50 Output Stage Transistors?

[thylantyr]

Nelson Pass has inspired me when he said there
is no limit on how many output stage mosfet
transistors you can parallel provided that you
can drive them

The squirrel cage in my head is turning, wondering
if using 25 - 50 output mosfets per rail is
doable with the attached schematic (av800).
I guess 25 per rail is good enough ... The design
has been built by Holton with 10 per rail, but
the crazy rabid fans of high powered amplifiers
need a "fix" and want more like a junky seeking
the next needle.

Why ?

Because.........

1. The muscle car fanatics know what I'm taking
about. Need for speed and engine rumbling.

2. It would be cool to drive the new breed of super
monster X-max woofers, typically DVC of 1 ohm
parallel the woofer coils for 0.5 ohm load with
a classic AB design, not the class D designs chosen
today by many.


How? Beats me, if I knew, I wouldn't be posting here.

Can the buffer stage (Q4, Q5) which is IRF610 and IRFP610
be replaced with IRFP240 and IRFP9240 and R9 be lowered
to give me more drive for the beefy output stage of
25 transistors per rail ?

Drools.....

Researching the archives revealed that Nelson was
gonna throw us a bone and give the fans a monster
AB design, a gleam in his eye...

Patiently awaits

Oh please

[sajti]

Hi,

for drivers, You can use IRF630/9630, because they are much more linear than IRFP240/9240.
But I think this idea is good if You have a lot of money, or You are the sponsor of the local electronic shop.
If the amplifier works with 10pairs, why do You wan to use 50 of them?

If You have 100m of heatsinks, You can build 1000W class "A" amplifier...

Sajti

[thylantyr]

The 10 pair version is only rated no less than 4 ohms continuous
at 1kw. I read some posts that said it's able to do 1600w
short term into 2 ohms.

http://www.aussieamplifiers.com/1kwamp.htm

But, it would be nice to drive those new 0.5 ohm woofers
at high power. Some of those woofers have 1-2kw power
ratings.

P.S. What happened to diy member "Kilowatt" ?
I read all his posts about his desire to build the
7200w amp running of the AC mains.

[millwood]

the datasheet for regular mosfets (like 540/640/etc.) usually indicate a 5x "pulse current" (for a duration of 10ms). so in theory a mosfet could output 25x of its rated power without sustaining permanent damages.

[thylantyr]

Thinks....
Scratches head....
Stares at wall .....

If you configure those buffer
mosfet (pre-output stage), ie,
IRF610 or whatever, you are
subject to the T0-220AB package heat
dissipation of 50W ? So, finding
a higher current rated transistor
in the same package might not
help you ?

True?

False?


I don't see said the blind man

[JoeBob]

If you're in the mood of matching 50+ transistors (that will be one hell of a job and you'd have to buy hordes of them), why not build a normal powered amp with massively paralleled smaller devices, like irf 620s, 630s, 520s, 530s etc (or even more smaller devices)? That'd be interesting so see.

[usekgb]

I'd like to see that myself

[sajti]

Matching 50 MOSFETs is a real pain. But if You want to use 25pairs of them, You can use large source resistors, such as 1ohm for each device. With this value the matching is not very critical...

Sajti

[thylantyr]

Sounds like fun, matching hundreds
of fets, time to make a transistor
matching tester (hehe)....

Actually, I saw a post on the subject
in the archives, but the link was
broken to the schematic (doh!)...

Trying to figure something out....

If you have two transistors in
the output stage that are perfectly
matched, then you don't need
the source resistor because current
sharing is equal.....

But, in reality it's not the case
so most people use 0.1 - 0.3 ohms.

Question is;

Take the same 2 transistors
and add 8 more in parallel,
unmatched.

If you are driving the same current
thru 10 transistors as you did
with 2 transistors, would not
the current split thru 10 transistor
paths instead of two ? Would this
not but a less of a burden on
the whole bank of 10 vs. 2 ?

(but if you increase current, that's
another story , heh)...

Is this right or wrong assumption?

Example,

If I drive 2 amperes thru the output
stage of two matched transistors,
1 ampere passes thru each one.

If I drive 2 amperes thru the
output stage of 10 unmatched
transistors *and* let say
one of those transistors is the
current hog and allows 1 ampere
to pass, then the other ampere
of current splits into the other 9
transistors, in which case, the
unmatched bank would still work?

FYI, I plan to match, but still
wonder.......

[sajti]

The source resistor not only for the current matching!
If You check some small power amplifier, with only one pair of output device, You can find source resistor too. Why? This resistor helps to the bias network, to keep the bias constant over different temperatures.
It also separates the emitter followers from capacitive loads, which can results oscillation.

Sajti