Hi guys
I use the amp on "Designing Audio Power Amplifier" from Bob Cordell
I changed some parts of it
I want you guys to tell me if there is any mistake in schematic
I run this amp and its sound is super clear and amazing
I test it in 1Khz sine wave under 3.9 ohms dummy load for about 2 hours and there were no problem at all
I appreciate to tell me your opinion about it or any improvement I can make
PDF
http://www.freeuploadsite.com/do.php?id=66395
I use the amp on "Designing Audio Power Amplifier" from Bob Cordell
I changed some parts of it
I want you guys to tell me if there is any mistake in schematic
I run this amp and its sound is super clear and amazing
I test it in 1Khz sine wave under 3.9 ohms dummy load for about 2 hours and there were no problem at all
I appreciate to tell me your opinion about it or any improvement I can make
http://www.freeuploadsite.com/do.php?id=66395
An externally hosted image should be here but it was not working when we last tested it.
That schemo looks very wrong. The input LTP has the collectors and emitters transposed. PNP types won't turn on if connected with the collectors going to a positive rail, and the emitters to the negative. The CB junction will be forward biased, and that's not what you want for amplification.
Also, 3K2 is way too much for emitter degen resistors. 320R would be more believable. Likewise, 3K2 is too much for the current mirror collector loads. Again, I'd expect to see something more like 320R there too.
Another problem is that the gNFB loop isn't DC isolated. Having a base DC return of 27K on one side of an LTP, and less than 1K on the other is just asking for a wicked DC offset at the output.
Also, 3K2 is way too much for emitter degen resistors. 320R would be more believable. Likewise, 3K2 is too much for the current mirror collector loads. Again, I'd expect to see something more like 320R there too.
Another problem is that the gNFB loop isn't DC isolated. Having a base DC return of 27K on one side of an LTP, and less than 1K on the other is just asking for a wicked DC offset at the output.
An externally hosted image should be here but it was not working when we last tested it.
Thanks Miles Prower for your tips
I drew LTP's bjts wrong in schematic
How about now??
I raised current tail from 150uA to about 1mA
and decrease the emiter degeneration resistors to 470 ohms
I also isolate the NFB from DC ground
Last edited:
do anybody have any idea why when I use Cascode VAS (the parts I've drew a red line across it) in my amp I get no output?
An externally hosted image should be here but it was not working when we last tested it.
The tail current for the LTP is on the thin side. I'd double it to 2.0mA. Even though the 2nd pre has an emitter follower buffer, a bit more current in the LTP can't hurt.
The DC isolation capacitor below the 1K resistor is way too small. This gives you a turn-over at 15.92Hz. 100uF would be better. Better still would be more like 470uF. That would drop that turn-over a couple of decades below the lowest audio frequency, and get those phase shifts out of the way.
As for the other problem (cascode 2nd pre) there's no tellin', other than something isn't turning on. Do you have this design on a proto board? If so, best to measure DC voltages to see if they're correct.
The biggest problem with active loads is that it leaves voltages indeterminate. A CCS will work with a wide variety of voltages across it. Usually, when actively loading the collectors of the LTP (I don't favour this myself, as I think passive collector loads there sound better) there's some provision for a reference voltage.
If everything was just peachy when the 2nd pre was just a single transistor loaded with a ring of two CCS, but stops working when a cascode transistor is added, then something must be different. If it stops playing and goes silent, then something is intercepting the signal.
From the schemo, I see a couple of possibilities. Q18 is turning off. It's getting its bias from two diodes in series. Are you sure the polarity is right? If one of those diodes is connected in the reverse direction, Q18's base will be left floating, and everything shuts down: no current for the 2nd pre, no bias for the finals. The other possibility is that Q13 is turning on when it should be off. If it's on, no signal. That's why I asked if this was a prototype physically on a proto-board, or if it was a sim. You need to probe around in there to check DC bias voltages to see what's on when it should be off, or what's off when it should be on. Regarding polarities, are you sure you're connecting the transistor correctly, and you haven't transposed the emitter and collector?
If all else fails, you may have to go with the original plan.
A passive collector load will drop whatever voltage it needs to in order to draw the design nominal collector current. You can't tell just from looking at a schemo what the actual collector voltage is if there's an active load. This fixes the collector current to one specific value. Since both the CCS and the collector set the current, something has to give, and the only thing left is the voltage.
You can't tell what current the emitter follower is drawing. If it's too much, then it will turn on Q13, and you lose your signal. Whether or not you cascode the 2nd pre, shouldn't make a difference, but strange things happen along the way from a preliminary schemo to finished design. That's why electronic design is part science and part art. The art part comes into play when you see unexpected results.
