120Hz hum, new caps, old tube unit, chassis-ground questions..

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
I had an ADCOM GFA-555 II amp, which when I had a source plugged into it - it would have a low hum/buzz like there was a ground fault.
I must have spent hours into days trying to figure out the issue (tried all sorts of grounding tricks)... So to trouble shoot the problem> I picked up a Pioneer AMP with similar specs, and swapped the ADCOM out, and surprise the hum was not present. I figured well I guess my other tech is just too advanced to run on this older ADCOM, so I sold it. Not long after something else happened (I don't recall what) but it suddenly dawned on me the problem... it turned out that I was using a "Power filter" (aka an ESP Digital QC Surge Protector & Noise Filter) similar to or if not this model "ESP D5131NT". It seems that my newer IC based Pioneer had no issue with it, but the older more analog Transistor based amplifiers did not like the power coming off that device, which makes me lean towards the idea that the power coming out of these devices has a "modified sine wave" instead of the typical "pure sine wave" and that is what it was picking up noise wise.

Not that I can hear what your amp is doing/sounding like (and not that I've read any replies here yet) but with a 120hrz hum that sounds an awful lot like it might be power related... like you are picking up ambient electrical fields off a power source - in which case try placing some metal shielding around various places to see if that alters the hum intensity.
Or as I suggested perhaps the power being supplied to the amp is oscillating in a way that is causing it. (Modified sine or Pure Sine)
 
I would keep the HP as a signal generator and then build a tube amp,

Mine is kind of finicky, it won't put out 20 CP's til it has warmed up for a while.

Those things have juice! People used to turn them into bootleg AM radio stations since they went a bit past 600 kc.
 
I did a single tube pre-amp in a small aluminium box with a short wired in umbilical cable to the power amp enclosure for heaters and HT ,
Signal is on jacks in the usual way ,
With around 25 cm of cable and careful placement I could avoid 60hz field from the mains transformer entirely in the first pre amp stage ,
The other thing you can avoid is vibrations from the mains transformer being induced microphonically through the chassis into the tube ,
That becomes an issue at high gain ,
Let alone speaker cab vibrations ,

Any recommendations on a pre fabbed umbilical cable that can handle HT power arrangements?
Might have to just strap together one out of high gauge wires and beefy panel mount connectors.

I think a remote power transformer, or at least extended location inside a wider chassis, is the best next step.

The hum is very low now. It’s already pretty great sounding and the best overdrive I’ve ever heard in any amp. No chance it remains a signal generator :)
 
You could go insane running down all the suggestions in this post. First thing is to replace main filter caps which I think you've done. Then measure resistors in output circuit, especially cathode resistors...should be matched to 2% and tubes must also be matched (but how do you know tubes are matched...cuz they were sold as such, or do you have tube tester?). Install a cathode balance pot ( 1 watt) if in doubt..can always remove it later....not reliable long term. 120hz hum on output should be measured and noted. Measure and note (write down) ripple on B+ supply, also measure and note filament voltage.

Next isolate that part of the circuit you want to re-use, for instance V3,V4,V5 and resolder pot side of C9 to grounc point of pot R17. Does that quiet output sufficiently? If not then it's time to sub in some power supplies (filament and B+) Substitute voltages have to be well grounded to circuit ground. But where exactly is circuit ground? Chassis is of course a good ground to some degree, what I mean by that is power transformer eddy currents and ripple currents in chassis make some points on chassis a good ground, others hummy-buzzy ground. If parts of original circuit use chassis as ground wherever convenient that could be a problem or not...and it could be manufacturer went to the trouble to check on that when laying out build. Or they didn't check except to choose a later stage ground point that was suffiently out of phase with earlier stage ground point to effectively more or less cancel cum in final output. So it may be necessary to interupt B+ feeds to any circuit upstream from the output stage so their ripple does not go into chassis.

