Transformer shot! (was scope SMPS/ capacitor venting)

Hi

If you had a reliable connection to the primary and all secondaries were in fact free (either the tranny completely removed from the circuit board or at least no diodes anywhere remaining), then that would mean the voltages are likely screwed up... But wait:

Are you sure that you have not mixed up the windings? Maybe the two 1.8V windings are actually the 2 symmetrical "innermost" ones, the 3V ones are the "medium" ones and the 15V are the "outermost" windings? Your measured winding voltage ratios "1.8:3.0:13.0 volts" and the schematic output voltage ratios "6.7:13.4:60.7 volts" (I've added a little compensation for 0.7V silicon diodes) are (from a purely ratiometric point of view) not very far away from each other. In fact they are so close that the differences between the smaller ones can be easily explained by your measurement errors (how accurate was that 0.8V measurement anyway?) and the possibly intended uneven loading of the power rails in the scope.

So, considering the winding connections slightly rearranged, the transformer looks just fine to me.

But once you have it out and disconnected, please make another test: apply ca. 15V RMS to the 12.7V winding (to the one where you measured

3V) instead of to the primary. And check if any isolation looks like breaking down. Note that the 15V value contains some compensation for the fact that the power supply uses inductors after the rectifiers (and therefore the normal winding voltage is higher than the normal output voltage). That would load the transformer close to its normal condition and any breakdown should become apparent.

Regards Dimitrij

I follow what you're saying, Dimitrij, but for that to be the case, the tranformer's internal wiring would have to be twisted and I can't see why they would do that. Admittedly the ground pin is in 'real life' at the far end of pinouts rather than the centre, but... well, I don't know. If you're right you must be some kind of genius, that's all I can say.

OK, you're the boss! I'll carry out that investigation tomorrow and report back. Maybe I can work out if the internal taps are out of sequence compared to the schematic by measuring the DC resistance of the winding at the various taps and... well you get what I mean. Intriguing idea certainly deserves to be fully explored. Many thanks.

A short that would drop the voltage on that side of the winding should drop the voltage on the other side due to coupling. Recheck everything, that's what I say.

In fact if you got a power amp that can throw some current maybe inject int o the winding that is giving the low voltage and see if it steps up in the other windings. Of course heavy current, but VERY low voltage.

Interestingly I just used my Phase Linear 400 Series Two as such an amp to inject a signal into, believe it or not, the SMPS transformer of a Phillips scope ! Bunch of coincidence, but of no help to you at the moment as it is a totally different model. This one kicks the voltage down to 24 VDC and t hen feeds the SPMS. I see no connection for a battery but I imagine it coul d be made to run on batteries.

Wouldn't be bad to have a scope run on like two laptop batteries...

Anyway, I learned to be very careful about condemning transformers. We lear n by mistakes and some of us are pretty fart smellers. I hit 191 on an IQ t est once, damn, how can I even be alive ?

Umm, I KNOW my IQ is not that high, it was the top I hit when I went on a k ick to take alot of online IQ tests. In fact my average was so good out of the about 20 of them I took, I doubt their validity. It was over 135 which is 1 % of the world so really, I doubt it.

At any rate, I would take an audio amp and feed that thing until it runs. Y ou got nothing to lose. reconnect it al and feed it from some nice maybe 10

0 WPC audio amp with a 10 KHz square wave or something and see what happens . You have no current limiting now so you follow the smoke.

And BTW, that hosting you're using SUCKS. It does not respond right to my z oom command and it also nags about my browser. I suggest a Dropbox account, AND USING THE /PUBLIC directory. What's more, on Dropbox all your stuff is private, no browsing nor web crawling can find it, you MUST give out the U RL by using "Copy Public URL". I highly recommend it. No ads or anything at all. Point the browser and the picture shows up, download it and I can zoo m like all hell.

The limitation is like 2GB. You can have really high res photos there. I ha ve had full length movies in mine.

Anyway, you need better resolution to see the diode circuit symbol numbers to know which winding is which because it does not appear to have pin numbe rs on the transformer.

I must admit the fact that Zs on my most powerful (voltage-wise) sig gen is 600 ohms was a concern. I would ideally like to have zapped the tranny with the same voltage and current as its working conditions would expect. There's always that nagging doubt in my mind about 'what if I'd had more power to throw at it? Would that show up something useful?'

One can train for an IQ test. A lot of people don't know that, though!

That's a very good idea. Must admit I hadn't thought of that dodge!

Sorry to hear that. I didn't chose Flickr (or whatever it is) I inherited it from an old Yahoo mail account. If it's that crap, I'll ditch it.

Well, at last there is a serious effort to actually record and report real measurements. However, you may be misleading yourself.

Are you sure of the pin locations and their function on the transformer? They will not likely correspond to the schematic arrangement - which is arranged for functional clarity alone.

The transformer pin numbers are not identified on the schematic.

This is why it is much easier to make accurate winding voltage measurements when the transformer is in-circuit, connecting to easily identifiable schematic components and circuit nodes.

The voltages you report would be normal if the pin functions were as listed below

13V.......60VAC 13V.......60VAC 3V........12V5 0V (gnd) 3V........12V5 1.8V......5V 1.8V......5V

The nonlinearity of the ratio is due to the increasing influence of forward diode drop at lower voltage and the proportional loading effects of differing currents on rectifiers, windings and output filtering components.

Recheck pin function before jumping to conclusions.

RL

Well I *was* until Dimitrij pointed out this possibility. He changed the thread title in his follow-up so I guess you missed it. So yes, it's something I need to further investigate and I shall report back here in due course with my findings.....

