Good idea, but I do not think it contributes to the quality. A quick comment, without deep calculations.
The diodes prevent the TDA to drive the TPI's. So the TIP's will be driven only by the C=0,1 uF and/or the R=2k4.
However, the C is much too little. f = 1/2PiRC = 53 kHz, because the R is only hFE x 1 Ohm = 30 Ohm. With 100uF it contributes from 53 Hz and up.
And the 2k4 delivers max 40V/2k4 = 16 mA. Where the TIP needs 2A/hFE = 66 mA in its basis.
A possible solution is to use 2 Darlingtons (TIP 142/147) and adjust a quiescent current of about 20 mA in their emitters. But this makes it all much more complicated.
Why not decreasing the 5,1 Ohm to 2 or 1,5 Ohm? I think this helps more.....
Or step to the inherent more complicated and less stable but reparable solution of driving the power transistors via de powerlines of the TDA and use their collectors as outputs. Elegant solution, without added distortion.
The diodes prevent the TDA to drive the TPI's. So the TIP's will be driven only by the C=0,1 uF and/or the R=2k4.
However, the C is much too little. f = 1/2PiRC = 53 kHz, because the R is only hFE x 1 Ohm = 30 Ohm. With 100uF it contributes from 53 Hz and up.
And the 2k4 delivers max 40V/2k4 = 16 mA. Where the TIP needs 2A/hFE = 66 mA in its basis.
A possible solution is to use 2 Darlingtons (TIP 142/147) and adjust a quiescent current of about 20 mA in their emitters. But this makes it all much more complicated.
Why not decreasing the 5,1 Ohm to 2 or 1,5 Ohm? I think this helps more.....
Or step to the inherent more complicated and less stable but reparable solution of driving the power transistors via de powerlines of the TDA and use their collectors as outputs. Elegant solution, without added distortion.
Be sure if I didn't test thoughtfully I didn't publish this schematic.
The TIPs are indeed driven by the 2.4K resistors. The diodes keep the TIPs in conduction at all times around crossover. Hfe at low currents is at least 30 so from 16mA they can drive almost half Amp. Capacitor is just to add a supplementary current at the very first moment of entry in conduction of transistor because the main reason I added this stage was the glitch around crossover.
I will take time today to post some oscillogrames. To convince you of the contribution of TIP I will post the voltage on the 0.75ohm emitter resistor. It is a half ellipse. You will see how the voltage on this resistor increase to 1.5V at the peak of the sinusoid (1.5A/0.75Ohm=2A).
I'm sorry I didn't take detailed pictures of the glitch of the original circuit at crossover, which disappeared TOTALLY. Maybe I will revert the prototype to take some pics just to have a comparison.
The TIPs are indeed driven by the 2.4K resistors. The diodes keep the TIPs in conduction at all times around crossover. Hfe at low currents is at least 30 so from 16mA they can drive almost half Amp. Capacitor is just to add a supplementary current at the very first moment of entry in conduction of transistor because the main reason I added this stage was the glitch around crossover.
I will take time today to post some oscillogrames. To convince you of the contribution of TIP I will post the voltage on the 0.75ohm emitter resistor. It is a half ellipse. You will see how the voltage on this resistor increase to 1.5V at the peak of the sinusoid (1.5A/0.75Ohm=2A).
I'm sorry I didn't take detailed pictures of the glitch of the original circuit at crossover, which disappeared TOTALLY. Maybe I will revert the prototype to take some pics just to have a comparison.
@andrewt Why don't you check datasheets before jumping to conclusion?
1N4007 forward voltage at 10mA is 0.7V, at 100mA is close to 0.8V.
Vbe of TIP41/42C is under 0.65V at any collector current under 300mA.
What is funny is that I have the prototype of this circuit working and under extended tests and you guys try to convince me that what I see is not true!
1N4007 forward voltage at 10mA is 0.7V, at 100mA is close to 0.8V.
Vbe of TIP41/42C is under 0.65V at any collector current under 300mA.
What is funny is that I have the prototype of this circuit working and under extended tests and you guys try to convince me that what I see is not true!