Circuit ground for sub'd in power supplies should be where V4 and V5 cathode resistors connect to ground or where or where C12 is grounded (connected to chassis), and ideally these will be the same point or only separated by an inch of chassis, cuz eddy currents or ripple current running in chassis between those two point could be a sosurce of hum in output.

Subbing in filaments supply (I think you did that) is easiest to do. If doing so does not fix adequately note the hum/buzz on B+ and output circuit and compare with previous notes. Maybe there was no improvement at all in those measurements....anyway just write it all down for future comparisons. If improvement from filament voltage sub gave no or inadequate improvement then it's time to sub B+, cuz what else is left at this point? If you don't have a good B+ supply for testing can you buy one...even a les expensive switching test bench supply from China will likely (maybe) be adequate for testing. Otherwise and anyway, while listening to output at sufficent level (on speaker you can sacrifice if that happens), and with filaments running on sub'd in dc filament supply, and upstream B+ power rails lifted, turn off the power to unit (but not to filaments). B+ voltage will take a couple secs to decay. You should observe the hum in output is gone for those couple secs if prob is from B+ supply or power transformer eddy currents or chassis ripple current. If that is the case then next step is to try co-locating all circuit ground points still relevant ( so not upstream stuff since B+ lifting previously mentioned)such as C12 grounds and V3,V4 and V5 circuit grounds. Doing so will eliminated as suspects the chassis ripple currents but maybe or maybe not power transformer eddy currents. If co-location does not fix sufficiently then only thing left to check is power transformer eddy current by it's removal and distancing from chassis by 6 inches. Note that if power transformer eddy currents are source of problem that might be revealed merely by removing it's mounting bolts and lifting it away from chassis by a 1/16" to 1/4".

If testing reveals power transformer eddy current problem then one will want to re-bolt power transformer to chassis and pick a chassis co-location ground point well away from power transformer, maybe at least 6 inches but only experimentation will know for sure. The ground of C12 is a good, and more or less mandatory ground co-location point. Having C12 (main filter cap) grounds on an arm of a star grounding design will only rarely work....barely even worth trying.

If you have or aquire a B+ test supply then sub it in and still the recommendations of previous 3 paragraphs hold but you will have more certainty and you won't have to deduce things during the 2 seconds waiting for B+ supply to decay to shutoff.

Bear in mind that even though HP was good test gear it might nothave had really low hum/buzz cuz some hum/buzz ws considered normal in those days and was just accepted....it sould have been hard to find a hi-fi amp that could amplifify signal into a speaker without adding it's own hum/buzz.

The above procedure is a pain in the a** and so time consuming it's hardly worth it but anyone who goes through this will have a new knowldege and skill others lack. This is akin to the torture one must endure to perfect the quad axel in figure skating as shown here:

Pain of perfection

Good luck.
 
You could go insane running down all the suggestions in this post. First thing is to replace main filter caps which I think you've done. Then measure resistors in output circuit, especially cathode resistors...should be matched to 2% and tubes must also be matched (but how do you know tubes are matched...cuz they were sold as such, or do you have tube tester?). Install a cathode balance pot ( 1 watt) if in doubt..can always remove it later....not reliable long term. 120hz hum on output should be measured and noted. Measure and note (write down) ripple on B+ supply, also measure and note filament voltage.

Next isolate that part of the circuit you want to re-use, for instance V3,V4,V5 and resolder pot side of C9 to grounc point of pot R17. Does that quiet output sufficiently? If not then it's time to sub in some power supplies (filament and B+) Substitute voltages have to be well grounded to circuit ground. But where exactly is circuit ground? Chassis is of course a good ground to some degree, what I mean by that is power transformer eddy currents and ripple currents in chassis make some points on chassis a good ground, others hummy-buzzy ground. If parts of original circuit use chassis as ground wherever convenient that could be a problem or not...and it could be manufacturer went to the trouble to check on that when laying out build. Or they didn't check except to choose a later stage ground point that was suffiently out of phase with earlier stage ground point to effectively more or less cancel cum in final output. So it may be necessary to interupt B+ feeds to any circuit upstream from the output stage so their ripple does not go into chassis.

