Merry Christmas, guys!
So I am in the mountains right now, Norwegian cabin holidays. But wanted to share an experience wrt that hum.
I cut three of my 7 per rail 0R47 PSU filter resistors, which originally gave a total filter resistance of 0R067, and the hum was reduced by quite a bit (total R of 0R117). Nelsons spec for the PSU says to use 1Rs, giving a total of 0R25 of filter resistance, so I figure I am going to replace my remaining four 0R47s per rail with 1Rs. I know the PSU business equates to complicated maths. But can I assume 50% decrease in ripple levels by doubling the R in the CRC, or is it not as easy as that?
PM’ed with Jan Didden, and he says this looks like a ripple issue since increasing R reduced the hum. And that increasing R is my best shot so long as the grounding is as good as possible with the PCBs used.
Figure I’ll give it a go. I have voltage to give away with my dual 24v secondaries. So I figure I can go even higher on the R, but do not have PSUD, and resistors are lying around anyways so why not try.
Specs for caps: https://no.mouser.com/datasheet/2/88/SLPX-32478.pdf
I use the 22.000uf version, on the Diyaudio PSU V3 boards.
Can I measure noise at the rails with my Fluke set to ACV btw? Would be nice to compare measurements before and after.
Hope you all are enjoying the holidays.
Regards, and cheers,
Andy
So I am in the mountains right now, Norwegian cabin holidays. But wanted to share an experience wrt that hum.
I cut three of my 7 per rail 0R47 PSU filter resistors, which originally gave a total filter resistance of 0R067, and the hum was reduced by quite a bit (total R of 0R117). Nelsons spec for the PSU says to use 1Rs, giving a total of 0R25 of filter resistance, so I figure I am going to replace my remaining four 0R47s per rail with 1Rs. I know the PSU business equates to complicated maths. But can I assume 50% decrease in ripple levels by doubling the R in the CRC, or is it not as easy as that?
PM’ed with Jan Didden, and he says this looks like a ripple issue since increasing R reduced the hum. And that increasing R is my best shot so long as the grounding is as good as possible with the PCBs used.
Figure I’ll give it a go. I have voltage to give away with my dual 24v secondaries. So I figure I can go even higher on the R, but do not have PSUD, and resistors are lying around anyways so why not try.
Specs for caps: https://no.mouser.com/datasheet/2/88/SLPX-32478.pdf
I use the 22.000uf version, on the Diyaudio PSU V3 boards.
Can I measure noise at the rails with my Fluke set to ACV btw? Would be nice to compare measurements before and after.
Hope you all are enjoying the holidays.
Regards, and cheers,
Andy
Last edited:
The increased resistance will work, but only to a point, because the series resistance becomes too much of a loss for the voltage in the PSU.
Leave it the way you have it now.
Leave it the way you have it now.
I‘m still dreaming around pretty unconstrained—about insralling a switch for changing the gain (you know, through that resistor)...
Since the amp has to be re-biased, I wonder (as a general question) if it would be a very stupid idea to install little voltmeters, so that all adjustments could be done from the outside?
Since the amp has to be re-biased, I wonder (as a general question) if it would be a very stupid idea to install little voltmeters, so that all adjustments could be done from the outside?
https://i.pinimg.com/originals/9c/db/f0/9cdbf07d8db6120e45a4375cdfb7df65.jpg
[emoji1787][emoji1787][emoji1787]
No it won‘t be anything like that...
(Gotta be careful, according to Mark I only got 10 clicks 😉 )
Last edited:
Are pilots over-represented in this forum? And those meters look way cool, fitting for a Pass amp I would say. Would the in-flight experience be inhibited if you moved some of those to your amps?
They‘re ... present, that’s for sure.
But I tbh I think a cockpit should be and stay in a plane, and a different approach should be used for audio-controls or controls of audio-controls.
Open for inspirations though...
