Amp's power MOSFET died. Can I just remove it?

[DaveC]

I recently repaired a couple of high-power MOSFET amps. Take a look to see
if there are low-value resistors in series with each MOSFET's drain, before
they are paralleled. If the sources and drains are paralleled without any
resistors, then you have to replace all three MOSFETs, with matched units;
otherwise one MOSFET will hog all the current and blow up. (For some
reason, for a period of time, designers had got it into their heads that
MOSFETs were immune to this problem, and so they stopped using the
precautions they had used for BJT's. Unfortunately, this was a mistake.)
The sad news is that to get three matched MOSFETs, you're going to have to
buy a dozen or so and then hand-pick the best-matched ones.

I notice that MOSFETs look like they've been replaced. Each set of 3 are
marked in pen with the same 2-digit number (ie, "21", "29", etc.). I presume
this indicates that they are a matched set.

Make sure that the gate resistors are still intact (and the rest of the gate
drive circuitry). Sometimes when a MOSFET fails it takes out the gate
resistor too; and then, when you connect up the replacement without fixing
the gate resistor, the new MOSFET has a dangling gate, meaning it waits a
random period of time and then turns full on, thus blowing itself up again.
Ask me how I know.

Very useful caution. I'll go slowly.

You may find that it is hard to get exact replacements to your MOSFETs: some
of the ones used in older audio amp designs are no longer commonly sold.
For one set of MOSFETs, I had excellent results buying equivalents from
Profusion PLC, in the UK (even though I'm in the USA, they were very
inexpensive and low-hassle, and shipped incredibly quickly).

I checked out Profusion. Indeed, they have MOSFETs that will work. As you
state, problem is they sell min quantity of 20. Which means it'll cost me
US$240 to replace one defective MOSFET (and it's matched siblings).

IF you're lucky enough that just one of the MOSFETs is blown, then yes, you
could just remove it and operate at lower power. Indeed, there's probably
not even any point in removing one from the other side.

Hoping for the best case scenario...

Thanks,

 

[Walter Harley]

I checked out Profusion. Indeed, they have MOSFETs that will work. As you
state, problem is they sell min quantity of 20. Which means it'll cost me
US$240 to replace one defective MOSFET (and it's matched siblings).

Hmm, peculiar; I think I only bought four, and they weren't that expensive.
(For the amp that I bought the MOSFETs from Profusion for, I didn't need
matching; it was a single push-pull pair per channel.)

But if you need them matched, you're going to need to buy at least a dozen
to get a good matched triple, IMO.

What kind of MOSFETs are they?

 

[DaveC]

Hmm, peculiar; I think I only bought four, and they weren't that expensive.
(For the amp that I bought the MOSFETs from Profusion for, I didn't need
matching; it was a single push-pull pair per channel.)

I e-mailed them and they did offer to sell individual quantities below their
minimum order. Yay!

They are recommending ECF20N25 and ECF20P25 as proper substitutes.

But if you need them matched, you're going to need to buy at least a dozen
to get a good matched triple, IMO.

But now here is a minimum quantity, "self-imposed". Boo!

I've found several references to MOSFET matching to avoid "current-hogging".
Most seem quite complex:

http://www.altavistaaudio.com/MosfetMatching.html
http://www.passdiy.com/howto/matching.htm

although this one seems pretty straightforward:
http://sound.westhost.com/transistor-matching.htm

I'm looking for similar D-S current flow at a given G voltage, correct? (Or,
as the latter URL suggests, the G voltage required to establish a quiescent
D-S current...)

What kind of MOSFETs are they?

Originals are 2SK1221 (TO-3) and its compliment (number not at hand right
now...)

Thanks,

 

[Walter Harley]

I've found several references to MOSFET matching to avoid

"current-hogging".

If they are in parallel it is quite necessary. If they have source
resistors, it is less necessary, because those provide a form of current
feedback.

I built myself a very simple rig to do the matching, but I don't remember
offhand the component values I used. I'll try to post the schematic later
tonight. It basically hooks the MOSFET up in a two-terminal current-sink
configuration (gate connected to drain) and sees how much current it draws.
In principle you're trying to match two curves, so you want to measure Ids
for a lot of different Vgs and get the best overall match. But in practice,
using a single point has worked okay for me.

Do you have a dummy load to test the amp against, once you've got it fixed?