Comparison between "classic" an "enhanced" schematic
First three oscillograms are done at the original configuration.:
First figure shows very clear the jump in the base of power transistors and the output glitch is also somehow visible
Second shows in detail the jump in base.
Third shows the glitch present even at 1KHz (By the way, zanden, reducing SYNC resistor the glitch accentuate).
Next two are done at the enhanced circuit:
Forth is the equivalent of first for the enhanced circuit.
Fifth is a detail of output and base (similar with second) for the enhanced circuit.
First three oscillograms are done at the original configuration.:
First figure shows very clear the jump in the base of power transistors and the output glitch is also somehow visible
Second shows in detail the jump in base.
Third shows the glitch present even at 1KHz (By the way, zanden, reducing SYNC resistor the glitch accentuate).
Next two are done at the enhanced circuit:
Forth is the equivalent of first for the enhanced circuit.
Fifth is a detail of output and base (similar with second) for the enhanced circuit.
Attachments
Contribution of TIP41/42C Stage
I measured signal over 0.1ohm resistor and 0.75ohm resistor.
First two pics are taken at -10dB of the maximum load.
From first pic is visible that at this power (8Vrms) the contribution of power transistor is very low, most of the power is delivered by the TIP stage and TDA itself.
second shows the contribution of TIP which is also reduced at this power.
Third pic is equivalent of second but near the full power. It shows the contribution of TIP transistors as a constant (as I designed) of about 1Amp during all half sin.
Forth and fifth pics are the equivalent of first two but at full power.
Forth picture shows the contribution of 2SA and TIP42C as being 9.2A at peak.
Fifth picture shows the contribution of TIP41C which is at peak of 1.02/0.75 = 1.36A.
Sixth is taken at same power and load as Forth and Fifth yellow is the amp output and blue is the TIP41 base.
Last is also the amp output and TIP base at a much lower power.
I measured signal over 0.1ohm resistor and 0.75ohm resistor.
First two pics are taken at -10dB of the maximum load.
From first pic is visible that at this power (8Vrms) the contribution of power transistor is very low, most of the power is delivered by the TIP stage and TDA itself.
second shows the contribution of TIP which is also reduced at this power.
Third pic is equivalent of second but near the full power. It shows the contribution of TIP transistors as a constant (as I designed) of about 1Amp during all half sin.
Forth and fifth pics are the equivalent of first two but at full power.
Forth picture shows the contribution of 2SA and TIP42C as being 9.2A at peak.
Fifth picture shows the contribution of TIP41C which is at peak of 1.02/0.75 = 1.36A.
Sixth is taken at same power and load as Forth and Fifth yellow is the amp output and blue is the TIP41 base.
Last is also the amp output and TIP base at a much lower power.
Attachments
hmm email notify have not been working and been a while since i was last here, have just read up on many pages.....
cmorariu dude, great job!!
I have to read the last 5-7 pages again when im a bit less tired then now, but so awesome to see some one with the time and equipment to do the test, pretty much confirming what i have been hearing for 5+ years now.
Again i have to say it.
It is now my design, i found it, it sounded awful, tweaked it for some time and got to the result you see today and when i posted (a few years after building it to be sure that i worked).
I would like to have it perfect if it is not as you describe with the crossover point, i have yet to hear any weird "noise" at that point spending literally 10´s of hours with my face in a speaker having volume below 2W to hear the crossover, it worked "best" with the 6.8ohm, go much outside this and you WILL hear the crossover as it sound likes someone turns the volume knob up 10% in 0.001sec and down again.
Oh and you will always have the users that will use 800hours doing "math" posting why it will not work etc.... if only they would apply a little time just making the damm thing as i described.... this is also why im rarely here....
But many have done and i am so pleased when i read someone doing it and loving it, if it is here or on facebook.
AND im especially pleased with you taking the time to not only build it, test it, but also making MORE tweaks to it and hopefully making it even better
.
ill try to check in a little more often.
cmorariu dude, great job!!
I have to read the last 5-7 pages again when im a bit less tired then now, but so awesome to see some one with the time and equipment to do the test, pretty much confirming what i have been hearing for 5+ years now.
Again i have to say it.
It is now my design, i found it, it sounded awful, tweaked it for some time and got to the result you see today and when i posted (a few years after building it to be sure that i worked).
