hailteflon said:
And "antennas", wires and ground connections and PCB ground loop, steel chassis, rectifier bridge, resistors, transformer, power cord.
The fact that it causes hum when the amplifier is fed by it, but even without its tubes installed, AND unplugged, should be a great clue.
What happens if you disconnect the transformer's center tap, under those conditions?
I can think of two possible causes for hum without power. The one I think is most likely is that the loop formed by the output lead and its return picks up magnetic fields. The other possibility I can see is that it's electrostatic fields picked up by the unshielded parts of the circuit. The output is high impedance if the amplifier is turned off. But if this is the case then attaching the lid to the box should make it quiet unless it's getting in through the mains cable - not very likely. Could be tested by holding the prongs of the plug to the case or putting the lead inside the enclosure before putting the lid on.
The box will not stop magnetic fields effectively though. Actually steel is pretty good because it shunts the magnetic flux around the circuit. Shielding by copper or aluminium works differently because the shielding in that case is only caused by currents induced. Sometimes this is used - for example the flux band around transformers, both mains freqency and SMPS.
Some weeks ago I built an active crossover and was worried that the plastic box would make it hum. But it didn't! It is interesting though that if the configuration switches are held in their middle position the 1 megaohm input of the next stage does pick up quite a bit of hum. After transporting the filter to my fathers place it didn't work any longer though - it just hummed. This was because the lead to the output jack had come loose. Although only a centimeter of it was unscreened it picked up hum. Having the lid on or off should make a great difference in garys preamp if it is electrostatic coupling that is the problem, especially when one considers the high impedances in tube circuits.
The box will not stop magnetic fields effectively though. Actually steel is pretty good because it shunts the magnetic flux around the circuit. Shielding by copper or aluminium works differently because the shielding in that case is only caused by currents induced. Sometimes this is used - for example the flux band around transformers, both mains freqency and SMPS.
Some weeks ago I built an active crossover and was worried that the plastic box would make it hum. But it didn't! It is interesting though that if the configuration switches are held in their middle position the 1 megaohm input of the next stage does pick up quite a bit of hum. After transporting the filter to my fathers place it didn't work any longer though - it just hummed. This was because the lead to the output jack had come loose. Although only a centimeter of it was unscreened it picked up hum. Having the lid on or off should make a great difference in garys preamp if it is electrostatic coupling that is the problem, especially when one considers the high impedances in tube circuits.
Hi Guys,
thanks Megajocke for looking this over.
I thought that this would create a loop between input-pcb-output. I'll move i/o closer together and connect their grounds in the fashion Chrish prescribed earlier. Having said this, I remember in the past connecting the other end of the output shield to the board and the hum goes wild. I'll try again.
Welcome back Mark,
I expect the caps from digikey tomorrow, I don't know why they are taking so long. I have yet to construct a series of resistors to step down the B+ and heater voltages to my sound card. I believe it will tolerate 3 volts. Tom has provided an algorithm to do this. PLEASE IGNORE THOSE SCOPE MEASUREMENTS! Tom noticed that my scope ground was not referrencing what I was measuring. I have since fixed that and have new measurements for the preamp out and amp speaker out. I am glad to report that they are much more in line with what I am hearing. I will post them later tonight.
12X4
Yeah, the preamp audibly hums without the tubes in. I have replaced the 12AU7s not the 12X4.
I looks like I have some work to do. I'll post pics and measurements later.
Cheers,
gary
thanks Megajocke for looking this over.
megajocke said:
I thought that this would create a loop between input-pcb-output. I'll move i/o closer together and connect their grounds in the fashion Chrish prescribed earlier. Having said this, I remember in the past connecting the other end of the output shield to the board and the hum goes wild. I'll try again.
megajocke [/i]Transformer center tap must go directly to the first filter cap (C101) too. Chassis is not right said:
Welcome back Mark,
I expect the caps from digikey tomorrow, I don't know why they are taking so long. I have yet to construct a series of resistors to step down the B+ and heater voltages to my sound card. I believe it will tolerate 3 volts. Tom has provided an algorithm to do this. PLEASE IGNORE THOSE SCOPE MEASUREMENTS! Tom noticed that my scope ground was not referrencing what I was measuring. I have since fixed that and have new measurements for the preamp out and amp speaker out. I am glad to report that they are much more in line with what I am hearing. I will post them later tonight.
megajocke said:
12X4
hailteflon said:
Yeah, the preamp audibly hums without the tubes in. I have replaced the 12AU7s not the 12X4.
