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 !

[isw]

In article <50aa7f6d$0$1938$426a...@news.free.fr>,

cLx <clx.kat@j'aimail.com.almost.invalid> wrote:

> 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'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.

Isaac

[Franc Zabkar]

On Mon, 19 Nov 2012 19:50:17 +0100, cLx
<clx.kat@j'aimail.com.almost.invalid> put finger to keyboard and
composed:

>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
--
Please remove one 'i' from my address when replying by email.

[cLx]

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]

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]

On Tue, 20 Nov 2012 21:28:14 +0100, cLx

<clx.kat@j'aimail.com.almost.invalid> put finger to keyboard and
composed:

>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/Web/Sections/ProductCatalog/Capacitors/FilmCapacitors/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).

[amdx]

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

[Arfa Daily]

"cLx" <clx.kat@j'aimail.com.almost.invalid> wrote in message
news:50abe7e4$0$1985$426a...@news.free.fr...

I'm with Franc on this. Replace both devices and all caps near them.

IGBTs are my least favourite power switching devices. I've always found them
hugely over-priced compared to FETs, and extremely fussy about their
operating conditions. I have also been told by a friend of mine that works
with them a lot, that they are *extremely* intolerant of having their gate
terminal touched when they are not in-circuit. He says that they can be
damaged as a result, so that's maybe something else that could have been a
contributory factor in the device re-failing.

Arfa

[Robert Macy]

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.

[Jon Elson]

Arfa Daily wrote:


> IGBTs are my least favourite power switching devices. I've always found
> them hugely over-priced compared to FETs, and extremely fussy about their
> operating conditions.

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

[Arfa Daily]

"Jon Elson" <jme...@wustl.edu> wrote in message
news:-OCdnSJfR6rY3DDN...@giganews.com...

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

[N_Cook]

Arfa Daily <arfa....@ntlworld.com> wrote in message
news:nMArs.896789$ti6....@fx20.am4...

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]

Jon Elson <jme...@wustl.edu> wrote in message
news:-OCdnSJfR6rY3DDN...@giganews.com...

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

[cLx]

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

[Arfa Daily]

"N_Cook" <div...@tcp.co.uk> wrote in message
news:k8ndaa$f6c$1...@dont-email.me...

I don't know. I must admit that at the time, I *did* consider trying exactly
that, but you know what it's like when you get one of these soul-destroying
jobs. I wrestled for several days with the idea of giving it one more go
with FETs in place of the IGBTs, but in the end, the owner said that if it
was going to be a lot of trouble, and absolute future reliability couldn't
be reasonably guaranteed, then he would just scrap it. I must admit that I
then drew a deep breath of relief, bundled it back together, and cheerfully
gave him it back. Next time, maybe ...

Arfa

[Spehro Pefhany]

On Fri, 23 Nov 2012 01:59:45 -0000, the renowned "Arfa Daily"
<arfa....@ntlworld.com> wrote:

> 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
--
"it's the network..." "The Journey is the reward"
sp...@interlog.com Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com

[N_Cook]

Spehro Pefhany <spef...@interlogDOTyou.knowwhat> wrote in message
news:kr63b8prd6sisedo1...@4ax.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]

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.

?-)

[Jon Elson]

Arfa Daily wrote:


>
> 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 ?

Below 400 V there is probably no benefit to an IGBT. At 400 V and above,
MOSFETs start to show a higher on-resistance. The higher the breakdown
voltage, the higher the on resistance. IGBTs have a constant Vce drop of
about 2V or better, up to several times continuous rated current. That
starts to look real good in serious power devices like motor drives.

IGBTs in an audio amp sounds really odd to me, I'll bet a set of
FETs could have been dropped into it perfectly.