I would like to have it perfect if it is not as you describe with the crossover point, i have yet to hear any weird "noise" at that point spending literally 10´s of hours with my face in a speaker having volume below 2W to hear the crossover, it worked "best" with the 6.8ohm, go much outside this and you WILL hear the crossover as it sound likes someone turns the volume knob up 10% in 0.001sec and down again.
Oh and you will always have the users that will use 800hours doing "math" posting why it will not work etc.... if only they would apply a little time just making the damm thing as i described.... this is also why im rarely here....
But many have done and i am so pleased when i read someone doing it and loving it, if it is here or on facebook.
AND im especially pleased with you taking the time to not only build it, test it, but also making MORE tweaks to it and hopefully making it even better
.ill try to check in a little more often.
Hum and overdrive with realy small signal
Dear folks,
very long thread so sorry for not reading everyting. I simply build an TDA7294 based amp (stereo) only to drive small speakers not need more than 5W.
It´s simple so I thought ther can´t be any issue on build it. I use the original example schematic from datasheed in hope everything will be fine.
But: un-mute the amp there is a big noise ( hum ). every realy small signal ( less than 20mv ) will overdrive teh amp and a rectangel output signal is produced.
It seemed, that there is no feedback control but I chekced the schemati and PCB more than 5 times. I see no mistakes..
So, if you can give me any hint whats going wron will be great.
Thanks a lot
Karsten
P.S. sorry for bad english
Dear folks,
very long thread so sorry for not reading everyting. I simply build an TDA7294 based amp (stereo) only to drive small speakers not need more than 5W.
It´s simple so I thought ther can´t be any issue on build it. I use the original example schematic from datasheed in hope everything will be fine.
But: un-mute the amp there is a big noise ( hum ). every realy small signal ( less than 20mv ) will overdrive teh amp and a rectangel output signal is produced.
It seemed, that there is no feedback control but I chekced the schemati and PCB more than 5 times. I see no mistakes..
So, if you can give me any hint whats going wron will be great.
Thanks a lot
Karsten
P.S. sorry for bad english
Comparison between "classic" an "enhanced" schematic - new measurements
I reversed my prototype to the original Dr. Frost schematic to take some measurements and then back to "enhanced" version for same measurements and now I'm able to post the results.
First I must state that I don't want to discourage anybody to build the original circuit which is really rewarding and hard to beat in terms of quality/simplicity. The "enhanced" circuit corrects some imperfections that might be impossible to detect by human ear. It does this at the expense of using some more components.
First I will re post the enhanced schematic which suffered minor but justified changes (pic1).
I used a +/-37V power supply which under max load went down to +/-32V. Load was resistive of 1.7ohm. All measurements were taken at 20.000Hz which is the most demanding for the circuit, at two different signal levels, maximum level (named +10dB) and a lower level by 10dB (named 0dB).
Pictures 2 and 3 shows the output of the (enhanced) amp at 0 and 10dB.
0dB corresponds to a RMS power of 19.7W, 10dB corresponds to a RMS of 208W, on the 1.7ohm load.
I concentrated on the slope near the crossover, the weak point of the classic circuit. The yellow waveform is the output of the amp, the blue one the voltage on 0.1ohm resistor.
Classic circuit:
Picture 4 shows the output of the classic circuit at 0dB using a 5.1ohm SYNC resistor, picture 5 using 2.5ohm SYNC resistor.
Picture 6 shows the output of the classic circuit at 0dB using a 5.1ohm SYNC resistor, picture 7 using 2.5ohm SYNC resistor.
Using 2.5ohm as SYNC strongly overheats TDA chip at full power I had to rise power and test for short periods.
Regarding the 10dB waveform keep in mind that the distortion is at about 2.5V and the amplitude of the waveform is of 30V! More noticeable will be at 0dB amplitude with a deformation of wawe between 1 and 2 volts at a max amplitude of about 6V.
Enhanced circuit:
Picture 8 shows the output of the enhanced circuit without 0.47uF capacitors at 0dB.
Pictures 9 and 10 shows the output of the enhanced circuit with 0.47uF capacitors at 0dB and 10dB.