Originals are 2SK1221 (TO-3) and its compliment (number not at hand right
now...)

Yeah, it's hard to find TO-3 replacements. You probably won't do better
than Profusion PLC. (And, I don't have any matched equivalents for those,
sorry!)

 

[DaveC]

I'll try to post the schematic later tonight.
Thanks!

Do you have a dummy load to test the amp against, once you've got it fixed?

Someone suggested a couple of Walmart electric heating coils in series. Sound
good?

Yeah, it's hard to find TO-3 replacements. You probably won't do better
than Profusion PLC. (And, I don't have any matched equivalents for those,
sorry!)

Profusion now says that the MOSFETs they will sell me are, due to good QC,
closely matched, and are good enough for parelleled service.

Thanks,

 

[John Woodgate]

Someone suggested a couple of Walmart electric heating coils in series.
Sound good?

No. You need an **8 ohm** resistor, of at least 250 W rating. (Assuming
you amp needs an 8 ohm load, which most of them do.)

 

[DaveC]

No. You need an **8 ohm** resistor, of at least 250 W rating. (Assuming
you amp needs an 8 ohm load, which most of them do.)

The person suggesting this solution stated that the heating coils (I think
they are actually electric tea kettles, IIRC) were each about 4 ohms. Not
exactly, but close, and a resistive load. Two in series seems quite a good
idea. And I can fill them with water to dissipate the energy, and make a pot
of tea, as well

I'll look around for a large wirewound resistor, but 200 W? I don't hold out
much hope, even here in Silicon Valley...

Thanks,

 

[Roger Hamlett]

The person suggesting this solution stated that the heating coils (I think
they are actually electric tea kettles, IIRC) were each about 4 ohms. Not
exactly, but close, and a resistive load. Two in series seems quite a good
idea. And I can fill them with water to dissipate the energy, and make a pot
of tea, as well

I'll look around for a large wirewound resistor, but 200 W? I don't hold out
much hope, even here in Silicon Valley...

Thanks,

You should find it easily. This sort of resistor is a lot more common than
you may think. However thick film types are tending to superceed wire wound
designs. Vishay, for instance. list a RPS500DH, which is a 500W metal film
module, commonly used in things like power inverters, and a 1000W wirewound
part.
Generally though, in 'one off' quantities, kettle elements are cheaper!.

Best Wishes

 

[John Woodgate]

The person suggesting this solution stated that the heating coils (I
think they are actually electric tea kettles, IIRC) were each about 4
ohms.

Are these for use on vehicles, from 12 V, then? I didn't see that
article.

You can buy high-power resistors from Farnell/Newark and they are not
costly. They are even in the horrible UK Digikey catalogue, as 'Aluminum
Housed Chassis Mount Resistors. Eight 1 ohm TMC-50 type in series give
you 8 ohms at 400 W if you bolt them to a big sheet of 1/8 inch
'aluminum'.

 

[Walter Harley]

The fixture I use to match MOSFETs looks like this:


Vdd

|
.-.
150R | |
25W | |
'-'
|
.-------o <-- Vgs
| |
.-. |
470R | | |
| | |
'-' ||-+
| ||->
'----||-+
|
|
===
GND
created by Andy´s ASCII-Circuit v1.22.310103 Beta www.tech-chat.de


I use Vdd = 15v, for MOSFETs comparable to what you're using. Measure Vgs
at the indicated point (or at the gate, but I like to keep the gate resistor
physically close to the gate). You should see something around 3 or 4V,
depending of course on your particular MOSFETs.

Note that the value will CHANGE substantially as the MOSFET heats up - for
TO247's I actually mount a bunch of them to a big hunk of aluminum, and then
test them all in succession a couple of times and throw out the outliers.
But for TO3's that's harder, so basically you just need to be aware of the
effect and try to treat them all the same.

Someone suggested a couple of Walmart electric heating coils in series. Sound
good?

What everyone else said. Personally, a while back I managed to find a deal
on a bunch of 50W 10R resistors, and soldered 'em up in series and parallel
to get two 500W 4R loads, which I can then put in series for a single 1000W
8R load if I want, or use as a stereo load. I mounted the whole deal in a
box with a fan on it so that it can do something with the heat. Nice hair
dryer. Use high-temp wire and be aware that it'll be thermal-cycling so
connectors may loosen.