Jon

[Phil Allison]

"Jon Elson"

>
> Below 400 V there is probably no benefit to an IGBT. At 400 V and above,
> MOSFETs start to show a higher on-resistance. The higher the breakdown
> voltage, the higher the on resistance. IGBTs have a constant Vce drop of
> about 2V or better, up to several times continuous rated current. That
> starts to look real good in serious power devices like motor drives.
>
> IGBTs in an audio amp sounds really odd to me, I'll bet a set of
> FETs could have been dropped into it perfectly.

** Switching PSUs in many high powered audio amps use IGBTs - often in
parallel groups. The topology is a driven, square wave inverter - no
feedback or regulation is needed.

The dodgiest time is at first switch on, when the filter electros present a
dead short to the DC output. Usually, the drive wave begins with a very low
duty cycle ramping slowly up to full square wave. Drive frequencies are in
the order of 100kHz to 150kHz.

Readily available, low cost mosfets are not in the game.



.... Phil

[Jon Elson]

Phil Allison wrote:

>
> "Jon Elson"

>> IGBTs in an audio amp sounds really odd to me, I'll bet a set of
>> FETs could have been dropped into it perfectly.
>
>
> ** Switching PSUs in many high powered audio amps use IGBTs - often in
> parallel groups. The topology is a driven, square wave inverter - no
> feedback or regulation is needed.

OK, in the off-line DC power supply section, this makes perfect sense!
Not in the audio amp section. (You can probably tell I rarely get into
late-model consumer audio gear.)

Thanks,

Jon

[Phil Allison]

"Jon Elson"

> Phil Allison wrote:
> "Jon Elson"
>
>>> IGBTs in an audio amp sounds really odd to me, I'll bet a set of
>>> FETs could have been dropped into it perfectly.
>>
>>
>> ** Switching PSUs in many high powered audio amps use IGBTs - often in
>> parallel groups. The topology is a driven, square wave inverter - no
>> feedback or regulation is needed.
>
> 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

[jurb...@gmail.com]

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 company money.

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

[jurb...@gmail.com]

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]

<jurb...@gmail.com>

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

[jurb...@gmail.com]

>"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 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.

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, need I 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.
>>

>> 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.

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.

On 29/11/2012 07:45, jurb...@gmail.com wrote:
> 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.

[William Sommerwerck]

> ...I solved a greyscale problem by adjusting the vertical height.

When this issue has calmed down... I want to hear about that.

I grew up reading the "Model Garage" stories in Popular Science and the ones
about an electronics technician (Art Margolis?) in Radio-Electronics. These
usually revolved around a difficult problem that did not have an obvious
cause.

[Franc Zabkar]

On Wed, 28 Nov 2012 19:22:28 -0800 (PST), jurb...@gmail.com put

finger to keyboard and composed:

>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?

[Phil Allison]

<jurb...@gmail.com

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

Of course it would. But not always the same transistor.

( snip loads of puerile shit )


** Bollocks.

You have thoroughly worn out you welcome here - pal.

Fuckwit GG posters are tolerable in small doses only.

Sooooo - YOU are now fair game.

Watch out.



.... Phil

[William Sommerwerck]

You have thoroughly worn out you welcome here - pal.
Fuckwit GG posters are tolerable in small doses only.
Sooooo - YOU are now fair game.
Watch out.

For what? You're going to throw a tub of spicy hummis at him?

[Phil Allison]

"William Sommerwerck"

** That joke is tad obscure.

My remark is fair warning that the GG poster has crossed the line and needs
to pull his head in.

Or I will bite it off .....



... Phil

[William Sommerwerck]

> You have thoroughly worn out you welcome here - pal.
> Fuckwit GG posters are tolerable in small doses only.
> Sooooo - YOU are now fair game. Watch out.

> For what? You're going to throw a tub of spicy hummis at him?

** That joke is tad obscure.

It is obscure. It's a line from a movie I haven't seen, "Paranorman".

> My remark is fair warning that the GG poster has crossed the line
> and needs to pull his head in. Or I will bite it off ...

If you mean figuratively... You won't be doing much actual damage.

