Ok, just a quick question. I'm building the 400 watt symmetrical amp and I'm wondering what the easiest way is to get the amp to be 2 ohm stable....or at least close. I have a 1kva toroid PER BOARD and lots of caps etc. which will give me about 75vdc on the rails. I hope. Anyway, a while ago Anthony said the amp can't handle very many more output transistor pairs...board has space for 4 pairs...I was thinking of doing 6 pairs (is this ok?). Also, which driver's do I need to heatsink? would this help drive more outputs if I heatsinked the Q8and Q9 drivers?
ok, to summarize
1) is 6 pairs of outputs too much? too little? (more is better :-D)
2) would heatsinking OR substituting some driver transistors help drive more output pairs
3) if I can sub a driver...(I don't mind using more expensive parts...like other output mosfets) what should I use?
thanks ya'll!
-Matthew K. Olson
www.aussieamplifiers.com/sym-sch.htm (I think thats the right link)
oh, and one other thing, building the av800 is a great idea, but I already spent $40 per board on these 400 watt boards...so I'd like to work with what I have for now...thanks guys
ok, to summarize
1) is 6 pairs of outputs too much? too little? (more is better :-D)
2) would heatsinking OR substituting some driver transistors help drive more output pairs
3) if I can sub a driver...(I don't mind using more expensive parts...like other output mosfets) what should I use?
thanks ya'll!
-Matthew K. Olson
www.aussieamplifiers.com/sym-sch.htm (I think thats the right link)
oh, and one other thing, building the av800 is a great idea, but I already spent $40 per board on these 400 watt boards...so I'd like to work with what I have for now...thanks guys
I built an amplifier similar to that design
a decade ago except it had;
*65V rails
*same transistors (can type), 8 per rail
*Four 500VA transformers
* fans
1/4" thick alumimum base plate
for transistors, then I mounted
heatsinks on the base plate.
I used two 110 VAC inputs
and did test tone load tests
at 0.5 ohms for quite a while
with no problems.
*****************************
more food for thought.
An amplifier typically can handle
less "ohms" when not doing
"tone testing" and just playing
music.
An amplifier can typically handle
less "ohms" if you limit the bandwidth,
such as bi-amp or tri-amp. Case in point, my Adcom 555II bridged
is rated for 600w @ 8 ohms, I
can operate a 4 ohm woofer but
it gets hot -- need fans..
But, I can drive eight parallel 8 ohm
tweeters for a 1 ohm load with no
problems... tweeter don't draw much.
I think you can easily drive 2 ohms
with that desgn if you have "can type"
transistors on massive heatsinks
playing music.. If you "tone test" the
amplifier, it will run hotter.
The orginal design used 4 transistor
per rail, I double the outputs with
no circuit mods.. it seems to work
well... I bet you can add a couple
more transistors per rail with no
big side effects.
a decade ago except it had;
*65V rails
*same transistors (can type), 8 per rail
*Four 500VA transformers
* fans
1/4" thick alumimum base plate
for transistors, then I mounted
heatsinks on the base plate.
I used two 110 VAC inputs
and did test tone load tests
at 0.5 ohms for quite a while
with no problems.
*****************************
more food for thought.
An amplifier typically can handle
less "ohms" when not doing
"tone testing" and just playing
music.
An amplifier can typically handle
less "ohms" if you limit the bandwidth,
such as bi-amp or tri-amp. Case in point, my Adcom 555II bridged
is rated for 600w @ 8 ohms, I
can operate a 4 ohm woofer but
it gets hot -- need fans..
But, I can drive eight parallel 8 ohm
tweeters for a 1 ohm load with no
problems... tweeter don't draw much.
I think you can easily drive 2 ohms
with that desgn if you have "can type"
transistors on massive heatsinks
playing music.. If you "tone test" the
amplifier, it will run hotter.
The orginal design used 4 transistor
per rail, I double the outputs with
no circuit mods.. it seems to work
well... I bet you can add a couple
more transistors per rail with no
big side effects.
Well...I'm using the plastic irfp240 and 9240 transistors...BUT...I have quite bit heatsinks. If I use 6 pairs...the 240's will be on one sink and the 9240's will be on the other. Each sink will dissipate 75 watts of heat easily. And, since this is monoblock design, thats 300 watts of heat total w/ the amps just coasting no fan...but I can put a fan if need be. Think that'll be enough for 2 ohms music? Should I sink the drivers? would that help drive more outputs?
-Matthew K. Olson
-Matthew K. Olson
I say you should try out your idea
and see how it goes..... I would
heatsink the drivers, the amp
I did, the drivers got hotter than
normal - hehe I mounted those
on the main heatsinks, works good.
and see how it goes..... I would
heatsink the drivers, the amp
I did, the drivers got hotter than
normal - hehe I mounted those
on the main heatsinks, works good.
Remember to carefully match outut devices, otherwise one or two will do the most of the job and will probably fail if driving low impedance loads.
Cheers
Andrea
Cheers
Andrea
What will sound aweful? If I don't match the outputs? Or I don't sink the drivers? I was planning on matching the outputs anyway...so no big deal there. 🙂 Sinking the drivers, thats a more hairy animal but I can do it.... 🙂 Ok, so basically what everyone is telling me is just go for it and find out. k. later
-Matthew K. Olson
-Matthew K. Olson
Andypairo said:
Won't the source resistors be sufficient to balance the load.
