# electro[epsco] power supply

#### nicklus

Joined Sep 11, 2017
8
Hello everyone,

I have a question concerning an old Epsco power supply. It's a model "NFB' is all I can find on it. It's 32 volt 15 Amp and someone has robbed the rheostat from it. I have called the company three times and e-mailed them once to get some information on it and they have not responded.

I would like to know the value of the rheostat to see if I can find one of appropriate size. There seems to be no information on the web that I can find. Any help is greatly appreciated as I am not electronically educated enough to maybe do the math and figure it out.

Joined Feb 20, 2016
3,253
The common thread on the net is these are unregulated very expensive boxes.
Grossly over priced. It may be worth junking to use parts for a real power supply but do nor try to run anything electronic on these.
I could not locate a circuit but by what I read, it should be pretty easy to trace the circuit out. That may help figure out the pot. These look to be used for electroplating, and are just a power "light dimmer" style control with no feedback. Very easy to blow something up with!

#### nicklus

Joined Sep 11, 2017
8

The common thread on the net is these are unregulated very expensive boxes.
Grossly over priced. It may be worth junking to use parts for a real power supply but do nor try to run anything electronic on these.
I could not locate a circuit but by what I read, it should be pretty easy to trace the circuit out. That may help figure out the pot. These look to be used for electroplating, and are just a power "light dimmer" style control with no feedback. Very easy to blow something up with!
It's not to be used for that. Mostly to be used for DC electrical projects. Definitely not for electronics. I think it's basically a oversized battery charger..... Two large transformers, a couple caps, bridge rectifier and should have the rheostat.
James

Joined Mar 10, 2018
3,781
If the rheostat looks something like -

The size is indicative of power handling in Watts. You could use an
ohmeter to measure it, unsolder one terminal, right or left, and measure
fdrom that terminal to the other terminal on opposite side. Not the center
terminal. Calibrate you meter leads shorted together and subtract that
value from value you measure between outer terminals on rheostat.

Then go shopping for value and ~ size. One other dimension would
be panel hole size for the shaft to consider.

If its an auto-transformer then current and size and line voltage ratings
will matter -

Regards, Dana.

#### ebp

Joined Feb 8, 2018
2,332
I think it is more likely to be a variable autotransformer ("Variac") than a rheostat.

#### nicklus

Joined Sep 11, 2017
8
If the rheostat looks something like -
That's the problem. I don't have the rheostat [auto transformer] as someone has robbed it from the chassis.

The size is indicative of power handling in Watts. You could use an
ohmeter to measure it, unsolder one terminal, right or left, and measure
fdrom that terminal to the other terminal on opposite side. Not the center
terminal. Calibrate you meter leads shorted together and subtract that
value from value you measure between outer terminals on rheostat.

Then go shopping for value and ~ size. One other dimension would
be panel hole size for the shaft to consider.

If its an auto-transformer then current and size and line voltage ratings
will matter -

Regards, Dana.

#### MisterBill2

Joined Jan 23, 2018
4,843
To discover which it is, rheostat or variable transformer, you will need to do a bit of circuit tracing, If it is on the AC power side of the rectifiers then it is a variable voltage transformer, while if it is past the rectifiers then it is probably a rheostat. The simple way to do this tracing is to connect an ohm meter to the AC line input connection, and then, with the device power switch set to ON, check for some degree of continuity between the line cord pins and the wires that were connected to the missing part. If there is any degree of continuity, up to about 1200 ohmns or so, then it was an AC variable transformer.

#### KMoffett

Joined Dec 19, 2007
2,784
Have you tried contacting the company? http://www.epscoinc.com/power_supplies.html

I have the 32V/30A version...all 113Lbs. I was in there once. As I remember, it was a variable auto-transformer (Variac) on the primary side of a step down transformer.

Ken

#### MisterBill2

Joined Jan 23, 2018
4,843
Based on the three bolts showing on the front panel in the photo, I would also think that it was either a Variac or a Powerstat, or some similar device. Those devices are variable transformers. So without them you have a fixed voltage power supply. Still potentially useful but not nearly as flexible in applications. Can you please post a photo of the inside of the device, and the back side of that panel?

#### nicklus

Joined Sep 11, 2017
8
Based on the three bolts showing on the front panel in the photo, I would also think that it was either a Variac or a Powerstat, or some similar device. Those devices are variable transformers. So without them you have a fixed voltage power supply. Still potentially useful but not nearly as flexible in applications. Can you please post a photo of the inside of the device, and the back side of that panel?
The power supply is at a friends house right now. I will see him in a few days and take a photo. Thank You

#### nicklus

Joined Sep 11, 2017
8
Have you tried contacting the company? http://www.epscoinc.com/power_supplies.html

I have the 32V/30A version...all 113Lbs. I was in there once. As I remember, it was a variable auto-transformer (Variac) on the primary side of a step down transformer.

Ken
I think they are basically the same. Yours is just a little heavier duty.

#### -live wire-

Joined Dec 22, 2017
901
I have a deep, burning hatred for old analog power supplies and measurement gauges. As well as that kind of analog design in general. There are often poor or no feedback systems, inaccurate readings, generally bad designs safety and efficiency wise, and they just look so outdated. You could, with $10-20 of common semiconductors and passive components, modify it to be much safer, accurate, better looking, and efficient. #### ebp Joined Feb 8, 2018 2,332 "You could, with$10-20 of common semiconductors and passive components, modify it to be much safer, accurate, better looking, and efficient."

No. Just flat out no.

