advertisement


Toshiba GT20D101/102 replacements?

chartz

If it’s broke fix it!
Hi,

I have this Hexfet Micromega Tempo 2 amp with those IGBT transistor pair to replace one channel (the rest being common European BC/BD components).
GT20D101/201 (the pair is € 35)
Does anybody know what else I could use?
IRF something come to mind, but which ones?
They will have to be cheap in case something bad occurs. Then we’ll see.

Schematics here: https://www.diyaudio.com/community/attachments/micromega_ver072021-pdf.969119

This is the lesser version though, with the IRF MOSFETS. But schematics are the same.

Cheers
micromega_ver072021-pdf.969119
 
AFAIK they were the only IGBT's ever made for audio applications! JLH was commissioned by Toshiba to write a report on using IGBT's as output devices... which showed them to be slightly less good than either bipolars or mosfets but they published it anyway!

It's kinda unknown territory then.... so yeah any of the "usual suspect" complementary vertical mosfets but you may have to tweak the design to accommodate different switch on voltage range (gate-source voltage). Check out any differences to compensation as well!
 
Cheers Jez. Will look into it.

Don’t you just love good old, trusty bipolars…
 
They were supposed to have the ease of drive of mosfets and the current gain of bipolars. Tosh had to do some engineering to avoid the latchup issues with IGBTs (which paved the way for modern IGBTs which are used in nearly all modern power electronics!).
They do have some issues: pole-splitting behaviour with low impedance loads, which means they have to be over compensated compared to mosfets. I can't remember if they need thermal compensation - I probably ought to read the apps note Jez mentioned again. I worked with Steve Langdon (I think Steve's name is on the AN) at Tosh when JLH did the IGBT testing. Wish I had loaded up on all the JFETs and bipolar audio devices at the time, instead of just getting a few for my own use. Wish I had met JLH as well!!

As far as replacements go - how about some of the Fairchild mosfets that the diyaudio chaps rave about?

I might have a couple of lightly used pairs of the IGBTs somewhere.


BTW, during "allocation" of semiconductors IGBT stands for "Ignore Great Britain Totally"! :D
 
They were supposed to have the ease of drive of mosfets and the current gain of bipolars. Tosh had to do some engineering to avoid the latchup issues with IGBTs (which paved the way for modern IGBTs which are used in nearly all modern power electronics!).
They do have some issues: pole-splitting behaviour with low impedance loads, which means they have to be over compensated compared to mosfets. I can't remember if they need thermal compensation - I probably ought to read the apps note Jez mentioned again. I worked with Steve Langdon (I think Steve's name is on the AN) at Tosh when JLH did the IGBT testing. Wish I had loaded up on all the JFETs and bipolar audio devices at the time, instead of just getting a few for my own use. Wish I had met JLH as well!!

As far as replacements go - how about some of the Fairchild mosfets that the diyaudio chaps rave about?

I might have a couple of lightly used pairs of the IGBTs somewhere.


BTW, during "allocation" of semiconductors IGBT stands for "Ignore Great Britain Totally"! :D

What are those then? I rarely visit that site these days.... I do have some of the nice Toshiba audio mosfets luckily, 2SK1530/2SJ201:)
 
Can't remember TBH. I think they are used in the "Alpha big boy amp", or something like that.
I also don't visit that site much, the noise to signal ratio is off the scale!
I don't feel the need for 40A+ devices so don't really pay attention. My CA15RLYs probably don't like anything more than 5A poked up them!

Sorry, not very helpful :(.
 
I'm sure google can help!:)

IIRC Sanken make some audio mosfets now also. I downloaded the datasheets a while back... but obvs can't remember what the type numbers are now:rolleyes:
 
654248-EC-141-E-48-F6-B5-F8-E54-DD657-A55-E.jpg


Actual amp. See the damage. Failed repair attempt by so-called tech. :p
No proper heatsink (yeah the box I know).
I see at least 6 kaput transistors without even testing…
 

There's various types in the series. The big problem with vertical power mosfets is that the seemingly amazing ratings are not what they seem as they are based on use as a high speed switch in SMPS applications and very much not on linear applications. It has generally got worse over the years as headline specs have improved and if you check out the datasheets for power mosfets you will find that most have NO rating for DC SOA. This series of (expensive!) devices from IXYS have extended DC SOA and are much more suitable for amplifiers, regulators and power sinks for testing PSU's.

