RE: DECstation 220 - Out of Ideas?

[Rob Jarratt via cctalk]

> -Original Message-
> From: cctalk [mailto:cctalk-boun...@classiccmp.org] On Behalf Of Rob Jarratt
> via cctalk
> Sent: 28 June 2017 19:12
> To: 'shad' <shado...@gmail.com>; 'General Discussion: On-Topic and Off-
> Topic Posts' <cctalk@classiccmp.org>
> Subject: RE: DECstation 220 - Out of Ideas?
> 
> I was thinking of trying with the chip removed, so I think I will do that. I 
> have a
> multimeter of course, which is how I have been tracing already. The scanner
> idea is a good one too, and if nothing else at least it gives me an idea of 
> where
> to probe without too much flipping of the board.
> 

With the 74LS125 removed I found the ROM output enable was still not going low 
enough to enable the ROM output. Doing a bit more tracing I reached a 74LS243 
quad transceiver.

Now here I think there *is* something odd going on. I noticed that the two 
inputs GAB and GBA are both at about 1.95V. I am not sure, from my reading of 
the datasheet whether that counts as low or high, it seems to be right on the 
borderline if I am reading the datasheet correctly. The ROM OE signal is going 
to the A4 pin (pin 6) of the transceiver, so I put the scope on the B4 pin and 
it is a steady 1.4V. That behaviour does not seem to correspond to GAB and GBA 
both being high or both being low, unless what is happening is that it is 
considering GAB and GBA to be both high and it is trying to send the B signal 
to the A side, in which case two signals are sourcing to the same line and 
possibly not letting the ROM OE signal drop. Perhaps the next step would be to 
remove this transceiver (or lift A4 or B4) and see if the signal does start to 
drop correctly. At which point it is either the transceiver that is bad, or 
more likely something is telling it to do the wrong thing.

Does that seem a reasonable analysis?

Regards

Rob

RE: DECstation 220 - Out of Ideas?

[shadoooo via cctalk]

Hello,
now you could try to check the voltages again, just to be sure the fault is
not on the 245.

Following the traces is of course obviously difficult, but if the board is
really only two-sided (no internal signal layers), components are not too
high, and you have a flatbed scanner, you could scan both sides, then
colorize top in green and bottom in red, flip bottom and do a properly
aligned overlay with a graphic editor like GIMP.
It's a matter of few minutes, but could improve the work.
When you are unsure, try with the multimeter.

Andrea

RE: DECstation 220 - Out of Ideas?

[Rob Jarratt via cctalk]

> -Original Message-
> From: cctech [mailto:cctech-boun...@classiccmp.org] On Behalf Of shad
> via cctech
> Sent: 26 June 2017 21:57
> To: cct...@classiccmp.org
> Subject: RE: DECstation 220 - Out of Ideas?
> 
> Hello Rob,
> if you are uncomfortable with soldering, I can replace the IC for you.
> Just ship me the board, I will desolder it.
> 


Thanks for the offer. Actually I am generally OK with desoldering these days 
(still struggle with through hole DIPs a bit at times though). I have not done 
any surface mount before except for some capacitors, but I have ordered some 
Chip Quik now and I am anxious to try it out.

Regards

Rob

RE: DECstation 220 - Out of Ideas?

[shadoooo via cctalk]

Hello Rob,
if you are uncomfortable with soldering, I can replace the IC for you.
Just ship me the board, I will desolder it.

bye
Andrea

Re: DECstation 220 - Out of Ideas?

[Jon Elson via cctalk]

On 25/06/2017 11:46, Rob Jarratt via cctalk wrote:


I have done a bit more tracing of components and have come across a
possibly anomaly. The signal that won't go low enough goes to the
input (pin 2) of a 74LS125 buffer.


OK, for SOIC parts, here's my technique.  (I've had to 
replace quite a few of them.)  First, use solder wick to 
remove solder from one side of the chip.  Then, use an 
X-acto knife under a pin, and while heating the solder pad, 
lever the knife to lift each pin outward and up, free from 
the solder.  Work your way down one side of the chip until 
all leads are free of solder.  Now, take the X-acto blade 
and score all the other leads right near the plastic body.  
Just a couple swipes of the blade is enough.  Now, tilt the 
body up, press down, and repeat a couple times and the body 
will break off of the leads.
Now, remove remaining leads by just sweeping them off the 
board, and clean up the pads with solder wick.

Takes maybe 2 minutes, and almost never damages the board.

