| Mindaugas |
I'm planning to build a medium power(60-100W for 8ohms) amp.
I found this project at
http://users.otenet.gr/~athsam/powe...hexfet_eng.htm
Now I wonder how this thing sounds. Is it worth to build it? Or it's just a waste of time and I must search for another better power amp schematics.
Any Comments would be appreciated |
|
|
| subwo1 |
| I tend to think that if the the mosfets are to be connected common source, then a circuit using only N-channel devices benefitting from their superior performance should do better. If an amp uses a P-channel device, it may as well use source follower configuration. The main hitch with source followers is that the gate drive circuit needs to operate about 15 volts beyond each drain power supply. |
|
|
| ojg |
| I have a feeling this amplifier will not work... The drain and source of Q12 need to be switched around. So it becomes a mirror image of Q13. See here for a similiar circuit. |
|
|
| peranders |
This volt thing is true but you must also consider how amperes at 15 V. With a nornmal load (speaker) you only have to have 5-7 volts gate-source voltage. With this design you loses a coulple of watts but in decibells it's almost nothing (1 dB?). Otherwise the design is pretty normal, nothing wierd. :)
The amp will very much likely work since it's already built. It seems that the output stage is having local current feedback. |
|
|
| ABO |
It is impossible to say beforehand if this amp will work.
This is an amp from Elektor magazine called hexfet amp. The problem with this amp is stability. It may, or may not be stable. The designers never figured out why it does work in one case and not in another.
I know because I e-mailed about it with the designer (t.giesberts@segment.nl).
use mica insulators for Q12,13
use ceramic insulators for Q10,11,8
very close matching of Q1-4, Q7,9, Q10,11
....and hope for the best. The sound is alright though.
I have its higher power brother with IGBT's instead of hexfets. This amp has the same stabitlity problems!
hope this helps |
|
|
| Rambi |
| quote: | Originally posted by ABO
It is impossible to say beforehand if this amp will work.
This is an amp from Elektor magazine called hexfet amp. The problem with this amp is stability. It may, or may not be stable. The designers never figured out why it does work in one case and not in another.
|
Not only difficult to get stable in the sense of not oscillating, but also in DC offset. I built one (channel) of these and couldn't get the offset stable; it would 'jump' up and down and personally I don't like an amplifier that does that :cannotbe:.
At least it didn't oscillate in my case.
If you like this 'kind' of amplifier (symmetric, mosfet output) maybe you should take a look at the cresendo ME
http://www.diyaudio.com/forums/show...&threadid=12885
A more thoughtful design, also from elektor, with a servo to keep DC offset within limits (some people don't like that either though).
Sound is supposedly OK, but I can't really comment on that since I never built/heard one.
| quote: |
I have its higher power brother with IGBT's instead of hexfets. This amp has the same stabitlity problems!
|
Probably because it is the exact same design with IGBTs instead of mosfets ;)
This amp was mentioned earlier on this forum as well, see
http://www.diyaudio.com/forums/show...?threadid=25384
And the IGBT version too:
http://www.diyaudio.com/forums/show...?threadid=23766
Remco |
|
|
| jcx |
depends, the basic topology could work ok with some changes for high frequency compensation
the pcb layout you point to sucks
so if you have a >100MHz 'scope and enough knowledge to cure the high frequency problems you may encounter with this design and you are capable of doing your own layout/prototype with good power, ground and RF practice - you wouldn't have to ask
bottom line, look elsewhere for a less problematic design |
|
|
| Tube_Dude |
| quote: | Originally posted by ojg
I have a feeling this amplifier will not work... The drain and source of Q12 need to be switched around. |
I see no reason for the amplifier do not work...the output stage is a regular compound complementar feedback pair where the output devices are Mos-Fets. (nothing new under the sun!!) ;) |
|
|
| ojg |
| quote: | Originally posted by Tube_Dude
I see no reason for the amplifier do not work...the output stage is a regular compound complementar feedback pair where the output devices are Mos-Fets. (nothing new under the sun!!) ;) |
Hmmm, you may be right. I was confused by the drawing of Q12, usually the gate "lead" is drawn towards the source, like in Q13. As long as Vgs is 2V for both Q12 and Q13 it should be fine. |
|
|
| AKSA |
Hi Mindaugas,
Nice to see you here! How the hell are you? Finished your new home?
I recognize this as the Elektor circuit too. I have built a couple of Elektors, notably the Crescendo and one other from Giesberts. All these amps had small, niggling problems.
First, it is over complex. Much more than it needs to be. (Per-Anders, I'm looking at you.......... grin!) Redolent with CCS and complementary input and VAS stages.
Second, instability problems are likely coming from the CFP output stage. My advice would be small mica caps, around 100pF, across the base/collector of Q10 and Q11. You might also lower gain of the driver by inserting a 0R47 source resistor in line with F1 and F2; this reduces feedback factor across the CFP.
Generally, CFPs don't work well in Class AB - they really prefer Class A.
Cheers,
Hugh |
|
|
| Upupa Epops |
| This amp have not any protection against overvoltage in gate ( Vgs max ) and will be dangerous. I short in output gate voltage ( Vgs ) will be app. 35 V and Vgs max of this type fets is only 20 V - fet will be reliable destroy. |
|
|
|