I discovered them about a year ago when needing a device as a series pass for a high voltage regulator which under some conditions would have to stand off around 300V whilst passing about 500mA for maybe 5-10 seconds. I found a device in this series that was rated for about 800 or 1000V and about 20A (not at the same time obviously!) and had the rated SOA to do it. It did the trick! At a price.... about £15 for the one device IIRC.
 
Interesting because as Jez said, modern VMOS are mostly useless for a Class AB.
Getting a P channel is even harder
 
So what should I use? Should I stick to the original IGBT, still available but unreliable and relatively expensive?
I really want to use MOSFETS here.
 
So what should I use? Should I stick to the original IGBT, still available but unreliable and relatively expensive?

That's the $64000 question, as they say.... (my post above is more about ratings and reliability of mos devices generically than a specific answer to your questions).

It was designed for the IGBT's and obvs if all else is working OK fitting the correct devices should get it going.

IF the unit you have is indeed basically the same as the schematic above then the IRF540/IRF9540/whatever is likely still available as are similar devices such as IRF9240 etc (and many others) but it is likely that many small differences are present in the two units re compensation and maybe bias generator/thermal compensation component values. If compensation is well out then it is very likely to blow up on switch on so you really could do with current limited bench supply whilst getting it right.

To cut to the chase.... if you just want to get it going again for quick resale then £70 (£35? not sure if both channels are blown) for the original output devices is likely the quickest/easiest/least hassle method.
 
No no I don’t sell anything. It’s an addition to my collection of broken/fixed hi-fi (call me Ebenezer if you will :D). Plus I have the Stage 3 player – a keeper. I also intend to actually use the thing now and then.
You know, when I got the big(ish) SE-A3, I wasn’t expecting much but once up and running it’s a nice power amp, so…
I can adapt the circuit I think. Only the last stage is different – the big ERO caps aren’t on the mosfet version for instance. The rest is totally identical, with higher voltage power rail that’s all.
Thanks for your thoughts.
 
No no I don’t sell anything. It’s an addition to my collection of broken/fixed hi-fi (call me Ebenezer if you will :D). Plus I have the Stage 3 player – a keeper. I also intend to actually use the thing now and then.
You know, when I got the big(ish) SE-A3, I wasn’t expecting much but once up and running it’s a nice power amp, so…
I can adapt the circuit I think. Only the last stage is different – the big ERO caps aren’t on the mosfet version for instance. The rest is totally identical, with higher voltage power rail that’s all.
Thanks for your thoughts.

Big electrolytics etc ain't important.... the killer differences are likely to be in small ceramic or film caps in the pF range plus the values of certain resistors. Note mosfet the wrong way round in schematic also!
 
Ah yes you’re right, I hadn’t seen this. Hawkeye!
Small compensation caps, yes.
So, Vishay or Japanese MOSFETS?
I’ve used IR IRFs in the past, but that was a long time ago (1985). They were pretty unreliable too.
 
Ah yes you’re right, I had overlooked this. Small compensation caps, yes.
So, Vishay or Japanese MOSFETS?
I’ve used IR IRFs in the past, but that was a long time ago (1985). They were pretty unreliable too.

I would be VERY wary of that schematic in fact.... There are many more film and ceramic (and tants) caps visible in the photo than shown on the schematic... and as to the filter rolling off everything above 1KHz at the input on the diagram:eek:
 
Yes, it’s just a working base on which I’ll redraw the other amp’s schematics, and decide on the MOSFET substitution feasibility.
C1 is 150 pF.
Wary is my other name, don’t worry :p
 


advertisement


Back
Top