Hi Sonal,
Thanks for the update. The gain is OK for me. I don't use a preamp, just a 50k pot. I only mentioned the gain as a possible reason for the oscillation. I realized that 22r would raise the gain later but too late to edit. I figured it wouldn't hurt for someone else to hear it if they decided to build this amp. Did you change 47pf to 68pf because you saw oscillation too? I'll leave the 15v zener because most of the ICs I have won't like the 24v. I'll report back after I make the changes.
Blessings, Terry
Changing the gain changes the feedback.
Changing the feedback changes the requirement for compensation.
If the amplifier is near unstable and you reduce the gain you get even closer to instability.
This can show in the overshoot on a sqw wave test signal.
Conversely, increasing the gain can result in a small reduction in the overshoot on fast signals.
Changing the feedback changes the requirement for compensation.
If the amplifier is near unstable and you reduce the gain you get even closer to instability.
This can show in the overshoot on a sqw wave test signal.
Conversely, increasing the gain can result in a small reduction in the overshoot on fast signals.
Hi Guys,
Ok, I tried 68PF, then added a 47pF to that, then 220pF. None of those work. The oscillation is the same. On the scope I am seeing oscillation just before clipping. If I push it into clipping the oscillation goes away. If I lower the rails to about +-28V it doesn't oscillate but that may be because it clips before it gets to the point of oscillation. I tried a few different ICs and still no change. It appears to be affected by voltage because the higher the rail voltage, the worse the oscillation. I suppose it may have to do with the outputs I'm using. As usual, I am open to suggestions.
Blessings, Terry
Ok, I tried 68PF, then added a 47pF to that, then 220pF. None of those work. The oscillation is the same. On the scope I am seeing oscillation just before clipping. If I push it into clipping the oscillation goes away. If I lower the rails to about +-28V it doesn't oscillate but that may be because it clips before it gets to the point of oscillation. I tried a few different ICs and still no change. It appears to be affected by voltage because the higher the rail voltage, the worse the oscillation. I suppose it may have to do with the outputs I'm using. As usual, I am open to suggestions.
Blessings, Terry
Hi Terry,
I have checked AX9 with 40v supply. BD139, BD140 get hot and 2SC4793, 2SA1837 get warm. I kept amp running for 30 minute BD139, BD140 temprature was stable but at 40v heatsink is must on both BD. 45v both driver and pre driver must be on heat sink.
I did changed to 100pF first then checked with value of 56pF, 68pF, 82pF and settled with 68pF.
Regards,
Sonal Kunal
I have checked AX9 with 40v supply. BD139, BD140 get hot and 2SC4793, 2SA1837 get warm. I kept amp running for 30 minute BD139, BD140 temprature was stable but at 40v heatsink is must on both BD. 45v both driver and pre driver must be on heat sink.
I did changed to 100pF first then checked with value of 56pF, 68pF, 82pF and settled with 68pF.
Regards,
Sonal Kunal
Last edited:
There is no predrivers on AX9 only VAS and drivers transistor, try with 100pF base to colector on drivers transistor like on predrivers in A33.
Tried 47p thru 150p on the drivers B-C still oscillating with rail over +-30. Going to try lowering gain and see it that helps.
EDIT: just lowered the gain and still there. +-30v and no oscillation. +-32v and it oscillates. Strange it is not doing it on Sonal's amp. I wonder if we have anything else different?
EDIT: just lowered the gain and still there. +-30v and no oscillation. +-32v and it oscillates. Strange it is not doing it on Sonal's amp. I wonder if we have anything else different?
Last edited:
I got it fixed. Sorry for running all around but after all the compensation did not fix the problem I had to take another look at the VAS transistors. I remembered that I had two different brands in there. Fairchild BD139 and NXP BD140. I pulled those and installed Toshiba 2SC3423-Y/2SA1360-Y. No oscillation so I pulled the caps I added to the drivers and put 68Pf back in the VAS and all is good. Running right now on +-37V without issue. Sounds really nice. I did run some square waves through it and they get pretty rounded starting around 10khz. Not the fastest but still very nice to listen to. I hope this helps someone who may face similar issues.
Blessings,Terry
Blessings,Terry
Last edited:
The original desing works with 3423/1360. It also works perfectly with KSC3503/KSA1381!
Sajti
Those were my other choice. I save those for higher rail stuff. I'll try lowering the cap values in the morning and see if those square waves will crisp up.
Blessings, Terry
It may takes lot of time, as I found, when I did my research. The perfect squarewave sound hars sometime. Small round on the 20kHz can make the highs much sweeter...
Sajti
I have seen that but I'm not convinced that is true. I'm not sure speed has to do with harsh. Of the amps I've built, the one that produces the sharpest square waves is the Ovation nx and it is not harsh at all. Neither are any of the Slewmasters I've built. The waves I'm seeing on this amp look more like ocean waves. I'll try to take some before and after pics just for the fun of it.
Blessings, Terry
Blessings, Terry
JLH was the only Designer I read that mentioned trying to adjust the sqw response to show a nice very slightly rounded shape and NO OVERSHOOT because JLH considered that gave good sound.
I have read a lot in the intervening thirty odd years and I am more convinced that JLH led the way to reading/understanding measurements and their impact on sound quality.
A little bit of overshoot may pass unnoticed but if a reactive load makes that little bit into a variable "lot of overshoot", then bad sound is usually the symptom.
I have read a lot in the intervening thirty odd years and I am more convinced that JLH led the way to reading/understanding measurements and their impact on sound quality.
A little bit of overshoot may pass unnoticed but if a reactive load makes that little bit into a variable "lot of overshoot", then bad sound is usually the symptom.
