Hi Ale,
Thanks for your reply!
The only other change was increasing R9 gate-stoppers on gyrators from 330R to 1K. At 20mA, I suppose the gate stoppers are not even needed on the BSH11BK.
I did also increase the grid-stoppers on the 2p29L from 330R to 1K as well. Forgot to mention that earlier. Can’t see the grid-stoppers causing the hum, but worth noting anyway.
All else remains the same as before the layout change.
Sounds quite nice. I could live with the hum, as it is not terribly loud, but it got me very curious as to what could be causing it.
Is it possible that the proximity of the filament regs to the gyrators are causing a reaction?
Or something about the filament regs sharing a common heatsink? (I am using mica spacers and measure open-line between the metal tab (drain) and the heatsink on the P16N06)
Best regards,
John
Thanks for your reply!
The only other change was increasing R9 gate-stoppers on gyrators from 330R to 1K. At 20mA, I suppose the gate stoppers are not even needed on the BSH11BK.
I did also increase the grid-stoppers on the 2p29L from 330R to 1K as well. Forgot to mention that earlier. Can’t see the grid-stoppers causing the hum, but worth noting anyway.
All else remains the same as before the layout change.
Sounds quite nice. I could live with the hum, as it is not terribly loud, but it got me very curious as to what could be causing it.
Is it possible that the proximity of the filament regs to the gyrators are causing a reaction?
Or something about the filament regs sharing a common heatsink? (I am using mica spacers and measure open-line between the metal tab (drain) and the heatsink on the P16N06)
Best regards,
John
Oscillation / instability will manifest in the LF range and perceived as hum. The gate stopper of the BSH111BK is mandatory, otherwise will tend to oscillate, particularly when anode current is below 25mA. Put a 330R, not higher. For the LND150 and top MOSFET 1k is ok.
For the 2P29L the grid resistor is 330. No need to start changing things when it was working before!
Cheers
Ale
For the 2P29L the grid resistor is 330. No need to start changing things when it was working before!
Cheers
Ale
Hello John,
Yes, chassis or heatsink can acquire quite a lot of noise, so grounding it is usually the best way.
The regulator boards do no emit much it the way of AC field, so putting them close to the tubes should not cause problems - in fact it should reduce hum.
The filament supply's input-voltage cable may benefit from being shielded all the way up to the regulator input terminals. At low currents like 120mA for 2P29L, using microphone cable or other audio twinax patch cable is perfect. This is a precaution for high-impedance filaments like 2P29L, since the regulator does not eliminate noise from external fields (for the sake of the sound - low differential impedance at the filament will crush & corrupt part of the music signal). The input-voltage cable can acquire common-mode noise from external fields, and shielding is helpful.
The input capacitor for the regulator - Is it stacked construction? Some of the large-size parts are made like a Swiss Roll, and are thus somewhat inductive. Stacked construction for this cap is certainly preferred. It is very remote from the music, so an audio type is not required. C3 must not be substituted.
Is the grid wiring well away from iron parts, or floating metal, or any filament wiring?
Ale mentioned the FET gate stoppers. It is certainly true that high values should be avoided, since the oscillation they aim to prevent may reappear by means of internal feedback around the FETs. This is actually increased if high-value stoppers are used, since the drain-gate feedback will see a higher impedance (looking out of the gate) at HF+. With the types of FET we use in audio builds, 100Ω is usually ideal. The resistor is not used to from a low-pass filter. It simply acts as a Q-spoiler for unintended resonances in the external circuit & wiring.
Yes, chassis or heatsink can acquire quite a lot of noise, so grounding it is usually the best way.
The regulator boards do no emit much it the way of AC field, so putting them close to the tubes should not cause problems - in fact it should reduce hum.
The filament supply's input-voltage cable may benefit from being shielded all the way up to the regulator input terminals. At low currents like 120mA for 2P29L, using microphone cable or other audio twinax patch cable is perfect. This is a precaution for high-impedance filaments like 2P29L, since the regulator does not eliminate noise from external fields (for the sake of the sound - low differential impedance at the filament will crush & corrupt part of the music signal). The input-voltage cable can acquire common-mode noise from external fields, and shielding is helpful.
The input capacitor for the regulator - Is it stacked construction? Some of the large-size parts are made like a Swiss Roll, and are thus somewhat inductive. Stacked construction for this cap is certainly preferred. It is very remote from the music, so an audio type is not required. C3 must not be substituted.
Is the grid wiring well away from iron parts, or floating metal, or any filament wiring?
