Mosfet output stage question

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.

thylantyr

Member
2001-02-19 10:38 pm
Mars
Refering to holtons amps (av800 or
1kw), he increased the output
stage mosfets from 14 to 20
to make the 1kw amp. (I think),
doesn't matter, what I'm asking
is......

What is stopping someone from
using..... oh let say 50 transistors
on the output stage?

Is this the general technique
used IF you want to make
an amplifier to run as low as .. oh lets
say 0.25 - 0.5 ohm (speaker load).

Do you need to tweak anything
in the schematic other than power supply and output stage paralleling ?

Only mosfet output stage I'm
referring too.....
 
skaara said:
Is it the same with bipolars? Like thylantyr said, if I have amp with 4 out. bipolars on each side, can i add 25 to each side to have 50 out bipolars (does bias increases by itself when i add them?)?

bipolars draw base current at idle so more xsistors means more DC from the front end....mosfets don't draw gate current at idle but due to the added capacitance need a lower Z front end. In practice both need a "hotter" drive as # of devices increase
 

thylantyr

Member
2001-02-19 10:38 pm
Mars
--Be sure to feed your devices, as
--each added device draws
--addditional current.

This is the part I need to understand
so bear with me - hehe

I'm looking at this schematic for reference.

http://www.aussieamplifiers.com/downloads/AV800.pdf

Do you mean each mosfet draws
additional gate current and
the "bias and buffer" stage needs
to be re-designed to supply more
current to feed the mosfets?

Or do you mean there is so much
input gate capacitance using 50
output stage mosfets which would
cause side-effects when sending
an AC signal which would require the
'bias and buffer" stage to have lower
output impedance to minimize the
effects?

AV800 schematic -
Is potentiometer "P1" in the schematic the bias adjustment?

What exactly am I thinking (lol) ?

Bipolar output stage questions -
Is this true/false statement --
In a bipolar design, the "pre-driver"
stage needs to be "beefed" up
since addition bipolars will draw
more base current so paralling
50 bipolars would be more work
since the pre-drive stage would
have to supply alot more current?
is this correct?

Last question for now ....
I thought mosfets didn't draw gate current so I was thinking that the original pre-drive stage in any
mosfet would be sufficient?

Thanks for any tips/comments,
I'm just compiling data, ideas,
trying to understand basic concepts
before I tackle a project. I built
a monster amp over 10 years ago
from Mark 5 electronics, it was
a 150w mosfet that I "beefed" up
using a 2KVA powersupply, 16 pairs
of mosfet outputs per channel with
240kuf of power supply capacitance.

I didn't really understand what
I was doing, it was more or less
"electronics hacking"... but it worked quite well. I was load testing it with 0.5 ohm load resistors in a bucket
of water - haha!

Also, I did something back then
that was kinda odd. The amplifier
had too much hiss.. what I mean is
when you placed your ear close to
the tweeters you can hear some hiss.
Back then I thought the amplifer had
too much gain so I cheated to reduce
the hiss. I changed the feedback
resistor so there was less gain,
to offest this, I made a single ended input pre-amp circuit using the 5534 opamp with gain = 2 (I think).. This
gave me similar input sensitivity
but with no hiss at all. What exactly
happened here? What is wrong
with doing this?

Instead of blindly working on my
next project, this time I'd like to
better understand on the inner workings of the circuit instead of
"hacking"......
 
thylantyr said:
--Be sure to feed your devices, as
--each added device draws
--addditional current.

This is the part I need to understand
so bear with me - hehe

I'm looking at this schematic for reference.

http://www.aussieamplifiers.com/downloads/AV800.pdf

Do you mean each mosfet draws
additional gate current and
the "bias and buffer" stage needs
to be re-designed to supply more
current to feed the mosfets?

Or do you mean there is so much
input gate capacitance using 50
output stage mosfets which would
cause side-effects when sending
an AC signal which would require the
'bias and buffer" stage to have lower
output impedance to minimize the
effects?

>Yes the AC signal....You have to charge and discharge the total gate capacitance

AV800 schematic -
Is potentiometer "P1" in the schematic the bias adjustment?

>yep

What exactly am I thinking (lol) ?

Bipolar output stage questions -
Is this true/false statement --
In a bipolar design, the "pre-driver"
stage needs to be "beefed" up
since addition bipolars will draw
more base current so paralling
50 bipolars would be more work
since the pre-drive stage would
have to supply alot more current?
is this correct?

>yep

Last question for now ....
I thought mosfets didn't draw gate current so I was thinking that the original pre-drive stage in any
mosfet would be sufficient?

>sufficient for dc but not ac

Thanks for any tips/comments,
I'm just compiling data, ideas,
trying to understand basic concepts
before I tackle a project. I built
a monster amp over 10 years ago
from Mark 5 electronics, it was
a 150w mosfet that I "beefed" up
using a 2KVA powersupply, 16 pairs
of mosfet outputs per channel with
240kuf of power supply capacitance.