Circuit ground for sub'd in power supplies should be where V4 and V5 cathode resistors connect to ground or where or where C12 is grounded (connected to chassis), and ideally these will be the same point or only separated by an inch of chassis, cuz eddy currents or ripple current running in chassis between those two point could be a sosurce of hum in output.

Subbing in filaments supply (I think you did that) is easiest to do. If doing so does not fix adequately note the hum/buzz on B+ and output circuit and compare with previous notes. Maybe there was no improvement at all in those measurements....anyway just write it all down for future comparisons. If improvement from filament voltage sub gave no or inadequate improvement then it's time to sub B+, cuz what else is left at this point? If you don't have a good B+ supply for testing can you buy one...even a les expensive switching test bench supply from China will likely (maybe) be adequate for testing. Otherwise and anyway, while listening to output at sufficent level (on speaker you can sacrifice if that happens), and with filaments running on sub'd in dc filament supply, and upstream B+ power rails lifted, turn off the power to unit (but not to filaments). B+ voltage will take a couple secs to decay. You should observe the hum in output is gone for those couple secs if prob is from B+ supply or power transformer eddy currents or chassis ripple current. If that is the case then next step is to try co-locating all circuit ground points still relevant ( so not upstream stuff since B+ lifting previously mentioned)such as C12 grounds and V3,V4 and V5 circuit grounds. Doing so will eliminated as suspects the chassis ripple currents but maybe or maybe not power transformer eddy currents. If co-location does not fix sufficiently then only thing left to check is power transformer eddy current by it's removal and distancing from chassis by 6 inches. Note that if power transformer eddy currents are source of problem that might be revealed merely by removing it's mounting bolts and lifting it away from chassis by a 1/16" to 1/4".

If testing reveals power transformer eddy current problem then one will want to re-bolt power transformer to chassis and pick a chassis co-location ground point well away from power transformer, maybe at least 6 inches but only experimentation will know for sure. The ground of C12 is a good, and more or less mandatory ground co-location point. Having C12 (main filter cap) grounds on an arm of a star grounding design will only rarely work....barely even worth trying.

If you have or aquire a B+ test supply then sub it in and still the recommendations of previous 3 paragraphs hold but you will have more certainty and you won't have to deduce things during the 2 seconds waiting for B+ supply to decay to shutoff.

Bear in mind that even though HP was good test gear it might nothave had really low hum/buzz cuz some hum/buzz ws considered normal in those days and was just accepted....it sould have been hard to find a hi-fi amp that could amplifify signal into a speaker without adding it's own hum/buzz.

The above procedure is a pain in the a** and so time consuming it's hardly worth it but anyone who goes through this will have a new knowldege and skill others lack. This is akin to the torture one must endure to perfect the quad axel in figure skating as shown here:

Pain of perfection

Good luck.

That quad axel struggle was real!
Thank you for this. It helps a lot, every bit of it.
I do have a nice HT power supply unit, restored / calibrated. It is the Precision Apparatus Company B-12, which i purchased from a very trusted individual. So I can do power subbing.
Yes i did replace all caps basically. Didn’t touch a couple small values in Osc but am deleting that anyways.
Yes the C9 input is grounded to its R17 pot ground. Doesn’t change hum. Severe high pitch squeal when nothing is connected to C9 input, but audio is identical when either grounding C9 to its ground or just hooking up a line input from external unit.
Yes i did sub the A/C heater supply with a 6V DC (actually measured 6.3 ironically, unloaded) lantern battery with lots of juice. The unit operated as normal, and no improvement in hum. I don’t think the heater supply is the hum.
I do not have a tube tester but it seems i am on the cusp of realizing i need to have one around since this isn’t the only project i have going on. The same individual above has offered a good price on a restored / calibrated one.
I only have a pair of known “great testing” 6v6’s. They are not guaranteed to be matching. I put them in and the hum was, by ear, identical.

