Indeed. And there is also a separate zenered supply derived from the main supply for the opamp. The original poster most probably hasn't realised that, but this is definitely a high power amp, and scalable. With a final stage gain of say, 5, and a supply of say 65VDC I would expect something like 200 watts @ 8 ohms?
Jan Didden
Jan Didden
Hi Guys,
Yup, looking too quickly. COFFEE.POT NOT FOUND .... TECHNICIAN HALTED. -Sorry all.
There is voltage gain but poor current gain. You need the drivers to be emitter coupled to give any reasonable damping factor and current output for high power. I'd recomend hooking your drivers up as emitter followers and add a pair of transistors in the original location.
I had assumed that you were connecting the drivers as emitter followers to achieve the required current gain and was looking for more parts. See what making assumptions can do?
-Chris
Yup, looking too quickly. COFFEE.POT NOT FOUND .... TECHNICIAN HALTED. -Sorry all.
There is voltage gain but poor current gain. You need the drivers to be emitter coupled to give any reasonable damping factor and current output for high power. I'd recomend hooking your drivers up as emitter followers and add a pair of transistors in the original location.
I had assumed that you were connecting the drivers as emitter followers to achieve the required current gain and was looking for more parts. See what making assumptions can do?
-Chris
current gain isn't so bad with good transistors, Hfe ~ 100 for ea Q over the reasonable Iout is practical (admittedly not a worst case design though typ driver could be 200 and output 100 Hfe for the devices under discussion) - 10K composite current gain is dropped a factor of a few by having driver Q collector R current a few times more than output Q base current - lets use a pretty conservative 2K composite Hfe
voltage gain in the cfp costs some too;
op amp Zload ~ (composite Hfe / Av ) * Rload
(2K / 5 ) * 4 Ohm = 1.6K, say 800-1600 Ohm for 2-4 Ohm Zload
this is in the range that a TL070 can drive but it will have some distortion with this load compared to many more modern op amps with higher output drive - opa134 would be my minimum choice but higher GBW op amps may require additional compensation
output Z in any transistor amp is a function of the total feedback around the output device, in this case a quick estimate would be:
(op amp Open Loop output Z) ~100 Ohm / (2K / 5 ) ~= 0.2-0.3 Ohms before op amp loop gain
- damping factor will be > 1K for f < 10 KHz when op amp loop gain is factored in
voltage gain in the cfp costs some too;
op amp Zload ~ (composite Hfe / Av ) * Rload
(2K / 5 ) * 4 Ohm = 1.6K, say 800-1600 Ohm for 2-4 Ohm Zload
this is in the range that a TL070 can drive but it will have some distortion with this load compared to many more modern op amps with higher output drive - opa134 would be my minimum choice but higher GBW op amps may require additional compensation
output Z in any transistor amp is a function of the total feedback around the output device, in this case a quick estimate would be:
(op amp Open Loop output Z) ~100 Ohm / (2K / 5 ) ~= 0.2-0.3 Ohms before op amp loop gain
- damping factor will be > 1K for f < 10 KHz when op amp loop gain is factored in
Hi JCX,
The problem is that you need to design with hFE in the outputs around 50, your drivers around 100. The peak current required is reduced at higher voltages with a resistor load, CCS being better here. High feedback = loss of dynamics so we don't really want high open loop gain. Your choice here. Damping factor without a driver will be terrible because the output stage gain would be 50 (100 in your example), maybe more depending on drive current. In any case, no where near what you might expect.
Look at peak current divided by output current gain. That's what you need as a standing bias current in the vas as a minimum. Otherwise the vas transistor is robbed of all it's current and the total current is not enough for the expected voltage across the load. Therefore the voltage drops and you have distortion (uncorrectable by any amount of feedback). Have I missed something here?
-Chris
The problem is that you need to design with hFE in the outputs around 50, your drivers around 100. The peak current required is reduced at higher voltages with a resistor load, CCS being better here. High feedback = loss of dynamics so we don't really want high open loop gain. Your choice here. Damping factor without a driver will be terrible because the output stage gain would be 50 (100 in your example), maybe more depending on drive current. In any case, no where near what you might expect.
Look at peak current divided by output current gain. That's what you need as a standing bias current in the vas as a minimum. Otherwise the vas transistor is robbed of all it's current and the total current is not enough for the expected voltage across the load. Therefore the voltage drops and you have distortion (uncorrectable by any amount of feedback). Have I missed something here?