That's most certainly the reason (or one of several) why
they are there already and if works for four devices it should
work equally well for any number of devices, I think.
Christer said:
I tried it myself yesterday, with 0.22R source resistors you have only a slight balancing effect. Even 50mV between Vgs of two mosfets brings to a dramatic difference in bias current.
I had accidentally blown a mos and tried to substitute it without matching it (I thought it wasn't worthwhile to desolder the others to measure them..): it carried almost twice the current compared to the others. So I took it out and it gave a reading of 3.78V @ 150mA.
Then I measured the others taking them out of the circuit and they were between 3.83 and 3.84.
So I kept on measuring ( I have about 35 pairs of them from the same batch) and I got Vgs values between 3.40 and 3.61V for the IRFP240s and between 3.65 and 3.95 for the 9240s.
With such Vgs voltage spread and a bias current of about 100mA per pair the degeneration voltage ( 0.1 * 0.22 = 22 mV) over the resistors can do little if you don't match the devices within a 10 mV range.
Cheers
Andrea
Mattyo5 said:
If you use the maximum admitted voltage you probably will have to put a small sink on the drivers, I run the amp at about 55V per rail so I don't use them. If you have doubts just put your finger on it: if it starts burning
you definitely need a heatsink 😉 Cheers
Andrea
Andrea,
Ok, you're right. I didn't quite think about that those are quite
high-transconductance devices so it would
probably require much larger resistors, which we
don't want.
I haven't had a close look at the design, but do those
resistors serve much purpose at all then, if they don't
give any reasonable balancing effect?
Ok, you're right. I didn't quite think about that those are quite
high-transconductance devices so it would
probably require much larger resistors, which we
don't want.
I haven't had a close look at the design, but do those
resistors serve much purpose at all then, if they don't
give any reasonable balancing effect?
Christer said:
It's a magnitude order matter... with unmatched devices those resistors can do very little, with matched ones they will help to tame the differences between the devices, which can't be equal (unless you buy 10.000 of them and match them within 1 mV at different currents...
).Cheers
Andrea
Andypairo said:
It's a magnitude order matter... with unmatched devices those resistors can do very little, with matched ones they will help to tame the differences between the devices, which can't be equal (unless you buy 10.000 of them and match them within 1 mV at different currents...
Cheers
Andrea
True, stupid of me.
Ok, well, so once again I ask, are you guys just telling me to go for it (w/ matching the transistors of course) and see what happens? I've been trying to figure out ways to heatsink the drivers. Will heatsinking the drivers allow me to run not only higher voltages (not that I want to go high) but also more transistors on the output?
-Matthew K. Olson
-Matthew K. Olson
Mattyo5 said:
Before trying to modify an existing (and tested) project I'd build one and see if is satisfies my needs. After that you can try (if you feel the need) to modify it, but note that it won't be the original amp anymore.
About heatsinking you can use regular TO220 sinks, no need to mount them on the main heatsink; if you plan to add more outputs the specs of the MJEs might be insufficient, I'd rather use the MJE15030/31 or similar.
Cheers
Andrea
If you look at the AV800 design, Holton added the buffer
stage between VAS and output to allow more output
transistors....
For the 400w amp, if you are a gambler, try 8 outputs
per rail max. vs. the standard 4... see if the VAS can
drive it.. /hehe
....but before I would do this, verify that the amplifier is functional
and stable with 4, then increase the number later. Make
accomodations to allow 4 more if need be.
If you don't want to match output Hexfets, use the original
mosfets shown in the schematic, those are lateral fets right?
if so, they current share at much lower current and matching
isn't so critical.
stage between VAS and output to allow more output
transistors....
For the 400w amp, if you are a gambler, try 8 outputs
per rail max. vs. the standard 4... see if the VAS can
drive it.. /hehe
....but before I would do this, verify that the amplifier is functional
and stable with 4, then increase the number later. Make
accomodations to allow 4 more if need be.
If you don't want to match output Hexfets, use the original
mosfets shown in the schematic, those are lateral fets right?
if so, they current share at much lower current and matching
isn't so critical.
Ok, MJE15030/1 Anyone 2nd that? For now I plan on doing it w/ the MJE350/340's but if someone else says go w/ the 15030/1's I'll use those. 🙂 Later all!
-Matthew K. Olson
-Matthew K. Olson
Not a bad idea to start w/ 4 pairs of outputs. I guess I could mount all of them and just hook up 4 and see where we go from there. I already have the irfp240's and 9240's...so I will have to match them. Thats no problem :-D If I could go to 8 pairs per channel that'd be great, but I don't think I'll risk that...maybe 6 pairs. Laters all
-Matthew K. Olson
-Matthew K. Olson
The MJE1503x can sustain much more current (8A vs 500mA) than the MJE340/50 while maintaining a similar Hfe; if you plan to use high voltage rails the voltage rating of MJE15030/31 might be pushed a little too hard, so MJE15032/33 seems the most reasonable choice (of course if you can't live without adding output pairs).
But, as suggested before, first go with the original project and then...
But, as suggested before, first go with the original project and then...
- Status
- Not open for further replies.