I looks like I have some work to do. I'll post pics and measurements later.
Cheers,
gary
"I thought that this would create a loop between input-pcb-output. I'll move i/o closer together and connect their grounds in the fashion Chrish prescribed earlier. Having said this, I remember in the past connecting the other end of the output shield to the board and the hum goes wild. I'll try again."
How was transformer center tap connected when this happened? Was this wild hum only with power on? The hum could be due to charging current from either the heater rectifier or HV rectifier flowing around the board.
"Tom has advocated this too. I'll get on it. Connect CT to ground track at filter caps (101,102). Cut pcb ground track before and after these caps and send a separate ground returning to the non-chassis side of the star ground. Do I have this right?"
If you look at:
http://home.comcast.net/~garyworld/preamp1.jpg
I think the center tap can be connected there, but then the track circling the board should be cut to the rear of the unit of where the center tap connects - so that peaky current can only flow between center tap and capacitors and not around the board. The connection from caps to the rest of the circuit can probably use the track going out to the left from the filter caps. The ground isolator network should connect to chassis and earth ground on one side and only to the common wire linking RCA barrels on the other in this case.
"I'll get a new fuse and make new connections with leads rather than soldering straight to it's terminals. "
Hmm, now I'm not following you. I meant that fuse in the picture i linked to. But I just checked through the thread and saw that you have moved the transformer out of the amp. I don't think you posted a picture of the insides of the other box so maybe there isn't a problem. I find the PCB layout a bit dangerous with the mains fuse there.
"12X4"
Does anyone know any data for this? Especially max h-k voltage.
"Yeah, the preamp audibly hums without the tubes in. I have replaced the 12AU7s not the 12X4. I looks like I have some work to do. I'll post pics and measurements later."
Good luck
Could you take a pic of the area around heater rectifer and smoothing caps? Are the tube heaters connected with individual traces back to their supply or do they use the same ground tracks as the rest of the circuit?
How was transformer center tap connected when this happened? Was this wild hum only with power on? The hum could be due to charging current from either the heater rectifier or HV rectifier flowing around the board.
"Tom has advocated this too. I'll get on it. Connect CT to ground track at filter caps (101,102). Cut pcb ground track before and after these caps and send a separate ground returning to the non-chassis side of the star ground. Do I have this right?"
If you look at:
http://home.comcast.net/~garyworld/preamp1.jpg
I think the center tap can be connected there, but then the track circling the board should be cut to the rear of the unit of where the center tap connects - so that peaky current can only flow between center tap and capacitors and not around the board. The connection from caps to the rest of the circuit can probably use the track going out to the left from the filter caps. The ground isolator network should connect to chassis and earth ground on one side and only to the common wire linking RCA barrels on the other in this case.
"I'll get a new fuse and make new connections with leads rather than soldering straight to it's terminals. "
Hmm, now I'm not following you. I meant that fuse in the picture i linked to. But I just checked through the thread and saw that you have moved the transformer out of the amp. I don't think you posted a picture of the insides of the other box so maybe there isn't a problem. I find the PCB layout a bit dangerous with the mains fuse there.
"12X4"
Does anyone know any data for this? Especially max h-k voltage.
"Yeah, the preamp audibly hums without the tubes in. I have replaced the 12AU7s not the 12X4. I looks like I have some work to do. I'll post pics and measurements later."
Good luck
Could you take a pic of the area around heater rectifer and smoothing caps? Are the tube heaters connected with individual traces back to their supply or do they use the same ground tracks as the rest of the circuit?
hailteflon said:
Well, it shouldn't float as capacitive coupling through the transformer can couple into the tubes. Grounding fixes that - but now the tubes have high h-k voltage stress which might not be too good. If the heater return back to filter caps share traces with signal ground then there will be a great risk for hum. It is probably best first made floating and then floated with a resistor divider up to maybe 100V or so. A cap to HV supply ground would be needed to attenuate noise coupled through the transformer.
