Another QC design by WhiteDragon with almost same component count but much better in performance.
Cool NanoFarad almost forgot that one.
I believe that thread the OP requested TIP35 specifically.
And also requested a simplified frontend
So it is why lower Ft drivers were chosen. And meet voltage requirements
Of course a more elaborate low distortion design is possible.
But was designed to meet the request of the builder.
For some reason they wanted the gain extremely high, so that could be changed as well.
Very high gain. Gain can be lowered reducing R8 and then R10 be matched to same value.
But yes it shows T4 as VBE for thermal tracking instead of diodes and has simple Zener diode based
current source for Diff.
Again simplified design, to meet request of builder for general purpose transistors.
Good reminder for this thread the 100 ohm degen resistors should be 1% which I noted.
Not specific to either design but specific for any amplifier using Degen in the differential
of course post #42 post looks " upside down" LOL using PNP transistors for input rather than NPN
But if OP for this thread chooses to add a simple current source with zener. This is a example.
Off course values would be adjusted for 12 volt rails not 26 volts.
I believe that thread the OP requested TIP35 specifically.
And also requested a simplified frontend
So it is why lower Ft drivers were chosen. And meet voltage requirements
Of course a more elaborate low distortion design is possible.
But was designed to meet the request of the builder.
For some reason they wanted the gain extremely high, so that could be changed as well.
Very high gain. Gain can be lowered reducing R8 and then R10 be matched to same value.
But yes it shows T4 as VBE for thermal tracking instead of diodes and has simple Zener diode based
current source for Diff.
Again simplified design, to meet request of builder for general purpose transistors.
Good reminder for this thread the 100 ohm degen resistors should be 1% which I noted.
Not specific to either design but specific for any amplifier using Degen in the differential
of course post #42 post looks " upside down" LOL using PNP transistors for input rather than NPN
But if OP for this thread chooses to add a simple current source with zener. This is a example.
Off course values would be adjusted for 12 volt rails not 26 volts.
Last edited:
Yes, i understand. I was the original poster. Anyway is it possible to reduce the gain to around 24dB by raising R18 & without making the amplifier unstable?
Yes correct R18 could be raised to lower the gain.
Stability non issue. Usually people stray away from high resistor values because of possibly noise.
But otherwise because C4 forms highpass filter.
R18 is made to be a higher value so C4 can also be a lower value so it is not so large on the PCB.
Stability non issue. Usually people stray away from high resistor values because of possibly noise.
But otherwise because C4 forms highpass filter.
R18 is made to be a higher value so C4 can also be a lower value so it is not so large on the PCB.
Then instead of raising R18 we can lower the value of R8 & R10 to 22Kohm for 27dB, which is kind of standard gain level (not too high & not too low). Anyway thanks for your replies.
Yes indeed.
Did you build the amp? love to see pictures.
Another fun simple amp. TIP35C use to be available in pretty large packages.
Seen them a lot in MI ( musical instrument) applications.
Low Ft like older 2n3055 but really good gain.
Did you build the amp? love to see pictures.
Another fun simple amp. TIP35C use to be available in pretty large packages.
Seen them a lot in MI ( musical instrument) applications.
Low Ft like older 2n3055 but really good gain.
Unfortunately no, but i promise i'll do it very soon (I'm kinda bored of everything happening in my life) & post pictures for you. I've everything needed for this amp so no problem. But it will be a vero board construction 🙂
What the serpentine line does not tell you: remove the bootstrap C4, and bridge D1, D2, R5 with a large clean-sounding capacitor.
And with 2 x 12 volts do use TO-220 as power transes.
He is designing specifically for his collection of T03 transistors.
Which is fine.
The bootstrap or any other constant current source wont even work without a voltage reference.
So no you don't remove the capacitor. It would defy the whole point.
Which is fine.
The bootstrap or any other constant current source wont even work without a voltage reference.
So no you don't remove the capacitor. It would defy the whole point.
geniuenly intrested which transostors would be better kd503/kd502, 2n3055, potentially knockoff bd911's or these things shown in the picture that i only have one pair of?
id increase the supply voltage to 14/15V rails (becuse thats what i would end up with after rectification) increase r2 to 1K to adjust for my parts selection, then make r3 and r17 220m for the same reason and make r1 and r14 an potentiometer.
Unfortunately, also transistors were and are not developed according to audio criteria. All that remains is to listen, because even different batches of the same type sound different. They are lucky finds. The easiest way to test the sound of active parts is to insert them into a power supply as a diode - connect collector and basis, then have you got two poles (cb - e = diode). It is this sound that the ear hears. And this sound is the same in all functions of this component, whether as a driver, power, power supply, diode, whatever.
