Induction Cooking Table : IGBT keeping to short !

[cLx]

Hello,

A day my induction cooking table did not work anymore (an Brandt TI302BS1).
Opened it, saw a couple of IGBT in bridge configuration (two IRGP4068D), and
the bottom one was shorted. Got some spares, replaced the shorted one, power
on : OK.

Two steaks later (in fact, 1 1/2 uses), got some unsuspected shutdowns, and
sooner, the same IGBT shorts again. What I need to check before daring to
replace the transistor and to retry to power on the beast again ?

I've traced a bit the schematics, but it's obviously incomplete :
http://clx.shacknet.nu/random/IMG_5802.JPG

And a little picture of the board :
http://clx.shacknet.nu/random/IMG_5803.JPG

Thanks !

 

[Franc Zabkar]

A day my induction cooking table did not work anymore (an Brandt TI302BS1).
Opened it, saw a couple of IGBT in bridge configuration (two IRGP4068D), and
the bottom one was shorted. Got some spares, replaced the shorted one, power
on : OK.

Two steaks later (in fact, 1 1/2 uses), got some unsuspected shutdowns, and
sooner, the same IGBT shorts again. What I need to check before daring to
replace the transistor and to retry to power on the beast again ?

I've traced a bit the schematics, but it's obviously incomplete :
http://clx.shacknet.nu/random/IMG_5802.JPG

I don't have any experience with these appliances, but my approach
would be to replace both IGBTs and all the capacitors, especially the
two 0.68uF film or polypropylene (?) types.

- Franc Zabkar

 

[cLx]

I'd go with the "experience tells me that many high-tech appliances are
designed by idiots" notion, and just replace the failed devices with
more robust ones.

I thought about that, but i'm afraid to choose something which would present
more gate capacitance than the originals IRGP4068D. I've closely inspected
the Kapton isolation, but new thermal paste under it already.

cLx

 

[cLx]

I don't have any experience with these appliances, but my approach
would be to replace both IGBTs and all the capacitors, especially the
two 0.68uF film or polypropylene (?) types.

Thanks for your advice. It's polypropylene caps I guess. There is a picture :
http://clx.shacknet.nu/random/IMG_5956.JPG
Any recommended source ?

I'll also change the two "snubber like" 47nF MKP capacitors (mounted from the
common node of transistors and coil to both supply rails).

 

[Franc Zabkar]

It's polypropylene caps I guess.

There is a picture :
http://clx.shacknet.nu/random/IMG_5956.JPG

Any recommended source ?

This appears to be the manufacturer's product page:
http://www.epcos.com/web/generator/...s/MetallizedPolypropylene/Page,locale=en.html

Curiously, the construction is "MFP" which is Metallized Polypropylene
Film, but the datasheet for the type, B32669, lists it as "MKP", and
the lowest capacitance is 1uF.

I'd try Mouser, Digikey, Farnell, RS Components.

I'll also change the two "snubber like" 47nF MKP capacitors (mounted from the
common node of transistors and coil to both supply rails).

- Franc Zabkar

 

[amdx]

Hello,

A day my induction cooking table did not work anymore (an Brandt TI302BS1).
Opened it, saw a couple of IGBT in bridge configuration (two IRGP4068D), and
the bottom one was shorted. Got some spares, replaced the shorted one, power
on : OK.

Two steaks later (in fact, 1 1/2 uses), got some unsuspected shutdowns, and
sooner, the same IGBT shorts again. What I need to check before daring to
replace the transistor and to retry to power on the beast again ?

I've traced a bit the schematics, but it's obviously incomplete :
http://clx.shacknet.nu/random/IMG_5802.JPG

And a little picture of the board :
http://clx.shacknet.nu/random/IMG_5803.JPG

Thanks !

Sometimes a call to the company or the service center will get an
answer like, " oh ya, you need to replace #%&#@ or it will keep shorting
the bottom one.
Mikek

 

[Robert Macy]

Hello,

A day my induction cooking table did not work anymore (an Brandt TI302BS1).
Opened it, saw a couple of IGBT in bridge configuration (two IRGP4068D), and
the bottom one was shorted. Got some spares, replaced the shorted one, power
on : OK.

Two steaks later (in fact, 1 1/2 uses), got some unsuspected shutdowns, and
sooner, the same IGBT shorts again. What I need to check before daring to
replace the transistor and to retry to power on the beast again ?

I've traced a bit the schematics, but it's obviously incomplete :http://clx.shacknet.nu/random/IMG_5802.JPG

And a little picture of the board :http://clx.shacknet.nu/random/IMG_5803..JPG

Thanks !