[Michael A. Terrell]

All Phil will ever do, is bore people to tears with his juvenile
whining.

[jurb...@gmail.com]

> All Phil will ever do, is bore people to tears with his juvenile whining.

Internet Turet's syndrome maybe. But I am not going to dwell on that. The only real power Phil has is that every once in a while he has something to say so people don't completely ignore him YET. Whatever.

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

No, that is a common misconception in the business. Do you remember the formula for capacitive reactance ? Apply it and just guess the frequency is over an octave above the sonic range.

The fact is that those caps are not being used as reactive components like in a tuned system, they are being used as coupling caps.

In that circuit they are effectively in parallel. Ground and the power supply rail are effectively at the same AC potential, so it's not 0.68, it's aready 1.36 uF. That is almost a piece of wire at 20 Khz. Almost, but we are dealing with a quite higher frequency here.

I got some pretty beefy 1 uF/400 V here but they are old and I can't be sure they can really handle the current. Actually I would like to use them in speaker crossovers someday.

Anyway, people also make this mistake working on SMPSes. Some use a coupling cap to keep DC off the transformer, OK, but that is a coupler. Think about it a sec., it LOWERS frequency to produce more output. Letting the capacitive reactance curve into that would fight against what you want to accomplish.

>"Now i'm stuck at finding the polypropylene caps. "

Digikey doesn't have anything ?

What are you dealing with here, rectified 240 volts ? At 320, 17 amps is 5,440 watts. Four burners would add up to 68 amps. What size breaker does this thing take ?

It is possible that they are special caps of course. Manufacturers love using special parts.

Either way, lots of that type of cap fail by exhibiting leakage under stress. Same with the snubbers.

When you get a circuit like that, treat is like a bridge rectifier. Balance. You change all four diodes in a bridge right ? You change both outputs in an audio amp right ? Same deal here.

Also everybody think about why SMPSes with the "totem pole" configuration usually use full wave rectification on the secondaries. Because balance is as important as it is in an audio amp. It may not seem so, but it is. If not they would save money on diodes.

[Franc Zabkar]

On Thu, 29 Nov 2012 09:31:06 -0800 (PST), jurb...@gmail.com put

finger to keyboard and composed:

>>"Wouldn't increasing the capacitance from 0.68uF to 1uF result in a 50%
>increase in cooking energy? "
>
>No, that is a common misconception in the business. Do you remember the formula for capacitive reactance ? Apply it and just guess the frequency is over an octave above the sonic range.
>
>The fact is that those caps are not being used as reactive components like in a tuned system, they are being used as coupling caps.
>
>In that circuit they are effectively in parallel. Ground and the power supply rail are effectively at the same AC potential, so it's not 0.68, it's aready 1.36 uF. That is almost a piece of wire at 20 Khz. Almost, but we are dealing with a quite higher frequency here.

Here is an induction cooker design where the capacitors and coil are
said to form a "resonant tank":
http://beaversource.oregonstate.edu/projects/44x201103/wiki/IGBT%2BCoil

The second IGBT's body diode allows the coil current to decay
gracefully when the first IGBT switches off. I forgot about that.

[Franc Zabkar]

On Fri, 30 Nov 2012 13:54:20 +1100, Franc Zabkar
<fza...@iinternode.on.net> put finger to keyboard and composed:

>Here is an induction cooker design where the capacitors and coil are
>said to form a "resonant tank":

>http://beaversource.oregonstate.edu/projects/44x201103/wiki/IGBT%2BCoil

The above design references the following document:

Induction Heating System Topology Review:
http://www.fairchildsemi.com/an/AN/AN-9012.pdf

[jurb...@gmail.com]

>"The second IGBT's body diode allows the coil current to decay
gracefully when the first IGBT switches off."

No. Don't you understand that ground is a human thing, not an electronic thing ? There is no up and down.