I would like to at some point install an output tube cathode pot and resistor to ground. Basically just create the R48/R47 circuit, which is not in my unit. My unit apparently had “change no 12” implemented, which is displayed below from the end of the service manual. There is no longer a “pair” of cathode resistors to measure and compare.

In all of these suggestions, i was expecting to see one in particular but didn’t….. I understand everything you’re saying about chassis currents from transformer and also C12 negative, and am going to inspect the layout a bit further, but what I’m starting to be most curious about is air-space distance from transformer to the tubes and EMF radiating. It seems stupidly obvious. You know, the kind of EMF noise you hear when you walk over to equipment with electric guitar in hand and bring guitar close to a power transformer. Isn’t this a factor inside tube amps? The power transformer is sitting just a couple inches from one of the output tubes, and maybe a few inches from the 6SN7 driver / inverter tube. You can see this in one of the earlier posts in this thread. But maybe you’re already referring to this and I’m not just not yet familiar with the terminology.

So my reasons for hooking up an external supply are vast, haha.. All of the physical contact noise/hum factors you mention are maybe going to be revealed, but maybe also just plain air-space distance between the transformer and the tubes will be revealed to be an issue. I could even try just disconnecting the factory PT and extending its leads, and sit it somewhere a couple feet away, then bring it closer to see if that’s what’s happening.

It’s really helpful to hear about the ground point for the external supply though. On the 201C, there is only around in inch between the output cathode ground point and the C12 ground point (which is also pin 9 of power transformer). So maybe that’s actually healthy. In any case, i will hook ground from supply to one of those, and they are close by each other as just said. I suppose i can choose one, so maybe that’s the transformer signal ground which is same point as C12 ground, that I’ll go with.

When i do hook up the external supply, I’ll be better able to isolate the audio section from the Osc. I’m concerned about rail voltage rising with the factory power supply, if i disconnect power to Osc section tubes. The PACO B-12 supply is regulated, so i won’t have that concern. If i wasn’t worried about voltage rising, i would already completely removed all Osc tubes and components.

I do have connection questions.. what about the PT’s own chassis ground? It’s mounted directly to 201C chassis. So, in the case of extending the factory PT on longer leads, i should connect the PT chassis to 201C chassis with an extra lead? And in the case of the B-12 regulated supply… The regulated output’s ground is fastened to the front face plate. So i guess in that case I’m just using one ground connection from B-12 for both “pin 9 ground” and chassis-to-chassis ground.

Lastly regarding connection.. these regulated supplies are supposed to…

1: Insert a bipolar +300/-300/ground to the 5Y3?
2: Insert a +350/ground to C12B?
3: Insert a +310/ground to C12C?

This isn’t a lot of work. This is absolutely worth the time because it already sounds musically excellent. It’s already teaching me more about tube layout and i have other projects i will apply this experience to.



IMG_0272.jpeg
 
I used a short lenght of 4 core and screen , more like a starquad , Lt-, LT+,Ht-,Ht+,sheild to chassis,
but make sure its rated for the voltage ,
No need for connectors , just use a cable gland either end ,

To make a great amp from the an existing Hp200AB
I have to concur with the OP on that CJ,
but maybe theres a way to have your cake and eat it ,

Retain all the original circuitry and functions only add the external pre stage and arrange to switch the functions of the other existing stages as required ,

The osscilator itself it worth having as a source of amplitude modulation , the 6bq5 can be switched to opperate as a reverb driver or low impedence passive tone stage driver , maybe even all three at the same time .