Profusion now says that the MOSFETs they will sell me are, due to good QC,
closely matched, and are good enough for parelleled service.

It's worth a shot. I'd still check the matching myself.

 

[Ban]

John Woodgate wrote:
|| (in <[email protected]>) about
|| 'Amp's
|| power MOSFET died. Can I just remove it?', on Wed, 19 Nov 2003:
||| The person suggesting this solution stated that the heating coils (I
||| think they are actually electric tea kettles, IIRC) were each about
||| 4 ohms.
||
|| Are these for use on vehicles, from 12 V, then? I didn't see that
|| article.

A 115V 1500W electric kettle apparently has 7.6 Ohms (when warm probably).
But you have to get an American one, difficult in Europe. Anyway, a simple
multimeter can verify this. Eventual inductive behaviour can be compensated
with a parallel R-C combination, but I do not think this will be needed in
the audio range. Of course the 200W will not heat up a lot of water, so put
just a minimum amount into it. Then you can make a nice Nescoffee after a
reasonably timed test. (Only English make tea).
Precision will not be possible, but for an amplifier test it will suffice.

ciao Ban

 

[DaveC]

The fixture I use to match MOSFETs looks like this:


Vdd

|
.-.
150R | |
25W | |
'-'
|
.-------o <-- Vgs
| |
.-. |
470R | | |
| | |
'-' ||-+
| ||->
'----||-+
|
|
===
GND

Thanks, Walter.

I'm a bit confused about the connection of the FET:

The drain and gate are tied together (per your earlier post). These are
connected to Vdd through the 150 ohm resistor?

But then you provide a gate voltage via 470 ohm resistor?

Source is tied to ground?

I think the ASCII is screwing me up...

 

[Walter Harley]

I'm a bit confused about the connection of the FET:

The drain and gate are tied together (per your earlier post). These are
connected to Vdd through the 150 ohm resistor?

But then you provide a gate voltage via 470 ohm resistor?

Source is tied to ground?

I think the ASCII is screwing me up...

View the ASCII with a fixed-width font. It should look fine.

Source is tied to ground. Gate is tied to drain through a 470R resistor.
Drain to power supply through a 150R power resistor. The only reason for
the 470R gate resistor is to eliminate high-frequency oscillations that
could skew the results. It's probably not necessary, but I just don't like
connecting a MOSFET gate directly to something like a test lead that could
have arbitrary inductances, capacitances, etc.

This circuit is a "two-terminal current sink". It is a self-contained
negative-feedback regulated system. If current flow increases, that makes
the voltage drop across the 150R increase, which drops Vgs, which decreases
current flow; and vice versa. So, Id is constant; all you have to do is
*measure* Vgs, you don't independently control it.

The number you get doesn't really tell you a lot about the MOSFET. But two
MOSFETs with the same Id-to-Vgs curve will show the same Vgs in this rig,
and that's all you care about.

If you wanted to actually plot the whole curve of Id versus Vgs, you would
need to separately control Vgs. But there's no reason to do that, because
you're not going to buy enough MOSFETs to be able to do matching on that
level, so you'd only make yourself unhappy

-walter

 

[James Sweet]

Are these for use on vehicles, from 12 V, then? I didn't see that
article.

You can buy high-power resistors from Farnell/Newark and they are not
costly. They are even in the horrible UK Digikey catalogue, as 'Aluminum
Housed Chassis Mount Resistors. Eight 1 ohm TMC-50 type in series give
you 8 ohms at 400 W if you bolt them to a big sheet of 1/8 inch
'aluminum'.

You could build a dummy load in a paint can (available at hardware stores)
filled with mineral oil to dramatically increase the power handling of the
resistors. The kettle elements sound like a good idea to me though.

 

[Jeff]

Are these for use on vehicles, from 12 V, then? I didn't see that
article.

You can buy high-power resistors from Farnell/Newark and they are not
costly. They are even in the horrible UK Digikey catalogue, as 'Aluminum
Housed Chassis Mount Resistors. Eight 1 ohm TMC-50 type in series give
you 8 ohms at 400 W if you bolt them to a big sheet of 1/8 inch
'aluminum'.

I like the Dale NH-250 series but they were not cheap last the time I priced
them.
A freind of mine used to use toaster coils from the
realy old ones that used ceramic insulators.
They would glow pretty good on a larger amp.
Jeff