The only possible difference is if one side of the coil has more capacitance to ground, which would make an inbalance. This COULD happen, but if it did, the bottom Xsistr failing would be such a common failure mode it would be on Fox fucking news. Switching the leads to the inductor could prove it, because then the top Xsistr would fail first.

We are talking about a ground fault condition here, without that, no anomality in the load could be imbalanced after it is running. Failing on startup is a different story, and this ain't it.

Get a grip, or a firm base of theory. Somehting. I'm surprised people can tie their shoes,,,, oh wait, they can't.

So much for that.

[cLx]

On 29/11/2012 11:23, Franc Zabkar wrote:
> On Wed, 28 Nov 2012 19:22:28 -0800 (PST), jurb...@gmail.com put
> finger to keyboard and composed:
>
>> 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?

It's a serie LC resonnance driven by a half bridge switcher, if I'm correct.

[William Sommerwerck]

I've been following this with curiosity and amusement.

I suspect this product is a marginal design that fails more-often than it
should, but not often enough to be considered outright "defective". (The
NuWave product advertised on late-night TV seems to be plagued with similar
problems.) There is probably no way to "fix" it, short (hmmm) of a complete
bottom-up redesign. It's a shame, because a counter-top induction "burner"
is a good idea. (I almost ordered the NuWave until I learned how unreliable
it -- and its seller -- are.)

The Infinity "SWAMP" switching amplifier from the late '70s is an example of
such a marginal design. It blew its output transistors far too often. The
designer later found the problem, and admitted that adding two cheap diodes
per channel would have prevented it.

[Franc Zabkar]

On Fri, 30 Nov 2012 09:31:06 +0100, cLx <clx...@jaimail.com.invalid>

put finger to keyboard and composed:

>On 29/11/2012 11:23, Franc Zabkar wrote:
>> On Wed, 28 Nov 2012 19:22:28 -0800 (PST), jurb...@gmail.com put
>> finger to keyboard and composed:
>>
>>> 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?
>
>It's a serie LC resonnance driven by a half bridge switcher, if I'm correct.

Yes, I see that now. I should have researched the topic instead of
relying on the OP's rough circuit diagram. The absence of the flyback
diodes from the drawing left me wondering how the coil current would
decay after the IGBTs switched off, so I assumed that the current had
to be zero when this happened.

[Franc Zabkar]

On Fri, 30 Nov 2012 00:18:28 -0800 (PST), jurb...@gmail.com put

finger to keyboard and composed:

I don't understand what triggered your "ground" rant. Of course the
circuit is balanced, and of course there is no connection to terra
firma. Did I suggest anything else?

Are you perhaps misunderstanding what I meant by "second" and "first"?
All I'm saying is that after the first IGBT turns off, the second
IGBT's flyback diode allows the coil current to decay. Then the second
IGBT turns on. After the second IGBT turns off, then the first IGBT's
diode allows the coil current to decay, and so on.

[Franc Zabkar]

On Thu, 29 Nov 2012 09:31:06 -0800 (PST), jurb...@gmail.com put

finger to keyboard and composed:

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

>No, that is a common misconception in the business. Do you remember the formula for capacitive reactance ? Apply it and just guess the frequency is over an octave above the sonic range.

>The fact is that those caps are not being used as reactive components like in a tuned system, they are being used as coupling caps.

XC = 1/wC = 1 /(2 x pi x 20000 x 1.36 E-6) ~ 6 ohms

My research would suggest that a typical inductance value for the coil
would be of the order of 50uH.

XL = wL = 2 x pi x 20000 x 50 E-6 ~ 6 ohms

So XL = XC, ie resonance.

>In that circuit they are effectively in parallel. Ground and the power supply rail are effectively at the same AC potential, so it's not 0.68, it's aready 1.36 uF. That is almost a piece of wire at 20 Khz. Almost, but we are dealing with a quite higher frequency here.

If those 0.68uF capacitors were coupling capacitors, then it wouldn't
matter how large they were. In fact the larger, the better. So let's
assume for the sake of analysis that they were infinitely large. This
means that the voltage at their junction would be constant (Vsupply /
2), irrespective of the induction coil current.

