No trigger waveform to an SCR controlled transformer primary issue.

[oldway]

I have now tried putting 3 known good boards in and whist it now works the max V out is still about 35 to 38V.

That's mean that this board (control of the pass transistors) is also NOT the cause of the problem....

Look for the connections ...(bad contact, loose or wrong  connection)
On schematics, it show that there are different connections possible for the voltage and current adjust.....there is probably something wrong with this...

[Armadillo]

Armadillo, thanks, I have checked MT2 outputs and they appear fine, as do the outputs from the "Supp;ies Board" after rectification.

The transformer I wrote is Transformer MT1, not MT2.

[Chris Wilson]

Armadillo: Oops, sorry, I can't find any issues with it, it's never been got hot, and never been got wet, but I will bypass the SCR's and isolate the pass transistors and check it can create the goods, thanks

[Chris Wilson]

I have now tried putting 3 known good boards in and whist it now works the max V out is still about 35 to 38V.

That's mean that this board (control of the pass transistors) is also NOT the cause of the problem....

Look for the connections ...(bad contact, loose or wrong  connection)
On schematics, it show that there are different connections possible for the voltage and current adjust.....there is probably something wrong with this...

The selection is made by eyelet ended jumper wires, they have NEVER been swapped since I got the supply and from when it has worked fine, and I have (several times) checked continuity between whichever letter terminals they are supposed to bridge in what the manual calls "normal operation". But I'll check again, I am obviously missing *SOMETHING*!  There is a diode shown near them, but I think it's irrelevant in "normal operation" but will double check. Any idea what D23 is actually for across the pass transistors? It seems fine but I am curious as to what it does. Thanks.


I am seeing what to me are odd voltages measuring between the base of the first driver transistor for the pass block, and its collector. Please clarify where I should measure voltages at max and min volts for:

A: The first and second driver transistors (which I have changed from 2N3055 to MJ15003)

B: The pass transistor block itself where bases, collectors and emitters are all effectively strapped together.

I want to be sure I am measuring voltages getting to there correctly. Thanks!

[Armadillo]

Armadillo: Oops, sorry, I can't find any issues with it, it's never been got hot, and never been got wet, but I will bypass the SCR's and isolate the pass transistors and check it can create the goods, thanks

10,000 is to 1, you never know. But worth a try.

If the transformer is able to supply 64Vrms [*loaded], then Do check the SCR also then check firing at wrong angles etc.

[Armadillo]

I am seeing what to me are odd voltages measuring between the base of the first driver transistor for the pass block, and its collector. Please clarify where I should measure voltages at max and min volts for:

A: The first and second driver transistors (which I have changed from 2N3055 to MJ15003)

B: The pass transistor block itself where bases, collectors and emitters are all effectively strapped together.

I want to be sure I am measuring voltages getting to there correctly. Thanks!

Can you picture your installation and soldering of the series pass transistors board particularly the mica insulation and spacers etc... More eyes easy to spot for you.

[Chris Wilson]

Mechanically it's a big job gaining access to the pass transistors, but I have some photos of the tunnel heat sink when I had it apart ages ago. The only thing I can think of that might be suspect in the soldering / mounting is if the white heat sink compound has got on the two pins and bridged them, but i don't think it's conductive is it? I am hoping not to have to go in there again, but of course will if needs must. The bases, emitters and collectors are all strapped together with hefty brass strips, and if I put my Peak DCA55 semi conductor tester on the three strips it says its a transistor and all looks well. I think if I have any doubts it might be the two external driver transistors, I need to triple check they are wired up 100% soundly and correctly. I will post voltages I see on the bases later, when I am sure if I need to measure base to emitter or base to collector. I am thinking base to collector.

I think it's just because it's a huge brute of a thing, but 6 times out of ten when I turn it on it trips the mains RCD, and sometimes it will when I turn it off, but a little bell says the SCR's are on the mains input side.... The smaller (25 Amp instead of 50 Amp @ 60V) sister supply never does that though.... Thanks Armadillo.

[Armadillo]

Armadillo: Oops, sorry, I can't find any issues with it, it's never been got hot, and never been got wet, but I will bypass the SCR's and isolate the pass transistors and check it can create the goods, thanks

10,000 is to 1, you never know. But worth a try.

If the transformer is able to supply 64Vrms [*loaded], then Do check the SCR also then check firing at wrong angles etc.