From memory, shorting a bipolar structure has significance. Like, over
current blows the 'emitter' wire off - OPEN. And, over voltage
avalanches between the collector and the emitter, welding a small
nonsemiconductor material connection - SHORT. Anything that makes you
have an overvoltage, like a snubber circuit, or ?? might be 'killing'
your part. If the esr goes way up on your DC filter cap...?

Plus, I've had too many designers count on the two components to be
somewhat matched, come from the same lot, have similar switching
characteristics. Thus, replace both, each time.

Don't forget proper handling of the component when you replace. Static
discharges, especially during winter months, can be pretty robust May
not kill the part today, just weaken it enough to die tomorrow.

 

[N_Cook]

What I've never really understood, is what supposed advantages IGBTs have
over FETs. A while ago, I had the misfortune to work on a switcher that was
in a Yammy powered speaker. It used a pair of IGBTs that were about eight
quid apiece from memory. Much like the OP's induction heater, it would run
fine for a while, and then the stupid things would just destroy themselves
again - and I mean violently, legs actually blown off, and all that good
stuff. I mend a lot of switch mode power supplies, and for the most part,
their failure modes and what needs to be done to obtain a reliable fix, hold
no mysteries for me. But after about the fourth set of devices that I put in
the Yammy, along with just about every other component in the surrounding
area, I gave up on it. Many switchers of similar size and ratings, use an
almost identical topology, but with a pair of FETs. Typically a couple of
quid apiece, and in my estimation, *much* more robust. So with power FETs
rated to 800 volts and staggering amounts of amps being readily available
almost for pence, why would anyone actually design with IGBTs ?

Arfa

Is there any rules for substituting powerFETs for IGBTs in such situations
that you had there, assuming you are not averse to going against the
designer's wishes?

 

[N_Cook]

Arfa Daily wrote:



The big problem with IGBTs is they are a parallel structure of many
bipolar transistors, with POSITIVE temperature coefficient. So, they
do not balance current across the die well. When driven hard into
saturation, they balance better. So, the secret an IR app engineer
told me years ago is you have to drive them hard into saturation,
and never allow them to stay in the linear region for more than
a few ns, if possible. (You can only do so much of this during
turnoff, however.)

Jon

Thanks for the insight, I'll try to remember the warning

 

[cLx]

Arfa Daily wrote:



The big problem with IGBTs is they are a parallel structure of many
bipolar transistors, with POSITIVE temperature coefficient. So, they
do not balance current across the die well. When driven hard into
saturation, they balance better. So, the secret an IR app engineer
told me years ago is you have to drive them hard into saturation,
and never allow them to stay in the linear region for more than
a few ns, if possible. (You can only do so much of this during
turnoff, however.)

Jon

I did not designed the gate control circuit on that applicance... I hope
there is no problem there.

 

[Spehro Pefhany]

So with power FETs
rated to 800 volts and staggering amounts of amps being readily available
almost for pence, why would anyone actually design with IGBTs ?

The main reason is that they're considerably cheaper when you need
both high voltage rating _and_ high current rating. That's because
they use less silicon die area. On consumer products the cheapest
component that will do the job acceptably well usually gets designed
in. Try pricing a 30A 800V MOSFET vs. a similar rating in a IGBT.. the
MOSFET will probably be 5x the price.


Best regards,
Spehro Pefhany

 

[N_Cook]

The main reason is that they're considerably cheaper when you need
both high voltage rating _and_ high current rating. That's because
they use less silicon die area. On consumer products the cheapest
component that will do the job acceptably well usually gets designed
in. Try pricing a 30A 800V MOSFET vs. a similar rating in a IGBT.. the
MOSFET will probably be 5x the price.


Best regards,
Spehro Pefhany

http://www.speff.com


Assuming there is space to put 2x powerFETs in place and the switching f is
not beyond sensible powerFET territory, other than adjusting the gate drive
, any other considerations ?

 

[josephkk]

http://www.speff.com


Assuming there is space to put 2x powerFETs in place and the switching fis
not beyond sensible powerFET territory, other than adjusting the gate drive
, any other considerations ?

Since IGBTs tend to be rather slow, power fet speed is not normally the
issue. Just the same IGBTs almost may be called IG-GTOs.

?-)

 

[Phil Allison]

"Jon Elson"

Phil Allison wrote:
"Jon Elson"


OK, in the off-line DC power supply section, this makes perfect sense!
Not in the audio amp section.

** But no body said that they were used there, you misread Arthur's post.

" A while ago, I had the misfortune to work on a switcher that was
in a Yammy powered speaker."

The "switcher " is the PSU.

..... Phil

 

[[email protected]]

I am not responding to you here Phil, but the whole bunch.

Everyone wants to get into all this esoteric shit here and really an induction top doesn't run at microwave speeds. Simple RF. I mean television SMPS RF.

The bottom keeps shorting, well lessee here, what can cause one transistor in a totem type pole circuit to short ? This is a switcher. Can an imbalance cause a problem ?