I picked up a couple of HP200CD models a few years back , one was Paulo Alto factory , the other made in West Germany
I never even noticed the model AB ,

So it appears we have a challenge, maintaining respect to the original circuit and its designers and the hands that assembled them ,while only adding functionality as a geetar amp ,

So youve used a 100v line to speaker ohmage transformer , that can just as easily be hidden away in the speaker cabinet ,
 
The line to speaker transformer is the Stancor A-3818, listed below. Everything is sounding great with it. 25 watts should certainly handle whatever this 201C can crank out. It all sounds very alive too. I was worried this larger iron might sound dead or stubborn following an amp that’s probably putting out around half that wattage. But it sounds huge.

Back to hum… The easiest thing to try first here is to snip the factory power transformer leads and extend them all with 3 foot 1000V/15A alligator leads ( that have great clamps, not cheap sets). Then place the transformer away to eliminate multiple types of local hum. See how it sounds.

But is having the [C12 negative / output tube ground] and the 5Y3 a few feet from [power transformer pin 9 ground] a terrible idea? I can make earth be ok by hitting AC/earth right to PT, then extend earth to 201C. But I worry that the 3 foot lead between PT ground and the big ground currents on unit.

If that is bad, I may need to hook up the regulated supply. But still am wondering exactly how that works in relation to 201C circuit, ie where it interrupts each power stage.




IMG_0273.jpeg
 
I guess it depends on what you want , a functional compact amp, contained in one box , thats portable , or something that requires a seperate PSU and high voltage cables all over the place .

Removing the mains transformer from the enclosure but then needing long wires to re-connect to the recitifier and filter caps and ground , thats a very bad idea for several reasons .

You said your pretty happy with the amount of hum remaining , so dont try and over think things ,
Balance up the flux in the OP transformer by adding the cathode balance adjust ,
Sure it requires adjustment , but need it not nessesairily involve HV measurements with a multimeter or oscilloscope, Ive found adjusting the cathodes for minimum hum in the speaker output to be highly sufficient .

I'd highly recommend you put a choke in the screen grid supply for the output tubes ,
all the great tube amps do this ,
Many modern tube amps suffice with a resistor , it doesnt make a whole lot of difference at lower volumes , but when you crank it , thats when the choke goes to work ,

Some good shots of the HP200 here ,
https://www.pa3esy.nl/meetinstrumenten/hewlettpackard/html/HP-200AB_set.html

and heres a version with a choke ,
https://people.ohio.edu/postr/bapix/HP200AB.htm
 
Well good news…

Decided to just skip the idea of distancing factory PT from unit and instead hook up tube supply PACO B-12. Noticed it is even less work.
“Cut” output of unit rectifier tube and provided 330 VDC to cap C12B, the first cap that normally follows rectifier. This insert point was my only option to retain the overall rails.
Connected DC ground to C12A/B/C negative.
Disconnected both 6.3VAC lines coming from factory PT and hooked up PACO B-12 6.3V to those.
Connected PACO B-12 chassis earth to 201C earth at same spot mains cable earth gets connected, at the terminal.

The hum is completely eliminated. With input grounded to itself as previously mentioned, all I get is the faintest little bit of high freq buzz, with audio monitoring dangerously cranked. I mean this tiny buzz must be like 80 decibels underneath regular sound. It’s inaudible, showroom quality I suppose.

With instruments connected to input, audio passes sounding great, seems exactly like before. I think that tiny buzz is gone too. It’s just dead silence other than minor guitar noises itself. Works the same with extra transformer and speaker connected too. Sounds phenomenal.

Not only does it sound musically just as amazing as before, it also is now the quietest amp I’ve ever heard. This is as good as it gets. I made a quick recording of a few guitar strums via line out, and will upload later.

I suppose the question is.. what next? The audio circuit isn’t the problem. The power is the problem. Proximity alone?
 