Oops! Sorry Chris, I missed your first post where you have swop the T1 transformer. Thanks for your patience.   

[oldway]

The first thing we would succeed to do is to let this power supply work correctly with the 3 known good working boards ....Later, we will worry with the failure of the other boards.

One thing we don't know is if these boards are exactly the same or not and if the components have the same values....so long I know, those boards had a lot of modifications, that's possible that output limit to 28V would be only a problem of values of components on the control board of the pass transistors.

D23 is only a protection of pass transistors against reverse polarities as it could happen if you connect a battery to the power supply switched off.

I would like to know if below 28V, the power supply is working properly with the boards known as good.

Does the current limitation work?

Is the voltage well regulated?

Does the voltage drop on the pass transistors vary greatly when the voltage is set at 5, 10, 15V and with different loads?

In short, is the 28V max output voltage the only fault?

[Armadillo]

OK I took the time to study the circuit.

I agree with Ian, VT4 and VT5 should be checked.

You need a digital logic probe or a oscilloscope.

Please confirm as per the diagram attached.

You need to answer "Is there pulses, is it alive?".

Vbe voltage is not enough to answer the question. thanks.

[Chris Wilson]

OK I took the time to study the circuit.

I agree with Ian, VT4 and VT5 should be checked.

You need a digital logic probe or a oscilloscope.

Please confirm as per the diagram attached.

You need to answer "Is there pulses, is it alive?".

Vbe voltage is not enough to answer the question. thanks.

With the three known good boards and it set to it's max of 25 to 28V out, no load, I took these scop shots using a X10 probe. One is pin 7 of IC2 another is on the collector of VT4, and finally on the collector of VT5. All used pin 13 of the SCR board as the scope ground.

Let me know if more scoping is needed, thanks you all

[Armadillo]

So there should be pulsing across the T1 at pin 6, 7 and 9 and 10.

Edit: Additional scope points attached. If you adjust P2, the duty cycle increase and the rectified voltage should follow to increase.

[Chris Wilson]

The first thing we would succeed to do is to let this power supply work correctly with the 3 known good working boards ....Later, we will worry with the failure of the other boards.

One thing we don't know is if these boards are exactly the same or not and if the components have the same values....so long I know, those boards had a lot of modifications, that's possible that output limit to 28V would be only a problem of values of components on the control board of the pass transistors.

D23 is only a protection of pass transistors against reverse polarities as it could happen if you connect a battery to the power supply switched off.

I would like to know if below 28V, the power supply is working properly with the boards known as good.

Does the current limitation work?

Is the voltage well regulated?

Does the voltage drop on the pass transistors vary greatly when the voltage is set at 5, 10, 15V and with different loads?

In short, is the 28V max output voltage the only fault?

The control amplifier board that seems to be the one that is faulty is the same part number as the one I replaced it with from the smaller 25 Amp supply. The schematic shows no alternative component values. Similarly the "Supplies Board" is identical. The SCR control board is borrowed from another 50 Amp supply and I cannot see any different component values on it, but the SCR and the Control Amplifiers boards do have some preset adjustment pots. The one on the Control Amplifiers board that should raise the max voltage level is already nearly on full, but that board in the 25 A supply it's from allows over 60V. I agree this is a bit hit and miss. The other thing is if I turn the supply on with the 3 good boards in it, and the volts on max, there's a pause of three or more seconds where a very low voltage shows on the V meter, then it suddenly jumps up to its max of about 25V.


I have just checked the supply with two off 1 Ohm 100W metal clad resistors in series as the load. Things are *NOT* normal. This is what occurs:

V and A knobs fully counter clockwise, no V or A show on the meters. I advance the V coarse know clockwise, leaving the A knobs at zero. No volts show. Leaving the V knob or knobs advanced a little, I rotate the coarse A knob clockwise. Current shows and so does V on the meters, however current follows the V knob, the A knob has no effect. If I rotate the A knob fully counter clockwise the supply should not output and V, but it still does and the current showing on the meter does not reduce.

Volts across the terminals 6 and 10 on the Control Amp Board change with the V control knob in so far as they reduce as V is increased. At the max attainable of about 25V I see about 6.5V, and as I reduce the V knob back to zero the V across 6 and 10 rises to about 40V and then decays, fairly fast at first, but slowing unti after about 30 seconds or so I see a resting V across pins 6 and 10 of about 17.2V  That seems very odd... I am hoping it means more to you than I

Thank you!