Jon

RE: DECstation 220 - Out of Ideas?

[Rob Jarratt via cctalk]

> -Original Message-
> From: Dave Wade [mailto:dave.g4...@gmail.com]
> Sent: 25 June 2017 20:26
> To: r...@jarratt.me.uk; 'Rob Jarratt' <robert.jarr...@ntlworld.com>; 'General
> Discussion: On-Topic and Off-Topic Posts' <cctalk@classiccmp.org>
> Subject: RE: DECstation 220 - Out of Ideas?
> 
> Rob,
>  I have a cheap hot air re-work station that you are welcome to borrow. You
> might need to order a nozzle as I do,t have many Dave
> 
> 

That is very kind Dave. I think I might first try some Chip Quik, but it would 
be interesting to have a go with a hot air rework station at some point.

Regards

Rob

Re: DECstation 220 - Out of Ideas?

[Anders Nelson via cctalk]

Heheh, that's a cool idea. I use a post-it note, sliding a non-sticky side
under the pins (from the rear of said pins) as I drag the iron across all
the pins.

=]

On Sun, Jun 25, 2017 at 11:27 AM Pete Turnbull via cctalk <
cctalk@classiccmp.org> wrote:

> On 25/06/2017 11:46, Rob Jarratt via cctalk wrote:
>
> > I have done a bit more tracing of components and have come across a
> > possibly anomaly. The signal that won't go low enough goes to the
> > input (pin 2) of a 74LS125 buffer.
>
> As soon as I saw "74LS125" I thought, "Oh, yes, another one".  That's a
> fairly early design tri-state buffer, and in my experience they're not
> very robust.  We used to replace lots of those, though admittedly I'm
> referring to the DIL version.  Worth checking it out.
>
> If you can't desolder all the SMD pins, try this trick: for each side,
> thread a piece of enamelled copper wire under the chip near the pins.
> Tie or wrap one one to some nearby component and hold the other end in
> your hand.  Starting at the end with the wire you're holding, heat each
> pin with your soldering iron, and drag the wire under it, thereby
> breaking its contact with the board.
>
> Failing that, apply a tiny amount of flux to the pins, and heat it up
> with a hot air gun, preferably a temperature-controlled one, and lift it
> off with forceps (or flip it over with a small screwdriver).
>
> --
> Pete
> Pete Turnbull
>
-- 
--
Anders Nelson

+1 (517) 775-6129

www.erogear.com

RE: DECstation 220 - Out of Ideas?

[Dave Wade via cctalk]

Rob,
 I have a cheap hot air re-work station that you are welcome to borrow. You 
might need to order a nozzle as I do,t have many
Dave


> -Original Message-
> From: cctalk [mailto:cctalk-boun...@classiccmp.org] On Behalf Of Rob Jarratt
> via cctalk
> Sent: 25 June 2017 11:47
> To: 'General Discussion: On-Topic and Off-Topic Posts'
> <cctalk@classiccmp.org>; 'Ian S. King' <isk...@uw.edu>
> Subject: RE: DECstation 220 - Out of Ideas?
> 
> 
> 
> > Thanks Ian. Unfortunately I don't know of any schematics. I suspect
> > you are right about it being a wire-OR because the documentation for
> > the P82C206 says its output is tri-state. As everything is soldered on
> > rather than socketed (except the ROMs) I am not sure I can measure any
> > currents, nor can I think of a way to check if the other chip
> > (Olivetti GA099-B) is trying to drive the output the other way :-(. I
> > will check the resistance between the outputs of the P82C206 and the ROM
> inputs and of the Vss pins.
> >
> 
> 
> I have done a bit more tracing of components and have come across a
> possibly anomaly. The signal that won't go low enough goes to the input (pin
> 2) of a 74LS125 buffer. The resistance across the input and output (pins 2-3) 
> is
> a few Kohm. Other input-output resistances on this chip are much higher.
> The same goes for Vcc to the input, the resistance is much lower than Vcc to
> other inputs. It could be that the 74LS125 is faulty, or something else on 
> that
> line has a low resistance path to Vcc, either way it would seem to be a
> possible reason for the signal never to go low enough.
> 
> I think I am going to have to get myself a cheap hot air rework station to 
> allow
> me to remove the 74LS125 (it is surface mounted) so I can check it out of
> circuit.
> 
> However, I was looking at this machine with a view to passing it on anyway. Is
> there anyone in the UK who would be interested in it (unrepaired)? It would
> be just for the cost of post and packing, or you can collect from the
> Manchester area.
> 
> Regards
> 
> Rob

Re: DECstation 220 - Out of Ideas?