Ale mentioned the FET gate stoppers. It is certainly true that high values should be avoided, since the oscillation they aim to prevent may reappear by means of internal feedback around the FETs. This is actually increased if high-value stoppers are used, since the drain-gate feedback will see a higher impedance (looking out of the gate) at HF+. With the types of FET we use in audio builds, 100Ω is usually ideal. The resistor is not used to from a low-pass filter. It simply acts as a Q-spoiler for unintended resonances in the external circuit & wiring.
Hi Ale, Hi Rod,
Thanks for your responses!
I had initially increased the grid and gate stoppers In an attempt to address some noise in the right channel. It just so happened that my 201s exhibited very similar noise on the same side when they were in that chassis.
It ended up to be a coincidence. Two bad tubes.
I will return them to their original value, which was Ale’s suggested 330R.
If that does not fix it, I will try screened cable for the supply going to the regs. Those wires are a few centimeters longer, now, so perhaps that is a contributing cause.
I really appreciate the help!
Best regards,
John
Thanks for your responses!
I had initially increased the grid and gate stoppers In an attempt to address some noise in the right channel. It just so happened that my 201s exhibited very similar noise on the same side when they were in that chassis.
It ended up to be a coincidence. Two bad tubes.
I will return them to their original value, which was Ale’s suggested 330R.
If that does not fix it, I will try screened cable for the supply going to the regs. Those wires are a few centimeters longer, now, so perhaps that is a contributing cause.
I really appreciate the help!
Best regards,
John
Hi Rod,
Also, yes, C3 is stacked construction, and the regs were hum free before the recent changes.
But if the noise persists through the above measures, I will try some more compact caps there.
I had somehow misplaced the original caps from your kit during construction, and had those in my parts box. A little big, but the right value.
Best,
John
Also, yes, C3 is stacked construction, and the regs were hum free before the recent changes.
But if the noise persists through the above measures, I will try some more compact caps there.
I had somehow misplaced the original caps from your kit during construction, and had those in my parts box. A little big, but the right value.
Best,
John
Sorry for the multiple posts.
Rod, what about trying a small common mode choke just before the filament regs instead of screened cable? I have some.
Maybe that way it could help address any noise picked up by the power supply umbilical, too? (Though, my umbilical is already shielded.)
Best,
John
Rod, what about trying a small common mode choke just before the filament regs instead of screened cable? I have some.
Maybe that way it could help address any noise picked up by the power supply umbilical, too? (Though, my umbilical is already shielded.)
Best,
John
Just to follow up:
I changed the gate and grid stoppers back to 330R, with no change in the hum.
Then I tried powering the filaments with 9v batteries which made the preamp nearly dead silent. This arrangement sounded good, too. I may explore this route at some later time.
This at least confirmed that the hum was caused by noise picked up by the raw filament supply wiring.
So I added 10000uf and a small common mode choke just before the supply input on the filament regulators. Hum is gone.
Thanks again for all your input!
Best regards,
John
I changed the gate and grid stoppers back to 330R, with no change in the hum.
Then I tried powering the filaments with 9v batteries which made the preamp nearly dead silent. This arrangement sounded good, too. I may explore this route at some later time.
This at least confirmed that the hum was caused by noise picked up by the raw filament supply wiring.
So I added 10000uf and a small common mode choke just before the supply input on the filament regulators. Hum is gone.
Thanks again for all your input!
Best regards,
John
Can you share what type of common mode choke you are using? The tiny toroid ones abound. And then there is LL1694 at $78 a pop.
I get more hum when using the 26 tube, with about 1 amp of filament current as opposed to the 01a, with a measly 250 mA. I have also traced it to the filament supply. I am going to try the shielded cable into the regs, as Rod suggested. The 60 Hz spike is 60-70 db below the fundamental, so, it is barely perceptible with the ear to the speaker. Nevertheless, it would be good to cut it down further.
I get more hum when using the 26 tube, with about 1 amp of filament current as opposed to the 01a, with a measly 250 mA. I have also traced it to the filament supply. I am going to try the shielded cable into the regs, as Rod suggested. The 60 Hz spike is 60-70 db below the fundamental, so, it is barely perceptible with the ear to the speaker. Nevertheless, it would be good to cut it down further.
Last edited:
Hi ra7,
This is what I used because it is what I had in hand:
Würth 744866563
It is 56mH @ 1.5A
Cheap. Not a toroid.
http://https://www.digikey.com/en/products/detail/würth-elektronik/744866563/5352954?s=N4IgTCBcDaIOwBYEA4BsqCsqDMIC6AvkA
I spoke a bit too soon, though. After about an hour of warm up the hum returned, albeit in a much diminished form. I will also try running mic cable all the way from the last cap on the PS straight to the regs and see if that helps.
