Circlotron's circuit is straight out of the NS Audio Handbook(out-of-print) and works great for low to medium frequency use.
If you ground the output transistors emitters, and float the center tap of the power transformer, you can get any power you need or want.
An LM1875 will work well on ±12V or so(it is spec'd for ±8V), and it will drive any number of outputs you may care to hook up.
A 1KW GC may be constructed with this method.
If you need good HF response the class B level shift may be dealt with by this method:
http://www.audiodesignguide.com/my/Cool_Follower1.GIF
The bias resistors and collectors of the level shift transistors should be returned to the same ±12V that drive the LM1875.
Any of the popular GC chips can be used, I mention the LM1875 because it is designed to run as low as ±8V and is spec'd at less distortion than the LM3886.
Any supply voltage may be used on the GC chip, whatever it is comfortable with, the current only needs to be the base current required by the outputs.
Don't forget to swap the ± inputs on the GC chip and move the feedback resistor to the center tap of the main transformer.
http://www.diyvideo.com/forums/attachment.php?postid=182299
If you ground the output transistors emitters, and float the center tap of the power transformer, you can get any power you need or want.
An LM1875 will work well on ±12V or so(it is spec'd for ±8V), and it will drive any number of outputs you may care to hook up.
A 1KW GC may be constructed with this method.
If you need good HF response the class B level shift may be dealt with by this method:
http://www.audiodesignguide.com/my/Cool_Follower1.GIF
The bias resistors and collectors of the level shift transistors should be returned to the same ±12V that drive the LM1875.
Any of the popular GC chips can be used, I mention the LM1875 because it is designed to run as low as ±8V and is spec'd at less distortion than the LM3886.
Any supply voltage may be used on the GC chip, whatever it is comfortable with, the current only needs to be the base current required by the outputs.
Don't forget to swap the ± inputs on the GC chip and move the feedback resistor to the center tap of the main transformer.
http://www.diyvideo.com/forums/attachment.php?postid=182299
I hadn't actually seen it there. I got the general ideal from Sandman's 1974 current dumping article in Wireless World. (thanks, janneman).
That approach will work too, but it is a very different concept. The transistors will be operating as common emitters, not emitter followers. Most importantly though, they will need to be biased like a conventional amplifier because they produce all of the power output whereas the cct I drew has the opamp producing the power output from zero to several percent, and until then the transistors are completely off.
Yep.
Not quite. Don't forget the Vbe still exists across the 1 ohm resistor so the opamp output current more or less levels out when the transistor starts to help. Any *excess* current above Vbe/Rbe is handled by the transistor. I really have to build one of these things. I have a couple of LM3875's coming my way in the next week or so. Should be good fun.
Hi!
Verrry interesting topic... got some power-OPs lying around, some power devices as well, a heatsink... but just no time right now...
One question though: Can I easily parallel more power devices? How do I achieve this (emitter resistors)? How well do they have to be matched?
And is one of the circuits described in this thread comparable to
?
(The Thel Music Amp is a discrete audio OP-amp... I have the circuit...)
OK, it has the drwabacks of adjusting the bias...
Bye,
Arndt
Verrry interesting topic... got some power-OPs lying around, some power devices as well, a heatsink... but just no time right now...
One question though: Can I easily parallel more power devices? How do I achieve this (emitter resistors)? How well do they have to be matched?
And is one of the circuits described in this thread comparable to
An externally hosted image should be here but it was not working when we last tested it.
?
(The Thel Music Amp is a discrete audio OP-amp... I have the circuit...)
OK, it has the drwabacks of adjusting the bias...
Bye,
Arndt
this particular topology is pretty typical for expanding op-amp's current output capabilities. If you bootstrap the op-amp, then you theoritically can build a very powerful amp.
The concern about non-linearity (looking backwards into the op-amp) is a valid one. Once there is more voltage drop over the current-sensing resistor, any incremental output current is provided by the transistors. In essence, for low current output, the op-amp sees a resistor network consisting of the current sensing resistor and the load. at higher current output, the op-amp sees a varistor whose value goes up and down with the output voltage to maintain constant current output from the op-amp. another way to put it, the op-amp sees a constant current load.
It is not clear to me how well the particular op-amp works driving such a load but in theory, it should work for either the lm3875 or any op-amp. If your op-amp cannot handle larger output current, you can always increase the value of the current sensing resistor. For example, if you want the transistors to kick in at anything above 10ma, the resistor should be 0.7/10ma=70ohm. so with a high-enough current sensing resistor, any op-amp will work in this topology (in theory).
the same principle can be used with mosfets as well, except that their threshold Vgs is quite big so the current sensing resistor will be adjusted upwards as well. Another concern is the op-amp's stability driving high capacitive load (the mosfets' gate capacitance).
The concern about non-linearity (looking backwards into the op-amp) is a valid one. Once there is more voltage drop over the current-sensing resistor, any incremental output current is provided by the transistors. In essence, for low current output, the op-amp sees a resistor network consisting of the current sensing resistor and the load. at higher current output, the op-amp sees a varistor whose value goes up and down with the output voltage to maintain constant current output from the op-amp. another way to put it, the op-amp sees a constant current load.