[Chuck Guzis via cctalk]

On 06/25/2017 11:41 AM, Rob Jarratt wrote:

> Thanks Chuck, I had seen some mention of Chip Quik. You say soldering a new 
> part is straightforward, but I am not clear how that works, is the solder 
> left behind on the pads so you just heat the pins?

As mentioned, I use the toothbrush to scrub the pads "clean" of any
leftover alloy+solder mixture, then place the part, tack the corner
leads into place with a bit of solder, then flow solder across all the
leads and clean up with a bit of solder wick.  I've done 0.5mm QFP parts
this way without issue.  Some use a bit of Blu-Tack under the package to
hold it in place, but I've never found that to be necessary.

There are several YT videos on this technique, including, I believe,
Dave's EEVBlog.  It's easier than it looks.

--Chuck

RE: DECstation 220 - Out of Ideas?

[Rob Jarratt via cctalk]

> When replacing components on an SMD board, I prefer to use a low-temp
> fusible alloy such as "Chip Quik":
> 
> http://www.chipquik.com/store/
> 
> A nice kit, but really, any low temp fusible alloy will work.  I use 
> Cerrobend as a
> powder, packed around the IC.
> 
> http://www.csalloys.com/specifications.html
> 
> I heat the board from the underside with a PAR38 halogen spotlight, which
> raises the temperature to around the point where it becomes uncomfortable to
> the touch.
> 
> The part then just slides off; the IC pads (and the IC itself if you want to 
> re-use
> it) can be cleaned with a toothbrush and the continued application of light
> heating.
> 
> All the parts around the IC are unaffected and soldering a new part is
> straightforward.
> 


Thanks Chuck, I had seen some mention of Chip Quik. You say soldering a new 
part is straightforward, but I am not clear how that works, is the solder left 
behind on the pads so you just heat the pins?

Regards

Rob

Re: DECstation 220 - Out of Ideas?

[Chuck Guzis via cctalk]

On 06/25/2017 09:24 AM, Rob Jarratt via cctalk wrote:

> Thanks Pete, so there is enough of a chance that I have found the
> culprit to make it worth the effort of desoldering it.
> 
> I have been wondering for a while about getting a hot air rework
> station. There are lots of cheap Chinese ones on Ebay. Given my
> likely light use it feels worth getting one at this point. Anyone
> tried one of these:
> http://www.ebay.co.uk/itm/Genuine-WEP-640W-Hot-Air-Gun-Soldering-Rework-Station-Temp-Control-Digital-/112453738508?
> At that price it feels worth the gamble.
When replacing components on an SMD board, I prefer to use a low-temp
fusible alloy such as "Chip Quik":

http://www.chipquik.com/store/

A nice kit, but really, any low temp fusible alloy will work.  I use
Cerrobend as a powder, packed around the IC.

http://www.csalloys.com/specifications.html

I heat the board from the underside with a PAR38 halogen spotlight,
which raises the temperature to around the point where it becomes
uncomfortable to the touch.

The part then just slides off; the IC pads (and the IC itself if you
want to re-use it) can be cleaned with a toothbrush and the continued
application of light heating.

All the parts around the IC are unaffected and soldering a new part is
straightforward.

FWIW,
Chuck

RE: DECstation 220 - Out of Ideas?

[Rob Jarratt via cctalk]

> -Original Message-
> From: cctalk [mailto:cctalk-boun...@classiccmp.org] On Behalf Of Pete
> Turnbull via cctalk
> Sent: 25 June 2017 16:12
> To: cctalk@classiccmp.org
> Subject: Re: DECstation 220 - Out of Ideas?
> 
> As soon as I saw "74LS125" I thought, "Oh, yes, another one".  That's a fairly
> early design tri-state buffer, and in my experience they're not very robust.  
> We
> used to replace lots of those, though admittedly I'm referring to the DIL
> version.  Worth checking it out.
> 
> If you can't desolder all the SMD pins, try this trick: for each side, thread 
> a
> piece of enamelled copper wire under the chip near the pins.
> Tie or wrap one one to some nearby component and hold the other end in your
> hand.  Starting at the end with the wire you're holding, heat each pin with 
> your
> soldering iron, and drag the wire under it, thereby breaking its contact with 
> the
> board.
> 
> Failing that, apply a tiny amount of flux to the pins, and heat it up with a 
> hot air
> gun, preferably a temperature-controlled one, and lift it off with forceps 
> (or flip
> it over with a small screwdriver).
> 


Thanks Pete, so there is enough of a chance that I have found the culprit to 
make it worth the effort of desoldering it.

