middle of a spindle synchronized move.


Spindle synchronized moves always start at index = spindle position 0,
and dealing with the spindle position "steps" is all handled in motion



--
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users



Peter Wallace
Mesa Electronics

(\__/)
(='.'=) This is Bunny. Copy and paste bunny into your
(")_(") signature to help him gain world domination.


--
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] For Peter C. Wallace

[andy pugh]

On 15 November 2017 at 14:14, Gene Heskett  wrote:

> 1. How long, probably in cycles, does it take motion to recover when it
> finds a huge change in either its position, or velocity inputs which the
> scale switching causes?

Unless it is actually using them at the time (being in the middle of a
feed-per-rev or spindle-synched move) it won't even be looking, so no
time at all.

> 2. at what point does the accumulated position overflow?

Never. (for practical purposes).
A 1024 count encoder at 10,000 rpm will overflow in 3 million years.

-- 
atp
"A motorcycle is a bicycle with a pandemonium attachment and is
designed for the especial use of mechanical geniuses, daredevils and
lunatics."
— George Fitch, Atlanta Constitution Newspaper, 1916

--
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] For Peter C. Wallace

[Gene Heskett]

On Sunday 12 November 2017 05:12:54 andy pugh wrote:

> On 12 November 2017 at 01:49, Gene Heskett  
wrote:
> > If I write a hal thing that changes the value pushed into the
> > encoder's scale input, how much of a time lag might there be to make
> > it effective?
>
> The scaling does not happen in the Mesa layer, it happens in the HAL
> driver.
>
> But, the encoder scale is a parameter, not a pin, so you can't change
> it in HAL.
>
> What I would do would be to pass the hm2_..._position and
> hm2..._...velocity through a pair of "scale" HAL functions with the
> scaling factor for both chosen by a "mux2" controlled by the gear
> selection before connecting to the motion... pins.
>
> You might be able to use the "gearchange" component for this (or,
> specifically, two more instances of the component) as this is
> basically what it does, ie switch between two scales.

That I can do, but it does bring up a related question. Actually 2 or 3.

1. How long, probably in cycles, does it take motion to recover when it 
finds a huge change in either its position, or velocity inputs which the 
scale switching causes?

2. at what point does the accumulated position overflow? or is position 
reset to zero on the index pulse? Seems to me like it ought to be, but I 
don't recall anyone saying.

I intend to temporarily set up a counter to be frozen and reset in one 
pulse, useing probably an edge module to tap the index pulse to transfer 
the counter to a sample-hold, then reset the counter when the edge 
recovers from a very short output pulse, with the sample-hold becoming 
the value for the 2 scale modules for that gear. By running the motor 
slow, I ought to be able to define the unknown, in this case the actual 
gear ratio when in each of its 2 gears. Or maybe even use it as the 
scale.n.gain input since its a pin, however I'd expect a count or more 
to be skipped at the higher motor speeds and that would confuse the 
scale values derived. Not a Good Thing.

Best to carve it into the ini file once determined.

Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page 

--
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] For Peter C. Wallace

[Gene Heskett]

On Sunday 12 November 2017 05:12:54 andy pugh wrote:

> On 12 November 2017 at 01:49, Gene Heskett  
wrote:
> > If I write a hal thing that changes the value pushed into the
> > encoder's scale input, how much of a time lag might there be to make
> > it effective?
>
> The scaling does not happen in the Mesa layer, it happens in the HAL
> driver.
>
I had assumed it was in the fpga. My bad.

> But, the encoder scale is a parameter, not a pin, so you can't change
> it in HAL.

Ooops, best laid plans, etc.

> What I would do would be to pass the hm2_..._position and
> hm2..._...velocity through a pair of "scale" HAL functions with the
> scaling factor for both chosen by a "mux2" controlled by the gear
> selection before connecting to the motion... pins.

If the parameter is changed for a 1.0 scale at boot time, then the other 
scale is the gear ratio's, right? But I'd better use the mux4's so the 
illegal combo's are skipped as they are now, in fact I've put an offset 
on inputs 0 and 3 of the selector mux4 that actually runs the motor at 
about 40 rpm as it makes the gear shifting instant and effortless, no 
grabbing the stopped spindle to get the gears to mesh. And I've disabled 
the pid that is unselected to get rid of the windup. Works very well 
indeed even if the spindle is running at high speed, Jons driver pulls 
it down to creep in a few milliseconds, the gear change is effortless 
and the spindle winds back up to full speed as the other switch finally 
closes long after the new gear is solidly engaged. 

> You might be able to use the "gearchange" component for this (or,
> specifically, two more instances of the component) as this is
> basically what it does, ie switch between two scales.

Or between straight thru and the gearchange(or scaled) version.  addf 
order sensitive I assume?

Thanks for the heads up Andy. But I'll not do anything until the encoder 
arrives, could be into December according to the shipping notice. I'll 
have to make an extension shaft to drive the encoder, to be fitted to 
the back of the motor, whose shaft ends at the top of its rear ball 
bearing. I'll run the motor backwards to drill the hole, first with a 
center drill to get a bigger socket, then tap it for a 3mm bolt that 
runs down the center bore of the extension shaft. I want the encoder in 
hand before I make the shaft extension. And a flimsy mounting bracket so 
the encoders bearings aren't taking a beating from gross mis-alignment.

Cheers Andy, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page 

--
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] For Peter C. Wallace

[andy pugh]

On 12 November 2017 at 01:49, Gene Heskett  wrote:

> If I write a hal thing that changes the value pushed into the encoder's
> scale input, how much of a time lag might there be to make it effective?

The scaling does not happen in the Mesa layer, it happens in the HAL driver.

But, the encoder scale is a parameter, not a pin, so you can't change
it in HAL.

What I would do would be to pass the hm2_..._position and
hm2..._...velocity through a pair of "scale" HAL functions with the
scaling factor for both chosen by a "mux2" controlled by the gear
selection before connecting to the motion... pins.

You might be able to use the "gearchange" component for this (or,
specifically, two more instances of the component) as this is
basically what it does, ie switch between two scales.

-- 
atp
"A motorcycle is a bicycle with a pandemonium attachment and is
designed for the especial use of mechanical geniuses, daredevils and
lunatics."
— George Fitch, Atlanta Constitution Newspaper, 1916

--
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

[Emc-users] For Peter C. Wallace

[Gene Heskett]

Hi Peter;

If I write a hal thing that changes the value pushed into the encoder's 
scale input, how much of a time lag might there be to make it effective?

The scenario is this:

I have ordered a 1000 line Omron encoder, which I intend to couple to the 
rear of the spindle motor shaft, feeding its a/b into the encoder in a 
5i25. But the index signal from the new encoder will not be used, 
remaining with the index in my shop made encoder, and therefore at a 
known location of the spindles rotation, with the desired end result 
being a different scale factor according to the gear setting in effect 
at the instant.

