Jestine Yong's Diode Removal Test

[JonathanAnon]

Just reading through Yong's book on SMPS repair again to try to get ideas for a particularly awkward SMPS I'm working on... He says that if you are having problems with an SMPS power cycling and want to narrow the problem down, then you can remove ONE of the secondary Schottky diodes, power on and see if the problem persists... This also comes with the warning NOT to remove all of the secondary diodes as it will "blow up"..

Anybody tried this method.. I was gonna give it a shot

[kc8adu]

Re: Jestine Yong's Diode Removal Test

just dont remove the one on the line the feedback comes from...

[JonathanAnon]

Re: Jestine Yong's Diode Removal Test

Hi kc8adu, that's what would make most sense but Yong never made specific reference to not taking out the diode for the B+ voltage, just not to take ALL of the diodes out at the same time.. I'm a bit confused as to why he wouldnt just say "dont take out the B+ diode".. His English is not great, so maybe that's what he meant to say?

If you take the example of the SMPS I'm currently working on, I've only two secondary dual schottky diodes, one for the 12V and one for the 5V... So by what is written (literally) in Yong's book, I could take out either and test but just not both.. But there doesnt seem any way that the feedback system wouldnt just shut it down immediately..

[budm]

Re: Jestine Yong's Diode Removal Test

The power supply uses negative feedback system, if the output tries to go up, the feed back (usually through Opto-isolator) will turn the LED in side the Opto (Depends on Current transfer Ratio of the Opto) on more or less which then will turn on the Phototransistor (more or less) to clamp down the drive in the primary side. So if you remove the neagtive feedback the power supply can run away without any regulation, just think of the steam engine with Governor control, without governor, you will have run away train.

[JonathanAnon]

Re: Jestine Yong's Diode Removal Test

So if you remove the neagtive feedback the power supply can run away without any regulation, just think of the steam engine with Governor control, without governor, you will have run away train.

okay, now that's interesting Bud.. Yong doesnt go hugely in to how the feedback system works. But I assumed that if there was NO feedback at all, that this would be seen as a fault, and the PWM IC would shut down. I guess probably the safest way to verify this would be to remove the feedback resistor and see if it still works.. ?

You see, Jestine Yong suggests that if you suspect there are problems with the feedback circuit, that you should check the feedback resistors, directly replace the Optocoupler and TL431 etc etc... I'm just wondering why he would suggest going to all this effort if the feedback circuit could simply be temporarily removed from the circuit in the manner we are suggesting..

[tom66]

Re: Jestine Yong's Diode Removal Test

I have to say, Jestine is wrong on a lot of things... and this is one thing.

Feedback is almost always negative in an SMPS... that is, a HIGH signal from the opto tells the power supply the output is too high, and it needs to drop the output. Let's imagine if it were positive... that is a LOW signal from the opto told the SMPS to turn the power supply outputs down. How will the supply start up with no current for the opto? The only way would be an "ignore feedback for N milliseconds" circuit, which has the risk of being wrong under low loads, or insufficient under high loads.

If you remove the schottky which is used for feedback rail, then the SMPS will go to its maximum output, which will likely lead to damage to the output capacitors and possibly the primary side MOSFET (especially if it is a flyback, as the reflected voltage is proportional to the output voltage.)

Most SMPS uses more than one rail for feedback, so one diode removal may not be enough to cause it to run out of control.

If you remove all of them, then of course you have no load to damage the output capacitors, but the SMPS now has an unloaded transformer. What will this do...? Probably very little... I don't anticipate any kind of damage to the drive electronics or the transformer. The Vcc will probably not rise quick enough due to the low load.

I don't see what removing the diodes does. If you do this, the SMPS will have no function... how will you test it? You can test the diodes out of circuit. If they are shorted, they are bad. If not, they are more than likely fine. I really wouldn't recommend removing a diode and running the supply without it.

[JonathanAnon]

Re: Jestine Yong's Diode Removal Test

I have to say, Jestine is wrong on a lot of things... and this is one thing.

There's a lot of good information in his books, but I cant understand why he didnt get a fluent English speaker to proofread them!! Loads of grammatical errors in sentences that NEED to be right..

I don't see what removing the diodes does. If you do this, the SMPS will have no function... how will you test it? You can test the diodes out of circuit. If they are shorted, they are bad. If not, they are more than likely fine.

The idea behind this was that if you suspected that the secondary diodes were failing / leaky under load that this would help to identify the issue.

[tom66]

Re: Jestine Yong's Diode Removal Test

Diodes go BANG and they are dead shorted. Leaky is possible but really only with small SMD diodes. The diodes would not fail under load, except to permanently short.

His English may not be perfect. That is not the issue. I actually mean his tests and so on are not the correct ways to do things, and may mislead you into making the wrong conclusion.

[smason]

Re: Jestine Yong's Diode Removal Test

Hi kc8adu, that's what would make most sense but Yong never made specific reference to not taking out the diode for the B+ voltage, just not to take ALL of the diodes out at the same time.. I'm a bit confused as to why he wouldnt just say "dont take out the B+ diode".. His English is not great, so maybe that's what he meant to say?.