First of all the drive. Now you might not be able to compare the drive to the top transistor without diff inputs or all that, but yuo should stil be able to tell. If the duty cycle is 50/50 then the E-B or S-G voltage will read the same on a voltmeter, which can be floated easily.

Once a 50 % duty cycle is conformed then the DC reading confirms equal drive. I ASSUME both transistors were changed at the same time. If not, shame on you.

There will be some snubber caps, and then there are the coupling caps. If it ain't drive it is load, PERIOD.

You got the bottom Xstr shorting, look at the TOP coupling cap. Leakage there would never bother the top Xstr. But the bottom might have a problem with it. The capacitor decides to be a resistor. C'mon you old folks, tell everyone about it.

Know what else ? If you can't get the same rating caps, so what ? Most likely they were chosen by price. Engineers might even make adjustments in the operating frequency to accomodate lower value caps, if it saves the companymoney.

You got 0.68uF and you are afraid to use 1uFs ? Just use them as long as the current capabilities are up to snuff. The value does not mean shit as long as it doesn't go too low.

Use whatever, within reason, just make them equal. They must be equal, otherwise the engineer would have just used one cap instead of two.

Kapeesh ? Think of it from their end.

J

 

[[email protected]]

To add;

The top snubber cap (or transistor) is just as likely to be a problem. This cap's value is a bit more critical. For snubbers I would use the original value. The couplers don't matter.

J

 

[Phil Allison]

<[email protected]>

I am not responding to you here Phil, but the whole bunch.

Everyone wants to get into all this esoteric shit here and really an
induction top doesn't run at microwave speeds. Simple RF. I mean television
SMPS RF.

The bottom keeps shorting, well lessee here, what can cause one transistor
in a totem type pole circuit to short ?


** An intermittent fault ( ie internal short) in the load.

That induction heating coil is a non- trivial device running at high voltage
& frequency.

The briefest short or arc between adjacent turns would take out one or both
IGBTs instantly.

BTW:

Stop being such a PITA jerk.



..... Phil

 

[[email protected]]

"An intermittent fault ( ie internal short) in the load. "

Of course it would. But not always the same transistor. As a matter of fact, if it is the final load doing it, both transistors would usually short. You saw the circuit. Or are you all talk and no analysis ? Can you REALLY read the fucking schematic to a basically clipping SASEPP audio circuit from the 1970s ?

Now if there is an imbalance in the coupling or snubbers or even the devices, the device on the good side will usually fail first. Do you disagree ?

I want to know if both devices were replaced. Don't you ? Before you battlewith me that is. I would hate to find out later I was right. It's not all that fair because it has happened so much on my life.

You wouldn't believe, but I got witnesses all over the place. Now be nice.

Look Man, on a fucking TV I solved a greyscale problem by adjusting the vertical height. This is no bullshit I got witnesses. It involved a Sony, needI say more ? Want the technical details just ask, but I am sure you, nor anyone else on this forum would have been able to solve it. Anyone.

I will admit I am a PITA, but I recommend you do not engage me. You are good, but if you want to take me on, pack a lunch. I can design the shit you try to fix.

J U R B Yeah that's me. Back from the 1990s and shit. Think I been forgetting shit all this time ?

T

 

[cLx]

An intermittent fault ( ie internal short) in the load.
I've got a look at the coils, seem OK, not a trace of burn, and the
turns are isolated by section to distribute the voltage equally.

I want to know if both devices were replaced. Don't you ? Before you
battle with me that is. I would hate to find out later I was right.
It's not all that fair because it has happened so much on my life.

On my first and only test, I've changed only the shorted IGBT, ie. the
bottom's one. Worked, but the top got tripping two or three times and
failed again. Now I've removed the both IGBT, the 3 polypropylenes caps,
and the both MKT snubbers for replacement. I've also inspected the
kapton isolation behind the IGBT, saw no damage here.

Now i'm stuck at finding the polypropylene caps. They are expensives,
and what i found can handle only 17 amps. Seem a bit underrated for such
that case.

Doing measurements is not easy as, without load or transistors, the
drive shuts itselfs (not detecting anything), just trying shorts pulses,
and i would not wasting my replacements parts.

 

[Franc Zabkar]

You got 0.68uF and you are afraid to use 1uFs ? Just use them as long as the current capabilities are up to snuff. The value does not mean shit as long as it doesn't go too low.

Wouldn't increasing the capacitance from 0.68uF to 1uF result in a 50%
increase in cooking energy?

Also, wouldn't each capacitor take longer to charge, and if the
capacitor wasn't fully charged when the IGBT switched off, wouldn't
this result in an interruption of the current in the coils, with a
potentially damaging back-EMF? Or am I completely misunderstanding how
this appliance works?

- Franc Zabkar