So given a couple turns of the spindle to allow the encoder to get 
re-synced with the new scale setting I ought to be able to do rigid 
tapping in either gear. Like 4-40's in high gear, and as much as 8mm in 
low gear.

Obviously I don't envision changing this in mid-tap.

Other than that potential screw-up on the part of the operator, me, do 
you see any problems with such a setup?

Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page 

--
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] For Peter C. Wallace

[Gene Heskett]

On Saturday 14 October 2017 00:52:06 Chris Albertson wrote:

> On Fri, Oct 13, 2017 at 4:56 PM, Gene Heskett  
wrote:
> > Always a consideration, Chris. But putting "Dell Dimension 745" into
> > their search box only came up with psu's & that sort of stuff.  No
> > mobo's. I suspect its too old.
>
> I meant that was a place holder
>
> Better way to search.  Go to link below and use search filters on the
> left then sort by price.  pcpartspicker is a kind of specialized
> search engine that will pointy to many different vendors.
>
> https://pcpartpicker.com/products/motherboard/
>
> All that said, the last PC I just bought came from a wholesaler who
> was selling HP computers that just came off lease. I needed something
> with alot of compute power
> so I bought an HP workstation with 16-core xeon and 64GB ECC RAM
> for $500.  I added an nVidia GPU and SSD and then Ubuntu Linux.
> It runs headless in a closet.  The GPU is used as a compute engine.
>
>
> If you really like the "745" look here.   It is one of the liquidators
> I was talking about.  $70 with 90 day warranty.
> pcliquidations.com/p19260-dell-optiplex-745-tower...
> rce=google_medium=cse_term=19260=CjwKCAj
>wyIHPBRAIEiwAHPS-GMcG42Q_lT9bPrMcv7bsH9R5XWWen0MB6M307ClsplsJ8s9DAW8_5R
>oCskkQAvD_BwE>

I might call and check the shipping tomorrow, if they are answering the 
phones on Saturdays.

Thanks Chris.

Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page 

--
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] For Peter C. Wallace

[Gene Heskett]

On Saturday 14 October 2017 00:41:04 Jon Elson wrote:

> > On Fri, Oct 13, 2017 at 7:43 AM, Gene Heskett  
wrote:
> >> Hi Peter;
> >> Or alternatively, can this Dell's supposedly blown mobo be repaired
> >> by replacing a known easily damaged part (that I've not found yet)?
>
> If the PS/2 ports stop working, it is very likely the 5V
> fuse has been blown.  Many mobos have thermal self-resetting
> fuses there, you might try that first, just let the machine
> cool for a few minutes and try with a different mouse/keyboard.
> If not, inspect around the ports for a fuse component.  Some
> mobos use picofuses, look like 1/4 W resistors, but all green.
>
> Jon
>
I got that fixed Jon, I could put the mouses rx button in any machine and 
kill it. Mouse, button both went in the round file. And 2 brand new, 
psycho painted ones are now playing in the garage.

But now I can't make a g33.1 work in a peck routine. See my earlier post 
with the code that won't run. But I was running out of play time & had 
to go round up some dinner, so I turned stuff off including 95% of the 
lights & came in to take care of the missus.  Tomorrow I'll try a 
different version of the code as I've made up a 1 pager for most of the 
common 'merican and metric sizes & tpi's. I just used a month ago, what 
s/b the same code set to do an 8mmx1.0, and it worked as expected around 
24 times, in both steel and alu when I made the cat head and spiders for 
the Sheldon.

Thanks and take care, Jon

Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page 

--
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] For Peter C. Wallace

[Chris Albertson]

On Fri, Oct 13, 2017 at 4:56 PM, Gene Heskett  wrote:

>
> Always a consideration, Chris. But putting "Dell Dimension 745" into
> their search box only came up with psu's & that sort of stuff.  No
> mobo's. I suspect its too old.


I meant that was a place holder

Better way to search.  Go to link below and use search filters on the left
then sort by price.  pcpartspicker is a kind of specialized search engine
that will pointy to many different vendors.

https://pcpartpicker.com/products/motherboard/

All that said, the last PC I just bought came from a wholesaler who was
selling HP computers that just came off lease. I needed something with
alot of compute power
so I bought an HP workstation with 16-core xeon and 64GB ECC RAM
for $500.  I added an nVidia GPU and SSD and then Ubuntu Linux.
It runs headless in a closet.  The GPU is used as a compute engine.


If you really like the "745" look here.   It is one of the liquidators
I was talking about.  $70 with 90 day warranty.
pcliquidations.com/p19260-dell-optiplex-745-tower...





-- 

Chris Albertson
Redondo Beach, California
--
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] For Peter C. Wallace

[Jon Elson]

On Fri, Oct 13, 2017 at 7:43 AM, Gene Heskett  wrote:

Hi Peter;
Or alternatively, can this Dell's supposedly blown mobo be repaired by
replacing a known easily damaged part (that I've not found yet)?


If the PS/2 ports stop working, it is very likely the 5V 
fuse has been blown.  Many mobos have thermal self-resetting 
fuses there, you might try that first, just let the machine 
cool for a few minutes and try with a different mouse/keyboard.
If not, inspect around the ports for a fuse component.  Some 
mobos use picofuses, look like 1/4 W resistors, but all green.


Jon

--
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] For Peter C. Wallace

[Gene Heskett]

On Friday 13 October 2017 13:47:26 Chris Albertson wrote:

> On Fri, Oct 13, 2017 at 7:43 AM, Gene Heskett  
wrote:
> > Hi Peter;
> >
> >
> > Or alternatively, can this Dell's supposedly blown mobo be repaired
> > by replacing a known easily damaged part (that I've not found yet)?
>
> Yes,  There is an easy to replace part.  Not knowing which Dell you
> have I can only show a generic example.  The big thing is to make sure
> the CPU socket and memory type is the same.   Cost less than $20.
>
> https://www.newegg.com/Product/Product.aspx?Item=9SIACTN65J1897
>
> This are cheaper to replace than prepare.  Repairing a 4 or 6 layer
> PCB is impossible as most of the traces are buried under ground and
> power planes and even if you could fix it the fix might change the
> length or impedance of a trace.

Always a consideration, Chris. But putting "Dell Dimension 745" into 
their search box only came up with psu's & that sort of stuff.  No 
mobo's. I suspect its too old.

