Unique variable DC power supply for opinions / comments

[Bzonic]

Maybe someone can help guide me a bit on this power supply.  It is partially working in that it powers up but then the DC voltage meter goes right to being pegged at 30 volts.  I check the output terminals and no voltage is present however.  and voltage is present. The previous owner told me that the guy who custom built this ( a silent key )worked for Marconi and used “extra” parts to do it.  Obviously there is not a schematic available but I hope to start eliminating possible causes.  The first thing I’m going to replace is the LM317 volatage regulator chip as I suspect it may be bad.  I thought it was a transistor at first but the datasheet told the story on it.  There are a few more semiconductors on the front fold down panel that I might be replacing as well – a 2n2647 transistor and a ECG 5455.  Other than those two I’m left with the bigger transistors along the back and I’m really hoping those are not bad.  Anyway I wanted to put this out there to get a feel for how familiar others might be with general PS operation as I might need to substitute components if I find something gone bad.  This thing is built like a tank and I’d love to be able to restore full functionality.  Thanks everyone!

Brian

[singapol]

First be very careful around such a heavy duty power supply.The current meter can read up to
60amps. Don't just start replacing components...hope you know how to use a multimeter to measure ac and dc voltages.Checking the condition whether components are shorted or open.


There is a big brown bleeder resistor across the filter capacitors.Measure the cap voltage to ensure the voltage has bled to "zero" BEFORE you dive in to check. I would check the front panel first as the controls are there.Trace where the ac/dc power supply are being fed to these circuits. If you say the voltage meter is pegged at 30V then check the circuit for faults.Probably something has shorted and the protection circuit has kicked in. There is a component mounted on a angle bracket after the current meter to the right...maybe a SCR check that it's for protection for over voltage or current. If you are satisfied the front panel circuits are good then proceed to the main transformer and asociated parts. You may have to disassemble the power transistor heatsink to check or access them.
                                       Good luck and be SAFE.

[nctnico]

I'd check the solder joints on that vero board and everywhere else.

[eas]

I've started my self-education in electronics by troubleshooting/repairing bench power supplies. The ones I've worked on all have service manuals, with a principle of operation, block diagram, and schematic. I haven't worked on enough of them to have a good sense of all the different variants, but, it seems that there will always be output elements, often old-school transistors, perhaps sometimes derated MOSFETS. There will be a control stage that compares actual output voltage to the set voltage. There may be a driver stage which is basically another little PSU, between the control and pass elements. Given the amount of current this thing can deliver, I'd assume it has a driver stage. There may be another, smaller, regulated power supply for the control and driver stage, and an unregulated DC supply for the output. Other bits are something to drive and calibrate the meters, something to control the current, overvoltage protection (crowbar?).

So, if this were mine, I'd start trying to figure out what does what, probably by starting to work out a schematic, by working inward from inputs, outputs, meters and major power elements (transformers, beefy heat-sinked semiconductors).

My SWAG, based on the symptoms you have described, is that the meter is driven off a signal in the control or driver stage, but that there is something wrong with the driver stage, or the unregulated DC to the output stage. It goes to full-scale because there is no feedback from the output.

Misc observations/questions from the photos:

• It looks like this uses classic transistors for the output elements. I see 8 on the heatsinks. Are there more on the other side? It seems like there should be more than 8 for 40+ amps of output. Or maybe there are I bunch of power resistors too?  I think I spotted one nestled on the backside of a heatsink
• Some of the transistors may be part of the driver stage. Try and figure out if the collector/output of some transistors connect to bases of the others.
  
• It looks like there are thermostatic switches mounted on the heatsinks too (near the top). Do these control the fans? Do they somehow cut the output if the temps get too high?
• I see a beefy internal heatsink near the back, what's mounted on that? Driver transistor? Rectifier?
• Forward of the beefy internal heatsink, what's that I see?  Is that a fan?  Or something else?
• Looks like the beefy switch on the right of the front panel switches the output. The nearby button is perhaps a way to reset an output crowbar without power-cycling.
• There is a beefy transformer, this surely steps down the AC that gets rectified to unregulated DC to feed the output stage. It might also have accessory taps for other purposes. How many taps come out of it?
• I also see a smaller transformer that probably provides stepped-down AC that powers the control and maybe driver stages.