Regards,
John
This is what I used because it is what I had in hand:
Würth 744866563
It is 56mH @ 1.5A
Cheap. Not a toroid.
http://https://www.digikey.com/en/products/detail/würth-elektronik/744866563/5352954?s=N4IgTCBcDaIOwBYEA4BsqCsqDMIC6AvkA
I spoke a bit too soon, though. After about an hour of warm up the hum returned, albeit in a much diminished form. I will also try running mic cable all the way from the last cap on the PS straight to the regs and see if that helps.
Regards,
John
Last edited:
Just as an aside, when I've experience hum with your topology, I found relief by providing a bit more headroom to the filament regulator. Maybe just another volt or so will do, but I know it's not necessarily easy to do! Use the lowest-loss diodes you can find and low-DCR wiring between the filter cap and reg.
jlevro-- Interesting. Easy enough to do. I actually drop the voltage a bit with the last RC section of the filament PS, so dropping a bit less is no problem.
I did notice when I tried the 9v batteries that the channel with the older battery at ~8V was a little noisy. However, the noise was not hum but a white-noise-like 'whooshing' sound.
euro21-- good question. I have tried it both ways and there is no difference I can detect.
Thanks for your replies!
Best,
John
I did notice when I tried the 9v batteries that the channel with the older battery at ~8V was a little noisy. However, the noise was not hum but a white-noise-like 'whooshing' sound.
euro21-- good question. I have tried it both ways and there is no difference I can detect.
Thanks for your replies!
Best,
John
If the preamp was hum-free before a layout change, some kind of field pickup is likely to be the cause. If it is common-mode pickup (most likely) the common-mode choke may help, but screening the cable may well work better. You can test for common-mode by removing the added capacitor, since this will have little effect on common-mode noise.
Differential-mode noise will be reduced by the (smaller) differential-mode inductance of the common-mode choke in combination with the new capacitor. But differential-mode noise should not be a problem, provided 7.0V+ is fed in, and the Raw DC is as per the manual, and with 10000µF caps.
Differential-mode noise will be reduced by the (smaller) differential-mode inductance of the common-mode choke in combination with the new capacitor. But differential-mode noise should not be a problem, provided 7.0V+ is fed in, and the Raw DC is as per the manual, and with 10000µF caps.
Batteries
Sorry to continue this topic in the "#26 pre amp" thread, but I thought I would post a follow up on my 2p29L saga.
I tried a few things to get rid of the hum I was experiencing:
1-Ran screened microphone cable straight from the last cap of the filament Raw PS al the way to the Filament Reg boards.
2-Tried common-mode chokes just before the Reg boards with and without 10,000uf caps
3-Bypassed the chokes in the Raw PS and made it CRCRC
(and finally)
4-Got some 18650 Li-ion cells and battery powered the filaments.
Numbers 1-3 had only a small effect on the hum (and the CRCRC really doesn't sound nearly as good as the LCRC-CMC-C supply).
I must be bombarded with some kind of radiation in this part of the house. Near by cell tower? Radio transmitter?
I have had to completely encase my 26 preamp--tubes and all. (The 26 pre sounds great LC coupled, and is very quiet in it's heavy steal coffin)
#4, however, has really done the trick for the 2p29L and sounds excellent, too. So far I have gotten 20 hours on a single charge. I think this may be the solution I will live with for this preamp.
3 18650 cells in series with 50R dropping resistor paralleled with a 100R wirewound pot to keep the voltage in range.
A little crude, I know. I need to check and reset the voltage at the end of each evening of listening, but it is no problem, really.
Regards,
John
Sorry to continue this topic in the "#26 pre amp" thread, but I thought I would post a follow up on my 2p29L saga.
I tried a few things to get rid of the hum I was experiencing:
1-Ran screened microphone cable straight from the last cap of the filament Raw PS al the way to the Filament Reg boards.
2-Tried common-mode chokes just before the Reg boards with and without 10,000uf caps
3-Bypassed the chokes in the Raw PS and made it CRCRC
(and finally)
4-Got some 18650 Li-ion cells and battery powered the filaments.
Numbers 1-3 had only a small effect on the hum (and the CRCRC really doesn't sound nearly as good as the LCRC-CMC-C supply).
I must be bombarded with some kind of radiation in this part of the house. Near by cell tower? Radio transmitter?
I have had to completely encase my 26 preamp--tubes and all. (The 26 pre sounds great LC coupled, and is very quiet in it's heavy steal coffin)
#4, however, has really done the trick for the 2p29L and sounds excellent, too. So far I have gotten 20 hours on a single charge. I think this may be the solution I will live with for this preamp.
3 18650 cells in series with 50R dropping resistor paralleled with a 100R wirewound pot to keep the voltage in range.
A little crude, I know. I need to check and reset the voltage at the end of each evening of listening, but it is no problem, really.
