Late 1970s SMPS Design Query

On a sunny day (Tue, 26 Jan 2016 08:58:29 -0000 (UTC)) it happened Cursitor Doom wrote in :

Yes, in the old days TVs used things like DY87 HV rectifier tubes, the heater was fed with one big insulated turn on the UI core horizontal output transformer.

Some good ideas again, thank you, Jan. Now let me run an idea by you whilst I nip off and check out your suggestions: suppose I completely remove the tranny from the board then feed its main primary winding with an old valve HP sig generator I have (set to say 50kHz). Assuming I can get the primary up to its normal working voltage level, I should expect to see a nice clean sine wave output from all the secondaries. However, if all that noise is *still* present, then it's clearly an internal breakdown of some sort within the tranny. I may well have to boost the HP's output via a different transformer to get it up to the necessary primary voltage for the tranny under test, though. Or possibly not; I don't know. Any problem you can see with this method?

On a sunny day (Tue, 26 Jan 2016 10:25:37 -0000 (UTC)) it happened Cursitor Doom wrote in :

Would probably work, use audio amp? What is the frequency? In audio range? In the old days some TV service technicians has some thing that would put a voltage on the transformer and check for load current, if it was shorted you would see it. Never used those myself. I say give it a try.

Maybe feed a function generator into the switching node? (can tolerate shorts, and you can measure the waveforms on the coils)

Cheers

Klaus

Upper end of, yes. Spec sheet for the controller chip states 50hz-100khz, but really I'd have to check what the fo of the actual resonance circuit is in this instance to be precise.

I've measured the Vcc on pin 1. It's about 11V but slowly rises for some reason; poor regulation clearly - but no noise. Can't find the 10uF cap that zener is supposed to be in parallel with. No decoupling cap across Vcc & gnd on that controller for some reason. Perhaps it's stable enough without one. :-/

I keep thinking about it....

I did suggest that method earlier, Klaus, but someone here poured cold water on the idea for reasons they didn't give. I may still try it later on, though.

.... Phil

It's 20khz (just measured it) at the pwm output of the contoller chip. Looks a lot cleaner today (but that's capricious nature of such faults I guess).

Hadn't thought about using one of those for windings before. Could be useful in future. The diodes would still have to be taken out-of-circuit to be sure of valid readings, though, AFAICS.

On a sunny day (Tue, 26 Jan 2016 12:53:53 -0000 (UTC)) it happened Cursitor Doom wrote in :

Yes, I would have guessed 15625 Hz (TV horizontal) :-) Not that far off.

Cleaner because HV multiplier disconnected? Did you disconnect all the loads?

There is no zener, externally on pin1.

When the supply reaches a point of regulation, this voltage should rise high enough to shut off the bias start-up section, and be limited by 12V zener V1813.

-pin 1 voltage can be low due to poor C1818 storage, burnt R1838 series limiter

The voltage may not rise higher because the control chip protection circuitry is seeing a fault that prevents continuous operation.

possible inhibitors

- pin 13 low input undervoltage (bad sj or series limiting parts in input power path)

- pin 11 high ( power switch over current / damaged high value current sensor R1811 / broken small signal ground path to controller section)

When no limiters are in effect the regulation point is pin3 of the chip ~3V7

RL

On a sunny day (Tue, 26 Jan 2016 09:45:01 -0500) it happened legg wrote in :

Oh yes there is.

The smps controller itself is an excellent impulse generator to observe the transformer's health. It seems to function, so use it.

If you want, you can apply a small signal source to any transformer winding, when power is removed from the supply, to observe ratios and gross faults/phasing/continuity etc. Using the safely self-protecting smps controller and functional power switch is many steps farther on in the troubleshooting sequence.

Applying an external signal source to the driver of the power switch when line power is present, even with the best of intentions, allows for a host of possible errors that can only result in more damage.

RL

Right. I note this elsewhere. When the circuit runs normally, the node will rise to the zener's ~12V and the start-up biasing will be inhibited. At present, this component is not showing any suspect or odd behaviour.

RL

This thread should probably be shifted sci.electronics.repair.

RL

On a sunny day (Tue, 26 Jan 2016 10:09:34 -0500) it happened legg wrote in :

Agreed, dunno why he cannot find the 10uF cap C1818 though.

I would unsolder the secondary rectifier diodes of the transformer, and the HV connections, and test again. Should only take 5 minutes. ?

On a sunny day (Tue, 26 Jan 2016 10:16:09 -0500) it happened legg wrote in :

I dunno, you can learn a lot from how this thing works.

The only cap I can find in parallel with that zener is a 30uF one - and it looks like it's original. This board seems to slightly vary in odd places from the published schematic in some respects.

For some reason the chip is supplying a really nice clean pulse train at a low duty cycle now. It's only at v1812's collector and thereafter that everything starts to get messy. The pulses on that collector are all really, really skinny and lack the bright tops of those on the base. Maybe this is normal for pulses being pulled through a transformer; I can't say. Your suggestion above is noted...

That's where it kind of started (under "venting capacitor").