And, while I thought a mouse was all I needed, but somethings still 
aglay, it will not run a g33.1 in a peck cycle. It starts down on the 
first stroke, cutting air, and spits out a following error. Slightly 
wierd, the first step down is about 16 thou. It starts down, and 
according to the dro, goes past the first move by about 2 thou and 
errors out, shutting the machine off.  And nothing I can do will make it 
go past the first step in the while loop. The short routine: 
tap3-8thsx16tpi-hole.ngc:

%
( tap3-8x16tpi-hole.ngc )
( measure in inch's )
G20
#<_z_depth> = -1.000
#<_tp_inch> = 16
#<_inc_per_peck>= [1.0 / #<_tp_inch>]
#<_z_dec>   =   [#<_inc_per_peck> / 4.]
S500 M3 ( start motor in low gear )
G1 f10 x0 y0 Z0 ( put it at touched off to 0. position )
#<_z_tmp>   =   0. ( start z_tmp at that Z location)
G1 F10 Z#<_z_tmp> ( and put it there to start the loop )
(debug,z-tmp=#<_z_tmp>)
(debug,inc-per-peck=#<_inc_per_peck>)
(debug,z_dec=#<_z_dec>)

o100 WHILE [#<_z_tmp> ge #<_z_depth>]

G33.1 Z#<_z_tmp> k#<_inc_per_peck>

#<_z_tmp> = [#<_z_tmp> - #<_z_dec>]

(debug,22 z-tmp=#<_z_tmp>)
G4 P4 (4 second pause to blow chips & buttercut the tap )

o100 ENDWHILE

M5

M2
%

And single stepping it, it never shows the line of code its executing in 
the code window, so I've no clue what line is the problem.

And at some point, before it actually starts the spindle, the little 
window that shows what modes are in effect, goes totally berserkers for 
about 5 seconds, then stabilizes back where it started.

By then it was time I rounded up some din-din for us, and shut things 
down for the night before my headache returned.

Anybody that see's what wrong with that code, speak up! More than likely 
a forehead slapper...

Thanks Chris & everybody.

Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page 

--
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] For Peter C. Wallace

[Chris Albertson]

On Fri, Oct 13, 2017 at 7:43 AM, Gene Heskett  wrote:
> Hi Peter;

>
> Or alternatively, can this Dell's supposedly blown mobo be repaired by
> replacing a known easily damaged part (that I've not found yet)?

Yes,  There is an easy to replace part.  Not knowing which Dell you have
I can only show a generic example.  The big thing is to make sure the
CPU socket and memory type is the same.   Cost less than $20.

https://www.newegg.com/Product/Product.aspx?Item=9SIACTN65J1897

This are cheaper to replace than prepare.  Repairing a 4 or 6 layer PCB is
impossible as most of the traces are buried under ground and power planes
 and even if you could fix it the fix might change the length or impedance
 of a trace.


-- 

Chris Albertson
Redondo Beach, California
--
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] For Peter C. Wallace

[Gene Heskett]

On Friday 13 October 2017 12:56:31 Gene Heskett wrote:
[...]
> But where the heck is my local keyboard and mouse?, neither works
> after a login and only the keyboard works on the login greeter.  And I
> have not played with what was a working configuration in re that in
> months.
>
> I'll go give that connector that may have shorted to that relay box, a
> coat of Go-2, aka goop, over the end face of the ribbon cable to help
> prevent that potential short again. I'll take the mouse and its button
> to the lathe and verify it works there. Or across to the other dell I
> use for mesaflashing to see if it works there. Back later with any
> more news about this headache.
>
> Thanks Peter, it helps if you name the tree I'm looking for. :)
>
> >Peter Wallace
>
> Cheers, Gene Heskett

This is crazy! The mouse had expired, and that killed it, keyboard and 
all.  With a known good tail-less rodent, its back to normal except for 
the gladeVCP error. So I need to go get me 2 fresh mice. One to take 
apart and heat warp the spring back into the button covers like I had to 
last week to this brand new one before it was usable, and one for a 
spare.

Does Excedrin have a number for this headache? :(

If the St Bernards are still within range of being whistled in, do it, 
and drink the whisky. I'm going after some mice.


Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page 

--
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] For Peter C. Wallace

[Gene Heskett]

On Friday 13 October 2017 11:09:33 Peter C. Wallace wrote:

> On Fri, 13 Oct 2017, andy pugh wrote:
> > Date: Fri, 13 Oct 2017 15:53:52 +0100
> > From: andy pugh <bodge...@gmail.com>
> > Reply-To: "Enhanced Machine Controller (EMC)"
> > <emc-users@lists.sourceforge.net>
> > To: "Enhanced Machine Controller (EMC)"
> > <emc-users@lists.sourceforge.net> Subject: Re: [Emc-users] For Peter
> > C. Wallace
> >
> > On 13 October 2017 at 15:43, Gene Heskett <ghesk...@shentel.net>
> > wrote:
> >
> > Plugging a new(er, I don't recall if it was programmed) 5i25 card
> > in, dmesg says it found a :5125 card.  Sounds like a blown pci
> > bus line. OTOH, TLMonster has a 5i25 in it, and it says its a 5125
> > too.  And its working ok. False Flag?
>
> 5125 is just a numerical version of 5i25 in a field which is numeric.
>
> Doesn't look like it can be a single PCI line.
>
> 1 = 0011001
> i = 01101001
> I = 01001001

And I can't reply to Peters addition because its below Andy's dash dash 
space \n sig separator.  So I'll copy/paste

>The VID:DID for a 5I25 is

>2718:5125
IIRC thats the correct of what I saw on screen, but I'll have to go plug 
a 5i25 back into it. Did, yes, correct.

>If you can read this its likely the PCI bus is OK

>Another clue to PCI bus issues is if LinuxCNC doesnt find the correct 
>HM2 cookie =
>0x55AACAFE
Can't test for this locally, no keyboard or mouse, but I'll go put a card 
in and try it from here with an ssh login. brb.

verified card had prob_rfx2.pin in it, put it in machine, powered up, 
collected coffee cup, and filled it on way back to this chair.

Linuxcnc -l runs, only fuss is a gladeXML incomplete message. so let me 
go power up the main machines psu & see if it will home. Yes, and I can 
jog it from here.

These are all the fusses I got in the login screen from here:

gladevcp: auxiliary dir: /usr/share/linuxcnc/aux_gladevcp/NativeCAM
gladevcp: 
importing: /usr/share/linuxcnc/aux_gladevcp/NativeCAM/pref_edit.py
gladevcp: importing: /usr/share/linuxcnc/aux_gladevcp/NativeCAM/ncam.py

(gladevcp:3704): libglade-WARNING **: Expected .  Got 
.

(gladevcp:3704): libglade-WARNING **: did not finish in PARSER_FINISH 
state
 GLADE VCP ERROR:With xml file: camview.glade : could not create 
GladeXML object

And I don't recall that I am using camview.glade!  However I was last 
working on making the camera work, as an edge detector yadda yadda. So 
its possibly gotten into the mix.

But where the heck is my local keyboard and mouse?, neither works after a 
login and only the keyboard works on the login greeter.  And I have not 
played with what was a working configuration in re that in months.

I'll go give that connector that may have shorted to that relay box, a 
coat of Go-2, aka goop, over the end face of the ribbon cable to help 
prevent that potential short again. I'll take the mouse and its button 
to the lathe and verify it works there. Or across to the other dell I 
use for mesaflashing to see if it works there. Back later with any more 
news about this headache.

