POWER DESIGNS PRECISION POWER SUPPLIES

[robrenz]

The second 2005 I got continued to blow the 1 amp fuse.  Three fuses later I see it happens when I move the current limit pot through its range and I noticed a spike in the needle on the amp gage. So it was the momentary open circuit on the pot that was spiking the current and blowing the  fuse.  So I took the pot apart to clean and lube but I kept getting a jump on my fluke87 ohms reading in roughly the same places. So half an hour later after ultrasonic cleaning and microscope inspection of every single turn of wire in the pot and still seeing the glitch in the pot travel. I finally realize the jump is the range change on the meter  (EDIT: from 400 Ohm to 4k Ohm range), not a open in the pot travel.  lock the range on the meter and no problem

[saturation]

Hi robrenz,

I'm back.  Lots of catching up to do.  When you mention range change, that's means moving from 0-10V to 10-20V or vice versa, but if you don't touch the switch all is well?  I presume the current limiting is set at a value far lower than the fuse rating? 

The second 2005 I got continued to blow the 1 amp fuse.  Three fuses later I see it happens when I move the current limit pot through its range and I noticed a spike in the needle on the amp gage. So it was the momentary open circuit on the pot that was spiking the current and blowing the  fuse.  So I took the pot apart to clean and lube but I kept getting a jump on my fluke87 ohms reading in roughly the same places. So half an hour later after ultrasonic cleaning and microscope inspection of every single turn of wire in the pot and still seeing the glitch in the pot travel. I finally realize the jump is the range change on the meter, not a open in the pot travel.  lock the range on the meter and no problem

[robrenz]

The range change was on my fluke meter as I swept through the whole range of the 10k 1k pot.  The current limit pot was dirty and did go open as you moved it through the range. But when I had the pot apart and was checking the resistance before I put it together again I was thinking the blink when the range changed was a open circuit. So for half an hour at least I was searching for a problem that was not there.  When I locked the range on the meter the  pot was fine after cleaning.  Quite embarrassing

[terabyte]

The range change was on my fluke meter as I swept through the whole range of the 10k pot.  The current limit pot was dirty and did go open as you moved it through the range. But when I had the pot apart and was checking the resistance before I put it together again I was thinking the blink when the range changed was a open circuit. So for half an hour at least I was searching for a problem that was not there.  When I locked the range on the meter the  pot was fine after cleaning.  Quite embarrassing

I bet you will never make that mistake again.  This is why experience in any field is so important.

[saturation]

A PM I received from Paul Rako regarding his Power Design unit:

Saturation,

You have received a new Private message from Paul Rako with the subject: Re: Power Designs 2005 Precision DC Supply.

"The only problem my old Power Designs supply had was oxidation on the switch contacts. I removed that with De-Oxit  (http://www.caig.com/). It worked fine after that, but I hate to tell you, I sold it to a friend. For very accurate voltage creation you can use to check A to D converters, see my pal Mark Thoren's presentation on buffering a Deka-pot to give a precise output voltage: http://www.edn.com/video/video.php/?bclid=1028763155&bctid=1026280202 "

http://www.rako.com/Articles/29.html

i think that's not bad at all for a unit >40 years old.

[alm]

The current limit pot was dirty and did go open as you moved it through the range.

Blowing fuses when a pot goes OC for a short time sounds like a design flaw to me. One should always assume that pots might go OC, especially the wire wound ones that might have a single bad turn.

For very accurate voltage creation you can use to check A to D converters, see my pal Mark Thoren's presentation on buffering a Deka-pot to give a precise output voltage: http://www.edn.com/video/video.php/?bclid=1028763155&bctid=1026280202 "[/i]

Video doesn't work here. I guess it's that nice cheap DIY design requiring a $4k+ pot? Seems Paul is a bit out of touch with the people that have limited budgets.

[saturation]

The current limit pot was dirty and did go open as you moved it through the range.

Blowing fuses when a pot goes OC for a short time sounds like a design flaw to me. One should always assume that pots might go OC, especially the wire wound ones that might have a single bad turn.

It might be, but a solution for this 1960-70s design was to protect the system with a fuse, rather than an active shut down.  Given how infrequently a bad turn happens in a precision pot it seems like a cost effective solution.  That the fuse didn't blow in years [ nor has my versions blown in all 3 units] and the units are 30+ years old is QED that design ethic was cost effecient and effective, I think.

For very accurate voltage creation you can use to check A to D converters, see my pal Mark Thoren's presentation on buffering a Deka-pot to give a precise output voltage: http://www.edn.com/video/video.php/?bclid=1028763155&bctid=1026280202 "[/i]

Video doesn't work here. I guess it's that nice cheap DIY design requiring a $4k+ pot? Seems Paul is a bit out of touch with the people that have limited budgets.

