Cound anyone help me to understand this power supply circuit ??

[Spanero]

Calling you expert power supply designers !!!

Here i have a dc to dc power supply circuit, It is out of a 100A output orbital welding system.
The specs

160vdc input ( i have not shown the rectification / capacitor on the input side)
expected 20A current draw mains side (as stated on the name plate, 110vac input rectified etc to the 2x 1400uf caps)
expected 100A max output @around 14v - 20v

When the machine is powered up all looks ok it will sit there not consuming much current at all.
When the thyristor is fired (i have tested it is fired about 3us) a large current consumption and the 25A mains trip trips out.
Looking at the drawing i see when the thyristor is fired a near on short through the inductor at the top of the page.

What i have done ....

I have tested all the components and all test ok.
the circuit idles fine and there is 160vdc across the 2x1400 capacitors
I have tested the output rectifier and even removed it during triggering the thyristor. same outcome trips the mains input.
I have put an external supply on the capacitors with current limit, this does current limit when the thyristor is fired.

I have noticed someone else has been inside the machine and i cant verify the wiring is correct, this is why i drew it out to try and check but i think i need help. !!!


I would love some other peoples input as i am now pulling my hair our.
I have basic electronics knowledge and most is self taught.

Many Thanks

Mark.


 

[Richard Crowley]

That 20A fuse in the upper right corner looks extraordinarily out of place. If that is how it is actually wired, then the symptoms seem right. The rectified mains power is delivered into a dead-short.  IMHO, either you have not drawn the circuit properly, or somebody has re-wired it improperly.

[rob77]

That 20A fuse in the upper right corner looks extraordinarily out of place. If that is how it is actually wired, then the symptoms seem right. The rectified mains power is delivered into a dead-short.  IMHO, either you have not drawn the circuit properly, or somebody has re-wired it improperly.

+1

i was about to highlight the same.. either your drawing is wrong or the circuit

[Spanero]

Hello all,

thankyou for the replys..

I confirm the drawing is correct, as i stated in the post i can see a short too. i thought that something magical happens with the inductor ...   

any ideas how it should be wired,
i will have a go and see what i come up with ...    BOOM..!

Thankyou

Mark.

[rob77]

actually it seems to be all messed-up (if the drawing is exact).
the best would be to google for some thyristor welder circuits and figure out the correct setup - let's hope the wiring was the only change someone did to the thing

all those circuits will be very very similar - there is not too much to "invent" with such a circuit.

[jlmoon]

Greetings,

You know in cases such as this, there is always the manufacturer.  They might actually have the scoop on how the device should be wired and could possibly sell you the correct replacement parts should it require any. 

JLM

[Jay_Diddy_B]

Hi,

+1 for the schematic being wrong.

I can see two of the 0.55uF 1000V capacitors in the picture and only one in the schematic.

Jay_Diddy_B

[Spanero]

Hello

I have left out one of the capacitors as i was simplifying the drawing, the other capacitor is just across the dc welding output. Good spot though.

sorry i am bad   

[megajocke]

Sure the position of the fuse seems a little funny, but it should work. The schematic looks correct to me, even though some of the text is unreadable due to low resolution of the photo.

It's a single-ended resonant power supply circuit, equivalent to the one shown in the attachment (taken from Switching Power Supply by Pressman et. al.). That the transformer and resonant capacitor series branch in this case is connected in parallel with the inductor instead of the power switch doesn't affect operation of the circuit.

Did you measure the inductances out of circuit? Transformer secondary shorted or open? There seem to be some inductance values given in the schematic, but there is no unit given (though I'd assume they are in µH). Are the two toroidal inductors closest to the cooling tunnel the transformer and inductor?

Be warned that this kind of circuit might not like to run without a load depending on how it was designed. Unsuitable load might cause commutation failure, shorting the input. What is the output supposed to be connected to? Directly to welding electrodes? Are there any large filter caps?

[Richard Crowley]

@megajocke, I don't see on your schematic diagram where is the dead-short (fuse) between L1 (or L2 if you wish) and (B) (the "common" node in your circuit).
I don't see anything remotely resembling that.  If the fuse were missing, THEN it would start to resemble your circuit.
Not to mention that the SCR is in a completely different part of the circuit. (In series BEFORE both of the inductors, not in parallel AFTER the first inductor)
If Spanero had posted YOUR circuit, then I could see how it works. But not the one he drew.  It just seems impossible to me.

In your circuit, it appears that when the SCR conducts, it "charges up" L3 which then "discharges" through the rest of the circuit.
But in Spanero's circuit, the SCR conducts and appears to send power through the entire the circuit, including "charging up" the inductor (at least until the fuse blows).

[Spanero]

Hello

I just wanted to say thank you for the input on this thread. i have been reading all the reply's and i have decided to do some new photos so the values can be seen.

I have also studied the diagram and made some alterations as i feel the wiring on the terminal block around pic 1 to maybe wired incorrect (i confirm my drawing is as the machine is wired!!).
Point A on the diagram looks to be a commutation circuit is it not ?

I do not have any inductor testing equipment but i have stripped down the transformer/inductor (which is housed in the heatsink) and all looks fine, the inductor on the blue ring in pic 1 is also looking good, there is one point i do not understand and that is the fuse 20A so some current is expected to flow that part of the circuit.