So when the upper IGBT is on, the voltage across the coil would be
Vsupply - Vsupply/2 = Vsupply/2. Similarly, when the bottom IGBT is
on, the coil voltage would be -Vsupply/2. This results in a linearly
increasing coil current, first in one direction and then in the
reverse direction, ie a symmetrical triangular current.

Vsupply/2 = L . dI/dt

I would think that a triangular current would be undesirable.

[Franc Zabkar]

On Thu, 29 Nov 2012 10:21:43 +0100, cLx <clx...@jaimail.com.invalid>

put finger to keyboard and composed:

>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.

I would measure the voltage at the junction of the two 0.68uF
capacitors, with the coil disconnected, and confirm that it sits at
half the DC supply voltage. This might identify any imbalance in the
capacitor characteristics, ie leakage, reduced capacitance.

[jurb...@gmail.com]

>"I don't understand what triggered your "ground" rant..."

I didn't mean it like it sounded I guess.

Anyway I understand what you say about the Xc, but I am assuming that the frequency is much higher than 20 Khz. I'd bet it's somewhere around 75 Khz which takes the tuned circuit out of the picture. The Xc of the caps would be much lower and the Xl of the coil much higher. If the whole shebang was 12 ohms total, one burner would pull almost 27 amps providing the supply is 320 DC. The numbers fall into line at around 75 Khz or so.

At that point, there is nothing tuned about it. I didn't mean to offend, I just know that so many have a few misconceptions about these things, actually not even these things pty se, SMPSes that use a similar configuration. If you don't have those misconceptions great.

I try to think from the engineer's standpoint. Why would I waste energy on Xc ? There is a reason to do it for example in the horizontal sweep ciscuit of a TV but usually it's all or nothing. Block the DC and pass the AC.

J

[Franc Zabkar]

On Fri, 30 Nov 2012 16:20:16 -0800 (PST), jurb...@gmail.com put

finger to keyboard and composed:

>Anyway I understand what you say about the Xc, but I am assuming that the frequency is much higher than 20 Khz. I'd bet it's somewhere around 75 Khz which takes the tuned circuit out of the picture.

That Fairchild design note I alluded to in an earlier post discusses
several topologies (including the OP's), all of them based on
resonance principles. The resonant frequency was chosen as 24kHz while
the operating frequency "of the system is set at 28kHz, which is
higher than the resonance frequency, in order to avoid noise generated
within the audio frequency band".

Wikipedia also suggests that 24kHz is a common design target:
http://en.wikipedia.org/wiki/Induction_cooking

The above article tabulates the skin depth of various materials at
24kHz.

However, Wikipedia also states that "Panasonic Corporation in 2009
developed a consumer induction cooker that uses a higher-frequency
magnetic field, and a different oscillator circuit design, to allow
use with non-ferrous metals."

As for some real examples, here are two service manuals for induction
cookers:
http://www.garland-group.com/docs/uploaded/gar/products/g_ec_sm_inducrtcs_4521635.pdf
http://www.garland-group.com/docs/uploaded/gar/document_catalog/Document%20Catalog%20Files/Service%20manual%20Fajita%20Heater.pdf

Both manuals refer to a diagnostic mode whereby the microprocessor is
accessible via an RS232 interface. There are several sample outputs
which indicate that the operating frequency is 20080 Hz.

[Franc Zabkar]

On Sat, 01 Dec 2012 16:27:24 +1100, Franc Zabkar
<fza...@iinternode.on.net> put finger to keyboard and composed:

>On Fri, 30 Nov 2012 16:20:16 -0800 (PST), jurb...@gmail.com put
>finger to keyboard and composed:
>
>>Anyway I understand what you say about the Xc, but I am assuming that the frequency is much higher than 20 Khz. I'd bet it's somewhere around 75 Khz which takes the tuned circuit out of the picture.

