Thanks again rongon for the very thorough reply. I will try the separate grounds as you describe above.
I had thought the AC heater wires under the volume pot might have been an issue, so I moved them around and closer too it. Nothing changed. Usually when there’s interactivity between components I can hear changes in the nature of the buzzing as things move around. I suppose I could try moving that last section of heater wires to the opposite side of the tubes, between the tubes and the OPTs, but then they’d be crossing near the cathode circuitry and plate wires.
I had thought the AC heater wires under the volume pot might have been an issue, so I moved them around and closer too it. Nothing changed. Usually when there’s interactivity between components I can hear changes in the nature of the buzzing as things move around. I suppose I could try moving that last section of heater wires to the opposite side of the tubes, between the tubes and the OPTs, but then they’d be crossing near the cathode circuitry and plate wires.
No worries. Explaining these concepts to someone else is a great learning experience for me. I've only recently gotten to the point where I feel I have a good handle on grounding schemes.
I hope you can get this hum problem sorted out.
How do you have your heater winding grounded? I see the black and red twisted pair running to the tubes, but I don't see a center tap or pair of resistors to make one. If the heater winding is left floating, that can certainly cause a harsh buzz in the output.
If that's not the issue...
The way you've described your grounding scheme, it sounds like you have it connected to chassis at one point only, at the IEC inlet socket. Then a single ground bus wire goes from there to the audio circuits at the other end of the chassis, right?
If that's correct, then I think what's happening is that the high current/high noise power supply parts have the shortest path to ground, and your audio circuits have a much longer path to ground, all the way down the bus wire back to the IEC socket. That would cause a slight resistance to appear between the audio circuits' ground points and where the ground bus is connected to chassis. Any current flowing through the ground bus wire would then appear as a voltage across this resistance, right at the audio circuits' grounding points. Since the high current parts have the shortest path to ground, they don't have that resistance (however tiny) appearing in series with their ground points. The audio circuits do have that tiny resistance in their ground path, though. That might cause the buzz you hear.
Maybe you can do a simple experiment?
- Leave the IEC socket safety ground connecting to chassis at the socket.
- Disconnect the ground bus from the IEC socket/chassis ground.
- Connect the 'far' end of the ground bus (the end that's nearest the inputs) to a bolt through the chassis with a sharp star washer, so that the shortest path to chassis ground is near the inputs.
Now any tiny resistance formed by the ground bus wire is seen by the power supply end (power transformer, rectifier, reservoir capacitor), while the audio circuits will now have the shortest path to chassis ground. Hopefully that will fix the hum/buzz.
--
I hope you can get this hum problem sorted out.
How do you have your heater winding grounded? I see the black and red twisted pair running to the tubes, but I don't see a center tap or pair of resistors to make one. If the heater winding is left floating, that can certainly cause a harsh buzz in the output.
If that's not the issue...
The way you've described your grounding scheme, it sounds like you have it connected to chassis at one point only, at the IEC inlet socket. Then a single ground bus wire goes from there to the audio circuits at the other end of the chassis, right?
If that's correct, then I think what's happening is that the high current/high noise power supply parts have the shortest path to ground, and your audio circuits have a much longer path to ground, all the way down the bus wire back to the IEC socket. That would cause a slight resistance to appear between the audio circuits' ground points and where the ground bus is connected to chassis. Any current flowing through the ground bus wire would then appear as a voltage across this resistance, right at the audio circuits' grounding points. Since the high current parts have the shortest path to ground, they don't have that resistance (however tiny) appearing in series with their ground points. The audio circuits do have that tiny resistance in their ground path, though. That might cause the buzz you hear.
Maybe you can do a simple experiment?
- Leave the IEC socket safety ground connecting to chassis at the socket.
- Disconnect the ground bus from the IEC socket/chassis ground.
- Connect the 'far' end of the ground bus (the end that's nearest the inputs) to a bolt through the chassis with a sharp star washer, so that the shortest path to chassis ground is near the inputs.
Now any tiny resistance formed by the ground bus wire is seen by the power supply end (power transformer, rectifier, reservoir capacitor), while the audio circuits will now have the shortest path to chassis ground. Hopefully that will fix the hum/buzz.
--
I disconnected the signal ground bus from the IEC and used alligator clips to try attaching the bus to a chassis star ground point in the center of the circuit (between the PS and signal components), so all had a short path to ground. I undid that and then used the alligator clip to attach the ground bus to the chassis near the inputs.
Neither scenario had any effect on buzz at all, even when no part of the bus was connected to chassis.
What’s intriguing to me at this point is that the buzz gets louder and softer with the volume control. So maybe it has something to do with how the VC was wired.
