Probably not:
Without output devices the open loop output impedance is about 5 ohms from R31//R131 (or 10 ohms if the driver stage goes out of class A) which forms a voltage divider with the RC link on the output. This makes the output lag behind what the VAS is doing. The reason for the overshoot is that the VAS needs some time to recover because of windup during the slew.
With such a high signal level as you are using, the VAS is probably even saturating. (you could take a look at its collector voltage to determine this) This would explain the suspect-looking bouncing when going negative. Does it look cleaner at lower signal level?
If you want more representative results, you could try disconnecting the R53 - C17 link for these tests. Another way is to let the driver transistors drive the output through a pair of the 0.33 ohm resistors. (if bias voltage can be set low enough) The overshoot should disappear.
I'd also like to suggest that you take a look at driver emitter-to-emitter voltage, to determine if the OPS bias setting circuit is working properly, before you connect the output stage.
Without output devices the open loop output impedance is about 5 ohms from R31//R131 (or 10 ohms if the driver stage goes out of class A) which forms a voltage divider with the RC link on the output. This makes the output lag behind what the VAS is doing. The reason for the overshoot is that the VAS needs some time to recover because of windup during the slew.
With such a high signal level as you are using, the VAS is probably even saturating. (you could take a look at its collector voltage to determine this) This would explain the suspect-looking bouncing when going negative. Does it look cleaner at lower signal level?
If you want more representative results, you could try disconnecting the R53 - C17 link for these tests. Another way is to let the driver transistors drive the output through a pair of the 0.33 ohm resistors. (if bias voltage can be set low enough) The overshoot should disappear.
I'd also like to suggest that you take a look at driver emitter-to-emitter voltage, to determine if the OPS bias setting circuit is working properly, before you connect the output stage.
Hey!
MJL21193
Is yours suggestion overthrown by megajockes one? if not what value range do you suggest?
megajocke
Removing Zobel worked. Signal was nice. So next step was adding one output pair. Done. Outputs are not properly biased, but that shouldn't matter? I am testing without heatsink. Predrivers and drivers also get hot, but I am making measurements quickly. Some strange curve appeared on clipping:
10 kHz square clipping:
10kHz sine clipping
Added the Zobel and WHOA, something is not good
10 kHz sine clipping:
And the nicest image 10kHz square went beserk above output voltage 70Vpp:
Getting more and more confused
Rait
MJL21193
Is yours suggestion overthrown by megajockes one? if not what value range do you suggest?
megajocke
Removing Zobel worked. Signal was nice. So next step was adding one output pair. Done. Outputs are not properly biased, but that shouldn't matter? I am testing without heatsink. Predrivers and drivers also get hot, but I am making measurements quickly. Some strange curve appeared on clipping:
10 kHz square clipping:
An externally hosted image should be here but it was not working when we last tested it.
10kHz sine clipping
An externally hosted image should be here but it was not working when we last tested it.
Added the Zobel and WHOA, something is not good
10 kHz sine clipping:
An externally hosted image should be here but it was not working when we last tested it.
And the nicest image 10kHz square went beserk above output voltage 70Vpp:
An externally hosted image should be here but it was not working when we last tested it.
Getting more and more confused
Rait
I didn't realize that the output stage wasn't connected when I suggested that. It's good to keep in mind for when the amp is up and running properly as it can tame overshoot. Start with a low value - 5pF and work your way up until you are happy with the way a 20khz squarewave looks on the scope.
You seem to be having a real problem there.
You seem to be having a real problem there.
Wow, that sure looks strange! It should not matter that the output transistors are not biased. (it would probably blow up if you tried biasing them without heatsink fitted; thermal instability would be extremely high)
Could it be that one of the grounds has gotten disconnected from the others? (signal generator, oscilloscope, input decoupling, output decoupling, input signal ground and PSU centerpoint)
(it would probably blow up if you tried biasing them without heatsink fitted; thermal instability would be extremely high)Could it be that one of the grounds has gotten disconnected from the others? (signal generator, oscilloscope, input decoupling, output decoupling, input signal ground and PSU centerpoint)
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.