I don't know how helpful that will be, I'm interested to see what other people come up with.

[Bzonic]

Thanks very much for the responses thus far.  Lots of good stuff for me to think about.  Hopefully I will get lucky and find the issue on the front panel and avoid any major tear down.  I will definitely be careful.  Once I have the LM317 regulator and a few other easy / cheap parts installed I will pop back in and provide an update.  Hopefully in a few days I will have uncovered the issue.  Thanks again

[dr.diesel]

Wow what a beast!  Be more careful to not drop it on your foot, LMAO.

[Bzonic]

Indeed a beast and it feels like all of 60 pounds. I was certainly thinking of the handle construction when I was taking down my stairs ha.
BTW, I wish I could answer the questions from EAS but I'm pretty new to learning all this stuff so the guesses of members here are much better than mine.  I can always take some closer pics if it comes to it though. 

[eas]

Indeed a beast and it feels like all of 60 pounds. I was certainly thinking of the handle construction when I was taking down my stairs ha.
BTW, I wish I could answer the questions from EAS but I'm pretty new to learning all this stuff so the guesses of members here are much better than mine.  I can always take some closer pics if it comes to it though.

I'm not much further down the road than you are. All of my questions are things I'd do to try and better understand how this beast is put together without benefit of documentation. Most of what I'd do to answer the questions for myself, is to look at the markings on components, look them up to figure out what they do, and try and figure out what's connected to what by visual inspection and use of the continuity function on my DMM.

[singapol]

Thanks very much for the responses thus far.  Lots of good stuff for me to think about.  Hopefully I will get lucky and find the issue on the front panel and avoid any major tear down.  I will definitely be careful.  Once I have the LM317 regulator and a few other easy / cheap parts installed I will pop back in and provide an update.  Hopefully in a few days I will have uncovered the issue.  Thanks again

Don't just replace LM317 and other cheap components...it's not needed. Measure the input and output voltage of LM317,that will give an idea of the supply voltage to this regulator.Trace the output to where it supplies this output voltage...probably to control circuits.Check the feedback control loop that regulates the power or driver transistors. It could be a simple fault or a short of one or more of the power transistors.The temperature cutoff switches seldom fail but you never know. I have seldom encountered a bad LM317, they are quite robust. About the scr it could also be a low/high side switch interfacing control circuit with the high power circuit for isolation of low voltage circuitry or it could be a protection device in case of catastrophic failure For sure there will be protection circuitry for such a huge power supply other than simple fuses.Perhaps the temperature cutoff switch is already performing this function.A cheap way but not foolproof.

                                                                                                   All the best.

[dom0]

If the voltage meter just goes to <full scale> but no output is present it just might be an issue with the meter amplifier (often analog meters with, for example, 1 mA f.s. are used for everything and driven by meter amplifiers (~like a non-inverting amplifier with the first feedback resistor replaced by the meter).

[Bzonic]

I had some time last night so I took a more detailed look into some things.  I am actually getting output that matches what the front panel meter indicates although it still goes right to about 27volts upon power up.  I reflowed most the solder joints that were easily reached and I noticed the pot seemed flakey that should control the voltage level.  I noticed that I could adjust it up from 27 to 30ish volts but that was it - only those last few volts.  The meter did respond to this as well.  A bad pot I'm wondering?  I took it apart completely and cleaned the wipers etc.  Same result after that.  According to my ohm readings I believe it may be bad.  Looks like a 2 ganged 25k ohm pot and I'm wondering where the best place to order one.  Also, the big object on the center brace is a relay that does in fact click over upon power up.  Does this likely mean that all my output transistors are ok?  Here are some more pics as well as some of the areas I'm not familiar with.  I have not yet replaced any components btw.

