Understanding a split-mode power supply.

[legg]

Hmm....

I further surmise that the failure of Q1 doesn't explain the failure of
the 8.2 Ohm resistor, because even with its specified maximum beta of
40, the switching transistor wouldn't pass enough current to burn out
the resistor.

The implication is that some other event has done for the both of them,
but surprisingly not destroyed the switching transistor.

<sigh> I thought I'd understood the sequence of events, but clearly not

Sylvia.

In the self-oscillatin flyback, base current is supplied
regeneratively by the feedback winding, to provide over-drive for
saturated switching. Beta is assumed to be somewhere between 4 and 12.

Note that the drive connection is drawn across the EB jn, so fusing of
the resistor would not turn the switch off - the voltage on the
emitter and base terinals could then be quite large - popping locally
connected parts like Q1.

Some re-examination of the actual board connections and further
examination of the feedback section (usually not very elaborate) may
offer further insight.

RL

 

[Sylvia Else]

How about Q1 being a thyristor as a crow bar on the bias of that second
transistor?

In which case, your symbol is incorrect. And would show the reason
why you're getting low ohm reading is my guess on what you call the
base-emitter and Collector being opened which is actually the M1
terminal etc..

That's just a guess of course.

The small transistor is a C1815 - an NPN, with an annoyingly difficult
to match pin-out.

Sylvia.

 

[Sylvia Else]

"No legg to stand on "




** Jesus Christ -

get fucking real you asinine fucking WOG WANKER !!

The WHOLE damn AC unit is metal encased, installed & MAINS earthed
!!!!

Leaving aside the abuse, I confess this was my sentiment as well. And
not only that, neither the outside nor inside units even get touched in
the normal way of things.

The entire board could go up in flames, if it were inflammable, and
still cause nothing more than a bad smell. Since it's outside even the
resulting fumes wouldn't be a threat.

Sylvia.

 

[legg]

The small transistor is a C1815 - an NPN, with an annoyingly difficult
to match pin-out.

Sylvia.

Looking at the package's flat surface, with leads pointing down, the
pin-out of 2SC1815 is ECB.

RL

 

[Jamie]

The small transistor is a C1815 - an NPN, with an annoyingly difficult
to match pin-out.

Sylvia.

Ok, now we're getting some where.
Just replace the 8.2R and transistor..

The tranny went because it's only rated for 50 Vceo, also
the the Vbe was taking a little hard this way with that
200KR driving it through the 560R.

when the 8.2R opened, it allowed the voltage to get
high and simply shorted it.

That's my story and I'm sticking to it!

The larger transistor is most likely ok.

http://webpages.charter.net/jamie_5"

 

[Arfa Daily]

Looking at the package's flat surface, with leads pointing down, the
pin-out of 2SC1815 is ECB.

RL

i.e. absolutely standard pinning for just about any and all Japanese TO92
transistors starting 2SA, B, C, or D ...

FWIW, the '1815 is about the commonest general purpose small signal NPN
Jappo transistor to be found anywhere, and anything similarly general
purpose will sub for it. It's generally not hard to rearrange the leads of a
differently pinned device, with a bit of sleeving on one or two legs to stop
them touching.

Arfa

 

[Sylvia Else]

Ok, now we're getting some where.
Just replace the 8.2R and transistor..

The tranny went because it's only rated for 50 Vceo, also
the the Vbe was taking a little hard this way with that
200KR driving it through the 560R.

when the 8.2R opened, it allowed the voltage to get
high and simply shorted it.

That's my story and I'm sticking to it!

Semi-plausible, though the small transistor should have been able to
sink enough current through the 200K to prevent its Vce rising outside
its limits.

Sylvia.

 

[Sylvia Else]

i.e. absolutely standard pinning for just about any and all Japanese TO92
transistors starting 2SA, B, C, or D ...

Well, to avoid P&P costs, I need to use what I already have, or what I
can buy from DSE or Jaycar.

FWIW, the '1815 is about the commonest general purpose small signal NPN
Jappo transistor to be found anywhere, and anything similarly general
purpose will sub for it. It's generally not hard to rearrange the leads of a
differently pinned device, with a bit of sleeving on one or two legs to stop
them touching.

Yes, I'd already concluded that given its position in the circuit, any
vaguely similar type will suffice, and that rearranging the leads is
practical, if not very elegant.

I've obtained a suitable resistor (not fuisible, but OK to test the
solution) and will give that a go later. Shame it's getting so hot outside.

Sylvia.

 

[legg]

Leaving aside the abuse, I confess this was my sentiment as well. And
not only that, neither the outside nor inside units even get touched in
the normal way of things.

The entire board could go up in flames, if it were inflammable, and
still cause nothing more than a bad smell. Since it's outside even the
resulting fumes wouldn't be a threat.

Sylvia.

I see that specific poster's dialog in your response, only. His actual
postings get filtered autiomatically here.

The unit is unlikely to go up in flames, for good reasons.