Thanks a lot for your answers
It helps a lot
I'll check the voltages when I add cascode in circuit and tell you the result
Thanks again
by the way
what kinds of protection this circuit needs beside fuses in power line and DC protection for loudspeaker??
I test this circuit with sine wave for 2 hours under full power (3.9 ohms - dummy load) and the temp of heat sink reached 90 degrees without any clipping in the output on oscilloscope
of course when I build the final version, I place fan on heat sink to turn-on in about 60 degrees
It helps a lot
I'll check the voltages when I add cascode in circuit and tell you the result
Thanks again
by the way
what kinds of protection this circuit needs beside fuses in power line and DC protection for loudspeaker??
I test this circuit with sine wave for 2 hours under full power (3.9 ohms - dummy load) and the temp of heat sink reached 90 degrees without any clipping in the output on oscilloscope
of course when I build the final version, I place fan on heat sink to turn-on in about 60 degrees
Thanks a lot for your answers
It helps a lot
I'll check the voltages when I add cascode in circuit and tell you the result
Thanks again
by the way
what kinds of protection this circuit needs beside fuses in power line and DC protection for loudspeaker??
I test this circuit with sine wave for 2 hours under full power (3.9 ohms - dummy load) and the temp of heat sink reached 90 degrees without any clipping in the output on oscilloscope
of course when I build the final version, I place fan on heat sink to turn-on in about 60 degrees
It helps a lot
I'll check the voltages when I add cascode in circuit and tell you the result
Thanks again
by the way
what kinds of protection this circuit needs beside fuses in power line and DC protection for loudspeaker??
I test this circuit with sine wave for 2 hours under full power (3.9 ohms - dummy load) and the temp of heat sink reached 90 degrees without any clipping in the output on oscilloscope
of course when I build the final version, I place fan on heat sink to turn-on in about 60 degrees
I add cascode again in certain right way but I got the same result as before
The voltage across Emitter-Collector of Q4 from 29V drops to about 1.5 Volts
and the voltage across Emitter-Collector of Q18 is about zero
I also don't think that Q13 is the problem because I disconnect its base connection but the problem was the same
I change the circuit back to design without cascode and my signal at output got back
by the way a strange thing
the negative rail voltage in cascode design drops to about 5 volts but the positive stay the same
The voltage across Emitter-Collector of Q4 from 29V drops to about 1.5 Volts
and the voltage across Emitter-Collector of Q18 is about zero
I also don't think that Q13 is the problem because I disconnect its base connection but the problem was the same
I change the circuit back to design without cascode and my signal at output got back
by the way a strange thing
the negative rail voltage in cascode design drops to about 5 volts but the positive stay the same
These transistors are going deep into saturation, and locking up the negative rail. If the finals are pulling the negative rail down to 5V, I'm surprised something hasn't poofed. Those MJL 21183s must be getting hotter than hell. Q18 bias chain becomes suspect here, if it's still working OK with Q4 only.
There is also the possibility that there's an oscillation or other instability that's causing negative rail lock-up, and the saturation of the second pre is a symptom, and not a cause. That would require some o'scoping while powering up.
If this was turning on, it should pull everything down to cutoff.
Hi to all
I finally built its pcb
But when I turn it on it had a lot of noise on it (loud horrible sound from speaker) and Q7 got real hot
But when I touch connection between R8 & R7 the loud noise decrease considerably
I change the R8 value to 220 ohms and R7 to 470 ohms and then the loud noise went away
But now I have a bit distortion on low volume (I think it's normal with this values of resistors and with voltage drop that made by them)
My question is here that what's the sourse of this noise and how can I change the resistors back to their first values?
By the way when I test this circuit on breadboard there was no such problem with original resistors values
I finally built its pcb
But when I turn it on it had a lot of noise on it (loud horrible sound from speaker) and Q7 got real hot
But when I touch connection between R8 & R7 the loud noise decrease considerably
I change the R8 value to 220 ohms and R7 to 470 ohms and then the loud noise went away
But now I have a bit distortion on low volume (I think it's normal with this values of resistors and with voltage drop that made by them)
My question is here that what's the sourse of this noise and how can I change the resistors back to their first values?
By the way when I test this circuit on breadboard there was no such problem with original resistors values
This looks like an oscillation problem. Touching adds enough stray reactance to damp out the oscillation. As for what's causing it, hard to tell. It could be due to the assembly on a bread board. These prototyping boards typically have a lot of stray capacitance between rows.
Maybe try some base stopper resistors? See if that helps.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.