You should ask him. From what I've read he's pretty responsive to his readers.

How's the book? It's one I've considered buying, just never pulled the trigger....

[tom66]

Re: Jestine Yong's Diode Removal Test

B+? Is this for CRTs? Or his name for the PFC/bridge rectifier voltage?

[JonathanAnon]

Re: Jestine Yong's Diode Removal Test

His English may not be perfect. That is not the issue. I actually mean his tests and so on are not the correct ways to do things, and may mislead you into making the wrong conclusion.

That's the reason I'm questioning this method in particular, Tom.. Most of the rest of the methods that he suggested make sense... But I thought this just didnt really make sense... I will just NOT use this diode removal method for troubleshooting SMPS.

B+? Is this for CRTs? Or his name for the PFC/bridge rectifier voltage?

Just looked back through the book and it seems that he is referring to B+ when referring to CRTs.. The term B+ is not actually defined in the book. This is the first reference to the term B+:

Each of these AC voltage produced by the secondary windings is then rectified, filtered, and regulated to produce a clean DC voltage. One of the main DC output voltages is the B+ voltage. The output from the B+ voltage supply is then connected, through a sampling error detection circuit and “feedback” loop back to the PWM IC. When the voltage from the B+ supply rises or drop a bit, the PWM IC will act to correct the output.

He seems to be referring to CRTs when he does this, so maybe I should stop using this term..

[JonathanAnon]

Re: Jestine Yong's Diode Removal Test

You should ask him. From what I've read he's pretty responsive to his readers. How's the book? It's one I've considered buying, just never pulled the trigger....

Yeah I mailed him my own "invention".



I couldnt understand why he was removing the main circuit fuse and soldering in a bulb.. instead of using a bulb tested like the one above .. I thought my suggestion was safer and more practical.. He responded with a positive and encouraging email..

The book is very easy to read... "accessible" I think is the word they use.. It's well worth a read.. Bit expensive for a download though..

[cheapie]

Re: Jestine Yong's Diode Removal Test

The power supply uses negative feedback system, if the output tries to go up, the feed back (usually through Opto-isolator) will turn the LED in side the Opto (Depends on Current transfer Ratio of the Opto) on more or less which then will turn on the Phototransistor (more or less) to clamp down the drive in the primary side. So if you remove the neagtive feedback the power supply can run away without any regulation, just think of the steam engine with Governor control, without governor, you will have run away train.

What happens if I yank out the optoisolator, output caps, and rectifiers (bridge them), and turn the transformer around backwards?

[tom66]

Re: Jestine Yong's Diode Removal Test

B+: Usually 100~200V (?) rail in CRT TVs used by the FBT, deflection circuits, and the tube. There is no real equivalent to B+ in modern electronics... maybe 24V on LCD power supply, as it is the most loaded power supply? And Vs on a plasma? Though most modern plasma power supplies split the Vs/Va and logic power.

What happens if I yank out the optoisolator, output caps, and rectifiers (bridge them), and turn the transformer around backwards?

If my calculations are correct, when this baby hits 88 miles per hour... you're gonna see some serious shit.

[JonathanAnon]

Re: Jestine Yong's Diode Removal Test

I see now that I completely misread Bud's post... I took him to mean that if I disconnected the feedback then the PSU would continue to work fairly normally, much like a plane when the autopilot disengages it would continue to "fly along the same path" .. I have reread the thread and see that this is not the case..

[PeteS in CA]

Re: Jestine Yong's Diode Removal Test

This can be done with circuit branches that don't mess up regulation. You can safely remove -12V and -5V rectifiers. If the +3.3V isn't part of the main regulator loop, its rectifier can safely be removed. In the ATX supplies I worked with ca. 2000, both the +5V and the +12V were sampled for the main regulator loop, so those could not be removed. And having a resistor load rather than a mother board as a load would be advisable (to avoid risk of damage to the mother board).

[tom66]

Re: Jestine Yong's Diode Removal Test

I see now that I completely misread Bud's post... I took him to mean that if I disconnected the feedback then the PSU would continue to work fairly normally, much like a plane when the autopilot disengages it would continue to "fly along the same path" .. I have reread the thread and see that this is not the case..

Well, technically, it would. For a couple of milliseconds. Its PWM output would then slowly drift upwards, causing the output voltage to skyrocket.

[PeteS in CA]

Re: Jestine Yong's Diode Removal Test

In principle, running open-loop could end up being benign:

* If you have a resistive load rather than a motherboard;

* If the +3.3V O/P has 6.3V or higher O/P caps;

* If the +5V O/P has 10V or higher O/P caps;

* If the +/-12V O/Ps have 25V or higher O/P caps;

* If the gate/base drive and O/P transformer and O/P inductor cores don't go into saturation at full pulse width.

OTOH, a 4V, 10V or 16V O/P cap on the +3.3V, +5V, or +/-12V O/Ps, respectively, might make an unpleasant noise and emit an unpleasant smell.

If the problem has to do with the regulator circuit not working right - a bad reference, bad divider resistors or bad compensation components - running open-loop won't detect that.