Hi guys n' girls.
I'm working on a new ampdesign as a diy project to the electronic technicians apprenticies at my work,
and would like Your oppinion on the schematic:
I deliberately chose 2N3055/MJE2955, since they're at hand, cheap and rugged
What do You think ??
I'm working on a new ampdesign as a diy project to the electronic technicians apprenticies at my work,
and would like Your oppinion on the schematic:
An externally hosted image should be here but it was not working when we last tested it.
I deliberately chose 2N3055/MJE2955, since they're at hand, cheap and rugged
What do You think ??
DC coupled at the input? OK...
The 500-ohm pot on the top differential pair, I assume for matching current between the two diff-pairs. So what if they are not perfectly matched? (not being difficult here, but never built an input stage like this with two diff-pairs, and was wondering about the effect of slight mismatch...)
The 500-ohm pot on the top differential pair, I assume for matching current between the two diff-pairs. So what if they are not perfectly matched? (not being difficult here, but never built an input stage like this with two diff-pairs, and was wondering about the effect of slight mismatch...)
Impressive output stage, but is does it require more current gain between it and the voltage amp?
The input stage is very similar to a prototype I designed about a year ago but I used shunt feedback to remove the reliance on high CMRR of the input pair for effective distortion cancellation (there was a thread on this recently). This would make the amp inverting and will need correcting at the speaker terminals.
Also either seperate supplies or some form of supply decoupling between voltage amp and current amp stages is always worthwhile.
The input stage is very similar to a prototype I designed about a year ago but I used shunt feedback to remove the reliance on high CMRR of the input pair for effective distortion cancellation (there was a thread on this recently). This would make the amp inverting and will need correcting at the speaker terminals.
Also either seperate supplies or some form of supply decoupling between voltage amp and current amp stages is always worthwhile.
This reiterates (more or less) some other comments:
At some pont try building a version without the pot on the dif pair. Compare performance to a version with it to see if matching the pair is really makes a diiference. Better still try setting the pot deliberatly to various degrees of mismatch to get a clearer idea of the effects of mismatch.
If you or the students ever decide to buld one for actuall listening, consider 2sa1943/2sc5200 for output devices. They don't cost that much especially as you may not need so many anyway.
What do you gain by tieing a leg of the the dif pair to the VAS network? It may be elegant but anychange you make to one section may impact the other. Being able to optimise each independantly has advantages at leasat for those of us who are less talented.
47uF caps across Q6, and Q8 can protect the CCs from any non-DC artifacts in the rails.
Another 47uF cap across the R18-R19-R11 series can keep AC signals away from the base of Q11 thus giving a more stable bias.
A small cap across R22 is sometimes used to reduce crossover distortion.
Since this is an instructional project, maybe these niceties can be optional to encourage the more curious to discover for themselves if they are helpful.
At some pont try building a version without the pot on the dif pair. Compare performance to a version with it to see if matching the pair is really makes a diiference. Better still try setting the pot deliberatly to various degrees of mismatch to get a clearer idea of the effects of mismatch.
If you or the students ever decide to buld one for actuall listening, consider 2sa1943/2sc5200 for output devices. They don't cost that much especially as you may not need so many anyway.
What do you gain by tieing a leg of the the dif pair to the VAS network? It may be elegant but anychange you make to one section may impact the other. Being able to optimise each independantly has advantages at leasat for those of us who are less talented.
47uF caps across Q6, and Q8 can protect the CCs from any non-DC artifacts in the rails.
Another 47uF cap across the R18-R19-R11 series can keep AC signals away from the base of Q11 thus giving a more stable bias.
A small cap across R22 is sometimes used to reduce crossover distortion.
Since this is an instructional project, maybe these niceties can be optional to encourage the more curious to discover for themselves if they are helpful.
EchoWars said:
The 500 Ohm pot is for adjustment of DC at the output.
Yes, the input transistors do require matching. The emitterresistors R1-4 somewhat lessens the importans I'm told !
Nelson Pass said:
Haven't tried the amp with a nasty load yet. I'll only put it in if it's nessecary
As told, the chosen ones are at hand, meaning in the house, and therefor a natural choise. It is though a place for improvement, I agree.sam9 said:
It forms a folded cascode, using both outputs of the diff-pair!
It's correct, though, that I'm not able to optimise the one, without affecting the other. I can live with that.
Worth trying
I'll make holes in the PCB, just in case.
Thanks for Your inputs, guys. Keep'em comming
Hi all.
I forgot to tell You, that the supply-rails are +/-25 Volts.
ADC:
Through Q9 and Q10 runs @16mA, and through Q12 and Q13 runs @70mA.
Bias is set to @80mA through EACH output device=>total bias @320mA.
With a 2 Ohm load at max out, the current in either of the transistors will vary more than a few mA. Therefor I don't think I another driver is needed. But I might try it out one day
When I started designing this circuit I did in fact have an extra driver in, but simulating it in PSpice, I convinced my self it wasn't needed.
I guess listening at it might show weather it'll be needed or not!
One thing is for shure, I'll put in the option for capacitors at the suggested places, to make experiments easier.
I forgot to tell You, that the supply-rails are +/-25 Volts.
ADC:
Through Q9 and Q10 runs @16mA, and through Q12 and Q13 runs @70mA.
Bias is set to @80mA through EACH output device=>total bias @320mA.
With a 2 Ohm load at max out, the current in either of the transistors will vary more than a few mA. Therefor I don't think I another driver is needed. But I might try it out one day
When I started designing this circuit I did in fact have an extra driver in, but simulating it in PSpice, I convinced my self it wasn't needed.
I guess listening at it might show weather it'll be needed or not!
One thing is for shure, I'll put in the option for capacitors at the suggested places, to make experiments easier.
on the folded cascade input stage, I simulated one a while back (2n5401 /5551) and didn't see much improvement over the traditional differential pair so I gave up on it.
Yes, 2955/3055 are cheap, very cheap. I am not sure why you would need that many of them but if you drive low-impendence load, you just may.
I agree that no additional predriver is needed.
I thought it is a well thought out design. A little bit complicated for me but let us know how it works out for you. Good luck.
Yes, 2955/3055 are cheap, very cheap. I am not sure why you would need that many of them but if you drive low-impendence load, you just may.
I agree that no additional predriver is needed.
I thought it is a well thought out design. A little bit complicated for me but let us know how it works out for you. Good luck.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.