Late 1970s SMPS Design Query

If you are doing protocol analysis or something, there's no simple way to do it.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com

I can't imagine I'll ever be doing that since I've never even heard of it.

Well, like snooping the serial characters in an RS232 link, or SPI, or USB or something. Or nasty stuff like PCI Express.

Digital storage is pretty helpful there, and some scopes will decode the serial data for you.

Digital scopes are also great when you want to blow something up and keep a record of what happened. I'm about to blow up an SSR, to see its limits, and the sad history will be preserved on the screen. I can photograph that and archive it in our PDATA folder for that part.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com

So far, it hasn't blown up. This SSR seems to current limit at about 6 amps.

I'll run it overnight at 1 Hz and see if anything happens.

It would be really hard to get a pic like that on an analog scope.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com

On a sunny day (Wed, 27 Jan 2016 15:28:20 -0500) it happened legg wrote in :

To repair.

I remember applying for a job in a service center one day, very big one. The guy asked 'how many sets can you repair in a week'? I said 30, with a peak of 60. That makes 20 minutes per set, and includes the paper work, getting it from storage, finding circuit diagrams if there were any, checking it, getting the parts from storage, fixing it, and bringing it back. Also I never ever used a variac except one time in some place we were, more out of curiosity... Only a scope. Not even a freaking signal generator for repair. In the service center the very exceptional case were you had to wobble a TV IF strip did happen, but had to be reported to get the time...

In the media, when things have to stop because say some camera does not work and they need 4 you better be faster, with a freaking out director, stressed artists, and / or black on the viewers screens if it is live. You have to know everything inside out, and be able to improvise on the spot if needed. Breaks later. The Show Must Go On. Even for recording sessions it is next to impossible to get all artist and other things together again at the same place and time. There is no such thing as 'cannot'. TV studio use per hour is expensive too.

So get on with it :-)

Dear viewers, our engineers are working very hard to bring back the show and are now having their meditative coffee break to calm their mind and gain insight into the deep mysteries of electronics white magic, thank you for your patience.

-- **** * * * * * * * * * * * *** * ** ***** *** * ***** ** *** **** * * ** * * * * * * * * * * * * * * * ***** * * * ***** * * * * * * * * * * * * ** * * * ** **** * ** * ****

• * * * * * * * * * * * * **** *** * ** ***** *** * ***** ** *** **** *** ** * * * * * * * * * * * ** * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * ***** * * * ***** * * * * * * * * * * * * * * * ** **** * * * * * * * * * * ** * * * * * * * * * * * * ** * ** * * * ** **** * ** * **** ** **** *** * * * * **

CPC1020_Imax.JPG

Put the whole footage up on Youtube under "Exploding SSR - EPIC". You'll be the most hated man on the toob. :-D

..........and then when you got home, Dad would dismember and eat you for dinner....... Tell that to the young snips these days and they won't believe you.

Apart from the usual job interview / workplace horror stories, I'm feeling pretty useless, in that, not so long ago, you'd be employing less experienced bodies and providing, indirectly, the learning experience that seems increasingly unavailable these days.

Even a recently as Y2K.

RL

On a sunny day (Thu, 28 Jan 2016 07:25:15 -0500) it happened legg wrote in :

Yea, maybe . I am who I am, and my upbringing was too? difficult for my parents, so they send me to some mil type drill living environment. Stay alive and ahead of the many. Maybe the side effect it had is that I learned at a very young age not to take any shit and have been a bit of rebellion ever since. Or it is just in the genes, daddy was a freedom fighter once. Eaten alive, nooo, that would have been short and simple.

When applying for the TV job, I think (IIRC) there were 30 applicants (every body wanted it, status, fun, work with artists, meet famous people, etc),

6 were selected, I was one of them and although I had every possible diploma we were put in the school banks for 6 month on company training, video, audio, satellite, RF, LF what have you, component level, systems, and the usual management and political structure lessons too. had to sign a contract to stay, career was laid out till retirement age. I left after 8 years to travel the world. Mama asked 'Do you not want to be a technical chief of the studio?' I replied 'no way'.