[Kleinstein]

The curcuit does not look very complicated. So it might be an option to just draw a circuit plan, at leat an approximate.

As with a full voltage at the output, chances are, that some of the output transistors are burnd. This a more comon fault. So it may be worth measuring. This might be possible with most of parts in place. A dead transistor usually has a dead short emitter to collector.

[Bzonic]

Thanks Kleinstein - hopefully my probes can reach the collector and emitter to verify them.  Despite the pot seeming like a suspect the previous owner said it just stopped working one day instead of acting erratic like a bad pot might cause.  Oh well - too much speculation (on my part) at this point so I will continue with it tonight.

[edpalmer42]

It's likely that the four modules on the back are in parallel.  If so, you might be able to disconnect one module at a time to see if one if bad.  Depending on how the transistors on the modules are mounted (nuts and bolts vs bolts threaded into the heatsink and sockets vs soldered) it might be possible to remove the transistors from the back and test them out of circuit.

[Bzonic]

Yes - good points.  Luckily this thing is somewhat disassembly friendly. 

[Bzonic]

I lucked out and discovered that all the output transistors were socketed... after I had half the heat sink off. Oh well, good thing is that 1 of the 8 transistors tested bad so I hope that a replacement will get this thing operational again.  Time to order.

[edpalmer42]

I lucked out and discovered that all the output transistors were socketed... after I had half the heat sink off. Oh well, good thing is that 1 of the 8 transistors tested bad so I hope that a replacement will get this thing operational again.  Time to order.

Maybe, but a shorted transistor wouldn't explain why the meter was going full scale with no voltage on the output.  I was looking at the picture of the front panel.  There are more switches there than I was expecting.  Do you know what they're for?  Could one be to enable the output?

Ed

[eas]

I lucked out and discovered that all the output transistors were socketed... after I had half the heat sink off. Oh well, good thing is that 1 of the 8 transistors tested bad so I hope that a replacement will get this thing operational again.  Time to order.

Poking around is great, and starting with an obvious fault (lack of output) is a good place to start. However, there is another obvious fault, the meter going full scale in the absence of any output voltage. I think working that angle might be more fruitful because my hypothesis is that it is closer to the root cause(s). 

Also, as I said before, its worth trying to figure out how those transistors on the heatsink are wired. Is the output of the bad transistor connected to the output, or is it feeding some of the other transistors?

Maybe, but a shorted transistor wouldn't explain why the meter was going full scale with no voltage on the output.  I was looking at the picture of the front panel.  There are more switches there than I was expecting.  Do you know what they're for?  Could one be to enable the output?

Ed

The switch on the right seems to be wired in-line with the output.

A shorted output transistor could explain the max reading with no output, couldn't it?  A shorted output transistor would lead to an unregulated output voltage, right? And if the voltage goes too high, over-voltage protection would cut in, disabling the output, right?

[Bzonic]

Thx again.  I need to mention that I initially said that I was getting no output from the front terminals.  I was incorrect in stating that.  I am (and have been) getting nearly the full 30v output.  My overall knowledge of electronics is limited to what I been reading in various text books for the last 3 months.  Needless to say I still have a long way to go before I can wrap my brain around concepts/functionailty.

[edpalmer42]

Thx again.  I need to mention that I initially said that I was getting no output from the front terminals.  I was incorrect in stating that.  I am (and have been) getting nearly the full 30v output.  My overall knowledge of electronics is limited to what I been reading in various text books for the last 3 months.  Needless to say I still have a long way to go before I can wrap my brain around concepts/functionailty.

Okay, that makes more sense.  I wouldn't normally expect an over-voltage protection circuit with a variable power supply.  But be warned, working on a high power unit like this with your limited knowledge and no service information is just asking for trouble.  Be very careful!