H-K max for a 12X4 is 450V negative
and 200V positive with respect to cathode.
Peak ratings.
Wouldn't this occur with AC on the filament? The reason I ask is because the 12X4 has AC on the filament from the same source as the DC for the other tubes. I was wondering if this would create hum.
and 200V positive with respect to cathode.
Peak ratings.
Wouldn't this occur with AC on the filament? The reason I ask is because the 12X4 has AC on the filament from the same source as the DC for the other tubes. I was wondering if this would create hum.
Hi megajocke,
I think that the easiest and best way to describe what you think should be changed, in the grounding scheme on the PCB, might be to indicate it on the following PCB Layout diagram (and/or some of the other diagrams and photos):
http://home.comcast.net/~garyworld/pcblayout.jpg
All of Gary's schematics and photos are best-accessed from this page:
http://home.comcast.net/~garyworld/site/?/photos/
This looks like one of the latest photos of the whole unit:
http://home.comcast.net/~garyworld/preampfeb.jpg
I think that the easiest and best way to describe what you think should be changed, in the grounding scheme on the PCB, might be to indicate it on the following PCB Layout diagram (and/or some of the other diagrams and photos):
http://home.comcast.net/~garyworld/pcblayout.jpg
All of Gary's schematics and photos are best-accessed from this page:
http://home.comcast.net/~garyworld/site/?/photos/
This looks like one of the latest photos of the whole unit:
http://home.comcast.net/~garyworld/preampfeb.jpg
Hmm that layout has its problems. The worst one is probably how the heater circuit is connected - the rectifier connection and that the heater return is shared with signal returns.
The rectifier bit is probably worse, I'd start with cutting the track as indicated in the picture and do the grounding changes. Cutting the track will keep the rectifier current out of sensitive areas at least.
I'd start with doing these modifications and then see how much it helps. If it still hums without power then something is fishy but we shouldn't be far from finding what causes it in that case. After these modifications hum when powered is quite possible though.
The rectifier bit is probably worse, I'd start with cutting the track as indicated in the picture and do the grounding changes. Cutting the track will keep the rectifier current out of sensitive areas at least.
I'd start with doing these modifications and then see how much it helps. If it still hums without power then something is fishy but we shouldn't be far from finding what causes it in that case. After these modifications hum when powered is quite possible though.
Attachments
gentlemen,
thanks for bearing with me. So I have made the changes megajocke detailed. I took a picture of the current state and inserted some comments. Please find it here:
http://home.comcast.net/~garyworld/preampfeb29.jpg
I had already scratched out both sections of the pcb ground around the filter caps when I saw megajocke's very nicely detailed picture. (I was inspired to do some photoshopping of my own.) I assume this necessitates a separate ground returning from this section to the star ground. Notice I put one in. I checked to make sure that there was no continuity between it and the rest of the pcb ground scheme before making the one pcb connection to star ground, (see pic above.) If this doesn't look right I can always tin in the cut section at C300. Also notice I connnected the CT at the first filter cap ground, attached the output ground/shield to the pcb, moved outputs up next to inputs, and bundled signal leads to be closer together. Let me know what else I can do to complete the starground and other grounding issues so that we may move on.
I deleted the erroneous scope measurements and replaced them with the following:
measured at preamp out, preamp on:
http://home.comcast.net/~garyworld/preampon.jpg
the little numbers indicate a 2.5V 120 Hz wave, a bit modulated (perhaps by small 60 Hz)
measured at speaker out of amp with both on:
http://home.comcast.net/~garyworld/bothon.jpg
this is a 25-30V 120 Hz wave with a more pronounced modulation by something else.
Once I corrected the probe ground problem Tom noticed, the amp out read flat when on and off, and the preamp measured flat when off but plugged in.