I found this interesting application note:
http://www.st.com/internet/com/TECHNICAL_RESOURCES/TECHNICAL_LITERATURE/APPLICATION_NOTE/CD00115561.pdf

"Induction cooking functions based on the principle of the series L-C
resonant circuit, where the inductance L is the cooking element
itself. By changing the switching frequency of the high voltage
half-bridge driver, the alternating current flowing through the
cooking element changes its value. The intensity of the magnetic field
and therefore the heating energy can be controlled this way."

It describes the OP's circuit as a resonant circuit consisting of an
L-C resonant tank. It states that "the IGBTs are driven by high
frequency complementary square waves with 50% duty cycle", and that
the frequency of "the PWM signal applied to the driver input pin ...
varies in a range between 19 kHz and 60 kHz". The two capacitors are
680nF 600V types, and the IGBTs are STGY40NC60VD.

The design has 9 power levels, with a frequency of 60kHz for lowest
power, and 25kHz for highest power. At 60kHz the plate current appears
to be roughly triangular, while at 25kHz it appears sinusoidal.

[Franc Zabkar]

On Sat, 01 Dec 2012 17:12:28 +1100, Franc Zabkar

<fza...@iinternode.on.net> put finger to keyboard and composed:

>It describes the OP's circuit as a resonant circuit consisting of an
>L-C resonant tank. It states that "the IGBTs are driven by high
>frequency complementary square waves with 50% duty cycle", and that
>the frequency of "the PWM signal applied to the driver input pin ...
>varies in a range between 19 kHz and 60 kHz". The two capacitors are
>680nF 600V types, and the IGBTs are STGY40NC60VD.

The datasheet for the STGY40NC60VD states that the IGBT is good for
"high frequency operation up to 50kHz". Hmmm ...

http://www.st.com/internet/com/TECHNICAL_RESOURCES/TECHNICAL_LITERATURE/DATASHEET/CD00003462.pdf

[jurb...@gmail.com]

Very interesting Frank. You do good research. However there is no real contradiction. The inductance must be higher than we assumed because if the Xl is only 6 ohms, it would pull a hell of alot more current. It had to be one or the other.

Even more interesting is how they're making it work with non-ferrous vessels. Technically it should work with anything that conducts electricity, coils in a transformer are not ferrous and you can induce current in them right ? It actually does work on non-ferrous I guess, just extremely inefficiently. They aimed to increase the efficiency. It says reduced efficiency and I saw a chart, but it wasn't boiled down to just HOW inefficient a given unit will be on specific cooking materials - as if the layman would usually even know. It would still be trial and error.

Now if they can make it work on glass cookware, they get my vote for the Nobel prize.


Very interesting, thanks for bringing that in.

[josephkk]

Making it work with glass would impress must anybody. However i wonder if
they can be made to work on slightly salty water (saline{?}).

Just a thought.

?-)

[jurb...@gmail.com]

Actually if they can really penetrate the non-ferrous they can heat the food directly. I gues an upside down microwave or something.

Hell, it might be all the rage range now because your steak heats the pan, not the other way around.

By now you know most of that "Visions" type cookware got recalled right ? The stuff exploded on a few folks and they went back to metal pots and pans.

Funny what it takes to induce common sense sometimes eh ? My Mother had them but got rid of them over that shit. Really, they would probably be alright for boiling and casseroles etc, but I think trying to fry crispy chicken in a glass pan might not be the smartest idea.

But then lead could be put in the glass, or barium. Maybe it would glow or something.

Colored glass I guess in a way could be considered "doped". A study of the various types' reaction to bombardment like this would be intereasting.