So the ideal is to have an amplifier or other device that runs on a single voltage. For example, a small SE or even a DAC or CD player - many simple, "cheap" ones run with just one voltage - audiophile, exemplary!
2n3055s transistors for 8 w amplifier? Bd439/440 would be enough, or even better a simple tda2030/40/50.
re built it in ltspice and made some "cosmetic" changes to adjust for my parts selection
EDIT: i forgot to change to the 2n3055 model
I'm sorry for even using the (wrong) term "RMS power" - apologies!
For effective sine power you need to use the rms max voltage swing and calculate the power from this value.
thus:
supply = 30 V (assumed value)
max peak voltage swing = +/- 13.5 V peak (supply/2 - 1,5 V)
max rms voltage swing = +/- 13,5 Vpk / sqrt(2) = +/- 9,5 V rms
max effective sine wave power = (9.5 V ^2) / 8 ohm = 11,3 W
alternatively and wrongly you could define the max power using a square wave:
supply = 30 V (assumed value)
max peak voltae swing = +/- 13,5V pk
max effective square wave power = (13,5 V ^2) / 8 ohm = 22,8 W
but, as I said, the latter is not a serious calculation method.
Cool, we are getting there. Least we have a Quasi output with the diode so its gonna be friendly.
Of course add the 2N3055 for outputs. Since gain and current and what is needed to drive them is
gonna be completely different than shown.
See we are at 14 volts now. VR1 is not constant current but sets the current regardless.
So variation of 12 to 14 volts is going change everything of course.
Using glass type diodes crossed my mind too so 4148 switching diodes be fine.
You can get bias with 2 diodes, but bias will jump like mad once you get it set with right resistor value.
Adjustment will be smoother and easier with 3 diodes.
Better to use 3 diode drops. 68 ohms wont bias the outputs now with 2 diodes.
likely in real life R4 needs to be adjustable. But the model helps you understand
a starting point.
Few issues, were learning. You lowered VR1 so current is going to go up.
Luckily using 3904/06 so they are very easy to drive. Q3 is easy load.
So we never ever need high current for Q1 /Q2.
In fact for stability which we haven't even got to.
Basically simplified, too much current is going to cause stability issues.
Were actually hoping to get to 1ma but 2ma was a tradeoff.
But you raised the current up by changing VR1 to almost 3ma
In order to get T1/ T2 to have equal current for no DC offset.
Were stuck with the limits of R2 and standard resistor values.
You raised it to 1K makes sense, trying to stick to 10% standard values.
But it needed to go down in value, since current was raised up
Anyways, stick to 2ma current for VR1 or were just shooting our own foots.
In fact we can even try to get down to 1ma even. Which will help with stability.
Q3 is easy to drive.
Show you screen shot of new circuit. Kinda went backwards
T1 and T2 not balanced. I'll give you a hint.
You will never ever get it without 5% or 1% resistors
R2 is going to make it happen no way around it.
You can vary R6 in your model or R8 in mine. or the load its driving to help a little.
But its limited to resistor values as well. I made sure to match my resistor numbers
to original model. Cant keep renumbering everything.
Last edited:
Quite a while ago I spent time trying different QC topologies in LTSpice and, IIRC, I got the best THD with the driver resistors / diode connected before the output resistors rather than directly to the output rail, i.e. in the schematic above, R12 to the emitter of T7 and R4/D4 to the collector of T8.
I was intrigued because I've seen at least three versions of the QC topology (see attached): as in your circuit (e.g. D. Self), with just the top resistor directly to the output rail (e.g. B. Cordell) and as described above (e.g. J. L. Hood). The difference wasn't huge and I'm sure it depends on the overall topology, specific transistors used, etc., but you guys may want to give these alternatives a try.
Another thing that seemed to make a big difference was the choice of diode, and this was very dependent on the specific circuit: some times the best was a transdiode (as shown by Self below), sometimes a 1N400x, sometimes something else. Again, worth trying different alternatives.
I was intrigued because I've seen at least three versions of the QC topology (see attached): as in your circuit (e.g. D. Self), with just the top resistor directly to the output rail (e.g. B. Cordell) and as described above (e.g. J. L. Hood). The difference wasn't huge and I'm sure it depends on the overall topology, specific transistors used, etc., but you guys may want to give these alternatives a try.
Another thing that seemed to make a big difference was the choice of diode, and this was very dependent on the specific circuit: some times the best was a transdiode (as shown by Self below), sometimes a 1N400x, sometimes something else. Again, worth trying different alternatives.