Neither scenario had any effect on buzz at all, even when no part of the bus was connected to chassis.
What’s intriguing to me at this point is that the buzz gets louder and softer with the volume control. So maybe it has something to do with how the VC was wired.
I remembered that I had an issue when this circuit was on the breadboard, that the angry insect buzz would go away when the volume pot is turned all the way up. I had assumed it was due to a dying cheap pot but the same thing happens with my new Alps. Buzz is loudest around 8 but goes away completely at 10.
I've got a 100k grid leak resistor connected from grid pin to ground bus. Pot ground pins are both soldered to left channel ground bus.
All connections were fine but I resoldered them anyway.
Lots of commentary about this in guitar forums but didn't find any solutions.
What else should I be checking?
I've got a 100k grid leak resistor connected from grid pin to ground bus. Pot ground pins are both soldered to left channel ground bus.
All connections were fine but I resoldered them anyway.
Lots of commentary about this in guitar forums but didn't find any solutions.
What else should I be checking?
My somewhat wild guess is that one or both of the 5842's are oscillating at RF. Normally you get an audible squeal if it is in the audio band, but 417A/5842 was meant to be a high gain, grounded grid, VHF/UHF low noise preamplifer, so oscillation could be far above the audible range, and angry insect buzz is an artifact of that.
If you can measure the current going through the tubes, and see if it changes as you advance the volume pot, that *might* give some insight into what is going on.
If you have a scope, or even a DMM that can read small signal AC in the presence of DC, and can see if the AC voltage changes as you advance the volume pot, that also *might* give you some insight into what is going on. Unfortunately, many (most) inexpensive DMM's are not good at that type of measurement.
edit: oscillation is generally positive feedback gone awry - looking at your pictures, you have an unshielded lead from your input in fairly close proximity to the lead to your opt, and that *could* create a feedback loop. Shielded inputs ( at one end only, to avoid a ground loop ) might help. Making your input wire cross at 90 degrees with the output might help. Fiddling with the load on
the grid(s) might help.
Or it might very well be a ground issue like rongon suggests. But the volume control observation suggests further inquiry to me.
BTW, Joplin hamfest, is the 24th.
Win W5JAG
Last edited:
Thanks. I will make re measurements you suggest later and look for shielded signal wire.
Another possibility just occurred to me. Someone on a guitar forum said he had the same behavior I’m experiencing and discovers that it was caused by reversed polarity on the output jack. Not quite the same here, but my OPTs have two 16 ohm secondaries that are wired in series to get 32 ohms. Maybe I connected them out of order? Would it matter?
Another possibility just occurred to me. Someone on a guitar forum said he had the same behavior I’m experiencing and discovers that it was caused by reversed polarity on the output jack. Not quite the same here, but my OPTs have two 16 ohm secondaries that are wired in series to get 32 ohms. Maybe I connected them out of order? Would it matter?
I spent a long time stripping teflon-insulated wire (ouch!) and cutting precise lengths for a tight and complicated space, so I'm not eager to undo all that work. But I will try it. The pin listing on the OPT's interior label lists secondary values in the following order (image attached):
4 -- 0 ohm
5 -- 16 ohm
7 -- 0 ohm
8 -- 16 ohm
It didn't occur to me earlier that maybe this indicates the order, with pin 8 being "hot" and pin 4 "return", and a jumper between 5 and 7 to connect them in series.
I did pin 5 as "hot" and pin 7 as "return," with the jumper from 4 to 8.
Again, maybe it doesn't matter, but as I've said already, it would have been nice to have some documentation with these OPTs, even if only pictures to show connections.
A friend suggested that my 100K pot might be too high for the input impedance of the 5842s, which could be why buzz is induced in the middle regions. He suggested removing the volume pot and putting a 25k resistor in series with the grid and a 25k grid stopper, to simulate a 50K pot set to half volume, to see if that gets rid of the buzz.
Come to think of it, the first amp these tube were in -- a Tubelab SE -- had a 50K pot and they worked just fine. The design this headphone amp is based on (by DHTRob) uses a 120k volume pot, though.
4 -- 0 ohm
5 -- 16 ohm
7 -- 0 ohm
8 -- 16 ohm
It didn't occur to me earlier that maybe this indicates the order, with pin 8 being "hot" and pin 4 "return", and a jumper between 5 and 7 to connect them in series.
I did pin 5 as "hot" and pin 7 as "return," with the jumper from 4 to 8.
Again, maybe it doesn't matter, but as I've said already, it would have been nice to have some documentation with these OPTs, even if only pictures to show connections.
A friend suggested that my 100K pot might be too high for the input impedance of the 5842s, which could be why buzz is induced in the middle regions. He suggested removing the volume pot and putting a 25k resistor in series with the grid and a 25k grid stopper, to simulate a 50K pot set to half volume, to see if that gets rid of the buzz.