Thanks Peter, it helps if you name the tree I'm looking for. :)
>Peter Wallace


Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page <http://geneslinuxbox.net:6309/gene>

--
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] For Peter C. Wallace

[Gene Heskett]

On Friday 13 October 2017 10:53:52 andy pugh wrote:

> On 13 October 2017 at 15:43, Gene Heskett  wrote:
> > Plugging a new(er, I don't recall if it was programmed) 5i25 card
> > in, dmesg says it found a :5125 card.  Sounds like a blown pci
> > bus line. OTOH, TLMonster has a 5i25 in it, and it says its a 5125
> > too.  And its working ok. False Flag?
>
> 5125 is just a numerical version of 5i25 in a field which is numeric.
>
> Doesn't look like it can be a single PCI line.
>
> 1 = 0011001
shouldn't this be a
  1 = 00110001
> i = 01101001
> I = 01001001

I'll go see if that computer shop over in Buckhannon has another of these 
Dells. I have another, but with a much older P4 in it.  Noticeably 
slower, I keep it around to use mesaflash to program this stuff.

Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page 

--
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] For Peter C. Wallace

[Peter C. Wallace]

On Fri, 13 Oct 2017, andy pugh wrote:


Date: Fri, 13 Oct 2017 15:53:52 +0100
From: andy pugh <bodge...@gmail.com>
Reply-To: "Enhanced Machine Controller (EMC)"
<emc-users@lists.sourceforge.net>
To: "Enhanced Machine Controller (EMC)" <emc-users@lists.sourceforge.net>
Subject: Re: [Emc-users] For Peter C. Wallace

On 13 October 2017 at 15:43, Gene Heskett <ghesk...@shentel.net> wrote:



Plugging a new(er, I don't recall if it was programmed) 5i25 card in,
dmesg says it found a :5125 card.  Sounds like a blown pci bus line.
OTOH, TLMonster has a 5i25 in it, and it says its a 5125 too.  And its
working ok. False Flag?


5125 is just a numerical version of 5i25 in a field which is numeric.

Doesn't look like it can be a single PCI line.

1 = 0011001
i = 01101001
I = 01001001

--
atp
"A motorcycle is a bicycle with a pandemonium attachment and is
designed for the especial use of mechanical geniuses, daredevils and
lunatics."
?? George Fitch, Atlanta Constitution Newspaper, 1916

--
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users





The VID:DID for a 5I25 is

2718:5125

If you can read this its likely the PCI bus is OK

Another clue to PCI bus issues is if LinuxCNC doesnt find the correct 
HM2 cookie =

0x55AACAFE

Peter Wallace
Mesa Electronics

(\__/)
(='.'=) This is Bunny. Copy and paste bunny into your
(")_(") signature to help him gain world domination.
--
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] For Peter C. Wallace

[andy pugh]

On 13 October 2017 at 15:43, Gene Heskett  wrote:

> Plugging a new(er, I don't recall if it was programmed) 5i25 card in,
> dmesg says it found a :5125 card.  Sounds like a blown pci bus line.
> OTOH, TLMonster has a 5i25 in it, and it says its a 5125 too.  And its
> working ok. False Flag?

5125 is just a numerical version of 5i25 in a field which is numeric.

Doesn't look like it can be a single PCI line.

1 = 0011001
i = 01101001
I = 01001001

-- 
atp
"A motorcycle is a bicycle with a pandemonium attachment and is
designed for the especial use of mechanical geniuses, daredevils and
lunatics."
— George Fitch, Atlanta Constitution Newspaper, 1916

--
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

[Emc-users] For Peter C. Wallace

[Gene Heskett]

Hi Peter;

My G0704's computer has upchucked. Boots slow, & last time the keyboard 
works is entering the passwd.  Then it takes at least an extra 30 
seconds to show the x screens. But no mouse on the login screen, and no 
keyboard.

Coming back in here and logging in via ssh, dmesg looks fairly normal 
with about 4 instances of it finding the logitech unifying receivers and 
registering them as mouse of keyboard. Xorg.0.log is silent on mouse and 
keyboard.

I rebooted several times and finally pulled the cover, finding a splash 
of smoke on an alu bar I'd added to the cover to hold the rear end of 
the 5i25 into the pci socket, but it has no evidence of an electrical 
contact on it, just the smoke from a totally destroyed ptc that would 
have been about 1/16" away. The alu was a short piece of 3/8" angle, 
glued to the top of a wooden bar, so it wasn't grounded, glued to the 
lid. All positioned to make sure the rear of the card couldn't work its 
way out of the pci socket. That seemed to have worked as it was solidly 
seated in the mobo socket. Removed the card, still no mouse or keyboard. 
I may have found the short, a box made of dbl-sided pcb was laying 
against the side of the cable from the db25 on the card, to the bob's 
db25 connector on top of the main motor controller/psu box, possibly 
shorting the end of the cable where it was clipped off nearly flush with 
side of the IDC connector. Most likely culprit at least.

Plugging a new(er, I don't recall if it was programmed) 5i25 card in, 
dmesg says it found a :5125 card.  Sounds like a blown pci bus line.
OTOH, TLMonster has a 5i25 in it, and it says its a 5125 too.  And its 
working ok. False Flag?

I've a 2nd rock64, still in the box in front of me, and I found a module 
while making a fully preemptable kernel for it, that makes a parport out 
of some gpio pins. Do you think it could be fast enough to talk to a 
7i90?  And how long do you think the cables maximum usable length might 
be? I believe the current cables are about 40" and theres very little 
slack.

Replacing the decade old Dell with the rock64 would reduce the 24/7 power 
draw by about 200 watts, a most desirable thing.

Or alternatively, can this Dell's supposedly blown mobo be repaired by 
replacing a known easily damaged part (that I've not found yet)?

Thanks Peter.

Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page 

--
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

[Emc-users] To Peter C. Wallace

[Gene Heskett]

Thank you for useing a common bolt hole pattern in those cards. I haven't 
laid it out for drilling yet, but I now have all 4 stacked in a 
stairstep pattern using nylon standoffs, that will clear or just touch 
the lid of my box, and the 50 pin cables needed will run from 1.25" to 
perhaps 3.75" long.  Not working on it tonight, a bit under the weather, 
overdid some hedge cutting and burning (about 5 or 6 years overdue) so 
the old fart is all stove up.

Probably tomorrow.

Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page 

--
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] To Peter C Wallace

[Gene Heskett]

On Friday 01 January 2016 11:23:35 Peter C. Wallace wrote:

> On Thu, 31 Dec 2015, Gene Heskett wrote:
> > Date: Thu, 31 Dec 2015 20:51:08 -0500
> > From: Gene Heskett <ghesk...@wdtv.com>
> > Reply-To: "Enhanced Machine Controller (EMC)"
> > <emc-users@lists.sourceforge.net>
> > To: "emc-users@lists.sourceforge.net"
> > <emc-users@lists.sourceforge.net> Subject: [Emc-users] To Peter C
> > Wallace
> >
> > I tried to hook up my soft start circuit tonight, but when I had it
> > wired in, the spindle didn't run. Checking dmesg, the card did not
> > sign in like it usually does.  Backed out of the mainboard socket at
> > least 1/8", again. By bumping the parport cables molded connector
> > against the wall.
> >
> > So I have an excedrin headache, number unk.  Everytime I turn that
> > computer around, to plug in a usb cable to charge my DS201, or for
> > any reason, the 5i25 comes far enough out of the pci socket to
> > disable something.  Or blow the card.  At 90 bucks a pop, that isn't
> > fun.
> >
> > Do you have a bracket that can somehow hold the card into the
> > socket?
> >
> > Its an LP bracket in an LP box, but it needs something to nail the
> > card down into the slot a lot firmer than the LP bracket is holding
> > it.
> >
> > Ideas, anybody?  Or do I have to make a x-crossed-brace to go across
> > the top of the pci sockets?
> >
> > Thanks all. :)
> >
> > Cheers, Gene Heskett
> > --
> > "There are four boxes to be used in defense of liberty:
> > soap, ballot, jury, and ammo. Please use in that order."
> > -Ed Howdershelt (Author)
> > Genes Web page <http://geneslinuxbox.net:6309/gene>
> >
> > 
> >-- ___
> > Emc-users mailing list
> > Emc-users@lists.sourceforge.net
> > https://lists.sourceforge.net/lists/listinfo/emc-users
>
> Whats is actually bending when the card works out of the socket?
> We have some standard height brackets with a rib on the edge
> to limit flexing, unfortunately we cannot get stock
> LP brackets with the rib
>
> If the case/bracket are not rigid enough to prevent card motion,
> one possibility is pad on the case top that presses on the connector
> finger area of the top card edge
>
I had previously bent the top of the bracket up a wee bit so the little 
plastic latch Dell uses for a holddown would  put more pressure on it, 
enough that the latch no longer blew around in a mild breeze. The card 
was not in fact out of the socket this time, but I did superglue a 
wooden strip to the lid, and added another 1/16" of some alu angle on 
top of that which should adequately restrict any jacking.

But its temporarily fixed, see the msg I just posted after battling with 
it quite a bit of the afternoon.  Its obvious I have a flaky BoB.

Thanks Peter.
>
> Peter Wallace
> Mesa Electronics
>
>
> --
> ___
> Emc-users mailing list
> Emc-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/emc-users


Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page <http://geneslinuxbox.net:6309/gene>

--
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] To Peter C Wallace

[Gene Heskett]

On Friday 01 January 2016 10:58:37 linden wrote:

> I just put a gob as far aft as i can get (calling the end where your
> cables come out forward) It is just like sticking things together with
> a wad of gum if you are using an elcheepo glue gut it dose not get hot
> enough to wick in between any thing. Its non conductive the clear
> sticks seem to work better as they do not bond quites as well and you
> can peal it off relatively easily. the yellowy cream colored sticks
> are a more permanent glue.
>
> On 15-12-31 06:53 PM, Gene Heskett wrote:
> > On Thursday 31 December 2015 21:27:52 linden wrote:
> >> Hi Gene,
> >>   I have used hot glue from a hot glue gun in the past (the
> >> Kind for arts and crafts) to hold cards in. It works well for high
> >> vibration and you can peal the glue off to get the card out if you
> >> need too. Not as good as a mechanical hold down but certainly
> >> better than nothing and not as permanent as potting compound. May
> >> be enough added stay puttedness for your application or untill you
> >> can come up with a better solution.
> >
> > Never gave that a thought, but it might be just the ticket. I don't
> > think its electrically conductive. I'm assuming you put a line of it
> > against both the socket and board?
> >
> > I can only access one side, without pulling the mainboard, which
> > might be less than optimum.
> >
> > Did you put anything against the board to keep it from wicking into
> > the socket itself?  Or does it set fast enough thats not a problem?
> >
> > Thanks for the idea Linden, I appreciate a fresh idea.
> >
> >> On 15-12-31 05:51 PM, Gene Heskett wrote:
> >>> I tried to hook up my soft start circuit tonight, but when I had
> >>> it wired in, the spindle didn't run. Checking dmesg, the card did
> >>> not sign in like it usually does.  Backed out of the mainboard
> >>> socket at least 1/8", again. By bumping the parport cables molded
> >>> connector against the wall.
> >>>
> >>> So I have an excedrin headache, number unk.  Everytime I turn that
> >>> computer around, to plug in a usb cable to charge my DS201, or for
> >>> any reason, the 5i25 comes far enough out of the pci socket to
> >>> disable something.  Or blow the card.  At 90 bucks a pop, that
> >>> isn't fun.
> >>>
> >>> Do you have a bracket that can somehow hold the card into the
> >>> socket?
> >>>
> >>> Its an LP bracket in an LP box, but it needs something to nail the
> >>> card down into the slot a lot firmer than the LP bracket is
> >>> holding it.
> >>>
> >>> Ideas, anybody?  Or do I have to make a x-crossed-brace to go
> >>> across the top of the pci sockets?
> >>>
> >>> Thanks all. :)
> >
> > Cheers, Gene Heskett

As it turned out when I had pulled it down to where I could look at it, 
the card was fully seated, my last bending of the bracket seems to have 
nailed it into the socket firmly enough.  I did add a wood and alu pad 
to the lid so that it can't come up more than 20 thou anyway.

That leaves the hal file I worked on to convert what was the axis a 
stepgen into usable gpio on the BoB's pin 8-9 and "wired" it up to the 
charge pumps outputs.  That part, FWIW, is working well since the logic 
highs coming out of the BoB are at 5.08 volts.  However, sitting off 
overnight emphasized that I need a bleeder on that supply, very badly.  
Before powering things up today, I measured the voltage on that bank of 
caps. Live, its 127, off overnight it was still showing 113 volts.  Very 
low leakage caps I think. :)

So I've printed a fresh copy of the hal file to see if I can find where I 
screwed up the pwmgen inputs, the enable is coming out, and reverse is 
coming out, both on the proper queue & pins on the BoB.  But theres no 
output on the BoB's pin 14, which is the pwmgen output, its sitting 
high, which is driving the - input on Jon's pwm-servo amp.  No leds on 
it, and the reset switch has no effect.  Humm, thinking out loud, I 
haven't checked the volts on the pwm inputs + pin of Jon's driver, brb.  
Yup, +5.08 as expected. And no pulses on the BoB's pin 14 with that wire 
disconnected.

On a hal-meter, pid-s looks normal and is outputing about 5x the input 
because there is no feedback from a stationary spindle. 
hm2_5i25.0.pwmgen.00.value matches.

