The whole SMPS thing...

[tvtech]

OK. We have very valuable Members here. Some extremely good minds. Good as in the best on the Internet. Par Excellence. In other words people only reply to stuff they know. Properly.

So the whole SMPS thing.....have any of you actually built a SMPS that runs off Mains??.

And not only one...but a few??. I am talking maybe to production. I ask this because SMPS these days is being touted as "easy". Every Tom, Dick and Harry thinks they can do it..

Trying to stop a stuff up before it goes Viral...

Regards,
tvtech

 

[Nigel Goodwin]

. I ask this because SMPS these days is being touted as "easy".

It's 'easy' as long as you have a suitable transformer

The rest is basically the manufacturers datasheet - it's sourcing the transformer that's the problem.

 

[jpanhalt]

Trying to stop a stuff up before it goes Viral...

Regards,
tvtech

Not quite sure where you are going with this thread. I have designed and built three one-offs, which were based on the LM2678, LM3489, and MC33167. Each used an isolation transformer and worked like the datasheet said. That was 6 or 7 years ago. I wasted quite a bit of time trying to do it without an inductor -- I even used a relay at one point. Once I convinced myself to follow conventional wisdom, it was easy. Moreover, if you follow the datasheets and design nomographs, it is easier than following a Julia Child's recipe.*

Are you asking about non-mains-isolated SMPS?

John
*

 

[tvtech]

Hi John

It's all about building 220/230V isolated SMPS. Like 220 in and isolated rails otherside...
Maybe 12V and 5V or whatever...

The question though....has anyone here actually built one..

A switcher that is???

 

[jpanhalt]

The question though....has anyone here actually built one..

A switcher that is???

See post #3:
Switcher: https://www.ti.com/product/lm2678
Switcher: https://www.onsemi.com/pub_link/Collateral/MC34167-D.PDF

The LM3489 is hysteric, which is similar.

John

 

[ronsimpson]

The question though....has anyone here actually built one..

YES

Many times yes. High volume.

You need UL CSA VDE and more. Now Tom, Dick and Harry can't do it.

 

[JimB]

I have never built an SMPS.

I have attempted to repair them several times, usually with no schematic available.
I may even have succeeded once.

My general approach to faulty SMPS is that they are like dead parrots...
... not repairable.
Maybe an overly pesimistic view, but one which saves me from battering my head on a brick wall too often.

I have of course used them in my day job many times.
But then they were always brand new, fresh out of the box.


JimB

 

[Nigel Goodwin]

My general approach to faulty SMPS is that they are like dead parrots...
... not repairable.
Maybe an overly pesimistic view, but one which saves me from battering my head on a brick wall too often.

I've repaired MANY thousands, since the first domestic one appeared in the Thorn 3000 series TV's

Not to difficult to repair (usually), but you need to understand that it's like a DC coupled transistor amplifier - and fails like a chain of dominoes

 

[tvtech]

Hi Guys

As a tvtech I repair these things on a daily basis too. Also thousands of them under my belt. Thing is on TV's, they are accessible, and not tightly packed with components like Computer PSU's.

No room to test or work on normal SMPS in any form. Indeed a throw away item once it goes faulty. The only SMPS I am prepared to work on are those that are part of a TV PCB and I can test/work and see what I am doing. Otherwise its a waste of time.

Jim has a valid point.

Regards,
tvtech

 

[Nigel Goodwin]

Thing is on TV's, they are accessible, and not tightly packed with components like Computer PSU's.

Computer PSU's are badly designed Chinese crap - obviously made to be 'throw away' components, and as they are so cheap to buy, that's what they are.

I have repaired occasional computer PSU's, a common problem (as with any SMPSU) was the startup resistors going high - but they are near impossible to work on.

 

[tvtech]

Computer PSU's are badly designed Chinese crap - obviously made to be 'throw away' components, and as they are so cheap to buy, that's what they are.

I have repaired occasional computer PSU's, a common problem (as with any SMPSU) was the startup resistors going high - but they are near impossible to work on.

Spot on Nigel. My thoughts exactly

Regards,
tvtech

 

[ronsimpson]

From what Nigel said:
The cap blows, shorting out the +12V output.
The diode(s) can't survive under this condition so they overheat and short out.
The MOSFET can't take the over current and pops it tip.
This leaves the current sense resistor across the power line and it opens up to save the fuse.
If you don't understand the 'chain of dominoes' then fixing two of the problems won't fix the supply.

Many times a diode has a transformer (more or less) across it. So good and bad diodes may show near zero ohms.
I have used a 100khz 10volt square wave from a signal generator across the primary of the power transformer to help hunt down shorted diodes. This is much like measuring "AC resistance".

Most electronics is not designed to be repaired. It is designed to be low cost. Many factories have a very low failure rate. To the point where the bad boards are not repaired. The PCB is short/open tested. Each component is tested before being placed on the PCB. The stuffed board is tested with out power applied. If there is a bad component the test machine points at it. At this point it is going to work.

Sorry these test machines are not at a repair shop in South Africa.

 

[fezder]

I have used a 100khz 10volt square wave from a signal generator across the primary of the power transformer to help hunt down shorted diodes. This is much like measuring "AC resistance".

hmm, care to tell bit more, like how excatly this test is made?

 

[ronsimpson]

Applying 110VAC to your broken power supply is dangerous.
I apply a signal from a signal generator (which current limits). This will not smoke parts.
I apply a signal across the primary. The secondaries should charge up to some small voltage if the diodes are good. If a diode is shorted you will not have smoke! The secondary will not charge up.

 

[fezder]

Thanks for tip!

 

[tvtech]

I always use a Variac when starting up a rebuilt supply. Saves a lot of blown spares and time if you have not nailed every faulty component in the course of the repair.

These things start oscillating at around 90 to 130VAC normally and then quickly stabilize.....

If you start it up with a Variac and you are at say 150VAC and buggerall is happening, you still have a fault somewhere. I switch off and re-check everything.

Saved lots of time and money over the Years by this simple approach.

Regards,
tv

 

[Nigel Goodwin]

I always use a Variac when starting up a rebuilt supply. Saves a lot of blown spares and time if you have not nailed every faulty component in the course of the repair.

I've never done that, as many supplies require a switch-on 'pulse' in order to start, something starting with a variac doesn't provide.

We did use a variac on one of the old Sony models, possibly the disastrous KV1810?.

 

[tvtech]

I've never done that, as many supplies require a switch-on 'pulse' in order to start, something starting with a variac doesn't provide.

We did use a variac on one of the old Sony models, possibly the disastrous KV1810?.

Oh well, luckily I am a tvtech that only really knows CRT SMPS's. The Variac works in all cases with both the latest and oldest SMPS we deal with.

That's all I can say.

Regards,
tv

 

[ronsimpson]

There is much talk about how SMPS (switch mode power supplies) are the thing that dies too much.

There are many reasons: poor design, hot parts, restricted air flow, cheep parts, bad solder, etc

In a TV the vertical amplifier runs at 30 volts all day and night. The horizontal pulse runs at 800 to 1000 volts consistently.
BUT
The power supply is designed for 110/220 volts. It is connected to a wire that runs directly outside the house and connects to the top of a pole. Then the wire spreads out over most of the country. God sits in a cloud and pounds the wire with 10,000,000 volts. AND Fools at the power company do (God only knows what) something to make spikes, interruptions, brown outs, and general chaos. (depends on which country)

It is amazing that power supplies survive at all.

 

[Nigel Goodwin]

Oh well, luckily I am a tvtech that only really knows CRT SMPS's. The Variac works in all cases with both the latest and oldest SMPS we deal with.

I was talking about CRT sets