I have been wondering for a while about getting a hot air rework station. There 
are lots of cheap Chinese ones on Ebay. Given my likely light use it feels 
worth getting one at this point. Anyone tried one of these: 
http://www.ebay.co.uk/itm/Genuine-WEP-640W-Hot-Air-Gun-Soldering-Rework-Station-Temp-Control-Digital-/112453738508?
 At that price it feels worth the gamble.

Regards

Rob

Re: DECstation 220 - Out of Ideas?

[Pete Turnbull via cctalk]

On 25/06/2017 11:46, Rob Jarratt via cctalk wrote:


I have done a bit more tracing of components and have come across a
possibly anomaly. The signal that won't go low enough goes to the
input (pin 2) of a 74LS125 buffer.


As soon as I saw "74LS125" I thought, "Oh, yes, another one".  That's a 
fairly early design tri-state buffer, and in my experience they're not 
very robust.  We used to replace lots of those, though admittedly I'm 
referring to the DIL version.  Worth checking it out.


If you can't desolder all the SMD pins, try this trick: for each side, 
thread a piece of enamelled copper wire under the chip near the pins. 
Tie or wrap one one to some nearby component and hold the other end in 
your hand.  Starting at the end with the wire you're holding, heat each 
pin with your soldering iron, and drag the wire under it, thereby 
breaking its contact with the board.


Failing that, apply a tiny amount of flux to the pins, and heat it up 
with a hot air gun, preferably a temperature-controlled one, and lift it 
off with forceps (or flip it over with a small screwdriver).


--
Pete
Pete Turnbull

RE: DECstation 220 - Out of Ideas?

[Rob Jarratt via cctalk]

> Thanks Ian. Unfortunately I don't know of any schematics. I suspect you are
> right about it being a wire-OR because the documentation for the P82C206 says
> its output is tri-state. As everything is soldered on rather than socketed 
> (except
> the ROMs) I am not sure I can measure any currents, nor can I think of a way 
> to
> check if the other chip (Olivetti GA099-B) is trying to drive the output the 
> other
> way :-(. I will check the resistance between the outputs of the P82C206 and 
> the
> ROM inputs and of the Vss pins.
> 


I have done a bit more tracing of components and have come across a possibly 
anomaly. The signal that won't go low enough goes to the input (pin 2) of a 
74LS125 buffer. The resistance across the input and output (pins 2-3) is a few 
Kohm. Other input-output resistances on this chip are much higher. The same 
goes for Vcc to the input, the resistance is much lower than Vcc to other 
inputs. It could be that the 74LS125 is faulty, or something else on that line 
has a low resistance path to Vcc, either way it would seem to be a possible 
reason for the signal never to go low enough.

I think I am going to have to get myself a cheap hot air rework station to 
allow me to remove the 74LS125 (it is surface mounted) so I can check it out of 
circuit.

However, I was looking at this machine with a view to passing it on anyway. Is 
there anyone in the UK who would be interested in it (unrepaired)? It would be 
just for the cost of post and packing, or you can collect from the Manchester 
area.

Regards

Rob

RE: DECstation 220 - Out of Ideas?

[Rob Jarratt via cctalk]

> -Original Message-
> From: Ian S. King [mailto:isk...@uw.edu]
> Sent: 25 June 2017 01:07
> To: r...@jarratt.me.uk; Rob Jarratt <robert.jarr...@ntlworld.com>; General
> Discussion: On-Topic and Off-Topic Posts <cctalk@classiccmp.org>
> Subject: Re: DECstation 220 - Out of Ideas?
> 
> Rob
> I don't recall from this thread - do you have prints?  I'd try to divine 
> whether
> this is a wire-OR bus (likely).  If so, I'd suspect the P82C206.  It may not 
> be toast
> itself - be sure it has sufficient power to drive the OC transistor that's 
> bringing
> down the bus, and that its connection to that bus doesn't have excessive
> resistance, and that its ground (earth) connection is also low-resistance.  A 
> long
> shot: it could be something else on that bus that's sourcing rather than 
> sinking,
> but the scenario seems unlikely.
> 
> I'd agree that it's probably not the ROMs, from the symptoms you're describing
> - unless, again, one of them has an internal short/leak to rail on its 
> read-enable
> pin.  A thought: can you measure the current and its direction on the P82C206
> pin?
> 
> That's my troubleshooting thinking given that I don't have it in front of me. 
>  :-)  I
> hope this is helpful -- Ian
> 
> 
> 