When I was thinking of the bias-control, I was reminded of those really cool biasing function in the top-line nak-tape-recorders (660zx-r), where you could use the vu-meter to bias the heads to every single tape...
But I tbh I think a cockpit should be and stay in a plane, and a different approach should be used for audio-controls or controls of audio-controls.
Open for inspirations though...
When I was thinking of the bias-control, I was reminded of those really cool biasing function in the top-line nak-tape-recorders (660zx-r), where you could use the vu-meter to bias the heads to every single tape...
On that note, and I probably could figure this out searching the forum, but: will the BA-3 complimentary with BA-2 OS clip above 1,9-2 volts input? As far as I have read, full output is reached around that input voltage, and I guess that full output means anything above it will make the amp clip. My current DAC (W4S) puts out 2.7 volts at full output, and it does not sound very pleasing creeping over 2v/towards max output.
So i keep the volume down to about 75% max. So I too have contemplated gaining down a notch, but figure Waynes Linestage can handle that input as I’ll put the DACs volume ctl to fixed, instead of variable as I now use it as a preamp temporarily.
Regards,
Andreas
So i keep the volume down to about 75% max. So I too have contemplated gaining down a notch, but figure Waynes Linestage can handle that input as I’ll put the DACs volume ctl to fixed, instead of variable as I now use it as a preamp temporarily.
Regards,
Andreas
Last edited:
Hi guys!
So finally there is some time to dig into this hum issue.
I have many things to measure yet, and tests to do including disconnecting FE channel by channel, and OS one at a time. But the following observations may provide some insight (measured with Fluke true RMS dmm):
1: I measure 1,5mVac at the right speaker output. Insputs shorted, no speakers connected. Funny thing is, when DC offset is off, Vac reading increases accordingly + 1,5mVac)
2: I measure 57mVac at both rails. But the neg rail jumps between 54 and 63mVac, for unknown reasons. Speakers connected, and turned off source (i know, not ideal).
3: The hum is a little bit louder in the left speaker than the right. Audible from 3 inches from the cone, 91-92db sens speakers. The hum in the left speakers has a slight hiss, but much less after twisting and redressing in December.
Measurements taken at full 4 amp per channel bias.
Like I said, I know far more tests and measurements are needed. But I post this in hope of someone pointing me in the right direction.
A few questions do arise:
1: I gather the measured ripple will not be a problem for the OS, it is the same ripple level that the XA-160 had according to Nelson, with no issues with a 200uV OS. But what about my BA-3 front end with approx 30db gain? Is it’s PSRR good enough to withstand 57mVac ripple?
2: What about the fluctuating ripple on the neg rail? Can that be one of the issues? I think this might have to do with a connections issue somewhere between IEC and bridges. Will check.
3: This is P/P. One of my N channels biases higher than the others. Might this worsen PSRR and be the cause? Have new fets but have had no time to replace yet.
Worth of note, is that reversing both FE and sinks, provided no change in audible hum. Speaker differences still the same.
Pictures and more measurements and tests, inspired by good help from Bonsai, Jan Didden, you guys and especially Dennis, will be provided as soon as I have the time. Tomorrow evening looks good.
I do not believe ripple is my only issue, if relevant at all. What I can say, is the hum levels decreased when I increased the R in the CRC. So next up is increasing it from 0R11 to spec of 0R25, having only half of that now.
I believe grounding is also an issue, and will investigate further. Also, dressing is on the agenda.
I know, pics pics pics will help. Will get on it!
Chasing ghosts? You betcha.
Regards,
Andreas
So finally there is some time to dig into this hum issue.
I have many things to measure yet, and tests to do including disconnecting FE channel by channel, and OS one at a time. But the following observations may provide some insight (measured with Fluke true RMS dmm):
1: I measure 1,5mVac at the right speaker output. Insputs shorted, no speakers connected. Funny thing is, when DC offset is off, Vac reading increases accordingly + 1,5mVac)
2: I measure 57mVac at both rails. But the neg rail jumps between 54 and 63mVac, for unknown reasons. Speakers connected, and turned off source (i know, not ideal).
