Re: [Emc-users] [was] Anybody know where to find a pdf on the HOYMK SSR40DA SSR?

[Gene Heskett]

On Sunday 20 December 2015 16:42:48 John Kasunich wrote:

> On Sun, Dec 20, 2015, at 03:45 PM, Gene Heskett wrote:
> > On Sunday 20 December 2015 04:27:39 Gene Heskett wrote:
> >
> > [... back to readable size, on my charge-pump detection to run the
> > SSR's]
> >
> > > Has anyone else concocted such a circuit, and were you able to get
> > > enough pump power out of it running the BoB on 5 volts, with the
> > > pump at 500 hz, to run an SSR without resorting to an external
> > > current booster?
> >
> > Now I am both bumfuzzled and confused. There really ought to be a
> > formula for calculating the size of the capacitors needed, but there
> > is absolutely zip in ITT #5, nor in the hambook of about the same
> > vintage.
> >
> > So, I guess its build it, and test it by the old fashioned cuss &
> > cry method. Seems to me there really ought to be, in the year 2015,
> > a plug in the blanks formula for that.  Sigh...
>
> My gut feeling is that you won't be able to do it.
>
> First, lets make sure we are talking about the same circuit.  A google
> image search found this schematic of a charge pump, which is the
> circuit I normally use.
> http://www.zen22142.zen.co.uk/Theory/diode_pump/Image152.jpg
> (refer to the circuit on the left)

Yes, thats the one John K., except the circuit ground is the + terminal, 
and the diode and cap pols reversed.  The idea is to make the 1st cap's 
available charge so low at powerup, that there will not be a 1 or 2 
cycle on time at powerup.

>
> Assume that the input is swinging from 0.0 to 5.0 volts.  Also assume
> that the loading is negligible.  When the input is low, the voltage on
> the left side of Cp is 0.0, and the right side is -0.7V due to the
> drop in D1.  When the input goes high, the left side of Cp is 5.0V,
> and the right side is 4.3V, assuming zero change in the voltage across
> Cp.  Because of the drop across D2, the output is 4.3-0.7 = 3.6V. 
> Pretty marginal.  And that is with zero load, and an output that
> swings a full 5V.  No real output will swing all the way down to 0.0
> or up to 5.0V, so you lose another couple tenths of a volt.

Unless cmos drivers are used, true.

> The above assumes that there is NO CHANGE in the voltage across Cp.
> That means no current flows into or out of Cp, and the circuit
> delivers no output current (I said "assume loading is negligable").
>
> To deliver any load current, the voltage across Cp has to change, and
> that will further drop the output voltage.  Say you need a load
> current of 10mA. That is 10 milli-coulombs per second.  If the input
> frequency is 500Hz, then each cycle needs to deliver 0.010/500 = 20
> micro-coulombs.  If you need an output voltage of 3V, and the ideal
> no-load output is 3.6V, then the voltage across Cp can only be allowed
> to drop 0.6V when 20 micro-coulombs of charge is extracted from it.  C
> = Q/V = 0.20 / 0.6 = 33uF.  That is a HUGE pump cap, and will
> badly load down whatever is driving it.

That is the reasoning behind what I was trying to do, but all my 
reference books are apparently too prehistoric to know about coulombs.

> When I make a charge pump, I usually think of something like Cr =
> 0.1uF, and Cp = 0.01uF.  Charge per cycle at 0.6V droop is 0.006
> micro-coulombs, and I run it at 10kHz instead of 500Hz, for a total
> load current of 60 micro- amps.  Put a 100K load resistor as a load
> and drive the gate of a logic level MOSFET and it works fine.  But
> getting milli-amp currents out of a low-powered charge pump is much
> much harder.

The mosfet isn't a problem, nor would a small sig npn transistor, but 
power for it is.  Not because its needed, but because of the external 
powers on-off transients feeding in and causing a few cycles of "on" 
before the rest of the circuit is ready.

That will obviously defeat the purpose unless that external supply is 
itself fully floating. I do have a smallish 12.6.ct.12.6 tranny, good 
for 200 mills or so, that I can make that psu with, and be completely 
isolated from a ground reference except for the capacitance in Cp, and 
of course the leakage capacitance in the transformer.  But theres no 
base thread, so I'm stuck at 500 hertz.