Yes, I think its the same video, here mirrored on youtube:



And yes, it does use that $4000 KV in a dial switch. 



http://www.ietlabs.com/voltage-divider/esi-dp-1211-dp1311-voltage-dividers.html


The PD and even the Krohn Hite EDC 106 designs use resistance dividers in a similar approach as the Linear design, but their DIY rotary switch based dividers provide a >~ accuracy and precision at much less cost. 

In addition, given PD precision PSU availability via eBay, one can get a working product that not only provides accuracy similar to Linear's design, fully assembled, burned in, cost ~ 3% of a KV divider alone and source current like a bench PSU.  The buyer has to beware units are aged and will need a going through to insure its functionality, which so far reported, its has been.

[MBY]

I have an model 2005 (old model), and I also had problem with the settings at first. The linearity was poor, a higher setting could result in a lower voltage for instance.

I cleansed the rotary switch contacts and now it works really well. It's linearity is good but the unit is uncalibrated. But I'm happy. The only problem I have is the current limitation, its not so stable. If I look on the schematics, its probably temp drift in the shunt resistor or the pot, as they are "non ovenized".

Do I recommend it? Yes. But only if you think its worth the bucks! I payed near $200 for the unit, shipping, customs, Posten's (Swedish postal services) ridiculous fees and have to walk for 1 km to lug it home. The unit in it self only cost me about $70 of the $200 if I remember correctly! 

[robrenz]

Calibration is very easy and covered in the manual. Then it will agree with your best meter at 0V and 20V which is better than no calibration at all.  Make sure you use the corresponding manual. AFIK the one in the Paul Rako article is for the version with the large horizontal meter (2005). There is another one for the unit with the thin vertical meter (2005A).  From my experience let it warm up at least an hour before calibration. If you have not had it on much since you bought it I would let it burn in a couple of days straight before calibration. FWIW

Regards, robrenz

[EEVblog]

I have a 110V PD 2005, and it seems to be in a bit crusty condition.
More to come on this I'm sure...

Dave.

[robrenz]

Anxiously waiting 

[MBY]

Calibration is very easy and covered in the manual. Then it will agree with your best meter at 0V and 20V which is better than no calibration at all.  Make sure you use the corresponding manual. AFIK the one in the Paul Rako article is for the version with the large horizontal meter (2005). There is another one for the unit with the thin vertical meter (2005A).  From my experience let it warm up at least an hour before calibration. If you have not had it on much since you bought it I would let it burn in a couple of days straight before calibration. FWIW

Regards, robrenz

I think I have the same unit, the meter is the large horizontal one. I did a "burn in" after I received the unit and then I tested the stability by dialing in 10.000 V (9.993 + vernier) measured with a Rigol DM3061 6½ digit DMM. After several days the drift was about 30 µV, amazing! Now I haven't used the PSU in some time, but the dial and vernier is untouched since last time. So, I just powered it on again, connected to the same DMM and I will check on it in a few hours!

I use a cheep 230-115 VAC 50 watt transformer by the way. When I measure the internals of the 2005 and the output voltage, the 2005 is slightly underpowered (95-105 VAC depending on load), but the small wal-wart type 230-115 transformer seems not to get too hot and the 2005 works really well.

For some reason, all the neon bulb was burned so I have them replaced. Always have exotic spare parts! (No, they where of course not an exact match. I needed lots of shrink tubing to make them fit in the old holders.)

Edit: When I think of it, I did a 0V calibration, but not the 20V one. Some rainy day... But I really wish I could come up with something to make it a really stable current source without external components/electronics. Maybe replace the current shunt resistor?

[robrenz]

Have you tested the Rigol with a shorting bar for its DCV stability over the same time period?

[MBY]

Have you tested the Rigol with a shorting bar for its DCV stability over the same time period?

No, I have not. I thought about that too, I basically measure the differential drift in two independent instrument so there is a source for error. However, when I measure things like this, the Rigol is only my "readout instrument". In parallell I usually have my Fluke 45 (100 000 counts) and my TTI 1906 "Computing Multimeter" (210 000 counts) to track. They are all within a few counts on the LSD on the Fluke.

But a shorting bar measurement is a good idea! Will do.

Edit: Me and English...

[saturation]

This might be the design flaw alm suggested.  In the 2005A schematic if current adjust pot's wiper arm in R35 were to go open circuit briefly, Q10 would not be properly biased.  See photo, Circle RED B + ARROW.  This would alter feedback on current limit section YELLOW CIRCLE, and could blow fuse F2 1A, RED A + ARROW.



Compare to the much later designed, 2020B, where if the wiper were to open, only the full resistance of R8 or R41 would be felt.

The current limit pot was dirty and did go open as you moved it through the range.

Blowing fuses when a pot goes OC for a short time sounds like a design flaw to me. One should always assume that pots might go OC, especially the wire wound ones that might have a single bad turn.

[krivx]

I have a 110V PD 2005, and it seems to be in a bit crusty condition.
More to come on this I'm sure...

Dave.