Thankyou People.
Mark

[Richard Crowley]

Your new circuit is starting to look more like the reference that megajocke posted.
It appears that the inductor is in parallel with the transformer primary, and the R/C circuit is also in parallel.
And the primary of the transformer is AC coupled.  So it appears to need the inductor and R/C part of the circuit to produce some AC oscillation.

I would question the drive to the SCR gate. And/or even the SCR itself.
Switchmode power supplies are always an order of magnitude (or two) more pain than working on linear supplies.
And you have the extra "excitement" of having such a high-power device.

[Spanero]

Thankyou Richard,

Yes megajocke looked to be on the right path, i am looking on google for the full design.

Sometimes i get so involved and stuck !!

I am posting the ac input part before the primary rectifier here for reference..
Thankyou

Mark.

[Richard Crowley]

I am posting the ac input part before the primary rectifier here for reference..

That only looks like a power-line filter to keep junk from going back out the power mains input.
I wouldn't expect there is a problem anywhere around that part.

[Spanero]

Ok..

on looking , reading and the help from you all......

I have looked at the thyristor and the firing, the firing is 3uSec according to my osc. The thyristor that is in takes 110uSec to turn off (this has been looked at sometime before and is a new thyristor this is why i thought it could have been wired wrong).
As this is in a resonant circuit i am feeling that this is too slow a turn off time and maybe is staying on which is a short across the inductor, 51 ohm resistor and capacitor, this then trips the trip (the trip too fast to blow the fuse).

I am now looking at a thyristor which turns off in 4uSec for inverters which makes adds up  now...  fingers crossed ...

Thankyou
 

[rob77]

thyristor turns off at zero crossing - once fired it stays turned on till the AK voltage drops to zero (more accurately below the forward voltage) or the current goes down below the holding current of that thyristor.
a special case if a GTO thyristor - that one can be turned off by negative gate voltage.

[megajocke]

Your new circuit is starting to look more like the reference that megajocke posted.
It appears that the inductor is in parallel with the transformer primary, and the R/C circuit is also in parallel.
And the primary of the transformer is AC coupled.  So it appears to need the inductor and R/C part of the circuit to produce some AC oscillation.

But there is no change in the primary circuit between the new and old schematics. The only change is that the schematic was redrawn. There was an inductor in series with the fuse in the original schematic too -- no dead short.

Through the sequence of transformations in the attached figure, each step being a circuit with equivalent operation save for the exact current waveform through the DC source, you can see that the circuits are actually the same. The last transformation, where the AC-grounded end of the resonant branch (transformer/capacitor series connection) is reconnected from one AC-ground to another, might be the least straightforward one, but circuit operation will be identical apart from the AC current component through the DC source.

The transformer leakage inductance substitues for inductances "L1" and "L2" in the figure from the book.

The purpose of the 51 ohm/3 nF RC snubber is to slow the voltage rise and damp ringing after the diode in antiparallel to the thyristor turns off. Without them the thyristor might parasitically turn on again when the anode-catchode voltage starts to rise.

The transformer/0.55 uF capacitor resonant branch is what creates a resonant current pulse that is supposed to turn off the SCR.

[megajocke]

I have looked at the thyristor and the firing, the firing is 3uSec according to my osc. The thyristor that is in takes 110uSec to turn off (this has been looked at sometime before and is a new thyristor this is why i thought it could have been wired wrong).
As this is in a resonant circuit i am feeling that this is too slow a turn off time and maybe is staying on which is a short across the inductor, 51 ohm resistor and capacitor, this then trips the trip (the trip too fast to blow the fuse).

I am now looking at a thyristor which turns off in 4uSec for inverters which makes adds up  now...  fingers crossed ...

Thankyou

What is the part number of the thyristor in there now? A "standard" thyristor like those often used for phase angle control is not going to work because of the long time needed to turn them off properly and the intolerance to high dv/dt. It could be hard getting anything apart from the designed-in correct part to work reliably.

[Jay_Diddy_B]

Hi,

I have put together a LTspice model of the circuit, so we can see the circuit working normally. The circuit is the same as the one in the original post.
There is no fast SCR models included in LTspice, so I made a model based on a switch.




Here are the key waveforms with the circuit operating normally:



The SCR turns off when the diode is conducting. The turn-off time of the SCR Tq must be shorter than the time the diode is conducting.

One of the features of this type of circuit, is they are inherently short circuit proof. I short the output the waveforms become:



The diode and the SCR currents are approximately equal in amplitude.


The main threat to these circuits is an open circuit. If the load voltage is increase, the diode current is reduced:



If you keep raising the output voltage you get below a critical level of diode current and the SCR is unable turn-off.

The fuse is in the correct position. If the SCR fails to turn off, current will ramp up in the inductor and the fuse will blow.

You need a fast inverter grade SCR in this circuit. These may be difficult to find. 20-25 years ago they were manufactured by RCA, GEC (Later Marconi Electronic Devices Limited 'MEDL')

I have attached the LTspice model in a zip file for anybody interested in playing with the circuit.

I have updated the LTspice file
The file I had attached was an old revision 

Regards,

Jay_Diddy_B

[Jay_Diddy_B]

Hi,

I had another quick look that this schematic. The leakage inductance in the transformer is a key component in this design. In the model I presented this is 'hidden' in the coupling coefficient.

Regards,

Jay_Diddy_B