3: The hum is a little bit louder in the left speaker than the right. Audible from 3 inches from the cone, 91-92db sens speakers. The hum in the left speakers has a slight hiss, but much less after twisting and redressing in December.
Measurements taken at full 4 amp per channel bias.
Like I said, I know far more tests and measurements are needed. But I post this in hope of someone pointing me in the right direction.
A few questions do arise:
1: I gather the measured ripple will not be a problem for the OS, it is the same ripple level that the XA-160 had according to Nelson, with no issues with a 200uV OS. But what about my BA-3 front end with approx 30db gain? Is it’s PSRR good enough to withstand 57mVac ripple?
2: What about the fluctuating ripple on the neg rail? Can that be one of the issues? I think this might have to do with a connections issue somewhere between IEC and bridges. Will check.
3: This is P/P. One of my N channels biases higher than the others. Might this worsen PSRR and be the cause? Have new fets but have had no time to replace yet.
Worth of note, is that reversing both FE and sinks, provided no change in audible hum. Speaker differences still the same.
Pictures and more measurements and tests, inspired by good help from Bonsai, Jan Didden, you guys and especially Dennis, will be provided as soon as I have the time. Tomorrow evening looks good.
I do not believe ripple is my only issue, if relevant at all. What I can say, is the hum levels decreased when I increased the R in the CRC. So next up is increasing it from 0R11 to spec of 0R25, having only half of that now.
I believe grounding is also an issue, and will investigate further. Also, dressing is on the agenda.
I know, pics pics pics will help. Will get on it!
Chasing ghosts? You betcha.
Regards,
Andreas
Last edited:
Update: Increased R to 0R235, almost to spec, ripple down to 34mVac RMS and 97mVac P-P. Hum almost completely gone now, need ear in cone at 91-92db sens speakers. Figure I found my ghost, but need to chase lower ripple. Hammond chokes under consideration. Figure I need to attenuate 12-15db of remaining ripple to make it silent up to 100db sens, approx.
Predominant now that hum is almost gone, is the slight buzz from the left speaker. It has been the same in level all along, around 100-120hz but audible pulses. reversing OS and swapping sinks did nothing to change it. So perhaps it get’s in before filtering, unless they are charging pulses. But funny only being audible in left speaker if charging pulses. Will redress and move bridges in an attempt to sort it, and move AC cable from left side to center.
In conclusion: it is getting pretty quiet now, but still not perfect.
Wish you all a good weekend.
Andy
Predominant now that hum is almost gone, is the slight buzz from the left speaker. It has been the same in level all along, around 100-120hz but audible pulses. reversing OS and swapping sinks did nothing to change it. So perhaps it get’s in before filtering, unless they are charging pulses. But funny only being audible in left speaker if charging pulses. Will redress and move bridges in an attempt to sort it, and move AC cable from left side to center.
In conclusion: it is getting pretty quiet now, but still not perfect.
Wish you all a good weekend.
Andy
Last edited:
So it looks like your problem is not due to cross channel ground loop - that's good news.
andynor
Lets see what happens when you rotate the tranny. Some times you can get a huge improvement.
How have you routed your cables from the input sockets to each amp module - see slide 64 in the GLP
(your message box is full)
andynor
Lets see what happens when you rotate the tranny. Some times you can get a huge improvement.
How have you routed your cables from the input sockets to each amp module - see slide 64 in the GLP
(your message box is full)
Bonsai: On it. Until then, see post 1 in this thread. Inputs routed almost exactly as 6L6s version. Neutrik original inputs so no contact with chassis, just to have said it.
Tranny-turning is the plan for tonight. Rather exited about that.