As a technician, you'd be advised to follow industry standards in
repair of any line-isolating assembly. One doesn't always know the
n'th reason why something was done.

A hipot is not impossible to arrange, locally, if you don't have the
equipment yourself.

RL

 

[Sylvia Else]

I see that specific poster's dialog in your response, only. His actual
postings get filtered autiomatically here.

The unit is unlikely to go up in flames, for good reasons.

As a technician, you'd be advised to follow industry standards in
repair of any line-isolating assembly. One doesn't always know the
n'th reason why something was done.

A hipot is not impossible to arrange, locally, if you don't have the
equipment yourself.

RL

My test repair with a non-fuisible resistor and an NPN transistor from
my stock seems to have worked (rather to my surprise), so I'll have to
source a fusible resistor now.

I'll take your comment about a hi-pot on board, though in fact I've only
worked on the high voltage side of the system - the transformer hasn't
been touched. Still, I suppose insulation breakdown in the transformer
could be implicated in the original failure.

Sylvia.

 

[Phil Allison]

"Sylvia Else"
"No legg to stand on "

My test repair with a non-fusible resistor and an NPN transistor from my
stock seems to have worked (rather to my surprise), so I'll have to source
a fusible resistor now.

** WES Components in Ashfield have them.

Code: " 1F8.2 " for 24 cents each.

I'll take your comment about a hi-pot on board, though in fact I've only
worked on the high voltage side of the system - the transformer hasn't
been touched. Still, I suppose insulation breakdown in the transformer
could be implicated in the original failure.

** Ignore "legg" - he is an ignorant twat.

There is no such industry practice as he alludes to in the repair of SMPSs
incorporated inside products.

Where the PSU is a module and plenty of stock is available at LOW cost to a
repair organisation - of course they prefer to just replace it, no matter
what the fault. Any excuse in the book is then dragged out to justify that.

Where such replacement is NOT possible or would be uneconomic - it get
repaired just like you have done.

BTW:

What does "legg" have to say about the way your SMPS was installed in the
outside unit of an air-con with no protection against water ingress,
condensation or invasion by any and all insect and rodent life ?

What does that do for the " safety isolation " angle - eh ???

Big joke.




...... Phil

 

[Sylvia Else]

"Sylvia Else"
"No legg to stand on "


** WES Components in Ashfield have them.

Code: " 1F8.2 " for 24 cents each.




** Ignore "legg" - he is an ignorant twat.

There is no such industry practice as he alludes to in the repair of SMPSs
incorporated inside products.

Where the PSU is a module and plenty of stock is available at LOW cost to a
repair organisation - of course they prefer to just replace it, no matter
what the fault. Any excuse in the book is then dragged out to justify that.

Where such replacement is NOT possible or would be uneconomic - it get
repaired just like you have done.

BTW:

What does "legg" have to say about the way your SMPS was installed in the
outside unit of an air-con with no protection against water ingress,
condensation or invasion by any and all insect and rodent life ?

What does that do for the " safety isolation " angle - eh ???

He might be unaware, since I posted that in a separate thread to
different groups. That thread was basically an opportunity for me to
winge about the quality of the circuit board, and give a heads up to
anyone considering buying an A/C unit.

I only posted this thread here when I had found something definitely
wrong with the PS, other than that it simply didn't work.

Legg FYI, that other thread referred to the fact that the board had
obviously suffered somewhat from being inadequately protected from the
environment and small creatures. I found a dead spider, some of its web,
and a good deal of dirt on the board.

http://members.optusnet.com.au/sylviae/cbCrop.jpg

Sylvia.

 

[Phil Allison]

"Sylvia Else"

Legg FYI, that other thread referred to the fact that the board had
obviously suffered somewhat from being inadequately protected from the
environment and small creatures. I found a dead spider, some of its web,
and a good deal of dirt on the board.

http://members.optusnet.com.au/sylviae/cbCrop.jpg

** Looks well cleaned up.

Perfectly OK to use in any device that has only FUNCTIONAL insulation
requirements.

( As opposed to "double insulation " or Class 2 safety insulation
requirements )



...... Phil

 

[Jasen Betts]

The small transistor is a C1815 - an NPN, with an annoyingly difficult
to match pin-out.

2SC1815

http://www.toshiba.com/taec/components2/Datasheet_Sync/50/6455.pdf

 

[Jasen Betts]

"Bob Parker"


** Cobblers ......

If you find the spider there you should'a looked before you sat.

 

[Sylvia Else]

Ok, as I mentioned elsewhere in this thread, I managed to get the A/C
working again by replacing Q1 and the 8.2 ohm resistor.

Actually, I'd misread it, and when I looked more carefully, it realised
it was 6.2 ohm, as is the other one on the board with the same markings
(i.e., a not blown one). Nice standard value that.

So now I have to replace the resistor with a fusible, but sourcing a 6.2
ohm fusible is problematic.

The situation is not made any easier because if the switching transistor
fails by shorting out, then after it blows the fusible will have 340V
across it.