That leaves my hal file, the BoB, or the 5i25. Or a broken conductor in 
the interconnecting db-25 cable.  It is screwed to the panel connectors 
at both ends.  But from the diameter of that cable, even 25 pieces of 30 
gauge would be bigger. That one is all of 42" long, and I don't have 
another more than 30".

Now its fixed, but I've NDI why/how.

I powered everything but the computer down, stopping linuxcnc.
Pulled the db25 off the breakout boards input, found pin 14 & hooked the 
ds201 to it.

Started linuxcnc up again, enabled the machine, hit the CW button and + 
several times.  On the ds201 it looks like it might be 10 Hz! But very 
slow attack and decay times, like it was pushing the ds201's bandwidth.

Scratching 

Re: [Emc-users] To Peter C Wallace

[Gregg Eshelman]

On 1/1/2016 4:29 PM, Jack Coats wrote:
> Back when I received computers via fedex with multiple 'boards' in
> them, the ones that did the best were ones where after each board was
> inserted, a bead of pure silicon caulk (the clear kind) was put along
> the top of the connector so it would contact both the connector and
> the board.  It stayed soft over time and was easy to pull off when the
> board is pulled for any reason.  Typically it was put on one side of
> the connector/daughter board (often the back) to ease installation.
>
> The idea of using hot-glue is similar, but hot-glue can set up
> 'harder' than the caulk over time. --- Still, it is the same thought.

I assume/hope aquarium silicone was used. The common stuff with acetic 
acid in it can be disastrous for electronics. If it smells like vinegar, 
you want it nowhere near anything electronic.


---
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus


--
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] To Peter C Wallace

[Jack Coats]

Back when I received computers via fedex with multiple 'boards' in
them, the ones that did the best were ones where after each board was
inserted, a bead of pure silicon caulk (the clear kind) was put along
the top of the connector so it would contact both the connector and
the board.  It stayed soft over time and was easy to pull off when the
board is pulled for any reason.  Typically it was put on one side of
the connector/daughter board (often the back) to ease installation.

The idea of using hot-glue is similar, but hot-glue can set up
'harder' than the caulk over time. --- Still, it is the same thought.

--
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] To Peter C Wallace

[Peter Blodow]

Andy, this idea is good and saves space, but has limitations as to the 
bus speed. Siemens built their desk top PCs (in fact, there were no 
others) like this in the 1990ies and dropped it as processing speed rose 
rapidly in these days. The L-shaped connector board had too much 
impedance for the megacycles.
Peter


Am 01.01.2016 12:08, schrieb andy pugh:
> My latest build has a different approach: 
> https://picasaweb.google.com/108164504656404380542/Holbrook#6234774182242152002
>  



---
Diese E-Mail wurde von Avast Antivirus-Software auf Viren geprüft.
https://www.avast.com/antivirus


--
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] To Peter C Wallace

[andy pugh]

On 1 January 2016 at 01:51, Gene Heskett  wrote:
> Ideas, anybody?  Or do I have to make a x-crossed-brace to go across the
> top of the pci sockets?

Are there any standoffs / tapped holes on the motherboard that you can
get a bracket to?

It seems to me to be a problem with the PCI port design, and has
caught me a few times.
My latest build has a different approach:
https://picasaweb.google.com/108164504656404380542/Holbrook#6234774182242152002

-- 
atp
If you can't fix it, you don't own it.
http://www.ifixit.com/Manifesto

--
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] To Peter C Wallace

[andy pugh]

On 1 January 2016 at 11:29, Peter Blodow  wrote:
> Andy, this idea is good and saves space, but has limitations as to the
> bus speed. Siemens built their desk top PCs (in fact, there were no
> others) like this in the 1990ies and dropped it as processing speed rose
> rapidly in these days. The L-shaped connector board had too much
> impedance for the megacycles.

This one is actually an "official" accessory for the motherboard.
It came with a suitable backplate, which I didn't use.

-- 
atp
If you can't fix it, you don't own it.
http://www.ifixit.com/Manifesto

--
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] To Peter C Wallace

[Gene Heskett]

On Friday 01 January 2016 06:08:31 andy pugh wrote:

> On 1 January 2016 at 01:51, Gene Heskett  wrote:
> > Ideas, anybody?  Or do I have to make a x-crossed-brace to go across
> > the top of the pci sockets?
>
> Are there any standoffs / tapped holes on the motherboard that you can
> get a bracket to?

TBD when I take it down here in a couple hours.  Twould need to be near 
the rear of the card to make it a simple one piece.

> It seems to me to be a problem with the PCI port design, and has
> caught me a few times.

Yes, and apparently caught the video card folks too so the AGP cards all 
have a latch at the rear of the socket.  Exactly what the PCI should 
have had from the gitgo IMNSHO.

> My latest build has a different approach:
> https://picasaweb.google.com/108164504656404380542/Holbrook#6234774182
>242152002

Looks good, but would need a full length pci riser I haven't seen at TD 
or NewEgg, and in an LP case, might not clear the connection tower 
sticking up off the mobo.  That and this Dell 745 doesn't have any pci-e 
slots that I've noticed enough to recall.

Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page 

--
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] To Peter C Wallace

[John Thornton]

I'll get a photo of what I did on my CHNC which has the motherboard just 
mounted to a panel...

JT

On 12/31/2015 7:51 PM, Gene Heskett wrote:
> I tried to hook up my soft start circuit tonight, but when I had it wired
> in, the spindle didn't run. Checking dmesg, the card did not sign in
> like it usually does.  Backed out of the mainboard socket at least 1/8",
> again. By bumping the parport cables molded connector against the wall.
>
> So I have an excedrin headache, number unk.  Everytime I turn that
> computer around, to plug in a usb cable to charge my DS201, or for any
> reason, the 5i25 comes far enough out of the pci socket to disable
> something.  Or blow the card.  At 90 bucks a pop, that isn't fun.
>
> Do you have a bracket that can somehow hold the card into the socket?
>
> Its an LP bracket in an LP box, but it needs something to nail the card
> down into the slot a lot firmer than the LP bracket is holding it.
>
> Ideas, anybody?  Or do I have to make a x-crossed-brace to go across the
> top of the pci sockets?
>
> Thanks all. :)
>
> Cheers, Gene Heskett


