EEVblog Electronics Community Forum
Products => Test Equipment => Topic started by: rwgast_lowlevellogicdesin on January 11, 2020, 01:42:55 am
-
A few years ago I did a lot of work for an audiophile guy that worked at dolby, and on the side would buy "vintage" stereo crap, then pay me to diagnose and fix it if he could make a profit on e-bay. Anyways he gave a PD2005 he brought home from dolby which had, Property of Hughes Aircraft and 1994 calibration stickers on it, so the psu has kind of a cool origin.
Cosmetically it is in very good condition, I sanded all that ugly ass old blue paint off the cover to give it a nice 100% stainless look also considered powder coating the cover peace just never got around to it. I replaced all the electrolytic caps in it, but am still in possession of the original sprauge brand cap which actually pass an ESR test fine. Next cleaned and no loxed all the pots and switches. The original incandescent bulb for the 10-20v range was burnt out so I just replaced it with a green LED. The AC plug was in bad shape so I just chopped it and added a new 102v male plug from home depot to the original AC cord. The panel meter has a crack that has been epoxeyed by someone (crack isnt in the way of read out), so I went on a hunt and got a boxed never used power designs panel meter, only one I could find was a 50v 500ma not 20v, so I haven't installed it yet because I haven't replaced it with the 25v scale and appropriate resistor. Shortly after I got the supply it crapped out and I had to rebuild the heater and regulator circuit inside the heater, I ended up replacing the two matched transistors with a more modern low noise variant and i may have replaced the third one too not sure, i also re insulated the heater and installed a new thermostat rated for the same temperature. The rest of the transistors on the main PCB are the original metal can to-3's and each has been pulled and had its gain and leakage tested, all still well with in spec! It is also dead on when compared to my calibrated 5v and 1ma references. I just plunged my DMM in to it and twisted the knobs randomly to make sure, and with the knobs set to 0.0327 I get a reading of 327.8mV on my DMM. As far as noise go when tested on my siglent scope I see about 10mv of ripple but im pretty sure that noise is the scope I see the same ripple on my battery powered DMM check reference! When the psu id plugged in to an ancient HP scope with a .1mv scale which is only 100,000nVs! I see about .3mv of noise, that number is probably not dead on, the scope needs some love too but its definately to far off.
I had planned to replace the entire power cord along with adding a switch to break the hot side of the power cable, this way the oven could be turned off. The supply draws about 12watts in the off position whenever it decides to turn the heater on. Its not just the power draw that annoys me when the supply is off its the fact that it will prematurely wear out the thermostat turning the oven on and off when i might not even turn the actual PSU on for two weeks. For now I usually just leave the thing unplugged and plug it in 10 minutes to warm up before I need to use it.
So im thinking about selling it I love the thing, but im thinking about something a little more modern even though it probably wont last 50 years and appear to be in cal. If i do sell it im not sure if should do any more work on it or keep it as original as possible... i know people want cars with original parts, not sure why you'd want that with electronics though. The thing is I really want to but one of these hantek or korad "precision" power supply's. That probably sounds crazy but I REALLY want to be able to program it to run tests and log the data, it would also be nice to have more than 500mA, which is totally fine for most any work with analog ic's and micro's. But for instance im working on a DC soldering station project.. these newer psu's would have allowed me to just start building it, and then worry about the power circuits last.
My biggest hang up with getting rid of the Power Designs is it is basically a 0-20v voltage reference that I can trust when where talking about opamp circuits and low noise design where that nV per root hertz number is a big deal, im really in to RF and precision DC stuff. The Hantek claims under 1mV of ripple at a load less than 3 amps and im not sure if I can count on that or if its a garbage PR spec. Ideally id like to get rid of the PA and maybe buy two Hanteks but im wondering if I should keep the the PA and just but a hantek a little later. Im also open to other psu's as long as the are programmable and have an API available or published protocol and commands, just not sure what I would be looking for around $100 bucks or so.
This is the hantek psu i have been looking at
https://www.circuitspecialists.com/pps2116a-programmable-dc-power-supply.html?otaid=gpl&gclid=CjwKCAiA3uDwBRBFEiwA1VsajJC9D95XXFL0RGh3aKH8N9cC-7vzV96-Bf285RlRxLEkMYlxnUs1WRoCh94QAvD_BwE (https://www.circuitspecialists.com/pps2116a-programmable-dc-power-supply.html?otaid=gpl&gclid=CjwKCAiA3uDwBRBFEiwA1VsajJC9D95XXFL0RGh3aKH8N9cC-7vzV96-Bf285RlRxLEkMYlxnUs1WRoCh94QAvD_BwE)
Lemme know what you guys think especially if your the owner of the hantek and have done some noise and accuracy measurements on it.