-Chris
Resistive collector load sucks (current and robs gm) – much more than my quick estimate assumed – I should have remembered working this out before
Also the local cfp voltage feedback is probably not workable at practical feedback divider R values which changes some more of my simplifying assumptions – to work as a local feedback controlled voltage gain block the cfp probably needs an extra transistor stage to become some variety of "triple"– much “cheaper” than more exact analysis
With driver collector current sources though the 2 transistor cfp would have adequate current gain - even 50 * 100 Hfe only requires 4 mA to control 20 A of output current
However I believe many people here miss the fundamental point about feedback and output Z – there is no practical method of achieving low output Z with controlled current sources (ie transistors, whether you prefer Ib, Vb or Qb control models) without negative feedback, emitter followers have gobs of (local) negative feedback reducing their output Z, but the same impedance result can be had with common emitter outputs if the same total feedback is looped around the output stage in a combination of local and global feedback loops - tell us how the JLH output works if you think different
Also the local cfp voltage feedback is probably not workable at practical feedback divider R values which changes some more of my simplifying assumptions – to work as a local feedback controlled voltage gain block the cfp probably needs an extra transistor stage to become some variety of "triple"– much “cheaper” than more exact analysis
With driver collector current sources though the 2 transistor cfp would have adequate current gain - even 50 * 100 Hfe only requires 4 mA to control 20 A of output current
However I believe many people here miss the fundamental point about feedback and output Z – there is no practical method of achieving low output Z with controlled current sources (ie transistors, whether you prefer Ib, Vb or Qb control models) without negative feedback, emitter followers have gobs of (local) negative feedback reducing their output Z, but the same impedance result can be had with common emitter outputs if the same total feedback is looped around the output stage in a combination of local and global feedback loops - tell us how the JLH output works if you think different
Hello? Who's there? 
I think the point was made that zox2003 should add some driver transistors. That's really all there was to it. I got off track a bit when I assumed the drivers referred to were drivers (my error) when they are in in fact a voltage amp stage.
This discussion goes to how much power should be expected since the peak output current is affected directly. Therefore helpful to zox2003.
I though jcx was explaining some things well that zox2003 may think about as an upgrade to his design. Hopefully we are still on track?
-Chris
I think the point was made that zox2003 should add some driver transistors. That's really all there was to it. I got off track a bit when I assumed the drivers referred to were drivers (my error) when they are in in fact a voltage amp stage.
This discussion goes to how much power should be expected since the peak output current is affected directly. Therefore helpful to zox2003.
I though jcx was explaining some things well that zox2003 may think about as an upgrade to his design. Hopefully we are still on track?
-Chris
consider the price of the advice
I’ve already suggested parallel chip amps, complete designs with pcbs can be found in the chip amps forum
Totally discrete designs that meet the requirements are also readily available and negligibly more complicated or expensive than a hybrid op amp-discrete design – and are well tested
Given the implied low performance required many of us would just buy some cheap Chinese PA amp for less than we could buy the xmfr and chassis for a diy project
With little in the way of motivation for discussing a design with such arbitrary “requirements” (and a poor op amp to start with) I think we're doing a great job of preventing the thread from dying of boredom
I’ve already suggested parallel chip amps, complete designs with pcbs can be found in the chip amps forum
Totally discrete designs that meet the requirements are also readily available and negligibly more complicated or expensive than a hybrid op amp-discrete design – and are well tested
Given the implied low performance required many of us would just buy some cheap Chinese PA amp for less than we could buy the xmfr and chassis for a diy project
With little in the way of motivation for discussing a design with such arbitrary “requirements” (and a poor op amp to start with) I think we're doing a great job of preventing the thread from dying of boredom
Hi zox2003,
It should work to some degree. Depends on resistor values and PCB layout. The only thing I can say is build it and see. It doesn't look like it will not work.
Listen, it wouldn't be the first amp I build that didn't work, or worked off the page. Unless it's a kit with premade boards there are always some variables at play. You know, I'm always surprised when I design something and it works first time. Then I start making it better if that's my goal.
As far as better is concerned, don't worry about it. Depends on the use, cost, performance .... Only you can determine those things.
-Chris
It should work to some degree. Depends on resistor values and PCB layout. The only thing I can say is build it and see. It doesn't look like it will not work.
Listen, it wouldn't be the first amp I build that didn't work, or worked off the page. Unless it's a kit with premade boards there are always some variables at play. You know, I'm always surprised when I design something and it works first time. Then I start making it better if that's my goal.
As far as better is concerned, don't worry about it. Depends on the use, cost, performance .... Only you can determine those things.
-Chris
Hey all,
I would like to know if I change the OP-AMP to something like OPA541
could I get 150w into a 4 ohm load?
Because I have a Sansui R-550 Receiver that distorts when i put the volume up, and thats with the bass on +5, and I have an QE hooked up also, and I want to take the amps out and put new ones in, with OPA541's and some transistor to boost the output, it has a 260VA transformer and 25-0-25 AC output (which is 35+/- DC) and that is perfect for OPA541. so could i do that?
BTW i would also like to see some pics of the amp!!
Thanks,
Kinser
I would like to know if I change the OP-AMP to something like OPA541
could I get 150w into a 4 ohm load?
Because I have a Sansui R-550 Receiver that distorts when i put the volume up, and thats with the bass on +5, and I have an QE hooked up also, and I want to take the amps out and put new ones in, with OPA541's and some transistor to boost the output, it has a 260VA transformer and 25-0-25 AC output (which is 35+/- DC) and that is perfect for OPA541. so could i do that?
BTW i would also like to see some pics of the amp!!
Thanks,
Kinser
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.