For what it's worth, with the changes I made today the preamp hum continues. Mark, the .33uF 400V caps arrived today, I will try to put them in tonight and report back.
thanks for bearing with me. So I have made the changes megajocke detailed. I took a picture of the current state and inserted some comments. Please find it here:
http://home.comcast.net/~garyworld/preampfeb29.jpg
I had already scratched out both sections of the pcb ground around the filter caps when I saw megajocke's very nicely detailed picture. (I was inspired to do some photoshopping of my own.) I assume this necessitates a separate ground returning from this section to the star ground. Notice I put one in. I checked to make sure that there was no continuity between it and the rest of the pcb ground scheme before making the one pcb connection to star ground, (see pic above.) If this doesn't look right I can always tin in the cut section at C300. Also notice I connnected the CT at the first filter cap ground, attached the output ground/shield to the pcb, moved outputs up next to inputs, and bundled signal leads to be closer together. Let me know what else I can do to complete the starground and other grounding issues so that we may move on.
I deleted the erroneous scope measurements and replaced them with the following:
measured at preamp out, preamp on:
http://home.comcast.net/~garyworld/preampon.jpg
the little numbers indicate a 2.5V 120 Hz wave, a bit modulated (perhaps by small 60 Hz)
measured at speaker out of amp with both on:
http://home.comcast.net/~garyworld/bothon.jpg
this is a 25-30V 120 Hz wave with a more pronounced modulation by something else.
Once I corrected the probe ground problem Tom noticed, the amp out read flat when on and off, and the preamp measured flat when off but plugged in.
For what it's worth, with the changes I made today the preamp hum continues. Mark, the .33uF 400V caps arrived today, I will try to put them in tonight and report back.
Hi Gary,
Is there any change in the amp hum with the preamp unpowered, or unpowered and unplugged?
By the way, is the filter caps' ground return wire resting against the center tap's wire?
Anyway, I'm not sure that everything was done as megajocke indicated.
(At this point, I think it would probably be better to forget most of what I've said and try to focus on megajocke's suggestions, instead.)
Note the last sentence, in the quote below.
[Does that still apply, megajocke?] (However, I'm not sure I understood how the filter caps' ground should be routed. And I'm left wondering where the 'ideal' center-tap connection point might be.)
------- QUOTE (from megajocke's post # 106):
I think the center tap can be connected there, but then the track circling the board should be cut to the rear of the unit of where the center tap connects - so that peaky current can only flow between center tap and capacitors and not around the board. The connection from caps to the rest of the circuit can probably use the track going out to the left from the filter caps. The ground isolator network should connect to chassis and earth ground on one side and only to the common wire linking RCA barrels on the other in this case.
--------
Here's MY current opinion:
It seems to me that if we can solve the hum problem for when the preamp is NOT powered, then we will probably also have mostly solved the problem for when it IS powered, and probably with the fewest modifications of the original equipment setup.
(After that is solved, if Gary were me, he would then also want to implement a complete, proper star-grounding scheme, if that has not already had to be done during the process of solving the original hum problem.)
-------
So, I'm wondering, after garyh and megajocke have made sure that the latest suggested wiring changes have been correctly implemented, are there any meter or scope measurements that could be performed, that might give important information?
Is there any change in the amp hum with the preamp unpowered, or unpowered and unplugged?
By the way, is the filter caps' ground return wire resting against the center tap's wire?
Anyway, I'm not sure that everything was done as megajocke indicated.
(At this point, I think it would probably be better to forget most of what I've said and try to focus on megajocke's suggestions, instead.)
Note the last sentence, in the quote below.
[Does that still apply, megajocke?] (However, I'm not sure I understood how the filter caps' ground should be routed. And I'm left wondering where the 'ideal' center-tap connection point might be.)
------- QUOTE (from megajocke's post # 106):
I think the center tap can be connected there, but then the track circling the board should be cut to the rear of the unit of where the center tap connects - so that peaky current can only flow between center tap and capacitors and not around the board. The connection from caps to the rest of the circuit can probably use the track going out to the left from the filter caps. The ground isolator network should connect to chassis and earth ground on one side and only to the common wire linking RCA barrels on the other in this case.
--------
Here's MY current opinion:
It seems to me that if we can solve the hum problem for when the preamp is NOT powered, then we will probably also have mostly solved the problem for when it IS powered, and probably with the fewest modifications of the original equipment setup.