-
Both Hantek and Korad are garbage by comparison. You'll want to aim a little higher if you plan to replace it with something decent.
What exactly are you looking for when you're saying more "modern"? An old supply like that would be suitable for a lot of stuff, as long as you take note of the value of the output capacitor.
-
like I said above I'm opened to used also, just not sure what to look for. Basically I want something that can be programed for testing and data logging hopefully with published protocol/command set. I'd like decent accuracy and low enough ripple for working with low noise analog and microwaves. The PD measures about 0.3mv ripple and the hantek specifies <1mv ripple under 3 amps, this is great if it's true. oh ya and I'd like 5amps, 500ma is fine for most things but sometimes you need more.
-
Greeting,
Please reply guys I need a good Bench Power Supply as well.
Thank you.
Yousif
-
Advice on maybe selling my Power Designs PD 2005A, and buying a hantek or korad? - DON'T, unless you can guarantee it goes to good home and gets used! You've put alot of time into the 2005 - and you'll regret it. The korad units i've seen are pants.
Power Designs stuff is achingly beautiful, satisfying to use and rarely bettered (drift/accuracy/noise). I have a 2005 from the 1960's and it is in-use most days. I get that remote control is useful but wouldn't bother with the cheap ones - worth holding out and getting an HP/Agilent, especially if you don't mind a fixer-upper.
-
I would recommend against selling it. I have a bunch of different modern power supplies (linear, and switch mode) but I find my go to supply is my PD4010. Even though I don't need the extra stability, or precision most of the time I use it, I still find the UI easier to use then my other supplies. Also it's just so much fun to use, everytime I use it I smile.
If you do sell it, I'd recommend seeing if anyone in the PD precision power supply thread wants them, that way it goes to a good home, because people love the pd supplies. I love them so much that I'm looking into getting a PD 6050 just so that it'll match with my PD4010
-
The PD is only one channel, while I often need at least two (for, e.g., bipolar supply for opamps). I'm slowly building my own power supply and I need 5 channels to test it. I rarely need more than 1A of power, 0.5A is enough 99% of the time for me. But when I need 1A+, there is no workaround. So for last four year my main unit is a DIY power supply with 2x15V 1A channels. I also have a few other DIY units for when I need more channels.
As of Korad, from what I've seen, they are good power supplies.
As of selling the PD, I have a similar dilemma. I wanted it badly, so I ordered 2020B from US. I played with it for 10 minutes, realized it needs some care and love and now it's collecting dust for more than one year. I also don't have enough space on my bench, and it needs a step down transformer as my unit is 110V. Don't know what to do... Probably I should either sell it, or restore and sell it, or restore and put on display :)
-
I wanted it badly, so I ordered 2020B from US. I played with it for 10 minutes, realized it needs some care and love and now it's collecting dust for more than one year. I also don't have enough space on my bench, and it needs a step down transformer as my unit is 110V. Don't know what to do... Probably I should either sell it, or restore and sell it, or restore and put on display :)
You can rewire/resolder the transformer primary to operate from 240V; put the two windings in series, not parallel. Remove 1-3 and 2-4, and add 2-3. Not a difficult job.
-
I would follow the specific instructions set out in the manual. If you power supply has a stability oven you want to make sure it's getting the correct voltage.
-
like I said above I'm opened to used also, just not sure what to look for. Basically I want something that can be programed for testing and data logging hopefully with published protocol/command set. I'd like decent accuracy and low enough ripple for working with low noise analog and microwaves. The PD measures about 0.3mv ripple and the hantek specifies <1mv ripple under 3 amps, this is great if it's true. oh ya and I'd like 5amps, 500ma is fine for most things but sometimes you need more.
Double what you want to spend however if it's low noise you want have a squiz at the SPD1000X models.
I posted some stuff in a thread about them trying to measure their noise but couldn't with a scope as it's down in the low 100's of uV and flaming hard to measure with a scope.
-
[..]
I get that remote control is useful but wouldn't bother with the cheap ones - worth holding out and getting an HP/Agilent, especially if you don't mind a fixer-upper.
It might be worth noting that the more-or-less affordable ones from HP/Agilent (I'm thinking of 6612C) have only a 12bit DAC (and have a fan), which might or might not suit your needs. Higher resolution DACs get into their SMUs, which are significantly more expensive even on the secondary market.