Thanks Ian. Unfortunately I don't know of any schematics. I suspect you are 
right about it being a wire-OR because the documentation for the P82C206 says 
its output is tri-state. As everything is soldered on rather than socketed 
(except the ROMs) I am not sure I can measure any currents, nor can I think of 
a way to check if the other chip (Olivetti GA099-B) is trying to drive the 
output the other way :-(. I will check the resistance between the outputs of 
the P82C206 and the ROM inputs and of the Vss pins.

Regards

Rob

Re: DECstation 220 - Out of Ideas?

[Ian S. King via cctalk]

On Sat, Jun 24, 2017 at 4:17 PM, Rob Jarratt via cctalk <
cctalk@classiccmp.org> wrote:

> As I mentioned in an earlier email, I have been trying to fix a DECstation
> 220. I have made what I think may be some progress, but I am out of ideas
> now.
>
>
>
> The problem appears to be that the 80286 CPU is shutting down and then the
> Olivetti chip is RESETting it repeatedly. The reason for this seems to be
> that it is failing to read the ROMs. The ROM is an ST M27512. The output
> enable signal (active low) does not appear to go below about 1.8V, although
> it does go up to 5V. It is as if the system is trying to read the ROM, but
> somehow the signal is not going low enough to actually make the ROM read.
> The signal is connected to DMAMEMR (DMA Memory Read) on a Chips P82C206
> controller, and to something on the Olivetti GA099-B, it may be connected
> to
> other components too, but I have not discovered any more so far. I have
> taken the ROMs out and read them in a reader, they seem to read back OK, so
> I don't think they are faulty.
>
>
>
> I am not sure what could cause the signal not to go low enough, apart from
> a
> bad P82C206 or a bad GA099-B. Does anyone have any suggestions or
> experience
> of common causes?
>
>
>
> Thanks
>
>
>
> Rob
>
> I don't recall from this thread - do you have prints?  I'd try to divine
whether this is a wire-OR bus (likely).  If so, I'd suspect the P82C206.
It may not be toast itself - be sure it has sufficient power to drive the
OC transistor that's bringing down the bus, and that its connection to that
bus doesn't have excessive resistance, and that its ground (earth)
connection is also low-resistance.  A long shot: it could be something else
on that bus that's sourcing rather than sinking, but the scenario seems
unlikely.

I'd agree that it's probably not the ROMs, from the symptoms you're
describing - unless, again, one of them has an internal short/leak to rail
on its read-enable pin.  A thought: can you measure the current and its
direction on the P82C206 pin?

That's my troubleshooting thinking given that I don't have it in front of
me.  :-)  I hope this is helpful -- Ian


-- 
Ian S. King, MSIS, MSCS, Ph.D. Candidate
The Information School 
Dissertation: "Why the Conversation Mattered: Constructing a Sociotechnical
Narrative Through a Design Lens

Archivist, Voices From the Rwanda Tribunal 
Value Sensitive Design Research Lab 

University of Washington

There is an old Vulcan saying: "Only Nixon could go to China."

DECstation 220 - Out of Ideas?

[Rob Jarratt via cctalk]

As I mentioned in an earlier email, I have been trying to fix a DECstation
220. I have made what I think may be some progress, but I am out of ideas
now.

 

The problem appears to be that the 80286 CPU is shutting down and then the
Olivetti chip is RESETting it repeatedly. The reason for this seems to be
that it is failing to read the ROMs. The ROM is an ST M27512. The output
enable signal (active low) does not appear to go below about 1.8V, although
it does go up to 5V. It is as if the system is trying to read the ROM, but
somehow the signal is not going low enough to actually make the ROM read.
The signal is connected to DMAMEMR (DMA Memory Read) on a Chips P82C206
controller, and to something on the Olivetti GA099-B, it may be connected to
other components too, but I have not discovered any more so far. I have
taken the ROMs out and read them in a reader, they seem to read back OK, so
I don't think they are faulty.

 

I am not sure what could cause the signal not to go low enough, apart from a
bad P82C206 or a bad GA099-B. Does anyone have any suggestions or experience
of common causes?

 

Thanks

 

Rob