Ok.
Maybe check for any DC offset on bad output first then....
Have you got a signal generator? Or a test cd. And have a look on the outputs with a scope and see if the bad output is balanced. If they are symmetrical the output splitter is hopefully working correctly and the distortion is probably before.
Disconnect the bad side of the stepped attenuator and test that with either
A signal generator and scope to make sure it's attenuating correctly (preferred method)
Just a multimeter carefully checking resistance of each step.
What sort of stepped attenuator is it? Hopefully a ladder or a series so the input to ground always measures about 100k. I suspect a shunt attenuator would work properly with this setup. Check this for each step.
if the input to ground is always 100k you could hook up a battery across it and measure the output voltage. A 9v should only deliver just under 0.1 mA but you could stick another 100k on the input to halve that and offer current limiting in case there is a short somewhere.
The output should then increase from 0v when off in increasing steps to 4.5v when up full.
Actually thinking about it. I would check the attenuator first. Lots of joints and a complex switch is more likely to fail than the rest of the circuit with no evidence of anything burnt.
What's the voltage drop on the the 2 power resistors? R4 I think. That's worth checking and that will also tell you if your CSS is working ok. Should be in the region 95 to 100v across them
Maybe check for any DC offset on bad output first then....
Have you got a signal generator? Or a test cd. And have a look on the outputs with a scope and see if the bad output is balanced. If they are symmetrical the output splitter is hopefully working correctly and the distortion is probably before.
Disconnect the bad side of the stepped attenuator and test that with either
A signal generator and scope to make sure it's attenuating correctly (preferred method)
Just a multimeter carefully checking resistance of each step.
What sort of stepped attenuator is it? Hopefully a ladder or a series so the input to ground always measures about 100k. I suspect a shunt attenuator would work properly with this setup. Check this for each step.
if the input to ground is always 100k you could hook up a battery across it and measure the output voltage. A 9v should only deliver just under 0.1 mA but you could stick another 100k on the input to halve that and offer current limiting in case there is a short somewhere.
The output should then increase from 0v when off in increasing steps to 4.5v when up full.
Actually thinking about it. I would check the attenuator first. Lots of joints and a complex switch is more likely to fail than the rest of the circuit with no evidence of anything burnt.
What's the voltage drop on the the 2 power resistors? R4 I think. That's worth checking and that will also tell you if your CSS is working ok. Should be in the region 95 to 100v across them
Last edited:
DC output is consistently near zero on both channels. There is about 7mV noise on the good channel and around 14mV and 21mV noise (VAC with hand-held cheap meter) on the bad channel at capacitor out.
The other side of the four output capacitors are at 262V and 97V consistent across both channels.
Voltage across R11 and R12 is consistently 97.2V across both channels.
For C2, again consistent - the output DC voltage is 90V and input is 140V.
The voltage drop across R4 (CCS 12.5K) for both channels is 95.2V and voltage drop across R8 is around 45V, again same across both sides.
All LEDs are lit.
It's a DACT attenuator with discrete SMD resistors. Pulling it will be fun, but would you guys assume at this point it is the likely suspect?
The other side of the four output capacitors are at 262V and 97V consistent across both channels.
Voltage across R11 and R12 is consistently 97.2V across both channels.
For C2, again consistent - the output DC voltage is 90V and input is 140V.
The voltage drop across R4 (CCS 12.5K) for both channels is 95.2V and voltage drop across R8 is around 45V, again same across both sides.
All LEDs are lit.
It's a DACT attenuator with discrete SMD resistors. Pulling it will be fun, but would you guys assume at this point it is the likely suspect?
Isn't the volume control connected via 6 wire to the main board? Can you not desolder from the top? My main board has through hole so I can heat the top for a second or 2 then the wire will be loose.
I would definitely check your attenuator is working correctly especially as there are relays and control stuff going on.
It looks as though the nuts and bolts if the main board are measuring correctly or at least the same. If one channel is coming out full blast then the noise floor may well measure higher.
Check the attenuator if at least to rule it out.
I would definitely check your attenuator is working correctly especially as there are relays and control stuff going on.
It looks as though the nuts and bolts if the main board are measuring correctly or at least the same. If one channel is coming out full blast then the noise floor may well measure higher.
Check the attenuator if at least to rule it out.
I think I found the problem - a cold solder joint on the left channel ground. Fixed and now just a few mV noise on all four capacitor outputs. Re-casing as we speak. Here is the culprit (far left pin). I could move the attenuator and see the entire pin could move. No continuity to ground.
