Here is a very cheap but surprisingly decent sounding little amp. It can probably be built with the parts in your junk box. I built it and it is stable on the scope.
Some explanations:
- The opamp output (pin 6) is not connected. The signal is taken from the compensation pin (pin 5). Doing so bypasses the op-amp's output stage. To avoid loading, you must use high hfe output transistors. The bc337-40 shown are really good for this use.
- The input pot is a 100K linear one. Interaction with the input resistors gives a usable (although not perfect) curve for audio. Input impedance is kind of low with a minimum of 13K. Offset varies between 13 and 3mv in one channel, depending of the position of the pot. The other channel is in between 9 and 0.8mv.
- The emitter resistors' value is a bit high at 33r. I picked the value according to the forward voltage of the common red leds I had in hand. It gives me a bias of around 12ma. It somehow limits the current in case of a short. Of course, you can go lower and increase the iddle current as long as you don't exceed the dissipation limit of the output transistors. Small TO92 can't be cooled very effectively.
- Power supply can be about anything in between +/-9 and +/-15VDC. Some additional decoupling isn't shown, apply as needed on the opamps power pins.
I've to give credit where it is due, the original idea is from Nuvistor here: http://www.diyaudio.com/forums/soli...reamp-circuit-4.html?postid=675703#post675703
Some explanations:
- The opamp output (pin 6) is not connected. The signal is taken from the compensation pin (pin 5). Doing so bypasses the op-amp's output stage. To avoid loading, you must use high hfe output transistors. The bc337-40 shown are really good for this use.
- The input pot is a 100K linear one. Interaction with the input resistors gives a usable (although not perfect) curve for audio. Input impedance is kind of low with a minimum of 13K. Offset varies between 13 and 3mv in one channel, depending of the position of the pot. The other channel is in between 9 and 0.8mv.
- The emitter resistors' value is a bit high at 33r. I picked the value according to the forward voltage of the common red leds I had in hand. It gives me a bias of around 12ma. It somehow limits the current in case of a short. Of course, you can go lower and increase the iddle current as long as you don't exceed the dissipation limit of the output transistors. Small TO92 can't be cooled very effectively.
- Power supply can be about anything in between +/-9 and +/-15VDC. Some additional decoupling isn't shown, apply as needed on the opamps power pins.
I've to give credit where it is due, the original idea is from Nuvistor here: http://www.diyaudio.com/forums/soli...reamp-circuit-4.html?postid=675703#post675703
Attachments
Good the see this come over from Headwize.
One of these days, I plan to build this after I get through with a couple of Kumisa III's
One of these days, I plan to build this after I get through with a couple of Kumisa III's
Any idea what the ( internal ) bias current is in the stage connected to pin 5 ?
What is the designed quiescent current through the BF245C ?
Wouldn't it be possible to use a darlington output transistor which would make the ( regular ) opamp output ( any opamp ) run in class A mode ?
This is again an externally power boosted opamp with a small twist ! Should sound good and hopefully "very good" ? No way to find out except by trying it out. Anyone tried this one already ?
What is the designed quiescent current through the BF245C ?
Wouldn't it be possible to use a darlington output transistor which would make the ( regular ) opamp output ( any opamp ) run in class A mode ?
This is again an externally power boosted opamp with a small twist ! Should sound good and hopefully "very good" ? No way to find out except by trying it out. Anyone tried this one already ?
Last edited:
It's shown properly in the NJM5534 data sheet. Wonder how they goofed that . Nobody checks the data sheets before publishing ?
Do you think it will work the same if I connect a bipolar supply FirstWatt B1 stage to pin 5 of NE5534?
Is there a Firstwatt BI stage circuit diagram on this forum ?
The intermediate stage of the NE5534 would need a very light load to make sure you do not derail it's performance. That would mean that it probably should have a load much less than 0.5 mA ? Sourcing or sinking any current should be a fraction of the standing current in that stage. Would the standing current be 2 to 3 mA ?
You can achieve a light load with a darlington or a Sziklai pair.
The intermediate stage of the NE5534 would need a very light load to make sure you do not derail it's performance. That would mean that it probably should have a load much less than 0.5 mA ? Sourcing or sinking any current should be a fraction of the standing current in that stage. Would the standing current be 2 to 3 mA ?
You can achieve a light load with a darlington or a Sziklai pair.
As suggested by Nuvistor, I used 5ma for the ccs.
The problem of loading isn't that bad: even pushing 100ma into a low impedance load, the output stage doesn't need more than 0.3ma as input current with the bc337-40. That's 6% of the standing current. Less if you handpick top of the range transistors.
What is more problematic is the fact that the voltage across the ccs varies quite a lot with the signal and this impacts strongly such a simple ccs. However, this only matters when you deal with high impedance loads (with higher voltage swings but less current demands), so the loading still stays completely marginal. According to sims, the ccs could drop as low as 2.2ma, but the loading (for a 300r load at clipping) would be a puny 0.06ma.
A bjt+led ccs would be more immune to such variations but would take more space onboard.
I forgot to add that the maximum output voltage is 11V, peak to peak or something like 4Vrms, for a 300ohms load.
PS: all calculations assuming +/- 9V supplies.
We could also have an opamp drive the output stage consisting of a complementary darlington pair and biased by red LED's. 7 resistors , 2 LED's and 2 darlington transistors . The darlington pair would keep the output of the opamp in the class A operation region. Opamps could be rolled. BDX33/34 or BDX53/54 output pair.
I'm sure this must have been tried by somebody. Might be interesting to compare this with the circuit in the first post to see if they sound different.
Anyone ?
I'm sure this must have been tried by somebody. Might be interesting to compare this with the circuit in the first post to see if they sound different.
Anyone ?
I just came across this thread, looks like a revival of a Headwize thread that got cut short by Headwize's demise.
Regarding 5534 pin 5, it's connected to part of the output stage, and will sink as much current as pin 6, but cannot source current. Look at the schematic in post 4, pin 6 is internally connected to pin 5 by a resistor and diode, which are for negative output current limiting, so pin 5 is not current limited. I have tried driving 300 ohm headphones directly from pin 5 with a 20ma current source, but output shorts will damage the IC, and bass is much better with the discrete output stage.
Regarding ccs, since pin 5 is a low impedance current sink, a resistor to +V in place of the JFET may be good enough, I used a JFET because I had them.
Regarding 5534 pin 5, it's connected to part of the output stage, and will sink as much current as pin 6, but cannot source current. Look at the schematic in post 4, pin 6 is internally connected to pin 5 by a resistor and diode, which are for negative output current limiting, so pin 5 is not current limited. I have tried driving 300 ohm headphones directly from pin 5 with a 20ma current source, but output shorts will damage the IC, and bass is much better with the discrete output stage.
Regarding ccs, since pin 5 is a low impedance current sink, a resistor to +V in place of the JFET may be good enough, I used a JFET because I had them.
Last edited:
- Status
- Not open for further replies.