One trick to limit the damage in a case like this is to wire a light bulb in series with the AC power to the unit.  It should only be drawing a small current with no load.  If something causes a large current surge, the light bulb will start glowing brightly.  This limits the current and hence, the damage.  Expect the bulb to be bright initially while the capacitors charge up.  You can try different size bulbs, but maybe a 60 or 100 watt bulb would be appropriate.  No CFL or LED bulbs!  Only a good, old-fashioned incandescent bulb.

Ed

[Bzonic]

Agreed Ed - I know this thing can bite me so I'm not really poking around willy nilly.  I made a dim bulb tester a while back for something else so yeah not a bad idea to use it. 

[Bzonic]

Overdue update to the power supply.  After identifying one bad output transistor I have it working now.... I think.  I now can get a better understanding the front panel.  There are two pots with one being a fine adjustment for voltage.  There are 4 switches total and it seems like some redundancy as well.  The far right switches the output on and the others seem to somehow be linked but I don't know in what way.  The main question I have now is why I'm only able to turn this thing up to 20 volts output.  Because this was basically homemade I'm wondering if the guy never intended it to go to 30 volts out but used a 30v meter because it is what he had on hand?  The output was initially pegged with the bad transistor at about 28 volts so that makes me wonder though.  Swapping the new LM317 voltage regulator with the new one makes no difference - both seem to work.  Also, I can not adjust it down below about 4 volts when it's on.  A working range of 4v to 20v will certainly work for nearly every thing I need so I'm posting this more as a way to learn than anything else.  Thanks for any info!!!

[edpalmer42]

I want to see the load test!!   

Seriously, without a schematic, it's almost impossible to say whether the 8 volt drop is reasonable.  It doesn't sound outrageous.  Monitor the voltage on the main capacitors with a 'scope as you crank up the load.  You'll see the typical sawtooth pattern emerge.  The lowest point of the sawtooth has to be high enough above the output voltage that the regulator can still function.  Note that without a schematic we don't know how much voltage the regulator needs.  The worst case will be with maximum output voltage since that's the closest to the input voltage.  Lower output voltages should have no problem.

The other serious test is thermal dissipation.  Set the output for the lowest voltage and then pull some current.  In this situation, most of the power being dissipated is in the supply itself.  Do a few tests working from low currents up to the maximum the unit can provide - if you can figure out a load that can take that!  Monitor everything carefully to look for parts overheating.  If it puts out 40 amps at 4 volts, the worst case internal dissipation will be something like 40 amps times maybe 15 - 20 volts for somewhere in the 700 watt range.  That's when you'll need those big heat sinks!

Ed

P.S.  that big output switch looks like it could also be a circuit breaker.  Can you see any ratings on it?  That would give you an idea what the rated output current is.  The builder might have also used whatever ammeter he happened to have on hand.

[Bzonic]

I'm following what you describe and would like to test it but I'm trying to figure out how I rig up a variable amp draw load.  Is there some obvious method or device I can hook up to simulate large amp flow? 

I did not look closely at that output switch to see if it has any integrated breaker - but will do so later.  The output terminal has a snap in piece with the two leads and there is an inline fuse holder on the positive.  Looks like you can stick a car fuse in there if you want.  Not sure if it was designed for a fuse to be used at all times or if it is just a safe guard that one can use when desired. 

[edpalmer42]

I'm following what you describe and would like to test it but I'm trying to figure out how I rig up a variable amp draw load.  Is there some obvious method or device I can hook up to simulate large amp flow? 

I did not look closely at that output switch to see if it has any integrated breaker - but will do so later.  The output terminal has a snap in piece with the two leads and there is an inline fuse holder on the positive.  Looks like you can stick a car fuse in there if you want.  Not sure if it was designed for a fuse to be used at all times or if it is just a safe guard that one can use when desired.

The closest thing I can think of for a load would be an automotive battery tester that was rated for 24V.  I think mine uses a couple of pieces of steel wire in parallel to get a 50A load at 12V.  Maybe something like a wire coat hanger?  Remember that it's covered with some kind of coating so it'll burn and smoke when you heat it up!

The fuse holder was probably to protect smaller loads.  Plug in whatever fuse was appropriate for the load.

Ed