This almost sounds like a good place for a cmos 7556 if I can figure out 
how to disable it until the psu is up and stable.

I am home from the woofs cataract surgery, which I was up well before 
breakfast to attend to, so I'll be later this afternoon before I am able 
to play. Right now I feel a catchup nap coming on. :)

Thanks for confirming my worries.

> John Kasunich
>
>
>
>
>
> --
> ___
> 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)
Some 

Re: [Emc-users] [was] Anybody know where to find a pdf on the HOYMK SSR40DA SSR?

[John Kasunich]

On Mon, Dec 21, 2015, at 01:17 PM, Gene Heskett wrote:

> The mosfet isn't a problem, nor would a small sig npn transistor, but 
> power for it is.  Not because its needed, but because of the external 
> powers on-off transients feeding in and causing a few cycles of "on" 
> before the rest of the circuit is ready.
> 
> That will obviously defeat the purpose unless that external supply is 
> itself fully floating. I do have a smallish 12.6.ct.12.6 tranny, good 
> for 200 mills or so, that I can make that psu with, and be completely 
> isolated from a ground reference except for the capacitance in Cp, and 
> of course the leakage capacitance in the transformer.  But theres no 
> base thread, so I'm stuck at 500 hertz.

It seems to me that the FET source needs to be common to the signal
and the load power.

Regarding the initial "power up" glitch, that shouldn't happen.  The
reservoir cap (Cr) needs to be much bigger than the pump cap (Cp).
I consider 10:1 as a minimum, 50:1 or even 100:1 can be used if 
needed.  A rough approximation is that if Cr is 50 times Cp, it will 
take at least 50 pulses to charge it up.  That should pretty much 
eliminate any mis-behavior at startup.

-- 
  John Kasunich
  jmkasun...@fastmail.fm

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

Re: [Emc-users] [was] Anybody know where to find a pdf on the HOYMK SSR40DA SSR?

[Gene Heskett]

On Monday 21 December 2015 13:43:23 John Kasunich wrote:

> On Mon, Dec 21, 2015, at 01:17 PM, Gene Heskett wrote:
> > The mosfet isn't a problem, nor would a small sig npn transistor,
> > but power for it is.  Not because its needed, but because of the
> > external powers on-off transients feeding in and causing a few
> > cycles of "on" before the rest of the circuit is ready.
> >
> > That will obviously defeat the purpose unless that external supply
> > is itself fully floating. I do have a smallish 12.6.ct.12.6 tranny,
> > good for 200 mills or so, that I can make that psu with, and be
> > completely isolated from a ground reference except for the
> > capacitance in Cp, and of course the leakage capacitance in the
> > transformer.  But theres no base thread, so I'm stuck at 500 hertz.
>
> It seems to me that the FET source needs to be common to the signal
> and the load power.
>
Absolutely.  It could even be a std 2n for that matter, and the 
shutdown lag wouldn't be quite as great as with the fet.  OTOH, that 
would all depend on the needed initial bias to make the fet conduct too.

But in the context of discussing the micro-coulombs per pulse pumped by 
Cp, should we not discount it somewhat because of the limiting source 
resistance the BoB's output stage represents?  Or are we assuming a Cp 
small enough that at 1 millisecond at each voltage level, t=rc is 
essentially infinite and Cp is fully charged to whatever differential 
voltage is available?

Also, and I am not sure I have two of them, and I know I don't have 4 for 
both circuits, so I expect I should order a small bag of schotkey 
diodes, which would raise the potential output voltage by approximately 
an additional volt.

> Regarding the initial "power up" glitch, that shouldn't happen.  The
> reservoir cap (Cr) needs to be much bigger than the pump cap (Cp).
> I consider 10:1 as a minimum, 50:1 or even 100:1 can be used if
> needed.  A rough approximation is that if Cr is 50 times Cp, it will
> take at least 50 pulses to charge it up.  That should pretty much
> eliminate any mis-behavior at startup.

I don't believe the junk box network I put together for the vacuum is 
more than 3 or 4/1.  Its driving a junk box to-220 of some sort, which 
is in turn pulling the 24 volt coil of a DPDT P octal plugin relay, 
whose contacts are paralleled in deference to the 6.5 amps running (at 
normal air flow, my lashup could even be higher because nearly all the 
plumbing but the radiator hose is nice smooth interior 1.5" PVC, and 
that 6.5 is probably measured with the 1.125" corrugated hose normally 
used. That stuff, while handy as a vac hose, is very high internal 
turbulance & a good flow restriction)

My "nap" lasted about 4.5 hours! I must have really been behind.