I didn't really understand what
I was doing, it was more or less
"electronics hacking"... but it worked quite well. I was load testing it with 0.5 ohm load resistors in a bucket
of water - haha!

Also, I did something back then
that was kinda odd. The amplifier
had too much hiss.. what I mean is
when you placed your ear close to
the tweeters you can hear some hiss.
Back then I thought the amplifer had
too much gain so I cheated to reduce
the hiss. I changed the feedback
resistor so there was less gain,
to offest this, I made a single ended input pre-amp circuit using the 5534 opamp with gain = 2 (I think).. This
gave me similar input sensitivity
but with no hiss at all. What exactly
happened here? What is wrong
with doing this?

>i don't know about the hiss.....maybe noisy input xsistors. nothing wrong with your approach as long as you don't worry about feedback related problems (TIM etc)

Instead of blindly working on my
next project, this time I'd like to
better understand on the inner workings of the circuit instead of
"hacking"......
 
--The most outputs I have used in a
--commercial product is 80. per channel.

I had an old old issue of Audio mag.. was it the Threshold that was reviewed and hooked
up to arc weld some metal ?

---80. per channel.
Ahhh.. the muscle car of amplifiers.
This is what I'm talking about,
for some reason building something
gargantuan just seems interesting
to me, the adiction to audio similar to
race car drivers and the need for
speed.

--muuuhahahahahaha.....
If I had the funds, I would buy
your amplifiers, how can I
refuse Nelson Pass especially
when he does the "muhahahah"
thing... heheh
 
Doomsday output stage for sure.

With presumably 40+40 fets per channel the connection to the gates is going to get quite long. Also you would have to be careful the resistance of the wire to the sources was all the same. What I would love to try that would suit TO-247packages at least, is to assemble a stack a bit like those old selenium rectifiers. Have a fet, then a sheet of aluminium, then another fet, then another sheet of aluminium.... Pass a (very?:)) long bolt through the whole lot so it looks like a sheeshkebab. The result would be very compact and have all the fet leads in close proximity and lined up to make paralleling a natural. Perhaps you could use the fins of a heatsink and assemble the fets to the fins like this?

GP.
muuuhahahahahahamuuuhahahahahahamuuuhahahahahahamuuuhahahahahahamuuuhahahahahahamuuuhahahahahahamuuuhahahahahahamuuuhahahahahahamuuuhahahahahahamuuuhahahahahahamuuuhahahahahahamuuuhahahahahahamuuuhahahahahahamuuuhahahahahahamuuuhahahahahahamuuuhahahahahahamuuuhahahahahahamuuuhahahahahahamuuuhahahahahahamuuuhahahahahahamuuuhahahahahahamuuuhahahahahahamuuuhahahahahahamuuuhahahahahahamuuuhahahahahahamuuuhahahahahahamuuuhahahahahahamuuuhahahahahahamuuuhahahahahahamuuuhahahahahahamuuuhahahahahahamuuuhahahahahahamuuuhahahahahahamuuuhahahahahaha...... ka-boom....
 
Amplifier "Laying Rubber"

You guys are getting my adrenaline going here!

Over some tasty vinyl and a bit of cognac, I test drove one of Mr. Pass's electronic dragsters, a Mighty Threshold replete with rows of MOSFETS...

Few amplifiers have the requisite horsepower to drive my "Frankenstein" Infinity speakers; their under 2-ohm appetite is a difficult one.

Well, as subsequent nods of approval were met with raised glasses, the gain was increased...

I wish I had purchased the pair of Threshold monoblocks that were originally paired with these oak trees in my living room; there is something wonderful in the sight of the rows of 10-inch drivers dutifully bouncing about as the music resounds with such authority.

Nelson, I raise my glass to you:

muuuhahahahahaha.....
 
--sheet of aluminium, then another fet, then another sheet of
--aluminium....

Ok, now the cat is out of the bag...... the can of worms is open....

I've been thinking about making my own pcb board, 4 layer,
using an established high powered design and making
the pcb modular.. Of course this is nothing unique, everyone
does this now. But what I want to do is make one circuit
board module be able to run with .... oh lets say 40-50
output transistors, but also be able to run with only 10.
Why? pcb's are expensive - hehehe About 7 years ago
(yep, been thinking and doing nothing for a long time - ha!)
I was thinking of the above method of stacking transistors
(not can type) on top of each other with 1/4" alumimum
sheet in between, but I thought it was a crazy idea.

So.. on the PCB board you design it similar to a car amplifier
pcb board, one rectangle with trannies running along the
edges... oh lets say 20 per edge....... then on the pcb
layout add another set of pads so you can solder another
20..... it gets kinda messy, I'm not sure if this would work - hehe

I'd rather make two pcb's instead of doing tihs stuff - hehe
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.