Please give us some notice if you do a video on it - you have a record of affecting the ebay prices of test gear

[MBY]

Thats a nice post, saturation!

I think I have the model 2005 (without "A", or is it implicit?). The model looks exactly like Paul Rakos: http://www.rako.com/Articles/29.html

And, according to him, the schematics:

So, this is yet another alterations of the current control. If I understand correctly, the front panel Adj pot is R25. Funny with the parallel R27 pot. I shall re-read this thread and pay more attention to the current discussion. Thanks for helping, great post!

Edit: Wrong link to schematics.

[robrenz]

Please give us some notice if you do a video on it - you have a record of affecting the ebay prices of test gear

If you are talking about US Ebay you better be looking and buying now. there are not many of the precision units currently listed and even less at sub $100 dollars.  I am partly responsible buying (2) 2020B's (2) 2005's and (1) 2010 in the last several months. They were inexpensive (all 5 totaled $465.00 including shipping) and I could not resist those awesome machined aluminum concentric knobs. ) There are plenty of the regular power designs supplies listed and they are also very good specs on paper.

[saturation]

You're welcome.  The thread hopefully will help other finds faults and fix them, as well as find the value.  There are several variations of the 2005, so check the rear panel for the specific version.  IIRC the younger versions no longer have ovens, for example, then the suffix was added as A, then later B.

The schematic you posted has several of the troublesome pot wiring techniques similar to the 2005B that can cause problems should the pot wiper fail open:

If R25 were to fail, Q8 would not be biased, its similar to the problem of the 2005B current control transistor circuit.



R27 and 25 are in parallel.  Its a technique for making fine adjustments.  If R27 were to fail the effective resistance would rise to ~< 2x, it should be wired as a modern 3 terminal pot rather than 2. 



R45 R46 is also of concern, if it opened, it would disconnect the output voltage select rotary switches.  It should be wired as a modern 3 terminal pot rather than 2. 

Thats a nice post, saturation!
I think I have the model 2005 (without "A", or is it implicit?).

[saturation]

Any PSU can provide adjustable current sourcing, if they have good voltage sourcing, and have adjustable current limiting.  You can convert any voltage source to current source effectively, as Norton's theorem implies, because neither of them are 'ideal' and thus are different ways of looking at the same phenomena, i.e., duality.

http://en.wikipedia.org/wiki/Dual_%28electronics%29

http://en.wikipedia.org/wiki/Voltage_source#Comparison_between_voltage_and_current_sources

In the case of the precision PD vs regular PDs, you should get very precise and stable current outputs from the precision PDs once the pots are cleaned! 

Edit: When I think of it, I did a 0V calibration, but not the 20V one. Some rainy day... But I really wish I could come up with something to make it a really stable current source without external components/electronics. Maybe replace the current shunt resistor?

[saturation]

Same here, looking forward to it.

I have a 110V PD 2005, and it seems to be in a bit crusty condition.
More to come on this I'm sure...

Dave.

Anxiously waiting 

[MBY]

Tank you again, saturation. I did not remember to check the pots when I opened the device. I wonder if they are mis-wired three terminal pots or only rheostats? If the former, a simple wire will improve the safety.

[saturation]

Yes, shipping outside the USA can be costly, but consider what the equivalent would have cost to get this much precision and accuracy; even at $200 its still a bargain, its both a PSU and VDC calibrator.

Again you are welcome.  Whether the pots are miswired 3 terminal or just 2 terminal pots, a change in the wiring method would be advised to avoid future problems.  While the fuse will the protect the PD supply, a pot failure may provide uncontrolled current output to any device being supplied by the PD and wired as a current source.

Do I recommend it? Yes. But only if you think its worth the bucks! I payed near $200 for the unit, shipping, customs, Posten's (Swedish postal services) ridiculous fees and have to walk for 1 km to lug it home. The unit in it self only cost me about $70 of the $200 if I remember correctly! 

Tank you again, saturation. I did not remember to check the pots when I opened the device. I wonder if they are mis-wired three terminal pots or only rheostats? If the former, a simple wire will improve the safety.

[robrenz]

I am running a CC test right now with my 2010 supply. I am running 200mA directly into my 8846A on the 400mA input and giving time for everything to get to thermal equilibrium before I start a statistical test of long term drift.  It still has not reached a stable value yet but the total change (dropping) over 80 minutes is 60uA. (I did turn the air conditioning on  ) I am watching on trend plot to see if it levels out at some value before I start.  I am thinking it will be the shunt in the meter and the shunt in the 2010 that need to get to steady state. I am definitely in CC mode with the flashing 0-10V light and have it set for 1V.  I will post what I get as soon as things settle out.

Regards, robrenz

[robrenz]

One of the last <$100 listings currently on US Ebay http://www.ebay.com/itm/Power-Designs-2005-A-Precision-DC-Power-Supply-/300713273173?pt=LH_DefaultDomain_0&hash=item4603e86755 2 available!