Since than I did every possible thing you can imagine, started my own company, had 6 people there at some point, then left it to travel again...

As to teaching others, I have been in front of the classroom twice. I am not sure it helped the students, I am non-conventional, but I was asked so why not give it a try. Where I learned electronics, at a well known private school over here, it was said: 'Only hobbyist make it through the exams'. And I think this was largely true.

I am not sure it is possible to teach electronics to just any person. Teaching fault finding is probably not just possible. I did read today that the guy who was working on AI (Artifial Intelligence) died (at age 88?). When we look at ourselves as a neural net, a lot of learning is not (most is not) in just scribbling down math, or doing simulations, but in building the structure in the neural net that 'knows' in a different way. Once I tried some neural net software, learning modes, weights, neurons, layers etc, and then you see how 'knowledge' is formed by the connections and weights of the net. We are like that, and without that proper trained net we are no good (at tronix). And it seems (I have read) that those pathways are formed at a very young age (before 4 years old). So and I grew up playing with tronix.. So what can you teach? Parroting is not learning, and that is what this generation is exposed to in 'schools'. Parrot Einstein and then get a piece of paper and the whole world thinks you are going to save it because you are now 'Dr So and so' or whatever and work at CERN and after 50 years still have no break even fusion, but a piece of paper with symbols only you and your other brain implanted manuscriptians can understand and oh what fun.

In history you see that it takes thousands of years and revolutions before new correct insights replace old wrong ones. flat earth to round earth, blating about gravity waves versus a better theory (my personal being a Le Saga), etc etc.

So what do we have? I remember being at the knobs in the studio when the moon landings were, relaying it to the people of the country (the pictures). Now US cannot even get up there, needs a Russian taxi. Politics overtook engineers (shuttle was a joke). Von Braun had a mars plan. And note that none of these highly integrated chips and ipads were used in the moon returns. So knowledge is lost between generations. if the neural nets ... well purely philosophically you can ask that although we humming beans (for human beans for the word blind readers) so if we humming beans need this ability to make tools (not much different from microbes making chemicals to survive) are we really the crown of creation? Mosquito larvae have survived on the outside of the ISS for weeks in vacuum. Maybe those will take over the universe, and we are just a food source.

But slightly going of topic, :-) I cannot teach anyone anything that their neural net is not ready to adapt to. But with a gun on their head I can make them parrot from a book. Insight has to come from inside out, and even then it is maybe already partly in the genes, I come from a family that had watch makers in it.

That brings me to that analogy of pendulum and the LC transformer, you can imagine that L1804 in that circuit diagram of that old scope to be a rubber band, pulled on by the BU208, now imagine the BU208 being a large slider with the band in it. When it pulls (at the right time) on the pendulum, just like you had a rubber band connected to your kid on a swing, it puts energy into the LC (swing, pendulum), when it releases the clamp in the slider that the rubber band is connected to, then that band snaps back to the end of the slider, and would perhaps break the end stop if it was not for the damping network V1808 R1814 that prevents it snapping back too strongly. etc, well. And there is L1806 / L1807 transformer too to help feed back energy to the supply. Or something similar. Al this does not help anyone because I wonder if anyone did make it this far reading this so .. So, in short, tronix probably needs to be in the genes, Like playing ball, catch one, no thinking. Or do the math and everybody has gone home when you got your first formula, and now need wind speed, humidity, angle, weight of the ball, temperature, air pressure, .. Got it?

Sorry if I missed that, it sounded like it was removed and being tested separate from the scope.

It was. But a few people suggested it be re-tested loaded. So it's been checked both loaded and unloaded and found to be still faulty.

I'd mothballed this project for the time being, but then thought about your post and tried checking the diodes on this board. Just one is giving odd readings. It's the BAW62 diode next to the opto-isolator chip. Testing with the ESR meter showed zero ohms & 34uF. Testing again with a conventional diode checker showed a pd of .56V in one direction and .92V in the opposite direction. Doesn't make much sense. Probably would be too optimistic to say this part is malfunctioning, but I'm going to remove it from the board tomorrow for a proper check over. Fingers crossed!