[cLx]

On 01/12/2012 06:27, Franc Zabkar wrote:
> On Fri, 30 Nov 2012 16:20:16 -0800 (PST), jurb...@gmail.com put
> finger to keyboard and composed:
>
>> Anyway I understand what you say about the Xc, but I am assuming that the frequency is much higher than 20 Khz. I'd bet it's somewhere around 75 Khz which takes the tuned circuit out of the picture.
>
> That Fairchild design note I alluded to in an earlier post discusses
> several topologies (including the OP's), all of them based on
> resonance principles. The resonant frequency was chosen as 24kHz while
> the operating frequency "of the system is set at 28kHz, which is
> higher than the resonance frequency, in order to avoid noise generated
> within the audio frequency band".

When it was brand new, i've measured the frequency at ~35KHz, modulated by
100Hz (double rectification of the mains' frequency).

See:
http://clx.freeshell.org/view.html?f=Tech/induction/img_0809.jpg
http://clx.freeshell.org/view.html?f=Tech/induction/img_0812.jpg
http://clx.freeshell.org/view.html?f=Tech/induction/img_0813.jpg

[William Sommerwerck]

> By now you know most of that "Visions" type cookware got recalled,
> right? The stuff exploded on a few folks and they went back to metal
> pots and pans.

I've been using Corning glassware and Corningware for years, and have never
had a piece break while cooking. Dropping them is a different matter.

[jurb...@gmail.com]

>"When it was brand new, i've measured the frequency at ~35KHz, modulated by
100Hz (double rectification of the mains' frequency). "

So they don't bother filtering, like a microwave. Hmmm.

>"I've been using Corning glassware and Corningware for years, and have never
had a piece break while cooking. Dropping them is a different matter. "

Visions were different. Some sort of superglass. The commercial showed them melting a conventional metal pan in their glass pan. I would imagine it was not stainless, IIRC the melting point is quite high compared to aluminum or steel, or cast iron.

But these things didn't just break, they exploded. They are no longer on the market, at least with the old formula.

It is pretty cool to be able to see what you're cooking like that. There was always Pyrex as well. One of the things visions touted was an extremely high thermal mass. In most cases if you wanted to cook say spaghetti, once it came to a boil again with the pasta in the water, you supposedly could shut the heat off and it would cook fully.

[Michael A. Terrell]

The older stuff was real 'Pyrex'. Now they use 'Soda Glass' and call
it 'Pyrex'. Soda glass will explode if heated unevenly.

[Michael A. Terrell]

I've done that with steel cookware on a stove, and in a glass
casserole in the microwave. The must be covered to help hold in the
heat, and let them sit about eight minutes, depending on the thickness
of the pasta. In fact, I cooked some medium sized shell pasta in my
microwave a few days ago.

[William Sommerwerck]

>> I've been using Corning glassware and Corningware for years, and have

>> never had a piece break while cooking. Dropping them is another matter.

> The older stuff was real "Pyrex" [borosilicate glass]. Now they use "soda
> glass" and call it "Pyrex". Soda glass will explode if heated unevenly.

I complained about this a few years ago, and was told that the newer glass
was as good. I doubt it.

Several years ago I dropped a Visions baking dish and it shattered like you
wouldn't believe. (I'm still finding the pieces.) Corning paid the postage
to return it, but they never told me what they found (if anything). They
made no offer to replace it.

See the Wikipedia article on Pyrex, section on Composition.

[Phil Allison]

<jurb...@gmail.com>

>"When it was brand new, i've measured the frequency at ~35KHz,

> modulated by 100Hz (double rectification of the mains' frequency. "

So they don't bother filtering, like a microwave. Hmmm.

** Microwave ovens have similar line frequency modulation and for the same
reason.

The high voltage PSU is not filtered.

Electronic transformers for 12v halogen lighting are the same too.


... Phil

[Cydrome Leader]

beware that anything made in the past some years isn't what it used to be.
it's just some sort of cheapo fake stuff and not the real pyrex as used in
labware.

You can even tell from the color of the stuff that it's nothing more than
melted bottles and windows.