As you can see all distortions disappeared!
I reversed my prototype to the original Dr. Frost schematic to take some measurements and then back to "enhanced" version for same measurements and now I'm able to post the results.
First I must state that I don't want to discourage anybody to build the original circuit which is really rewarding and hard to beat in terms of quality/simplicity. The "enhanced" circuit corrects some imperfections that might be impossible to detect by human ear. It does this at the expense of using some more components.
First I will re post the enhanced schematic which suffered minor but justified changes (pic1).
I used a +/-37V power supply which under max load went down to +/-32V. Load was resistive of 1.7ohm. All measurements were taken at 20.000Hz which is the most demanding for the circuit, at two different signal levels, maximum level (named +10dB) and a lower level by 10dB (named 0dB).
Pictures 2 and 3 shows the output of the (enhanced) amp at 0 and 10dB.
0dB corresponds to a RMS power of 19.7W, 10dB corresponds to a RMS of 208W, on the 1.7ohm load.
I concentrated on the slope near the crossover, the weak point of the classic circuit. The yellow waveform is the output of the amp, the blue one the voltage on 0.1ohm resistor.
Classic circuit:
Picture 4 shows the output of the classic circuit at 0dB using a 5.1ohm SYNC resistor, picture 5 using 2.5ohm SYNC resistor.
Picture 6 shows the output of the classic circuit at 0dB using a 5.1ohm SYNC resistor, picture 7 using 2.5ohm SYNC resistor.
Using 2.5ohm as SYNC strongly overheats TDA chip at full power I had to rise power and test for short periods.
Regarding the 10dB waveform keep in mind that the distortion is at about 2.5V and the amplitude of the waveform is of 30V! More noticeable will be at 0dB amplitude with a deformation of wawe between 1 and 2 volts at a max amplitude of about 6V.
Enhanced circuit:
Picture 8 shows the output of the enhanced circuit without 0.47uF capacitors at 0dB.
Pictures 9 and 10 shows the output of the enhanced circuit with 0.47uF capacitors at 0dB and 10dB.
As you can see all distortions disappeared!
Attachments
More about "enhanced" circuit
All pictures are related to schematic and test conditions of previous post.
Waveforms in 2SC5200 & 2SA1943 emitters versus output:
1. Voltage in absence of signal. As you can see voltage is about 25mV, which means the current through class AB stage formed by TIP41/42C is about 250mA in absence of signal. This current can be adjusted by modifying 2.4K resistors. I don't advise more than 300mA nor less than 100mA.
2. Voltage at 0dB signal. Pink waveform is the sum of first two and show the current through load. You can see at current crossover points the class A behavior of the circuit.
Waveforms across 0.1ohm resistor (reversed) and across 0.47ohm resistor:
3. At 0dB signal. (no 0.47uF capacitors)
As you can see at this power TIP41C drives 1A (0,47V/0.47ohm). The total peak current through both transistors is 0.34V/0.1ohm = 3.4A.
4. At 10dB signal. (no 0.47uF capacitors)
As you can see at this power TIP41C drives 1.2A (0,57V/0.47ohm). The total peak current through both transistors is 1.18V/0.1ohm = 11.8A.
(Voltage on 0.1ohm don't show the contribution of TDA to the total current delivered to load).
5. & 6. Same situation as 3 & 4 but with 0.47uF capacitors present.
Power contribution of TIP stage is enhanced (the benefit of capacitors in fidelity is seen in previous post).
Considerations:
The AB stage with TIP transistors not only correct the crossover distortions but contribute with its share to the total power delivered by the amp.
As tested the enhanced amp is able to continously deliver 208Wrms into 1.7ohm at high fidelity. (I measured FFT of the input and output signal at full power and they are absolutely similar) In my experiment power was limited by the poor transformer. Using a good transformer (more than 400VA) able to deliver 32 - 34 Vac at least 300Wrms can be delivered on 1.7ohm load. That means 600Wrms on 3.5ohm on a bridged configuration, which is quite impressive!
At this point I will start building the double bridged stereo amplifier I desired. I'll come back with pics and measurements when ready.
All pictures are related to schematic and test conditions of previous post.