Come to think of it, the first amp these tube were in -- a Tubelab SE -- had a 50K pot and they worked just fine. The design this headphone amp is based on (by DHTRob) uses a 120k volume pot, though.
Attachments
Because the pot wiper lead goes past some buzz source.
It is worst at half resistance (around "8" on an Audio taper pot) because then the wiper impedance is highest.
Lower value pot is a long downhill battle.
You want to re-route and shield the connection from pot wiper to grid.
OK, I'll try PRR's suggestion of rerouting the heater wires out from under the volume pot (where they currently measure 1" distance from it, btw). Or were you saying to leave the heater wires and move the signal wires?
If I take the heater wires to the opposite side of the tube, they'd be far from the volume pot but would cross the wires that go from plate to OPT primary and would also be close to the ground bus. Would that be likely to induce hum too?
If so, they could alternatively go 90* "in the air," which would bring them close-ish to the cathode resistor and bypass cap but far from everything else.
Any recommended products for shielded signal wire?
*******************************
I disconnected the pot and did three tests with grid stopper and leak resistors that I have on hand to simulate the resistance difference in various pots lower than 100k. I used my computer as source and controlled the volume with its volume control. In all cases the buzz level stayed the same and did not rise or fall with the source volume, unlike the behavior I got with the pot.
GS - 27k, GL - 27k (a 54k pot) = loud hum
GS - 18k, GL - 15k (a 33k pot) = loud hum
GS - 3.9k, GL - 4.7k (a ~9k pot) = substantially mitigated hum but still audible
I stopped there, but with these numbers it seems like I'd be heading for at most a 5K logarithmic stereo pot, which does seem to exist although it's far outside the norm for these kinds of builds.
If I take the heater wires to the opposite side of the tube, they'd be far from the volume pot but would cross the wires that go from plate to OPT primary and would also be close to the ground bus. Would that be likely to induce hum too?
If so, they could alternatively go 90* "in the air," which would bring them close-ish to the cathode resistor and bypass cap but far from everything else.
Any recommended products for shielded signal wire?
*******************************
I disconnected the pot and did three tests with grid stopper and leak resistors that I have on hand to simulate the resistance difference in various pots lower than 100k. I used my computer as source and controlled the volume with its volume control. In all cases the buzz level stayed the same and did not rise or fall with the source volume, unlike the behavior I got with the pot.
GS - 27k, GL - 27k (a 54k pot) = loud hum
GS - 18k, GL - 15k (a 33k pot) = loud hum
GS - 3.9k, GL - 4.7k (a ~9k pot) = substantially mitigated hum but still audible
I stopped there, but with these numbers it seems like I'd be heading for at most a 5K logarithmic stereo pot, which does seem to exist although it's far outside the norm for these kinds of builds.
Last edited:
That kills it both as an oscillator and an amplifier, so I'm not all that sure this tells you much, other than that additional noise is not being conducted into the output wiring.
The induced hum from the filament wiring theory seems easy to test. As a temporary measure, rectify and filter the filament circuit so they are running off DC and see what effect, if any, this has.
Win W5JAG
edit: on a couple of occasions, I have seen bad electrolytic caps cause unexpected problems. It is probably a long shot, but noise can also be coming in on your B+ line. If you have some conventional filter caps, you could try disconnecting the ones you have, and tacking in substitutes and see if that changes anything. Maybe reversing the connections on those appliance capacitors might help?
Last edited:
Since this is troubleshooting, you could always find a cheap pair of RCA interconnects and cut them up. The local Big Lots and Dollar General stores around here sell cheap RCA cables and Y-cables (1/8" stereo to dual RCA).
In my build, I took 22-gauge solid-core hookup wire and made twisted pair. Red and black for the right channel, white and black for the left channel. I connected the black (ground wires) to the chassis-connected star ground point near the input RCA jacks. I left the black wires disconnected at the end nearest the tube grids.
Also, you never answered the question of where your heater winding center tap is grounded, or if it's grounded. Did you ground one end of the heater winding, so its red wire = 6.3VAC and black wire = 0V?
--
Thanks all for your help. I really appreciate it.
This morning I plugged the amp into a CD player and turned off the computer to rule out ambient electrical noise from that (no effect). I then resoldered all connections to make sure that everything was tight. Also moved the inter-tube heater wiring to the other side (attached) and in various orientations, and it had no effect on noise, even when I put it right up against the body of the volume pot.