It is not clear to me how well the particular op-amp works driving such a load but in theory, it should work for either the lm3875 or any op-amp. If your op-amp cannot handle larger output current, you can always increase the value of the current sensing resistor. For example, if you want the transistors to kick in at anything above 10ma, the resistor should be 0.7/10ma=70ohm. so with a high-enough current sensing resistor, any op-amp will work in this topology (in theory).
the same principle can be used with mosfets as well, except that their threshold Vgs is quite big so the current sensing resistor will be adjusted upwards as well. Another concern is the op-amp's stability driving high capacitive load (the mosfets' gate capacitance).
Hi!
How well - compared to using a power-OP like LM3875 - would that work? But still the voltage should not be higher than the OP-amps voltage rating? SO that would be the biggest limit for using a small OP like OPA604?
I ask because of course it would be nicer (and cheaper) if one could simply use a small OP amp...
Bye,
Arndt
millwood said:
How well - compared to using a power-OP like LM3875 - would that work? But still the voltage should not be higher than the OP-amps voltage rating? SO that would be the biggest limit for using a small OP like OPA604?
I ask because of course it would be nicer (and cheaper) if one could simply use a small OP amp...
Bye,
Arndt
Cradle22, the answer is no. the topology is only useful if you want to expand the current output capability of an op-amp. so no matter how many pairs of transistors you use and what transistors you use, the voltage swing is still limited to that of the op-amp (minus the voltage drop on the current sensing resistor - which is quite small).
For voltage expansion, you need to bootstrap the op-amp.
For voltage expansion, you need to bootstrap the op-amp.
A freind helped me make this amp and another came around to listen to it.
The unbiassed one said it sounds exactly the same as your GC, the constructor freind and myself said it sounded better?
In the end we ended up switching off the transformer that supplies the transistors and it made no difference to the sound?
Im not sure if the afterburner actually supplies any current in this circuit?
Any one else built this yet?
The unbiassed one said it sounds exactly the same as your GC, the constructor freind and myself said it sounded better?
In the end we ended up switching off the transformer that supplies the transistors and it made no difference to the sound?
Im not sure if the afterburner actually supplies any current in this circuit?
Any one else built this yet?
Luke said:
Luke, your experience makes perfect sense: the "afterburner" makes a difference only if the load is above and beyond what the IC can handle. in the case of the LM3875, it is above 10A peak and 5A consitnues. For an 8ohm load, that means the (theoretical) voltage on the load would be in the 40v-80v range. Given that the IC is supplied by at maximum 40v rails, the afterburner will not help on 8ohm load. That doesn't mean that the afterburner isn't working.
If you are adventurous, try to listen to the same set-up with 4-ohm speakers. or 2-ohm speakers if you are really adventurous.
Hi Millwood,
that makes sense. I recall that the 100pF capacitor in Circlotrons original circuit added terrible hum and was removed.
I reccall that when pushed hard the amp made a funny distortion as I noted in a previousd post that was not clipping (cant prove that as I have no scope trace) but i now suspect it was the output stage kicking in. Is the design flawed or did I do something wrong here?
Luke, I am not sure what amp was giving that funny noise. If you push the straight lm3875 hard, the spike circuitry may kick in and causes distortion tho I have not heard personally how that sounds.
I am not sure how the particular circuitry sounds under load but it is unlikely that the distortion is caused by the IC (when mated to the "afterburner") as it outputs very small voltage.
The advantage of using the lm3875 in this particular set-up is their high rail voltage. A normal op-amp can only go up to 18v per rail and with the voltage drop over the output devices will further limit power of such a set-up.
so if one is to use a normal op-amp, a high-output voltage one is more desirable.
I am not sure how the particular circuitry sounds under load but it is unlikely that the distortion is caused by the IC (when mated to the "afterburner") as it outputs very small voltage.
The advantage of using the lm3875 in this particular set-up is their high rail voltage. A normal op-amp can only go up to 18v per rail and with the voltage drop over the output devices will further limit power of such a set-up.
so if one is to use a normal op-amp, a high-output voltage one is more desirable.
The QUAD 405 and later current dumping amplifiers have an inductor behind the class C section which is chosen such that the distortion generated in the class C section is cancelled at the output. Basically, the integrating voltage to current transfer of the inductor should just cancel the integrating voltage to current transfer of the class A part with its feedback capacitor and output resistor. This can give 20dB or so of improvement, or even more if you are prepared to manually adjust the bridge balance. It is explained in detail in the articles Peter Walker wrote about it. See for example home.jps.net/~shilohz or the other sites you find when you search the Web for QUAD 405.
Hi I tried this circuit again as Millwood suggested, with an 80 ohm resistor replaing the 1 ohm. This was so as the transistors would kick in at around 10ma.
In a previous post Italked about a strange distortion when I turned the amp up loud. Now I had that horoble distortion at all volumes, it was unbearable and had some faint music in the background.
I guess if I were to perservere with this I would try the inductor as MarcelvdG has suggested. I really doubt it would clean this amount of distortion quite honestly.
In a previous post Italked about a strange distortion when I turned the amp up loud. Now I had that horoble distortion at all volumes, it was unbearable and had some faint music in the background.
I guess if I were to perservere with this I would try the inductor as MarcelvdG has suggested. I really doubt it would clean this amount of distortion quite honestly.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.