#### KMoffett

Joined Dec 19, 2007
2,784
Well -live wire-, they do what they do. Right now mine is powering two high performance race car 12V radiator fans connected in series for 30V/15A for a vertical wind tunnel. Yes, I also have an switch mode, constant voltage, constant current, 30VDC/20A supply. But that's overkill for this job. Sometimes heavy metal is all you need...especially if you already have it.

Ken

#### -live wire-

Joined Dec 22, 2017
901
"You could, with $10-20 of common semiconductors and passive components, modify it to be much safer, accurate, better looking, and efficient." No. Just flat out no. WRONG. For CV you need a$1-2 IGBT or mosfet, 2-3 dollars of caps, a common PWM IC for a dollar or two (similar to the LM2587), and some a feedback resistor divider and shunt. You just need some potentiometers and very cheap logic chips for current limiting and setting the voltage. It could also work for constant current. It is the same principal, actually, as in this video. It would just need higher value components.

#### ebp

Joined Feb 8, 2018
2,332
I have designed numerous industrial power supplies, including controllers for very high power cathodic protection systems that operate in constant current, constant voltage or constant potential (regulating to an electrochemical reference half-cell). I've done solar-input switchers for charging military lithium ion batteries. I've done numerous buck and boost converters for assorted industrial applications. I've done a controller for a three-phase variable frequency permanent magnet generator at a couple of kilowatts. I know where the poles and zeros need to go. I know about right half-plane zero issues in certain classes of switchers. I know how to design switching transformers and the power inductors. I know how to read safe operating area curves for semis and the issues involved in paralleling them. And I know you can't take a diddly little board and hook up a few bits to it and expect to make a regulated CC/CV power supply at tens of amps and tens of volts. Oh, and I know how to solder.

#### -live wire-

Joined Dec 22, 2017
901
I am talking about for maybe 30V 10A max. Not for anything greater. And I am not suggesting that such a board can be used for higher power applications, but rather that you could use similar components and use higher values and still have minimal expenses but good regulation.

#### ebp

Joined Feb 8, 2018
2,332
Well, I watched the video. He has some vague notion of what is going on. Sort of.

One of the scope waveforms shows the unmodified regulator to be either unstable or have very bad dynamics. His description of the the fur on the voltage waveform is wrong - it is either switching noise or improper frequency compensation; the image wasn't good enough to be certain of anything other than it was bad (my guess is that it is switching noise made to look worse because of bad probing technique). He made no mention whatsoever of the frequency compensation for the switcher, which is critical for stability and decent dynamic performance.

The add-on current limit circuit is a hysteritic ("bang-bang") circuit. It works, but is very crude. Now I know how all the cheap switchers boards on ebay do current control. You would never find that in a good design for a bench supply. IWatching numbers on a digital meter says very little about performance other than that an average is "right", but it does usually impress those who don't understand the issues.

The circuit he was using is not the type he described but most likely a SEPIC which is the offspring of the wedding of a boost converter and and an inverting (flyback) converter. They are very useful in that the output voltage can be higher or lower than the input voltage without change of sign but they do not easily scale for high power because 100% of the energy between input and output is coupled through an capacitor that is worked very hard and must have very high ripple current rating for any hope of reasonable reliability.

#### MisterBill2

Joined Jan 23, 2018
4,843
What ebp did not mention is that getting all of those things right for a switching power supply is not a simple task! While the basic principles are fairly simple the actual details are quite complex. The fact is that in a switching power supply with any sort of adequate performance the whole system includes all of those nasty unanticipated reactances, both series inductance and resistance, and the capacitance between things close to each other, as well as the actual characteristics of the semiconductor devices. A whole lot of the success comes from some deeper insights, which is why I choose to use switching supplies from trustworthy sources.
A simple variable supply like the one in this discussion does not even need any regulation to serve the intended purposes. There are still lots of applications that don't require. tight regulation.

#### -live wire-

Joined Dec 22, 2017
901
Well, I watched the video. He has some vague notion of what is going on. Sort of.

One of the scope waveforms shows the unmodified regulator to be either unstable or have very bad dynamics. His description of the the fur on the voltage waveform is wrong - it is either switching noise or improper frequency compensation; the image wasn't good enough to be certain of anything other than it was bad (my guess is that it is switching noise made to look worse because of bad probing technique). He made no mention whatsoever of the frequency compensation for the switcher, which is critical for stability and decent dynamic performance.

The add-on current limit circuit is a hysteritic ("bang-bang") circuit. It works, but is very crude. Now I know how all the cheap switchers boards on ebay do current control. You would never find that in a good design for a bench supply. IWatching numbers on a digital meter says very little about performance other than that an average is "right", but it does usually impress those who don't understand the issues.

The circuit he was using is not the type he described but most likely a SEPIC which is the offspring of the wedding of a boost converter and and an inverting (flyback) converter. They are very useful in that the output voltage can be higher or lower than the input voltage without change of sign but they do not easily scale for high power because 100% of the energy between input and output is coupled through an capacitor that is worked very hard and must have very high ripple current rating for any hope of reasonable reliability.
I understand that there are many limitations and flaws with that specific design (ripple, poor response time, etc.), but it would still be accurate enough for most situations. And it was to demonstrate that you can use digital ICs and a few other components to cheaply achieve a decent degree of regulation for medium power applications. You also could use a cheap, high clock speed (100MHz-1+GHz) microcontroller with PID or other more intelligent control. This would certainly help with regulation. And you can use IGBTs for the high power switching that could then go to a filter and use BJTs with a varying voltage applied to get reliable constant current.