--
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] To Peter C Wallace

[John Thornton]

the photo

http://gnipsel.com/images/chnc/CHCN-5i20.jpg

On 1/1/2016 6:05 AM, John Thornton wrote:
> I'll get a photo of what I did on my CHNC which has the motherboard just
> mounted to a panel...
>
> JT
>
> On 12/31/2015 7:51 PM, Gene Heskett wrote:
>> I tried to hook up my soft start circuit tonight, but when I had it wired
>> in, the spindle didn't run. Checking dmesg, the card did not sign in
>> like it usually does.  Backed out of the mainboard socket at least 1/8",
>> again. By bumping the parport cables molded connector against the wall.
>>
>> So I have an excedrin headache, number unk.  Everytime I turn that
>> computer around, to plug in a usb cable to charge my DS201, or for any
>> reason, the 5i25 comes far enough out of the pci socket to disable
>> something.  Or blow the card.  At 90 bucks a pop, that isn't fun.
>>
>> Do you have a bracket that can somehow hold the card into the socket?
>>
>> Its an LP bracket in an LP box, but it needs something to nail the card
>> down into the slot a lot firmer than the LP bracket is holding it.
>>
>> Ideas, anybody?  Or do I have to make a x-crossed-brace to go across the
>> top of the pci sockets?
>>
>> Thanks all. :)
>>
>> Cheers, Gene Heskett
>
> --
> ___
> Emc-users mailing list
> Emc-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/emc-users


--
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] To Peter C Wallace

[linden]

I just put a gob as far aft as i can get (calling the end where your 
cables come out forward) It is just like sticking things together with a 
wad of gum if you are using an elcheepo glue gut it dose not get hot 
enough to wick in between any thing. Its non conductive the clear sticks 
seem to work better as they do not bond quites as well and you can peal 
it off relatively easily. the yellowy cream colored sticks are a more 
permanent glue.

On 15-12-31 06:53 PM, Gene Heskett wrote:
> On Thursday 31 December 2015 21:27:52 linden wrote:
>
>> Hi Gene,
>>   I have used hot glue from a hot glue gun in the past (the Kind
>> for arts and crafts) to hold cards in. It works well for high
>> vibration and you can peal the glue off to get the card out if you
>> need too. Not as good as a mechanical hold down but certainly better
>> than nothing and not as permanent as potting compound. May be enough
>> added stay puttedness for your application or untill you can come up
>> with a better solution.
>>
> Never gave that a thought, but it might be just the ticket. I don't think
> its electrically conductive. I'm assuming you put a line of it against
> both the socket and board?
>
> I can only access one side, without pulling the mainboard, which might be
> less than optimum.
>
> Did you put anything against the board to keep it from wicking into the
> socket itself?  Or does it set fast enough thats not a problem?
>
> Thanks for the idea Linden, I appreciate a fresh idea.
>
>> On 15-12-31 05:51 PM, Gene Heskett wrote:
>>> I tried to hook up my soft start circuit tonight, but when I had it
>>> wired in, the spindle didn't run. Checking dmesg, the card did not
>>> sign in like it usually does.  Backed out of the mainboard socket at
>>> least 1/8", again. By bumping the parport cables molded connector
>>> against the wall.
>>>
>>> So I have an excedrin headache, number unk.  Everytime I turn that
>>> computer around, to plug in a usb cable to charge my DS201, or for
>>> any reason, the 5i25 comes far enough out of the pci socket to
>>> disable something.  Or blow the card.  At 90 bucks a pop, that isn't
>>> fun.
>>>
>>> Do you have a bracket that can somehow hold the card into the
>>> socket?
>>>
>>> Its an LP bracket in an LP box, but it needs something to nail the
>>> card down into the slot a lot firmer than the LP bracket is holding
>>> it.
>>>
>>> Ideas, anybody?  Or do I have to make a x-crossed-brace to go across
>>> the top of the pci sockets?
>>>
>>> Thanks all. :)
> Cheers, Gene Heskett


--
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] To Peter C Wallace

[andy pugh]

On 1 January 2016 at 12:05, John Thornton  wrote:
> I'll get a photo of what I did on my CHNC which has the motherboard just
> mounted to a panel...

That's what I have. The question is how to stop the PCI card from
falling out of the motherboard.


-- 
atp
If you can't fix it, you don't own it.
http://www.ifixit.com/Manifesto

--
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] To Peter C Wallace

[Peter C. Wallace]

On Thu, 31 Dec 2015, Gene Heskett wrote:

> Date: Thu, 31 Dec 2015 20:51:08 -0500
> From: Gene Heskett <ghesk...@wdtv.com>
> Reply-To: "Enhanced Machine Controller (EMC)"
> <emc-users@lists.sourceforge.net>
> To: "emc-users@lists.sourceforge.net" <emc-users@lists.sourceforge.net>
> Subject: [Emc-users] To Peter C Wallace
> 
> I tried to hook up my soft start circuit tonight, but when I had it wired
> in, the spindle didn't run. Checking dmesg, the card did not sign in
> like it usually does.  Backed out of the mainboard socket at least 1/8",
> again. By bumping the parport cables molded connector against the wall.
>
> So I have an excedrin headache, number unk.  Everytime I turn that
> computer around, to plug in a usb cable to charge my DS201, or for any
> reason, the 5i25 comes far enough out of the pci socket to disable
> something.  Or blow the card.  At 90 bucks a pop, that isn't fun.
>
> Do you have a bracket that can somehow hold the card into the socket?
>
> Its an LP bracket in an LP box, but it needs something to nail the card
> down into the slot a lot firmer than the LP bracket is holding it.
>
> Ideas, anybody?  Or do I have to make a x-crossed-brace to go across the
> top of the pci sockets?
>
> Thanks all. :)
>
> Cheers, Gene Heskett
> -- 
> "There are four boxes to be used in defense of liberty:
> soap, ballot, jury, and ammo. Please use in that order."
> -Ed Howdershelt (Author)
> Genes Web page <http://geneslinuxbox.net:6309/gene>
>
> --
> ___
> Emc-users mailing list
> Emc-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/emc-users
>


Whats is actually bending when the card works out of the socket?
We have some standard height brackets with a rib on the edge
to limit flexing, unfortunately we cannot get stock
LP brackets with the rib

If the case/bracket are not rigid enough to prevent card motion,
one possibility is pad on the case top that presses on the connector
finger area of the top card edge


Peter Wallace
Mesa Electronics


--
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

[Emc-users] To Peter C Wallace

[Gene Heskett]

I tried to hook up my soft start circuit tonight, but when I had it wired 
in, the spindle didn't run. Checking dmesg, the card did not sign in 
like it usually does.  Backed out of the mainboard socket at least 1/8", 
again. By bumping the parport cables molded connector against the wall.

So I have an excedrin headache, number unk.  Everytime I turn that 
computer around, to plug in a usb cable to charge my DS201, or for any 
reason, the 5i25 comes far enough out of the pci socket to disable 
something.  Or blow the card.  At 90 bucks a pop, that isn't fun.

Do you have a bracket that can somehow hold the card into the socket?

Its an LP bracket in an LP box, but it needs something to nail the card 
down into the slot a lot firmer than the LP bracket is holding it.

Ideas, anybody?  Or do I have to make a x-crossed-brace to go across the 
top of the pci sockets?