I'll see if I can scrounge up some diodes and caps to try DC rectification. My 6.3v winding is rated for 2.5A, though, so the current penalty for DC conversion may generate too much heat. 2(2 x 0.3A) = 1.2A, added to the 1A for the EZ81 brings us to 2.2A. If I also put the EZ81 on DC, that would be 3.2A and one very, very unhappy PT.
I still had the noise problem on the breadboard, though. At the time, I attributed it to the bad volume pot (a cheap Alpha) I was using there. So my hunch is that it won't change anything but at this point I just need to try everything.
I might try rewiring the OPT secondaries to see if order matters. Switching them around would lead to a shorter jumper, which is probably just a smarter way, anyway, and they'd conform to the way it's written out on the interior label.
This morning I plugged the amp into a CD player and turned off the computer to rule out ambient electrical noise from that (no effect). I then resoldered all connections to make sure that everything was tight. Also moved the inter-tube heater wiring to the other side (attached) and in various orientations, and it had no effect on noise, even when I put it right up against the body of the volume pot.
I'll see if I can scrounge up some diodes and caps to try DC rectification. My 6.3v winding is rated for 2.5A, though, so the current penalty for DC conversion may generate too much heat. 2(2 x 0.3A) = 1.2A, added to the 1A for the EZ81 brings us to 2.2A. If I also put the EZ81 on DC, that would be 3.2A and one very, very unhappy PT.
There is a two-part electrolytic (a JJ 100uF + 100uf) in the PS. I had the same noise issue with multiple other known-good electrolytics when this amp was on the breadboard, so I doubt it's that, but I'm going to try everything now and will give it a shot.
This is what I'm going to try next. The only difference between the current amp and the breadboarded one is that I switched the tabs used on the 25uF PS caps since the red-marked ones may have indicated outer foil. On the breadboard I attached the red-marked tabs to the positive of the HT line; on the chassis build they go to ground.
I still had the noise problem on the breadboard, though. At the time, I attributed it to the bad volume pot (a cheap Alpha) I was using there. So my hunch is that it won't change anything but at this point I just need to try everything.
True. I'll get some cheap shielded coax by the foot at Home Depot. I was also interested in recommendations for premium, final-build quality products if anyone has any.
Sorry, I missed that. Unless I'm mistaken, each wire (red and black) is carrying 3.15VAC. As Win noted, I posted a ways back that I tried a virtual center tap and one or two other techniques. It had no effect on noise so I put it back to the original state. Perhaps I should try grounding one heater lead?
Could be, but unfortunately this is the only pair of 5842s I have.
I might try rewiring the OPT secondaries to see if order matters. Switching them around would lead to a shorter jumper, which is probably just a smarter way, anyway, and they'd conform to the way it's written out on the interior label.
Attachments
Last edited:
If you get to the point where you think it could be a tube issue, I can mail you over some known good 417A's. 417A/5842's are all over the place, and these are on the cool side so a little easier to work with.
I wouldn't want to let go of them permanently but you are welcome to experiment with them if it gets to that.
Hopefully shielding your inputs as PRR suggests is the cure. My post yesterday was a little cryptic - connecting the shield at only one end is effective and won't make a new ground loop.
Good luck with this.
Win W5JAG
I wouldn't want to let go of them permanently but you are welcome to experiment with them if it gets to that.
Hopefully shielding your inputs as PRR suggests is the cure. My post yesterday was a little cryptic - connecting the shield at only one end is effective and won't make a new ground loop.
Good luck with this.
Win W5JAG
jdrouin said:
Can you still get parts pretty readily in Tulsa?
Maybe elevate the heaters at DC with a voltage divider and ground them for AC through a large value cap, like 0.5 uF or so.
We've seen my (lack of) technical skills in your other thread I trainwrecked, so maybe rongon or PRR can comment on this. I am no expert on heater wiring. Yours looked pretty good to me.
Win W5JAG
Attachments
Thanks for the offer to send test tubes, Win, but I’d rather not risk damage to your 417a’s. If it gets to that point I’ll buy another pair.
Just to clarify, in your drawing would that elevation circuit be for the heater of just one tube? You wouldn’t do that for all of them, right? My instinct would be to do it on the rectifier since it’s first in the chain, although it’s also a long distance from the B+ rail.
If chain locatiion doesn’t matter, then a 5842 heater would be closest to the B+ rail.
My friend lent me a signal generator today, so I’ve got that to play around with now.
Just to clarify, in your drawing would that elevation circuit be for the heater of just one tube? You wouldn’t do that for all of them, right? My instinct would be to do it on the rectifier since it’s first in the chain, although it’s also a long distance from the B+ rail.
If chain locatiion doesn’t matter, then a 5842 heater would be closest to the B+ rail.
My friend lent me a signal generator today, so I’ve got that to play around with now.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.