[Armadillo]

The SCR control board is borrowed from another 50 Amp supply and I cannot see any different component values on it, but the SCR and the Control Amplifiers boards do have some preset adjustment pots.

The duty cycle determine the firing of the SCRs and hence the rectified voltage hence the duty cycle is important. If there is no P2 pot on the SCR board, then it is not the same board and those bias setting will be different also.

it would be better to put back all the original boards and troubleshoot from there then to mixed and match with wrong boards causing confusion.

[Chris Wilson]

Armadillo:

Scope shots attached, Pin5 IC1 is the same point as you wanted to see if the saw tooth waveform is.

Note I change V scaling between captures.

The files are too big for the forum to accept as all attachments so I have zipped them up and put them on my server as http://www.gatesgarth.com/armadillo.zip  I think the file names should be descriptive enough, thanks. I will now adjust the control and see what hapens

[Chris Wilson]

So there should be pulsing across the T1 at pin 6, 7 and 9 and 10.

Edit: Additional scope points attached. If you adjust P2, the duty cycle increase and the rectified voltage should follow to increase.

Adjusting P2 appears to make no difference to the rectified voltage at any given V out level.

[Armadillo]

Check VT2, your sawtooth swing is very weak. And if your reference point is Pin 13, why does it go negative?
I cant see very clearly but how many volts is the swing. Is this the original board.?

Measure voltage of C4 positive with respect to Pin 13 and C4 negative with respect to Pin 13. (Z1 and Z2 voltages)

Edit; The voltage at the ramp comparator is important, can you probe again using DC coupling.

Edit: Take out C5 [0.47uF], measure ESR. [Ramp Generator]

[oldway]

.....it would be better to put back all the original boards and troubleshoot from there then to mixed and match with wrong boards causing confusion.

Ok, I agree, boards seems not to be compatibles from one power supply to another.

From the start: I am getting little or no voltage on the secondary or the primary of the mains transformer,

that's indeed a problem of the scr's control board.
Let's remember how it works: it control only the voltage drop on pass transistors....It must stay between 5 and 12V ...So, if the voltage drop on the pass transistors is withing this values, the pre-regulator is working fine.

Let's now check the pulses at the same test points you have measured on the good working scr's control board.


P2 is an ajust of the max firing angle of the scr's.

P1 is an ajust of the voltage drop on the pass transistors.

[Chris Wilson]

Check VT2, your sawtooth swing is very weak. And if your reference point is Pin 13, why does it go negative?
I cant see very clearly but how many volts is the swing. Is this the original board.?

Measure voltage of C4 positive with respect to Pin 13 and C4 negative with respect to Pin 13. (Z1 and Z2 voltages)

Edit; The voltage at the ramp comparator is important, can you probe again using DC coupling.

Edit: Take out C5 [0.47uF], measure ESR. [Ramp Generator]

Hi again, you are being very kind spending so much time with this! C4 + to pin 13 SCR board shows 15.76V  and C4 - to pin 13 SCR board shows 30V


I hopefully attach the revised scope shot in DC coupling mode and 5V per division scaling. The scaling automatically compensates for the X10 probe setting

I lifted a leg of C5 and with my PeakESR70 ESR meter it shows "open circuit or low capacitance" I lifted the leg of the same capacitor on my borrowed SCR board from the same 50A supply and that one measures o.56uF with zero Ohms ESR What type of cap can I replace it with, I can see what I have, and what voltage rating please? I have 470nF poly film ones rated at 250V DC?

[Armadillo]

Check VT2, your sawtooth swing is very weak. And if your reference point is Pin 13, why does it go negative?
I cant see very clearly but how many volts is the swing. Is this the original board.?

Measure voltage of C4 positive with respect to Pin 13 and C4 negative with respect to Pin 13. (Z1 and Z2 voltages)

Edit; The voltage at the ramp comparator is important, can you probe again using DC coupling.

Edit: Take out C5 [0.47uF], measure ESR. [Ramp Generator]

Hi again, you are being very kind spending so much time with this! C4 + to pin 13 SCR board shows 15.76V  and C4 - to pin 13 SCR board shows 30V


I hopefully attach the revised scope shot in DC coupling mode and 5V per division scaling. The scaling automatically compensates for the X10 probe setting

I lifted a leg of C5 and with my PeakESR70 ESR meter it shows "open circuit or low capacitance" I lifted the leg of the same capacitor on my borrowed SCR board from the same 50A supply and that one measures o.56uF with zero Ohms ESR What type of cap can I replace it with, I can see what I have, and what voltage rating please? I have 470nF poly film ones rated at 250V DC?