This is a normal reading across a paralleled back-to-back diode and opto-emitter. This circuit produces line frequency triggering signals and has nothing to do with psu function.

RL

Oh, right. Thanks for that, Legg. I've actually started work on another Philips scope (a PM3240) which suddenly 'died' one day recently. I'd initially suspected the smps section, as power went in but didn't come out, but it turns out to be a part (or parts) on another board that smps feeds which has gone low-res and is causing the psu to shut down. Hope to track that down later today. Curiously the chopper transistor in this smps has no heat-sinking at all, which struck me as very odd, but it seems to have been designed that way. :-/

** The OP's ESR meter readings are not one tiny bit normal.

Obviously he does NOT own a ESR meter intended for in-circuit electro cap testing.

... Phil

Well this is the one I use:

And it specifically claims to be able to measure ESR in-circuit @ 100kHz. This is the first time I've *ever* seen it report 0.00 ohms ESR! So I deduce from your comment that I should remove it from circuit and test it by itself; fine I'll do that.

There are no caps in the viscinity of these diodes to which an esr test could be applied. The esr or cap reading on these nodes is meaningless.

RL

Legg, I think you must have missed an earlier post in which Phil mentioned that a cap ESR tester is also a useful tool for checking diodes. Anyway, I've removed that particular diode from the board and it tests fine as I'd kind of suspected it would, so the hunt will go on - down the road a way, though.

** You are correct - the OP has taken my advice out of context.

" You can use a capacitor ESR meter to check for shorted diodes on SMPS transformers. The combination of high frequency ( ie 100kHz) and low voltage used in the test is ideal - good diodes do not conduct and windings appear as open circuits. "

Seems he found no shorted diodes that were connected to transformer windings - which is useful info.

In any case, I would be very unhapppy about an ESR plus capacitance meter that showed a small signal diode as being a cap of 34uF with zero ESR.

The Atlas ESR70 meter he linked has enough voltage output to forward bias a diode which can lead to false readings.

The Bob Parker meter reads open ( all zeros) if connected to a resistance over 99ohms, a cap less than 15nF, an inductor over 150uH or a semiconductor junction.

.... Phil

The images you posted, if we are talking positive supply here, indicate it maxing out. That would account for the heat.

If indeed, the FIRST pic in your post is the B-E, and the SECOND is the C-E there is a severe overload. And if you can't find it elsewhere it is proba bly the transformer. Good luck.

Well, I have to back off a bit on that because that thing has shaping compo nents n shit and any one of them could be shorted and cause something like this.

Is it possible for you to get a like 200 watt audio amp and a generator and feed the transformer from that ? You could measure the current, and turn i t on and off and watch which sources drop the quickest. (if you use an audi o amp put a fan on it but you probably know that already) This could lead y ou to a short which once disconnected might have the PS up and running no p roblem.

But I doubt it. I don't like those waveforms, but you have to try. Hope I a m wrong.

Anyway, when you got the generator plugged in to the Ampzilla with a 15 KHz sine (forget square) you are controlling the amplitude. Monitor voltages a nd don't go beyond what they call for. But when you switch that off immedia tely and you see one source drop liker a rock while the others decay in a l eisurely fashion, this indicates an overload. Disconnect that and maybe you get proper operation again of the PS. If so you know where to go.

Sorry to bore you if you already did this stuff, one way or the other.

Once you really eliminate problems on the secondary, it is all them coils n whatnot or something in that circuit, or the transformer.

Those SMPS transformers do go bad that way, no shorts but their cores crack or actually pulverize in there. They lose their inductance. That means the current is ramping up but it is so inefficient it is not going to work, al so, the waveform going to any rectifiers is so fast that they are not effic ient, they were not designed for this.

So anyway, excite the transformer yourself. Find all this out without blowi ng their circuit, probably at least half unobtainium or close. Fuck all tha t.