(After that is solved, if Gary were me, he would then also want to implement a complete, proper star-grounding scheme, if that has not already had to be done during the process of solving the original hum problem.)
-------
So, I'm wondering, after garyh and megajocke have made sure that the latest suggested wiring changes have been correctly implemented, are there any meter or scope measurements that could be performed, that might give important information?
Gary,
I have taken the liberty of posting some of the details of your preamp's hum problem in the sci.electronics.design and the rec.audio.tubes Usenet newsgroups, with a request for any ideas that might be of help. The message-traffic of those groups is accessible through http://groups.google.com , where it will also be archived.
I also posted a link to this diyaudio.com thread, and a link to your thumbnail-images web page, in those newsgroup threads.
I may try to post, here, any significant new ideas about possible solutions, from those sources, if the posters from those forums do not contribute directly to this thread.
The sci.electronics.design thread about this preamp problem is at:
http://groups.google.com/group/sci....b16bf44591d/28e4c3665610bf53#28e4c3665610bf53
The rec.audio.tubes thread about this preamp problem is at:
http://groups.google.com/group/rec....55b8584b92a/89805ada6da5b556#89805ada6da5b556
I have taken the liberty of posting some of the details of your preamp's hum problem in the sci.electronics.design and the rec.audio.tubes Usenet newsgroups, with a request for any ideas that might be of help. The message-traffic of those groups is accessible through http://groups.google.com , where it will also be archived.
I also posted a link to this diyaudio.com thread, and a link to your thumbnail-images web page, in those newsgroup threads.
I may try to post, here, any significant new ideas about possible solutions, from those sources, if the posters from those forums do not contribute directly to this thread.
The sci.electronics.design thread about this preamp problem is at:
http://groups.google.com/group/sci....b16bf44591d/28e4c3665610bf53#28e4c3665610bf53
The rec.audio.tubes thread about this preamp problem is at:
http://groups.google.com/group/rec....55b8584b92a/89805ada6da5b556#89805ada6da5b556
OK. Here is a post that looks like it contains some new ideas, which was posted in the sci.electronics.design group thread (by user 'legg', signed 'RL'). Maybe some of the experts, here, will see some good clues, in this:
----- quote:
Sounds (reads) like you've tackled a lot of issues, so far. There's always a problem with iterations, that makes it possible to accidentally pass over or defeat solutions.
First thing to fix is hum when off. Ignore all other issues.
With hum present, disconnect all other equipment. Leave disconnected until powered-off hum is solved. This includes your 'computer scope'.
These connections will just complicate things. Use your ears to measure the hum. When you can't hear objectionable levels, the problem is on the road to being solved.
trannybox - the orange wire should follow the brown wire from the AC line socket to the switch. Twisting uses the least hardware. You are switching the neutral wire in this photo. This could explain a 60Hz hum when the preamp is off.
You said that you reversed wires going to the switch. This wording is subject to misinterpretation. The live terminal of the IEC inlet (brown wire in your trannybox photo) should go through the fuse, to the switch. From the switched terminal the returning connection should run to the transformer primary.I am surprised that both transformer primary wires are colour-coded brown. The orange wire on the polarized neutral of the IEC inlet should follow the brown wire to the switch and return unbroken to the transformer primary. Twisting it around the live wire uses less hardware to accomplish this routing.
This is actually how the wiring is laid out in the 'preamp1' photo,
using printed wiring, but without the polarized inlet. The 2-pin 120V hardware can also be polarized - the wider blade is supposed to be neutral. Old Dynaco hardware may predate this convention.
There is no ground wire connection. Assuming plastic hardware, this would still have connected to the transformer case, followed by a lead to the chassis of the target load, to allow for leakage current return to safety earth. If not present, connect the transformer case to the chassis, anyways, even from a remote plastic box.
Check the line cord phasing of the power amplifier, it should match that of the preamp, even if line and neutral are not properly wired into your wall socket.
The power supply schematic does not use the same transformer you are using. The original has a colour-coded primary, a grounded core/screen. If you have an isolation transformer, use this temporarily for the preamp, power amp and combined set to see if there's a difference one way or another.
Ignore other issues until powered-off hum is defeated.