Right; now re-checked. DC measurments proved (unsurprisingly) too close together so I re-tested using 100khz instead. These are the impedances WRT ground of the output taps of the long winding in the order they actually come out of the transformer: GND, 0.17ohms, 0.17ohms, 0.26ohms, 0.28ohms, 3.7ohms, 3.8ohms. So this doesn't seem to tally up with the schematic. Or does it? I need a pint of strong coffee to kick-start my head on this one. :-/ Anyway, later...

If the transformer is removed from the board, which seems to be the case, you can probe the PCB for continuity between known component leads/schematic nodes and empty PCB transformer pin lands.

Using a logical physical numbering scheme (if one is not allready present on the actual transformer body), you can assign numbers to the board and the schematic, for reference.

RL

Inductive reactance goes up with the square of the winding's nominal voltage, so you take the square roots of your impedance values.

If you then take into account that the lower impedance values are strongly dominated by the DC resistance (which stays linear and does not square) and the upper one is mostly dominated by the AC reactance (which does square), the ratios seem to look just fine (well, so far as I can see, within a reasonable margin of error).

But the ratios don't tell the whole story. Even if there is a winding short, all impedances will be very low (which they sort-of are, I would have expected higher values everywhere, but then 100 kHz is maybe too high, try testing at 10 kHz and see...), but the ratios between the windings would be still be mostly correct.

Try to run it on higher voltage (like 15 V applied to 12.7 V secondary), and see if it pulls excessive current and warms up. That would indicate damage more clearly.

Dimitrij

The transformer secondaries go (on the left side viewed from the top of the transformer).

1.5kV 1.0kV Gap/no pin HV Common 60 60 12 12 6 6 0

As others have pointed out this PSU will not run happy without a load and I don't know what would be suitable. When I worked on these I always just left the psu connected to the scope. Lets face it, the scopes been turned on at some point with the psu connected so its not going to do much more damage and at least you will know the loading is correct. The EHT multipliers on these break down internally on these.

In one of your pictures there are a couple of diodes that look messed up (V1809 and V1811) near the bridge. They are supposed to be BY208-1000 (1000v rectifiers), I can see "40" on one, maybe 1N4007?

All noted, thank you, gentlemen. I'll have to check those tips out tomorrow or a divorce will be in the offing. Until I report back tomorrow then, thanks...

I managed to find a PM3264 PSU to try out.

Unloaded it squeals as expected so I tried a makeshift load with what I had lying around, 6 x 470R 5Watt w/w resistors.

You can pull one of the connectors out of the scope for a connection. (See photos)

Just for fun (and I'm running this off an isolation transformer), pull V1812 and scope T2 with T1 as probe ground. you should see a nice drive waveform for a few seconds and you can check the frequency is 20KHz.

Incidentally the core on L1806 on this board was loose (came apart) and also caused squealing but of a different note.

If you want me to take any readings let me know, nothing too time consuming though :)

Photos of load (It gets hot so take care)

Well on that basis there may be nothing wrong after all.

It's not possible to test this board with it connected to the scope. On this model, it slots inside the two main signal boards which make access under proper, full working conditions impossible. Just *another* obstacle I've faced with this repair. The EHT multiplier has been totally disconnected all through my tests except where explicitly stated otherwise.

I like your thinking! But no, the one nearest the bridge is a BY208-1000 alright, the other one to the side of it is a BY134. They both tested fine out of circuit.

I was really struggling trying to match up the pins to their particular outputs; fortunately JC has has posted the pin-outs for this transformer and saved me some brain cells (I can't afford to lose any more). Seems the voltages I'm getting are not far off what they should be after all.

I did just try this a moment ago, Dimitrij, but doing this just flattens the output from the sig gen, I'm sorry to say. Hardly surprising since it's a 600ohm unit and the 12.7V tappings are 0.52ohms 'apart'! To perform this test properly I'd have to adopt the work-around suggested by another chap here who said use an audio amp to get the current up. I may well have to do this if it comes to it. The other problem is, my oscilloscope current probe is lacking a termination unit so it's readings will be meaningless and I can't use the true RMS current range on my DVM because it's probably going to be out of its bandwidth at this frequency range. :(

Have you posted this before on another thread related to this? Either you have or it's deja vu. Sorry if you did, I didn't note it properly.

Yeah, those Stocko connectors. On another Philips manual I have for a different scope, they actually publish the proper values for a dummy load which would be really helpful to have on the current problem I face.

Not sure why you say pull V1812, but here's the waveform I got between those two points when I did this test a few weeks ago:

But then it goes downhill. Here's V1812's C/E junction:

And the B/E junction of the main chopper transistor:

Not surprising it doesn't work properly with control voltages like that!!!

Hi, Its been some time since I worked on these but I'm pretty sure we ran these with the board out, turned round so you can get the connectors on and I guess without the HT connected. Alternately put a suitable load on the PSU.

That might be one problem, the sine voltage around T1801 is 800v, your BY134 is a 600V diode. Also HV diodes can go reverse leaky, try a high ohmsmeter on it (10-20 meg range). Shouldn't be any reverse leakage.

I guess you saw my next post on this? Try a load on the board before you do any more work. It will tell you if the PSU runs silent or not under load. The one I tried was screaming like heck then silent with a load.

Pulling V1812 lets you see the clean output from the driver chip without all the crap feedback from the transformers. Your frequency looks good.

I got the same crap and ringing/distortion on my PSU without a load. Put a load on it. switchers don't work off load.

Will a DVM suffice or should I do this with my faithful old analogue AVO?

You're a late-comer to this party, so you will be unaware that even when tested under full working conditions with all the loads plugged in, this twitcher/switcher still hisses and the 20 Ohm power resistor R1814 (just below right from the chopper transistor on the schematic) quickly starts to burn up.

I take your point on the dummy load, though. I must rig one up before doing any more live testing.