another bench supply challenge for you

Discussion in 'General Electronics Chat' started by sleepyjz, Jul 23, 2012.

  1. sleepyjz

    Thread Starter New Member

    Jul 23, 2012
    6
    0
    Hi just new to the forum. I hope my first post doesn't qualify for a TL;DR, but I am wondering if someone can help me with this problem which has been slowly eating away at me for weeks. I have a 30v, 0-12 A linear DC benchtop, toroid based power supply whose opamp and voltage reference was blown by a short when the PSU was delivering full current. The first sign after this was the supply delivering a controllable but unregulated ~ 50 v. The op amp and various components were replaced (the 7812ct, both 1458M opamps, all 6.2 and 12v zener's, LM336 voltage reference, TIP31A, all 6 heatsinked 2n3055's, both 2n5551's which control the 2 relays). The supply is automatic switching with relays activated at 10v and 20v. The LM336-2.5 is a 3 pin voltage regulator and is acting as a zener diode with the "-" and "adj" pins joined together, which I think is correct. The problem is this: A sign that may help indicate whats wrong is that the green constant voltage LED flickers when dialing the 10-turn voltage potentiometer. The red current limiting LED turns on when the current-limiting control pot is turned down all the way, which is normal, but the voltage needle bumps up a bit when the constant voltage is activated, by dialing up the current limiting (where the CV takes over with no load). A slight negative voltage is present before the bump-up. The main thing is, the 7812CT voltage regulator (which is heatsinked and thermal pasted) and the 2 2n5551's on the relay part of it get extremely hot even without a load, especially when setting the voltages to the higher limits of the PSU. They would likely fail within 10 minutes (get to about 70 C, didn't allow it to go further). No shorts as far as I know, and no resistors are noticeably charred, and I've tested the ones that I think receive a heavier current - they are working properly. Any help would be more than appreciated, this is kind of my last resort.
    many thanks,

    - sleepyjz

    PS: Circuit diagram is attached. The X,Y and Z terminal are for remote programming but are shorted, as recommended, by shorting bars.
     
    Last edited: Jul 23, 2012
  2. crutschow

    Expert

    Mar 14, 2008
    13,000
    3,229
    So it would appear the relay part of the circuit is drawing excessive current somewhere. Check the relay coil load for the transistors (the diode across the coil for example) to make sure everything is OK there.
     
  3. Dodgydave

    Distinguished Member

    Jun 22, 2012
    4,974
    744
    Looking at the circuit, i would remove ic201 which is the 7812 regulator feeding the relay pcbs, this should give you a variable voltage upto 10volts with variable current limit.

    If that works replace the 7812 and check the output from the 7812 regulator, then the problem lies on the relay pcbs.
     
  4. sleepyjz

    Thread Starter New Member

    Jul 23, 2012
    6
    0
    Hmm I have replaced at least 2 7812's before but never considered running it without. It seems like the only thing giving me hell is the relay PCB's. The 6.2 v relay zeners fail often as well.

    Thanks will try, I don't think I have replaced those D203, or D204, though I might remember them testing alright.

    I see the inside of this power supply when I close my eyes, it's ridiculous. Like some circuit board tetris effect :eek:
     
  5. grahamed

    Member

    Jul 23, 2012
    99
    11
    As has been said there is excess current through the relay drive transistors. Now the circuit operation is that the transistor should be either off or fully on, neither should lead to excess current.

    In the off condition (i.e. low output voltage) there should be essentially zero current coming through the 78L12 and in the on condition the current should be limited by the collector loads (i.e. the relay coils). Looking at the collector loads the only other current path is is through the (reverse biassed) diode. So check the the resistance from the 12V rail to the collector - it should be in the high tens to the low hundreds of ohms (depending on the actual size of the relay - look for something similar in a catalog). Lift one end of the diode and check its resistance (OK its voltage drop) when measured both ways; one way open circuit and one way a resistance - depends on your meter. Basically compare with almost any known good diode. If necessary replace (with a higher voltage rated) 1N400X diode.

    I am sorry if this comes across as a bit simplistic - I have no idea of your level of knowledge - but you seem to be struggling a bit.

    Tell us what you find!
     
  6. sleepyjz

    Thread Starter New Member

    Jul 23, 2012
    6
    0
    OK so tested all diodes on the relay PCB and they seem fine, replaced D203 and 204 just incase (all 1n4002). The 2 relay coils resistances differed by a couple of ohms, but were both around 80 Ω (@ grahamed & crutschow) . I cleaned and de-tarnished the terminals of the switch mechanism of the relay. Unfortunately the problem is still there. Does the constant voltage LED issue indicate something? it brightens, dims, and flickers when dialing the voltage pot up or down. Also, I have to ignore the markings on the PCB for the 2n5551s, the "E", "B", and "C" do not line up with the 2n5551's pinout. When I took the 7812 out, I could actually get near 20v adjustable (so not 10v @ dodgydave) out of the supply. I measured the voltage at the input, output and ground terminals for the 7812 on the pcb and got almost 28 v. The LED flickered just as much without the IC. Any other suggestions? Thanks for the help, all, I appreciate it.
     
  7. grahamed

    Member

    Jul 23, 2012
    99
    11
    Hi again

    Oh dear, not going well is it! Do not despair - yet.