Quite a few old analogue precision PSUs have 'remote programming' inputs ...
-
Sorry I forgot about this post for a minute, but I decided I am going to keep the PD, I put a lot of effort in to restoring the thing, and seeing the noise output from it at the 100uv/division on my HP scope is just to satisfying to give up. I have a slim mini rack like aluminum enclosure that matches the PD look 100% so I think im going to use it to build an add on box that uses the PD's output as a reference in to some opamps and power transistors to make higher voltages with more current when needed and switch in a TLE2426 (http://www.ti.com/product/TLE2426) behind the new power stage when I need a bipolar supply.
As far as remote automation goes, I have read through the manual thoroughly and just can not figure out a way to do it using the remote/sense terminals on the back. To remotely set the voltage you need to connect a resistor to the remote terminals with a 1kohm value per 1 volt of output at the terminals, well there is no good way to do this digitally I can think of, digipots dont have enough resolution and using a fet as a resistance device doesn't seem to accurate. Literally the best way I can think to control the voltage through a computer/micro would be to motorize a a 20k 10 turn pot.... that seems a little ridiculous, or a relay controlled decade box, also ridiculous. Maybe there is a better way to solve the problem im just not seeing? Also the current setting has no remote control terminals or sense terminals so to control the current you would have to install a relay across the set button and once again digitally control the current setting pot somehow.
I figure if I want real automation for component testing etc, my best bet is to buy a korad or used higher end supply (if there are any recommendations on a good older model psu that can be programmed preferably via serial please list them). Another option that isnt a whole ton of work is to build a PSU based off this design http://tuxgraphics.org/electronics/201005/bench-power-supply-v3.shtml (http://tuxgraphics.org/electronics/201005/bench-power-supply-v3.shtml). Not only is it pretty easy to build this thing but the atmega8 code could easily be ported to an atmega328 running arduino so you could quickly do things like swap out the r2r ladder for a 16 bit dac along with using low noise transistors and stable voltage references, hell I could even make an option to use the super drift free output of the PD as the voltage supply's voltage reference. Still not sure if that option is better than a korad though I cant seem to find any body showing actual measurements of its output on a low noise instrument, the specs say <1mv output noise which is still pretty far from the PDs 200uV but perfectly acceptable for most anything, the only thing is I dont belive marketing until i see it interdependently tested.
-
Another option that isnt a whole ton of work is to build a PSU based off this design http://tuxgraphics.org/electronics/201005/bench-power-supply-v3.shtml (http://tuxgraphics.org/electronics/201005/bench-power-supply-v3.shtml). Not only is it pretty easy to build this thing but the atmega8 code could easily be ported to an atmega328 running arduino so you could quickly do things like swap out the r2r ladder for a 16 bit dac along with using low noise transistors and stable voltage references, hell I could even make an option to use the super drift free output of the PD as the voltage supply's voltage reference.
I would recommend against that design. The micro-controller is not just providing a reference voltage, it's output is directly coupled to the pass transistor. This means that the micro-controller is part of the feedback loop, and it is responsible for handling any input or output transient not handled by the capacitors. The micro-controller only samples at a few kS/s, and the software uses a very naive control algorithm that makes the supply a piss-poor slew rate, which translates in poor regulation, ripple rejection, etc. Especially at higher frequencies. Even 50Hz ripple is a challenge for it.
-
[..]
As far as remote automation goes, I have read through the manual thoroughly and just can not figure out a way to do it using the remote/sense terminals on the back. To remotely set the voltage you need to connect a resistor to the remote terminals with a 1kohm value per 1 volt of output at the terminals, well there is no good way to do this digitally I can think of, digipots dont have enough resolution and using a fet as a resistance device doesn't seem to accurate.
I have no first hand experience, but I'd use the FET as a current controller, while simultaneously measuring voltage across a precision shunt resistor in series with that FET and voltage across both. That way a resistance can be simulated. Isn't that what calibrators and active loads are doing?
Literally the best way I can think to control the voltage through a computer/micro would be to motorize a a 20k 10 turn pot.... that seems a little ridiculous, or a relay controlled decade box, also ridiculous. Maybe there is a better way to solve the problem im just not seeing?
I'm not suggesting that, but I read that such had been done in the past.
Also the current setting has no remote control terminals or sense terminals so to control the current you would have to install a relay across the set button and once again digitally control the current setting pot somehow.
[..]
Well, that's a pity.