From the suppliers I now know about (thanks, Trevor), I can get a 0.5
watt 340V 4.7 ohm fusible. I propose to put it in series with a 1.5 ohm
1 watt non-fusible. My reasoning is that the 1.5 ohm resistor will have
only 1/3 the power dissipation, and will handle twice as much. Therefore
the 4.7 ohm will go open circuit before the 1.5 ohm could get hot enough
to be a problem.

Does this stand up?

Sylvia.

 

[legg]

He might be unaware, since I posted that in a separate thread to
different groups. That thread was basically an opportunity for me to
winge about the quality of the circuit board, and give a heads up to
anyone considering buying an A/C unit.

I only posted this thread here when I had found something definitely
wrong with the PS, other than that it simply didn't work.

Legg FYI, that other thread referred to the fact that the board had
obviously suffered somewhat from being inadequately protected from the
environment and small creatures. I found a dead spider, some of its web,
and a good deal of dirt on the board.

http://members.optusnet.com.au/sylviae/cbCrop.jpg

Sylvia.

Unless it was x-posted to s.e.d. I won't see it. It's also -10C here,
with not much call for AC at the moment. I clean my central AC heat
exchanger each fall - I assume that discrete units might benefit from
similar treatment.

I think you mentioned that the unit was 10 years old; can't have been
too unsuitable an installation, if you're only having trouble with it
now. Funny how people religiously clean their computer cpu heatsinks
and fans, but expect AC units to clean themselves.

Glad to hear you've had success reviving the unit.......fusible or
flame-resistant resistors are not hard to get, with sources like
DigiKey, on-line.

By the way, what kind of scope were you using, that had Huntron-like
branch stimulation capabilities?

RL

 

[Sylvia Else]

By the way, what kind of scope were you using, that had Huntron-like
branch stimulation capabilities?

It was a low end Hameg 20Mhz that I bought in France about 25 years ago.

Sylvia.

 

[Arfa Daily]

Ok, as I mentioned elsewhere in this thread, I managed to get the A/C
working again by replacing Q1 and the 8.2 ohm resistor.

Actually, I'd misread it, and when I looked more carefully, it realised it
was 6.2 ohm, as is the other one on the board with the same markings
(i.e., a not blown one). Nice standard value that.

So now I have to replace the resistor with a fusible, but sourcing a 6.2
ohm fusible is problematic.

The situation is not made any easier because if the switching transistor
fails by shorting out, then after it blows the fusible will have 340V
across it.

From the suppliers I now know about (thanks, Trevor), I can get a 0.5 watt
340V 4.7 ohm fusible. I propose to put it in series with a 1.5 ohm 1 watt
non-fusible. My reasoning is that the 1.5 ohm resistor will have only 1/3
the power dissipation, and will handle twice as much. Therefore the 4.7
ohm will go open circuit before the 1.5 ohm could get hot enough to be a
problem.

Does this stand up?

Sylvia.

Are you absolutely sure that it is definitely 6.2 ohms ? That is a *very*
odd value, particularly for a fusible type ... FWIW, I really don't think
that the circuit would give a damn if you replaced it with a 6.8 ohm, which
is a standard value. Did you actually measure the one that's ok with an
accurate low ohms meter, and get a reading of 6.2 ? Seems to me that a grey
band and a blue band might easily be misread one for the other, with some of
the banding paints I've seen used over the years. I find that it is often
very easy to misread red for orange or brown, especially if the resistor
runs warm in normal use.

Otherwise, if it definitely is 6.2 ohms, and you really want to replace it
with exactly that value, your reasoning with making such a value in the way
that you suggest, would be quite valid.

Arfa

 

[Sylvia Else]

Are you absolutely sure that it is definitely 6.2 ohms ? That is a *very*
odd value, particularly for a fusible type ... FWIW, I really don't think
that the circuit would give a damn if you replaced it with a 6.8 ohm, which
is a standard value. Did you actually measure the one that's ok with an
accurate low ohms meter, and get a reading of 6.2 ? Seems to me that a grey
band and a blue band might easily be misread one for the other, with some of
the banding paints I've seen used over the years. I find that it is often
very easy to misread red for orange or brown, especially if the resistor
runs warm in normal use.

Blue Red Gold Gold = 6.2 ohms, 5%. It is in the E24 series, but I've
certainly never seen one before.

When I made up an equivalent, my meter gave the same reading (allowing
for tolerance) for the equivalent as it does for the identical resistor
on the board.

So, yes, I'm pretty sure.

As for whether I could substitute a 6.8, maybe I could, though I'd have
trouble getting one that has a 350 volt rating. The only supplier I know
of that purports to have them doesn't give the rating. The suppliers
that give ratings don't carry that value.

If I understood the circuit better, I'd be more comfortable about
changing the value. The mere fact that an unusual value has been used
gives me pause - maybe it's the value that's required there. Even if
another value worked, I'd not know the ramifications.

Sylvia.