RL
----- end quote
----- quote:
Sounds (reads) like you've tackled a lot of issues, so far. There's always a problem with iterations, that makes it possible to accidentally pass over or defeat solutions.
First thing to fix is hum when off. Ignore all other issues.
With hum present, disconnect all other equipment. Leave disconnected until powered-off hum is solved. This includes your 'computer scope'.
These connections will just complicate things. Use your ears to measure the hum. When you can't hear objectionable levels, the problem is on the road to being solved.
trannybox - the orange wire should follow the brown wire from the AC line socket to the switch. Twisting uses the least hardware. You are switching the neutral wire in this photo. This could explain a 60Hz hum when the preamp is off.
You said that you reversed wires going to the switch. This wording is subject to misinterpretation. The live terminal of the IEC inlet (brown wire in your trannybox photo) should go through the fuse, to the switch. From the switched terminal the returning connection should run to the transformer primary.I am surprised that both transformer primary wires are colour-coded brown. The orange wire on the polarized neutral of the IEC inlet should follow the brown wire to the switch and return unbroken to the transformer primary. Twisting it around the live wire uses less hardware to accomplish this routing.
This is actually how the wiring is laid out in the 'preamp1' photo,
using printed wiring, but without the polarized inlet. The 2-pin 120V hardware can also be polarized - the wider blade is supposed to be neutral. Old Dynaco hardware may predate this convention.
There is no ground wire connection. Assuming plastic hardware, this would still have connected to the transformer case, followed by a lead to the chassis of the target load, to allow for leakage current return to safety earth. If not present, connect the transformer case to the chassis, anyways, even from a remote plastic box.
Check the line cord phasing of the power amplifier, it should match that of the preamp, even if line and neutral are not properly wired into your wall socket.
The power supply schematic does not use the same transformer you are using. The original has a colour-coded primary, a grounded core/screen. If you have an isolation transformer, use this temporarily for the preamp, power amp and combined set to see if there's a difference one way or another.
Ignore other issues until powered-off hum is defeated.
RL
----- end quote
Yes, like gootee says, does it still hum when unplugged and/or turned off? That should be solved first and then we can try to find the source of the hum when it's powered.
Could you check if it hums with the rectifier tube pulled? If it does that then it's definitely the heater return current. If it doesn't it could still be. In that case check if there is hum if you short the black and green twisted wires together where they connect to the board. If it still does then I'd guess heater return current or ripple on anode voltage.
The new connection looks better but I think it actually might better to bridge the lower cut and use that trace for power supply return. I'd connect the ground isolator to the jacks but I doubt it will make any difference. The heater circuit looks suspicious though.
Could you check if it hums with the rectifier tube pulled? If it does that then it's definitely the heater return current. If it doesn't it could still be. In that case check if there is hum if you short the black and green twisted wires together where they connect to the board. If it still does then I'd guess heater return current or ripple on anode voltage.
The new connection looks better but I think it actually might better to bridge the lower cut and use that trace for power supply return. I'd connect the ground isolator to the jacks but I doubt it will make any difference. The heater circuit looks suspicious though.
I just read the newsgroup posts, and some interesting points are made. Especially the one about hum when turned off and high impedance state. If it is due to the high impedance and electrostatic coupling then this hum should not matter as off state performance is quite unimportant as was pointed out. It should diminish in strength greatly though by putting the cover on. But if it is magnetic coupling into loops in the circuit then it will probably not change much when turned on. It will be easier to find the problem though if it doesn't hum when switched off.
It should diminish in strength greatly though by putting the cover on. But if it is magnetic coupling into loops in the circuit then it will probably not change much when turned on. It will be easier to find the problem though if it doesn't hum when switched off.
Assuming the circuit isn't defective (no open/loose connections, etc.):
1) Make a separate power supply and keep it at least a couple feet away from the preamp. Connect the power supply to the preamp by a length of cable long enough so you can keep the power supply far enough away from the preamp. Don't use and/or remove the power supply from the preamp.
1) Make a separate power supply and keep it at least a couple feet away from the preamp. Connect the power supply to the preamp by a length of cable long enough so you can keep the power supply far enough away from the preamp. Don't use and/or remove the power supply from the preamp.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.