Waveforms in 2SC5200 & 2SA1943 emitters versus output:
1. Voltage in absence of signal. As you can see voltage is about 25mV, which means the current through class AB stage formed by TIP41/42C is about 250mA in absence of signal. This current can be adjusted by modifying 2.4K resistors. I don't advise more than 300mA nor less than 100mA.
2. Voltage at 0dB signal. Pink waveform is the sum of first two and show the current through load. You can see at current crossover points the class A behavior of the circuit.
Waveforms across 0.1ohm resistor (reversed) and across 0.47ohm resistor:
3. At 0dB signal. (no 0.47uF capacitors)
As you can see at this power TIP41C drives 1A (0,47V/0.47ohm). The total peak current through both transistors is 0.34V/0.1ohm = 3.4A.
4. At 10dB signal. (no 0.47uF capacitors)
As you can see at this power TIP41C drives 1.2A (0,57V/0.47ohm). The total peak current through both transistors is 1.18V/0.1ohm = 11.8A.
(Voltage on 0.1ohm don't show the contribution of TDA to the total current delivered to load).
5. & 6. Same situation as 3 & 4 but with 0.47uF capacitors present.
Power contribution of TIP stage is enhanced (the benefit of capacitors in fidelity is seen in previous post).
Considerations:
The AB stage with TIP transistors not only correct the crossover distortions but contribute with its share to the total power delivered by the amp.
As tested the enhanced amp is able to continously deliver 208Wrms into 1.7ohm at high fidelity. (I measured FFT of the input and output signal at full power and they are absolutely similar) In my experiment power was limited by the poor transformer. Using a good transformer (more than 400VA) able to deliver 32 - 34 Vac at least 300Wrms can be delivered on 1.7ohm load. That means 600Wrms on 3.5ohm on a bridged configuration, which is quite impressive!
At this point I will start building the double bridged stereo amplifier I desired. I'll come back with pics and measurements when ready.
Attachments
Dear Cmorariu,
I was on long holidays...
You did really a great job. I overlooked the way you did the enhancement, it is a simple, good, stable and robust improvement.
Most I can understand now.
One point I still cannot explain/understand. Why are the glitches more penetrant if you diminish the "feed back resistor" from 6,8 Ohm to 2,5 Ohm? I have thought longtime about it, but can not get the clue. Dr. Frost's experiments prove that 6,8 Ohm is the optimum. And why would I deny real live experiments? But why is that optimum at that specific value?
I would like to have a theoretically explanation for that. Any idea?
Kind regards,
Jan
I was on long holidays...
You did really a great job. I overlooked the way you did the enhancement, it is a simple, good, stable and robust improvement.
Most I can understand now.
One point I still cannot explain/understand. Why are the glitches more penetrant if you diminish the "feed back resistor" from 6,8 Ohm to 2,5 Ohm? I have thought longtime about it, but can not get the clue. Dr. Frost's experiments prove that 6,8 Ohm is the optimum. And why would I deny real live experiments? But why is that optimum at that specific value?
I would like to have a theoretically explanation for that. Any idea?
Kind regards,
Jan
the thing about crossover distortion is that it sounds horrible to me with efficient speakers. in the early days of solid state there were very few solid state amps that sounded decent to me when driving a klipsch la scala or grf corner horn. they also sounded terrible to me when driving speakers that had terrific transient response such as quad electrostats. they did sound ok driving typical midfi 2 cuft sealed box speakers like an ADC brentwood or AR2.
at higher currents you may see slew rate effects that limits the chip's ability to respond to feedback and clean up the mess. just a guess. i haven't tested for it.
Im not talking about crossover distortion, this amp does VERY well in that area, im talking about the crossover from TDA to Transistors, the point where they have to match so that the transistors helps along/takes over.
i was responding mainly to cmorariu and i simply adopted the nomenclature he was already using (see the quote).
in any event, the nature of the cmoraiu waveform anomlay is quite similar (though differently located) to 0-v crossover distortion no matter what anyone may name it.
so i am implicitly positing (without deciding--for others to test) that cmoratu-crossover-distortion may sound as bad as 0-v crossover distortion.
of course it is just theory until tested.