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)
Some mill pix are at:
Genes Web page 

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

Re: [Emc-users] [was] Anybody know where to find a pdf on the HOYMK SSR40DA SSR?

[John Kasunich]

On Sun, Dec 20, 2015, at 03:45 PM, Gene Heskett wrote:
> On Sunday 20 December 2015 04:27:39 Gene Heskett wrote:
> 
> [... back to readable size, on my charge-pump detection to run the SSR's]
> 
> > Has anyone else concocted such a circuit, and were you able to get
> > enough pump power out of it running the BoB on 5 volts, with the pump
> > at 500 hz, to run an SSR without resorting to an external current
> > booster?
> 
> Now I am both bumfuzzled and confused. There really ought to be a formula 
> for calculating the size of the capacitors needed, but there is 
> absolutely zip in ITT #5, nor in the hambook of about the same vintage. 
> 
> So, I guess its build it, and test it by the old fashioned cuss & cry 
> method. Seems to me there really ought to be, in the year 2015, a plug 
> in the blanks formula for that.  Sigh...
> 

My gut feeling is that you won't be able to do it.

First, lets make sure we are talking about the same circuit.  A google image
search found this schematic of a charge pump, which is the circuit I normally
use.
http://www.zen22142.zen.co.uk/Theory/diode_pump/Image152.jpg
(refer to the circuit on the left)

Assume that the input is swinging from 0.0 to 5.0 volts.  Also assume that
the loading is negligible.  When the input is low, the voltage on the left side
of Cp is 0.0, and the right side is -0.7V due to the drop in D1.  When the input
goes high, the left side of Cp is 5.0V, and the right side is 4.3V, assuming 
zero change in the voltage across Cp.  Because of the drop across D2, the
output is 4.3-0.7 = 3.6V.  Pretty marginal.  And that is with zero load, and 
an output that swings a full 5V.  No real output will swing all the way down
to 0.0 or up to 5.0V, so you lose another couple tenths of a volt.

The above assumes that there is NO CHANGE in the voltage across Cp.
That means no current flows into or out of Cp, and the circuit delivers 
no output current (I said "assume loading is negligable").

To deliver any load current, the voltage across Cp has to change, and that
will further drop the output voltage.  Say you need a load current of 10mA.
That is 10 milli-coulombs per second.  If the input frequency is 500Hz, then
each cycle needs to deliver 0.010/500 = 20 micro-coulombs.  If you need
an output voltage of 3V, and the ideal no-load output is 3.6V, then the voltage 
across Cp can only be allowed to drop 0.6V when 20 micro-coulombs of
charge is extracted from it.  C = Q/V = 0.20 / 0.6 = 33uF.  That is a
HUGE pump cap, and will badly load down whatever is driving it.

When I make a charge pump, I usually think of something like Cr = 0.1uF,
and Cp = 0.01uF.  Charge per cycle at 0.6V droop is 0.006 micro-coulombs,
and I run it at 10kHz instead of 500Hz, for a total load current of 60 micro-
amps.  Put a 100K load resistor as a load and drive the gate of a logic 
level MOSFET and it works fine.  But getting milli-amp currents out of a
low-powered charge pump is much much harder.

John Kasunich





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

Re: [Emc-users] [was] Anybody know where to find a pdf on the HOYMK SSR40DA SSR?

[Gene Heskett]

On Sunday 20 December 2015 04:27:39 Gene Heskett wrote:

[... back to readable size, on my charge-pump detection to run the SSR's]

> Has anyone else concocted such a circuit, and were you able to get
> enough pump power out of it running the BoB on 5 volts, with the pump
> at 500 hz, to run an SSR without resorting to an external current
> booster?

Now I am both bumfuzzled and confused. There really ought to be a formula 
for calculating the size of the capacitors needed, but there is 
absolutely zip in ITT #5, nor in the hambook of about the same vintage. 

So, I guess its build it, and test it by the old fashioned cuss & cry 
method. Seems to me there really ought to be, in the year 2015, a plug 
in the blanks formula for that.  Sigh...

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)
Some mill pix are at:
Genes Web page 

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