I can allready hear the improvements of lower ripple. Much cleaner and less constricted sound. Plus more R means room for higher IQ =)
Tranny-turning is the plan for tonight. Rather exited about that.
I can allready hear the improvements of lower ripple. Much cleaner and less constricted sound. Plus more R means room for higher IQ =)
Wait a minute. There is a possible significant cross channel loop here. Havent heard it before (hmm, either I am an idiot, or some other change has lead to it). I do remember when the ripple was higher and grounding worse, there was a slight increase w/shorted inputs, but this increase is really significant. With inputs not shorted, the noise increases by maybe 10 times. With only one input shorted, the other channel is as noisy as before, and the other way around.
Finally something to work with, I say, or maybe I should’nt.
Finally something to work with, I say, or maybe I should’nt.
Attachments
Last edited:
Hei, guys!
Just wanted to chime in with a hum update. I solved the open inputs issue. It is now just as quiet with open as with shorted inputs. The hum is now also the same volume in both speakers. Also, it is almost gone now. The ear must be into the cone to hear it. So getting there. Next off is to rotate the tranny, using Bonsais headphone trick. I will also redress both AC and DC wiring to their final positions using the headphone trick. Also, I wanna reduce the loop area between rectifiers and PSU board. We will see how it goes. I am at a point of diminishing returns, and nearly there. Just have to plan a bit ahead, and find the time. Two kids and one on the way does not make for infinite time for the hobby.
It seems as though my issues were a result of the grounding scheme, but also some common impedance issues.
I did increase the filtering to Nelsons spec of 0R25, and ddecreased the bias a little bit. Ripple now seems to have much less effect on the sound, and it is both cleaner and tighter. Trying different grounding schemes wrt wite placement on the board, was also helpful. Also, grouping FE wires and OS wires in the twist, as helped. This has been a really valuable experience, so not nothing for nothing to have issues in the first build.
I am also contemplating CLC, and a better transformer. But those air core hammonds cost quite a penny, and Toroidy does not come cheap, so we will see.
Once I am happy with the noise levels, The plan is to make this baby look it’s best, and redo/swap all parts that have been soldered and unsoldered too many times. And get an LED in the front panel.
I would like to thank you all for assistance and good advice along the way! Great people.
Regards,
Andreas
Just wanted to chime in with a hum update. I solved the open inputs issue. It is now just as quiet with open as with shorted inputs. The hum is now also the same volume in both speakers. Also, it is almost gone now. The ear must be into the cone to hear it. So getting there. Next off is to rotate the tranny, using Bonsais headphone trick. I will also redress both AC and DC wiring to their final positions using the headphone trick. Also, I wanna reduce the loop area between rectifiers and PSU board. We will see how it goes. I am at a point of diminishing returns, and nearly there. Just have to plan a bit ahead, and find the time. Two kids and one on the way does not make for infinite time for the hobby.
It seems as though my issues were a result of the grounding scheme, but also some common impedance issues.
I did increase the filtering to Nelsons spec of 0R25, and ddecreased the bias a little bit. Ripple now seems to have much less effect on the sound, and it is both cleaner and tighter. Trying different grounding schemes wrt wite placement on the board, was also helpful. Also, grouping FE wires and OS wires in the twist, as helped. This has been a really valuable experience, so not nothing for nothing to have issues in the first build.
I am also contemplating CLC, and a better transformer. But those air core hammonds cost quite a penny, and Toroidy does not come cheap, so we will see.
Once I am happy with the noise levels, The plan is to make this baby look it’s best, and redo/swap all parts that have been soldered and unsoldered too many times. And get an LED in the front panel.
I would like to thank you all for assistance and good advice along the way! Great people.
Regards,
Andreas
Considering from the output stage, like tou recommended. I did try that once, but no reduction in hum or buzz. Also, that means I have to split FE return from the twist, making a larger loop area than nescessary. What would you recommend, Michty ZM, and why?
Regards,
Andy
Regards,
Andy