Firstly C4- and pin 13 why you get -30V... it should be 10V? Check Z2.

I checked with the resistance in series to the capacitor. it will charge only at 5.4v at 100Hz.  So the ramp is actually correct in your DC coupled diagram.

The circuit will not care whether its positive cycle or negative cycle, it will pulse both SCR together. We need another scope as below

Make sure above voltage is correct then;

Scope 1: Pin 6 and Pin 7 Channel A, Trigger at Channel A, Pin 5 IC1 with respect to Pin 13 Channel B.

Next;

Scope 2: Pin 8 and Pin 9  Channel A, Trigger at Channel A, Pin 5 IC1 with respect to Pin 13 Channel B.

We need to know when the SCRs are triggered after zero crossing.

Edit: Adjust P2, and you should see the trigger point follow [the wave shifted]

Edit Measure across Z2. what is the voltage?

[Chris Wilson]

C5 fell apart when messing with it, so I have fitted a 470nF 250V DC working polyester film cap. It has made little difference, nothing measurable, to the saw tooth wave form. I returned to looking at the secondaries of T1 on the original SCR board I just renewed C5 on. I am sure things are not correct here. With my probe set X10 and a vertical scope scaling of 5V per division I see the attached. But when I probe pin 8, on the other secondary, even a 50V per division scaling isn't enough. I had to add an automotive attenuator to the input socket of the USB scope, a device I use for looking at injector waveforms and coil on plug waveforms on engines. With X10 probe and 20:1 attenuator, and a vertical scaling of 10V per division I get the other waveform hopefully attached. I suspect this latter one is NOT correct and may point to a faulty SCR??

[Armadillo]

So you found it.
But you better probe in pair like 6 as common and 7, and then 8 and 9 as common. since there are isolated.

You need to probe properly using correct common before can make the call.

So the Z2 voltage is 10 or 30?

[IanMacdonald]

C5 (0.47uF) charges through 47k from 15v. The initial charge current will be 0.3mA (300uA). 1uA will charge 1uF at 1v/s, so so the cap will charge at a rate of approx 600v/s. In 10ms (a half cycle) it will charge by 6v. The actual value would be a bit less because of diminishing current as the cap charges, and you'd need integral calculus to get a precise answer. (Too late at night for that sort of thing!) This is good enough for our purpose anyway. Thus the sawtooth should be about 5v amplitude as a ballpark.

The problem with measuring the SCR firing pulses is that you can't measure a pulse of a volt or two superimposed on AC ripple from the off-load transformer primary, which will be ~700v p-p and too much even for a x10 probe on a 5v/div range.*  If you want to check the pulse transformer T1 outputs, disconnect the SCRs and use a 33R (or thereabouts) test load on each pulse trans secondary.

(* Similarly I keep telling the climate change afficonados that you can't possibly measure a  fraction of a degree change superimposed on 50C day/night and summer/winter changes, in which case the 'ripple' isn't even a sinewave but highly random. A similar measurement problem to this, and I'd call the IPCC's bluff over the notion that they could do so accurately.)

[Chris Wilson]

So you found it.
But you better probe in pair like 6 as common and 7, and then 8 and 9 as common. since there are isolated.

You need to probe properly using correct common before can make the call.

So the Z2 voltage is 10 or 30?

Across Z1 I see 15.4V and across Z2 I see 10.0V

Could I ask you how EXACTLY to probe the T1 secondaries properly to be sure it's something up with an SCR please? I did resistance tests of them both, gate to cathode and saw both were about 35 ohms, if I can get the darned things out (another nightmare job by the looks of it...) is there any definitive test on the bench? Peak make an SCR tester, but I don't have one, so would need to rig something up. Thanks again everyone. I wonder if this could explain it so often tripping the mains panel RCB when either turned on or turned off?

[IanMacdonald]

SCRs are easy enough to test. The gate-to-cathode resistance (of a high power type) is a few tens of ohms as measured on a DMM, and it appears like a resistor rather than a diode. Anode to cathode should be O/C both ways.

As a test circuit, use a 12 DC anode supply with a lamp in series as a current limiter.  Applying 50mA or so to the gate should fire it. (say 220R from a 12v supply) After which it will conduct until the power is cut. If it passes this test it is unlikely to be faulty.