Thanks all. :)

Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page 

--
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] To Peter C Wallace

[linden]

Hi Gene,
 I have used hot glue from a hot glue gun in the past (the Kind for 
arts and crafts) to hold cards in. It works well for high vibration and 
you can peal the glue off to get the card out if you need too. Not as 
good as a mechanical hold down but certainly better than nothing and not 
as permanent as potting compound. May be enough added stay puttedness 
for your application or untill you can come up with a better solution.



On 15-12-31 05:51 PM, Gene Heskett wrote:
> I tried to hook up my soft start circuit tonight, but when I had it wired
> in, the spindle didn't run. Checking dmesg, the card did not sign in
> like it usually does.  Backed out of the mainboard socket at least 1/8",
> again. By bumping the parport cables molded connector against the wall.
>
> So I have an excedrin headache, number unk.  Everytime I turn that
> computer around, to plug in a usb cable to charge my DS201, or for any
> reason, the 5i25 comes far enough out of the pci socket to disable
> something.  Or blow the card.  At 90 bucks a pop, that isn't fun.
>
> Do you have a bracket that can somehow hold the card into the socket?
>
> Its an LP bracket in an LP box, but it needs something to nail the card
> down into the slot a lot firmer than the LP bracket is holding it.
>
> Ideas, anybody?  Or do I have to make a x-crossed-brace to go across the
> top of the pci sockets?
>
> Thanks all. :)
>
> Cheers, Gene Heskett


--
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] To Peter C Wallace

[Gene Heskett]

On Thursday 31 December 2015 21:27:52 linden wrote:

> Hi Gene,
>  I have used hot glue from a hot glue gun in the past (the Kind
> for arts and crafts) to hold cards in. It works well for high
> vibration and you can peal the glue off to get the card out if you
> need too. Not as good as a mechanical hold down but certainly better
> than nothing and not as permanent as potting compound. May be enough
> added stay puttedness for your application or untill you can come up
> with a better solution.
>
Never gave that a thought, but it might be just the ticket. I don't think 
its electrically conductive. I'm assuming you put a line of it against 
both the socket and board? 

I can only access one side, without pulling the mainboard, which might be 
less than optimum.

Did you put anything against the board to keep it from wicking into the 
socket itself?  Or does it set fast enough thats not a problem?

Thanks for the idea Linden, I appreciate a fresh idea.

> On 15-12-31 05:51 PM, Gene Heskett wrote:
> > I tried to hook up my soft start circuit tonight, but when I had it
> > wired in, the spindle didn't run. Checking dmesg, the card did not
> > sign in like it usually does.  Backed out of the mainboard socket at
> > least 1/8", again. By bumping the parport cables molded connector
> > against the wall.
> >
> > So I have an excedrin headache, number unk.  Everytime I turn that
> > computer around, to plug in a usb cable to charge my DS201, or for
> > any reason, the 5i25 comes far enough out of the pci socket to
> > disable something.  Or blow the card.  At 90 bucks a pop, that isn't
> > fun.
> >
> > Do you have a bracket that can somehow hold the card into the
> > socket?
> >
> > Its an LP bracket in an LP box, but it needs something to nail the
> > card down into the slot a lot firmer than the LP bracket is holding
> > it.
> >
> > Ideas, anybody?  Or do I have to make a x-crossed-brace to go across
> > the top of the pci sockets?
> >
> > Thanks all. :)

Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page 

--
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] To Peter C Wallace

[Gregg Eshelman]

On 12/31/2015 7:27 PM, linden wrote:
> Hi Gene,
>   I have used hot glue from a hot glue gun in the past (the Kind for
> arts and crafts) to hold cards in. It works well for high vibration and
> you can peal the glue off to get the card out if you need too. Not as
> good as a mechanical hold down but certainly better than nothing and not
> as permanent as potting compound. May be enough added stay puttedness
> for your application or untill you can come up with a better solution.

Hot glue is very easy to remove. Simply pour a little rubbing alcohol on 
it. Preferably 91%. It pretty much instantly breaks its bond and will 
even get it out of fabric.


---
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus


--
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] To Peter C Wallace

[rayj]

On an aviation list I get many people use the glue E-6000 to pot 
electronics in.  Not sure if wicking would be an issue, but no 
conduction or corrosion problems.  It might bond a little better than 
hot glue, for better or worse.  It will sag if put on thick and takes 24 
hr to cure for most applications.

I have used hot glue for many semi-permanent bonding jobs and It doesn't 
seem to wick much but it will sag and leaves little threads when you 
pull the nozzle away.

Raymond Julian
Kettle River, MN

The things we admire in men, kindness and generosity, openness, honesty, 
understanding and feeling are the concomitants of failure in our system. 
And those traits we detest, sharpness, greed, acquisitiveness, meanness, 
egotism and self-interest are the traits of success. And while men 
admire the quality of the first they love the produce of the second. 
-John Steinbeck, novelist, Nobel laureate (1902-1968)

On 01/01/2016 02:53 AM, Gene Heskett wrote:
> On Thursday 31 December 2015 21:27:52 linden wrote:
>
>> Hi Gene,
>>   I have used hot glue from a hot glue gun in the past (the Kind
>> for arts and crafts) to hold cards in. It works well for high
>> vibration and you can peal the glue off to get the card out if you
>> need too. Not as good as a mechanical hold down but certainly better
>> than nothing and not as permanent as potting compound. May be enough
>> added stay puttedness for your application or untill you can come up
>> with a better solution.
>>
> Never gave that a thought, but it might be just the ticket. I don't think
> its electrically conductive. I'm assuming you put a line of it against
> both the socket and board?
>
> I can only access one side, without pulling the mainboard, which might be
> less than optimum.
>
> Did you put anything against the board to keep it from wicking into the
> socket itself?  Or does it set fast enough thats not a problem?
>
> Thanks for the idea Linden, I appreciate a fresh idea.
>
>> On 15-12-31 05:51 PM, Gene Heskett wrote:
>>> I tried to hook up my soft start circuit tonight, but when I had it
>>> wired in, the spindle didn't run. Checking dmesg, the card did not
>>> sign in like it usually does.  Backed out of the mainboard socket at
>>> least 1/8", again. By bumping the parport cables molded connector
>>> against the wall.
>>>
>>> So I have an excedrin headache, number unk.  Everytime I turn that
>>> computer around, to plug in a usb cable to charge my DS201, or for
>>> any reason, the 5i25 comes far enough out of the pci socket to
>>> disable something.  Or blow the card.  At 90 bucks a pop, that isn't
>>> fun.
>>>
>>> Do you have a bracket that can somehow hold the card into the
>>> socket?
>>>
>>> Its an LP bracket in an LP box, but it needs something to nail the
>>> card down into the slot a lot firmer than the LP bracket is holding
>>> it.
>>>
>>> Ideas, anybody?  Or do I have to make a x-crossed-brace to go across
>>> the top of the pci sockets?
>>>
>>> Thanks all. :)
>
> Cheers, Gene Heskett
>

--
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users