    In general my advice is to break down the circuit into modules - as cruthschow says the rest of the circuit will work - in a limited fashion without the relay drive circuit - and I would say that the relay drive circuit can work without the rest of the circuit.

    So isolate the parts - the only connections are the output voltage being fed through R203 and its counterpart in R2 drive. I would lift the top end of the resistors and connect them to the 12V rail. Take out the relays and replace the coils with a 100ohm resistor, lift one end of D0101/2 to depower the circuit completely.

    The relay circuit should now be stable with the no excess current with the relay drivers either on or off. If it is not then there must be a problem within this limited area as the rest is not connected.

    If there is still excess current then you must find it. Obviously it is coming through 7812 and the drivers. It should not matter whether the driver is on or off as the 100ohm will limit the current to 120mA or so, which is . If the current is not so limited then there is another path from the collector to the rail. Short out the base-emitter. Pull up the collector. Remove the transistor completely. Lift the diode. Find that current.

    The circuit could be oscillating but in general that should not make the current high in this kind of limited current circuit.

    The voltage at the centre terminal IC201 should be the same as the negative output terminal - which I assume is your reference - it should be wired to it! Check again, check resistance resistance to the various 0V points in the circuit.

    All this takes longer to type than do!
     
  8. sleepyjz

    Thread Starter New Member

    Jul 23, 2012
    6
    0
    Thanks for the suggestions while more zener diodes and other parts are on their way. It's just frustrating that someone improvised with the parts inside of it, and it doesn't line up with the schematic fully. Could it actually be the torroid itself? I doubt it but don't know for sure. You mentioned oscillating. There was a small MP capacitor across the AC terminals of the bridge rectifier that wasn't in the diagram (it was for sure across the AC's not DC's, what would this accomplish?). This did not survive the bridge rectifier replacement :(. Iv'e checked the 2n3055's multiple times and they are OK. all 1n4002 diodes and all resistors are good too. Zener's were bad but as I said they are being shipped now. I'll test the other suggestions when I get some parts. I'm not trying to push it, but as I have said a couple of times, the constant voltage LED flickers when dialling the voltage pot up or down. What would cause this instability?
    Thanks for the replies everyone.
     
  9. bountyhunter

    Well-Known Member

    Sep 7, 2009
    2,498
    507
    It indicates to me the voltage control loop is oscillating or not working correctly.

    This design uses a very dangerous method of voltage control: the power pass transistors are run "wide open" and the voltage control amp has to "steal" their drive current to control them. I would never use that method since it means if something goes wrong in the control loop, the output is always going to slam up to 50V.

    The CV amp shown in the schematic does not appear to have any compensation, there are no R-C components feeding from the output back to the inverting input (as the CC amp does). I have no idea how it is supposed to be stable running uncompensated.
     
  10. sleepyjz

    Thread Starter New Member

    Jul 23, 2012
    6
    0
    Yeah to be honest this whole supply lasted 2 weeks, and I considered it "working" then. I have spent way to much time on this and it's not worthwhile anymore. Do you have any recommendations for stable lab PSU's, with somewhat similar current and voltage ranges (0-30v or higher, 12 amps w/limiting)?
    Many thanks,

    - sleepyjz
     
    Last edited: Aug 16, 2012
  11. bountyhunter

    Well-Known Member

    Sep 7, 2009
    2,498
    507
    Actually, if you are into learning, a blown power supply is a good project since the hard part is done: you have the chassis with meters and heatsinks and you can install a new controller board. But getting 360W is more than you will practically get from a linear supply.

    If you are looking for a bench PS, ebay has some that are pretty cheap but probably made in China so you take your chances. Ebay also sells used units that are from name brand companies.

    The best way is probably to buy a H-P supply or other good brand used from a used equipment dealer. There are lots of them on the net. The one you have looks like a standard "rack mount" unit and there are thousands of that type to choose from of various height and weight and cost, up to 5kW.

    At Power Ten we made some nice 19" rack supplies about 2" high that put out about 1kW with many choices of output voltage and current ranges. You might find some used ones.

    http://www.capovani.com/search_results.cfm?mfr=Power Ten Inc


    http://www.teknetelectronics.com/Search.asp?p_Var1=Power Ten Power Supply (DC)


    http://www.bizitestequipment.com/used-power-supplies.asp

    http://www.testequity.com/categories/DC+Power+Supply/?filter=used

    http://www.labx.com/v2/newad.cfm?CatID=149

    http://www.metrictest.com/catalog/views/used_power_supplies.jsp

    http://www.ebay.com/sch/Power-Supplies-/50967/i.html
     
    Last edited: Aug 16, 2012
    sleepyjz likes this.
  12. sleepyjz

    Thread Starter New Member

    Jul 23, 2012
    6
    0
    very true - I just finished adding components to a pcb for a lm723 based supply. I'll use the 2n3055's and the TIP31 (instead of the BD139). I am hoping the heat dissipation of the unit will suffice. I will most likely keep the relay pcb instead of hooking everything up on a single tap on the torroid. We will see how that goes. Do you know if I can keep the 68000 μf capacitors with this circuit? Would that be a bad idea? there are about 5 of them in parallel. Also, what are the benefits of using more 2n3055's (e.g., all six of them instead of four as seen in the schematic) ?

    [​IMG]

    Thanks for the suggestion and the links. For the time being, I am using two ATX supplies in series to power things.
     
Loading...