I currently run the Impasse into an Ashly crossover and then split to subs and full-range speakers. The Ashly has input attenuation and output attenuation, but I run both at unity.
I think I can handle the extra heater current, but not sure how well 6BX7 or 6BL7 might work to lower the overall gain? Otherwise, maybe need something with a different attenuation curve (custom).
I think I can handle the extra heater current, but not sure how well 6BX7 or 6BL7 might work to lower the overall gain? Otherwise, maybe need something with a different attenuation curve (custom).
I should have the means to measure this - so, my source is 5.0 Vrms at 1kHz and 10k load and has output impedance of 20 Ohms.
Looking at my attenuator datasheet, the first four steps are: 42, 46, 50 and 60. Lower than that the spousal external limiter kicks in.
So, that is 10**(-60/20)*5 Vrms = 5 mVrms, 15.8 mVrms, 25.1 mVrms and 39.7 mVrms. I would set the attenuator to "max volume" and then put these waveforms onto the input (so to the transformer). Would this work? My analyzer can only put out 1 Vrms and this seems an easy way to measure actual values I am using...
Sounds like a good project after I get back from Germany
Looking at my attenuator datasheet, the first four steps are: 42, 46, 50 and 60. Lower than that the spousal external limiter kicks in.
So, that is 10**(-60/20)*5 Vrms = 5 mVrms, 15.8 mVrms, 25.1 mVrms and 39.7 mVrms. I would set the attenuator to "max volume" and then put these waveforms onto the input (so to the transformer). Would this work? My analyzer can only put out 1 Vrms and this seems an easy way to measure actual values I am using...
Sounds like a good project after I get back from Germany
i think i may have found the problem.
After replacing just about everything ive noticed that the jumper on my impasse board is open.
On further investigation i think this is supposed to be closed as currently i dont think the centre tap on the heaters is tied to ground.
Can i ask, under what circumstances would this jumper be left open?
Am i right in thinking that i should stick a jumper across this?
Thanks, I hope finally this will be the solution to my noise!
After replacing just about everything ive noticed that the jumper on my impasse board is open.
On further investigation i think this is supposed to be closed as currently i dont think the centre tap on the heaters is tied to ground.
Can i ask, under what circumstances would this jumper be left open?
Am i right in thinking that i should stick a jumper across this?
Thanks, I hope finally this will be the solution to my noise!
Was looking at wrong tube diagram, thought it was the 6922 but wasn't (thanks Google). So it's a shield not a centre tap. Tried connecting it but hasn't helped.
The noise does seem to be before the volume knob (I originally thought it wasnt. Anyway.
I'm now thinking it's something to do with grounding scheme or wire dressing.
My scopes not great and have been struggling to track the noise with it as it's quite noisy itself.
The audio ground is independent of each other for left and right on the input. Should these be shorted inside the amp? It seemed odd that the audio ground wasn't connect on the PCB but the wiring scheme I found suggested that haven't separate bus bars for the left input and right input was the thing to do.
However, when you connect a source, presumably they get connected via the source which seems likely to cause trouble.
This audio ground on the input also has no path to chassis ground. I'm guessing the input transformers are meant to isolate the signal.
Should I also twist ground and signal wires from the selector? I've read various different thoughts on this.....
The noise does seem to be before the volume knob (I originally thought it wasnt. Anyway.
I'm now thinking it's something to do with grounding scheme or wire dressing.
My scopes not great and have been struggling to track the noise with it as it's quite noisy itself.
The audio ground is independent of each other for left and right on the input. Should these be shorted inside the amp? It seemed odd that the audio ground wasn't connect on the PCB but the wiring scheme I found suggested that haven't separate bus bars for the left input and right input was the thing to do.
However, when you connect a source, presumably they get connected via the source which seems likely to cause trouble.
This audio ground on the input also has no path to chassis ground. I'm guessing the input transformers are meant to isolate the signal.
Should I also twist ground and signal wires from the selector? I've read various different thoughts on this.....
Made some progress.
Turns out i hadnt connected the input shield to the output shield. Id missed that somehow.
Thats pretty much made the switching on my stepped attenuator silent or at least very quiet.
Im still getting the fizzing sound though. However this is back to the situation where the volume pot is not affecting it.
Ive replaced the depletion mosfets (having matched them etc)
Leds
Swapped Tubes etc.
CHanged alps pot to stepped attenuator (100k constant load ladder with 2 resistors per step)
Connected the 6922 shield
Replaced filtering with a better unit.