Regards,
John
It's curious that it worked OK before moving things around. I'm happy to continue offering suggestions, if you have the energy to try them!
Another data point arose recently, and may be worth a thought - the suggestion that the Hammond EI power transformers have taken quality short-cuts.
See here:
DHT driver for triode wired SE EL84, 6V6 or EL34
Hammond 266 are (or were) my standard recommendation for power trafos, as they are EI construction, and easy to buy from any location.
Naturally, mechanical vibration of power trafos is often audible in itself.
I suspect that mechanical vibration from the suspect Hammonds may be a sign of saturation - and power transformer saturation will be a disaster for preamps, because the external field strength around the trafo will increase greatly.
External fields are certainly the cause of the problem, if a low impedance filament feed masks the problem. I say masks, because applying the low resistance across the filament bypasses the fields, but it also shorts the low-gm side of the filament to the high-gm side, and this usually produces audible degradation.
If you find that LCRC in the Raw DC make a big difference, that that is a clue, because the power trafo current-peaks are smaller in LCRC, and this in turn affects the strength and spectrum of the external fields.
Locating the power trafo and raw DC some distance from the signal wiring can be helpful.
Looking again at the photos, wiring the input/grid and output in twinax (2-core shielded audio cable or Microphone cable) in cases where the wiring and signal parts are not fully enclosed in a chassis may help protect against E-fields.
Do send me some email, if I can help though.
Another data point arose recently, and may be worth a thought - the suggestion that the Hammond EI power transformers have taken quality short-cuts.
See here:
DHT driver for triode wired SE EL84, 6V6 or EL34
Hammond 266 are (or were) my standard recommendation for power trafos, as they are EI construction, and easy to buy from any location.
Naturally, mechanical vibration of power trafos is often audible in itself.
I suspect that mechanical vibration from the suspect Hammonds may be a sign of saturation - and power transformer saturation will be a disaster for preamps, because the external field strength around the trafo will increase greatly.
External fields are certainly the cause of the problem, if a low impedance filament feed masks the problem. I say masks, because applying the low resistance across the filament bypasses the fields, but it also shorts the low-gm side of the filament to the high-gm side, and this usually produces audible degradation.
If you find that LCRC in the Raw DC make a big difference, that that is a clue, because the power trafo current-peaks are smaller in LCRC, and this in turn affects the strength and spectrum of the external fields.
Locating the power trafo and raw DC some distance from the signal wiring can be helpful.
Looking again at the photos, wiring the input/grid and output in twinax (2-core shielded audio cable or Microphone cable) in cases where the wiring and signal parts are not fully enclosed in a chassis may help protect against E-fields.
Do send me some email, if I can help though.
Hi Rod,
Thanks, as always, for your help!
First of all, let me state the obvious: This is not any fault of your Reg boards. They are great. It is something going on at my house. Like I said, my type 26s will howl and moan like an amorous banshee if they are not fully shielded. I ended up with a heavy-gauge steel box to enclose the whole circuit, and a shielded PS umbilical.
It is also not uncommon to pick up some RF noise when I hook up my solid state amps.
I forgot to mention in my last post that I also tried the filament regulators back in their original position, and still got the hum.
I think that perhaps I just didn't notice it before I reworked the layout.
All the signal wiring is coax. I even tried a shielded anode wire. None of it helped.
I'm actually not that picky about a little hum. But once I did notice it, I thought it might be good to try and get rid of it.
My HV and LV PSUs are on a single chassis. Maybe I should try to split them up?
The induced noise would be from HV to LV (is this the right way to put it?), I think, because when I use the battery powered filaments with the LV supply powered (but idle) there is no hum.
Thanks, again, for your input.
Best regards,
John
Thanks, as always, for your help!
First of all, let me state the obvious: This is not any fault of your Reg boards. They are great. It is something going on at my house. Like I said, my type 26s will howl and moan like an amorous banshee if they are not fully shielded. I ended up with a heavy-gauge steel box to enclose the whole circuit, and a shielded PS umbilical.
It is also not uncommon to pick up some RF noise when I hook up my solid state amps.
I forgot to mention in my last post that I also tried the filament regulators back in their original position, and still got the hum.
I think that perhaps I just didn't notice it before I reworked the layout.
All the signal wiring is coax. I even tried a shielded anode wire. None of it helped.
I'm actually not that picky about a little hum. But once I did notice it, I thought it might be good to try and get rid of it.
My HV and LV PSUs are on a single chassis. Maybe I should try to split them up?
The induced noise would be from HV to LV (is this the right way to put it?), I think, because when I use the battery powered filaments with the LV supply powered (but idle) there is no hum.
Thanks, again, for your input.
Best regards,
John