[Cydrome Leader]

William Sommerwerck <grizzle...@comcast.net> wrote:
>>> I've been using Corning glassware and Corningware for years, and have
>>> never had a piece break while cooking. Dropping them is another matter.
>
>> The older stuff was real "Pyrex" [borosilicate glass]. Now they use "soda
>> glass" and call it "Pyrex". Soda glass will explode if heated unevenly.
>
> I complained about this a few years ago, and was told that the newer glass
> was as good. I doubt it.
>
> Several years ago I dropped a Visions baking dish and it shattered like you
> wouldn't believe. (I'm still finding the pieces.) Corning paid the postage

I wattched a Visions lid from the large pasta pot get dropped recently
onto a wooden floor.

I would not doubt there was more energy stored in that lid than just it
falling into the ground. It din't break, it became hundreds of pieces, but
sharp ones unlike a windshield.

We noticed that those 1980s white corningware plates with the yellow
printing on the edges were unbreakable when new, but became extremely
brittle over the years. They seemed to heat in the microwave oven, so
maybe that has something to do with it.

The extra plates and bowls from the set that never really saw use can
survive a kitchen floor drop test. The rest just shatter.

> to return it, but they never told me what they found (if anything). They
> made no offer to replace it.

Did you send it to the real corning or the fake place called "world
cookware" or something like that that now just licenses the corning name?

[William Sommerwerck]

>> Corning paid the postage to return it, but they never told me

>> what they found (if anything). They made no offer to replace it.

> Did you send it to the real Corning or the fake place called "World
> Kitchen."

The latter.

It's not actually fake. According to the Wikipedia article, World Kitchen
was originally Corning's Consumer Products division.

I'm bothered by companies buying up trademarks, then applying them to
products that have no connection with the trademark's original usage. Pyrex
is a good example of this bad practice. So is Accutron. Modern Accutron
watches do not use a tuning fork.

[cLx]

Phil Allison wrote:
> <jurb...@gmail.com> wrote:

No big DC filter capacitors means you'll not have huge peaks two times by
mains cycle. Better power factor, in fact, it's almost itselfs a PFC ;)

[jurb...@gmail.com]

Actually I have a pretty novel idea I am thinking of trying to implement. An audio amp with only one output device hopefully, if I can find a fast enough triac type device (a few kinds would work). Would advance the phase of the triggering pulses according to input and switch off automatically when the HF power cycles. No filters or even rectifiers. If I can't do it a triac I can use a bridge rectifier and do the same thing with a fast SCR, or even a tansistor. The only advantage to a transistor though is that it would be switching off at zero crossing all the time. Still no filters.

It would have all the problems of a class D amp but less components and you could probably use slower output device(s) or a higher frequency.

I almost decided to patent it but I found some things that are just too damn close. Convertors for VFDs running off three phase mains, with no recitification. It was interesting anyway.

I might take a crack at it one day, the problem is designing the trigger circuit, there is not going to be a chip that'll do what I want - yet.

But we are talking more efficient than class D if I can pull it off. Also in my [discouraging] research I found there is another type of class D amp out there now, it uses active filters almost instead of analog outputs. The class D part always keeps a pretty constant voltage across the analog devices keeping their dissipation down. The manufacturer claims efficiency nearly as high as class D but without the sonic anomalies that those golden ears dudes can hear.

The days of linear operation may really be at an end. Don't even bother to rectify or filter ? Chop chop chop, nary a watt dissipated.

Enough commentary now, we can start a thread for that. I would never mention my idea except I DID find out that it is unpatentable. If perchance I ever do build the thing I will get a provisional patent and see which way the wind blows. It's only a hundred bucks.

Of course I think in that case every Eurasian engineer out there will look at it and by the time I get to market there will be three products like it, two challenges to the patent and more legal bills than I could ever make on one product.

Such is life. I need to look for something even wierder to build.

Speaking of which, I happen to own one of those induction cooketops and I would like to do something with it. Maybe I could make a weapon ? I got some microwave parts hanging around for the same reason.