Better mains cables with ferrit beads etc
Increased my 12k to 12.5k resistors so i get a full 100v drop plus heatsinks on Q2 to make sure theyre not getting hot.
Resoldered just about everything on the psu and main pcb.
Now using different speakers and different dac
Done the usual cable swapping.
Tried having only the amps plugged into the mains.
The F4 amp by itself seems to be silent.
The fizzing isnt always in the same channel or on both channels at once so im pretty sure its on the main pcb and not the psu board or noise on the heaters from the power transformer.
Jackinnj, in the kit you sent me the parts were listed as metal film resistors. Is it possible that some of them were carbon comp? Ive not seen the matt red/brown ones before? Just want to eliminate that possibility of damp carbon comps. Having said that, i think they all got replaced when i did the mosfets with the possible exception on R8 680k 1W. Unfortunately i cant see on the photo of what ive got of my board and the amps just got back in the rack.
The other thing i'm wondering is if i have some dodgy caps.
Particularly C106 and C107 decoupling the heater supply.
How likely are these ceramics to be duff? Or even the output films?
As the noise is fairly quiet im thinking its probably after the 6sn7.
Im open to any suggestions and im really pulling my hair out now.
Thanks
Turns out i hadnt connected the input shield to the output shield. Id missed that somehow.
Thats pretty much made the switching on my stepped attenuator silent or at least very quiet.
Im still getting the fizzing sound though. However this is back to the situation where the volume pot is not affecting it.
Ive replaced the depletion mosfets (having matched them etc)
Leds
Swapped Tubes etc.
CHanged alps pot to stepped attenuator (100k constant load ladder with 2 resistors per step)
Connected the 6922 shield
Replaced filtering with a better unit.
Better mains cables with ferrit beads etc
Increased my 12k to 12.5k resistors so i get a full 100v drop plus heatsinks on Q2 to make sure theyre not getting hot.
Resoldered just about everything on the psu and main pcb.
Now using different speakers and different dac
Done the usual cable swapping.
Tried having only the amps plugged into the mains.
The F4 amp by itself seems to be silent.
The fizzing isnt always in the same channel or on both channels at once so im pretty sure its on the main pcb and not the psu board or noise on the heaters from the power transformer.
Jackinnj, in the kit you sent me the parts were listed as metal film resistors. Is it possible that some of them were carbon comp? Ive not seen the matt red/brown ones before? Just want to eliminate that possibility of damp carbon comps. Having said that, i think they all got replaced when i did the mosfets with the possible exception on R8 680k 1W. Unfortunately i cant see on the photo of what ive got of my board and the amps just got back in the rack.
The other thing i'm wondering is if i have some dodgy caps.
Particularly C106 and C107 decoupling the heater supply.
How likely are these ceramics to be duff? Or even the output films?
As the noise is fairly quiet im thinking its probably after the 6sn7.
Im open to any suggestions and im really pulling my hair out now.
Thanks
Ive now replaced the neons, all the poly caps and the heater decoupling.
Ive made a quick recording of the noise with my phone and converted to mp3. The background hum is more the phone mic, its the pops im interested in. I normalised to -10db so shouldnt go crazy but may be best to turn your speakers down a little.
https://1drv.ms/u/s!AqeRWzApq3EmoW6w2SjAYnSmN1q5
Im hoping someone will recognise the characteristic of this sound.
Thanks
Ive made a quick recording of the noise with my phone and converted to mp3. The background hum is more the phone mic, its the pops im interested in. I normalised to -10db so shouldnt go crazy but may be best to turn your speakers down a little.
https://1drv.ms/u/s!AqeRWzApq3EmoW6w2SjAYnSmN1q5
Im hoping someone will recognise the characteristic of this sound.
Thanks
ok, last one i promise!
I wondering if ive not taken adequate static precautions. Are the depletion mosfets particualrly sensitive? And if so, presumably we can go back to the first diagnosis of the noise being caused by the current source?
With a compnent like this are they more likely to get damaged in a reduced performance/added noise sort of way or break outright?
Thanks, I know i must be wearing eveyones patience thin by now.
I wondering if ive not taken adequate static precautions. Are the depletion mosfets particualrly sensitive? And if so, presumably we can go back to the first diagnosis of the noise being caused by the current source?
With a compnent like this are they more likely to get damaged in a reduced performance/added noise sort of way or break outright?
Thanks, I know i must be wearing eveyones patience thin by now.