Looking forward to hearing the results ,

If you look closely at the model 200CD you'll see that they have moved the mains , output TX and choke in to seperate compartments of the chassis , there plenty of metal between each iron cored component ,
that minimises any unwanted interaction .
It still wont result in noise as low as you have it now , with an external transformer and regulated PSU ,

https://www.radiomuseum.org/r/hp_audio_oscillator_200cd.html#

If you happen to be testing the unit out open on the bench be aware the hum and noise levels may change when you replace the outer cover ,

You may be able to reduce one step further the noise , by adjusting the cathode balance , for the sake of the addition of the wirewound pot its well worth a try . Id make the adjustment available via a small hole drilled in the outer enclosure for the same reasons I mentioned above .

Id highly reccomend you use REW and an audio interface to look at the noise floor at the output .
that will reveal a much clearer picture of what noise you are dealing with and any adjustments you make can be visually inspected for their effect.

Adding smaller foil bypasses on the 2nd and 3rd PSU filter caps might be worth a try ,
I found that helps reduce fizzle in the noise floor . Sometimes HV electrolytic caps can make fizzling noises while they form up , this usually subsides after a breaking in period .

I included the Paco for clarity ,
whats your ground referencing arrangement on the 6.3v supply rail ?
via the centre tapped winding on the PSU or the pair of 27 ohm resistors in the HP200?




Paco.JPG
 
Last edited:
Ive gotten superb results with two closely matched 100 ohm resistors making the heater /HT- reference at the PSU end , that means any residual noise is nulled as best it can be within the PSU chassis , then the signal amp then doesnt have to bear any of these spurious currents flowing in its metal work .

I would recommend you to try a smaller amount of feedback again for noise reduction and the change in tone it gives , if you can make an arrangement where the FB is applied via a pot it should make a very effective volume control over a certain range , the original circuit applys some 30db of feedback without osscillation ,so a further reduction of upto 30db of noise is at possible at the expence of needing a larger drive voltage at the input .
 
Here’s the audio using external supply. Don’t mind the music’ing, just some strums to listen to sound.
Audio is recorded from the 600 ohm line out of 201C. So it’s got that DI sound. Through the 600:8 transformer and into a Jensen P12Q from the 1950’s, it sounds incredible.

https://www.dropbox.com/scl/fi/om2k...nput.aif?rlkey=i823o7h3qvo0qff23m9egw6r3&dl=0

The first several seconds is a recording with zero input to 201C. Input is grounded to itself. You’d have to turn it up a lot in some DAW software in order to hear the faint high buzz I had mentioned. Note there is like no hum and no white noise. This is the 201C fully active. I am amazed.
A few seconds in, I loudly clap my hands near tubes which you can hear as a high clink sound twice, and then I bang the table hard twice which you can faintly hear as two crunch sounds since whole unit is jostled enough to cause disruptions. Basically just proving to myself that audio was even on.

Around 7 seconds, this is with the external preamp, which has guitar going into it, connected to 201C. This is actually as bad as I can get it to sound because I didn’t have the tube preamp set very wisely, which is a UA Solo 610 and this unit is a bit hummy and noisy itself. So if you crank this guitar recording you will hear some hum and white noise. This is why I included the first 7 seconds, to show what the 201C is sounding like on its own. Also, this is all low volume stuff so that the 201C stays clean. If I crank some other audio source into the 201C, like any clean line level audio other than a relatively noisy preamp and guitar, it’d all pretty devoid of any 201C noise whatsoever. It’s like a hifi guitar.

In another recent thread i started in this forum, someone helped with a variable negative feedback design. It works great. But I am so happy with how quiet this sound is without any NFB, I may not even end up implementing it on this amp.

As for the power connections.. I think I just got lucky on the first try..
I decided to let the 201C’s pair of 6.3VAC resistors to ground be enough, and didn’t provide a center coming from the PACO B-12. I just fed the pair of 6.3VAC lines right to where the factory transformer feeds them. It sounds great, so not sure that would need to change. But I like your idea, so maybe that’s best.

The DC negative / 0 goes to C12 neg, and the PACO chassis earth goes to 201C earth input where mains earth hits it usually.

Obviously the PACO has everything filtered on board, so when I build an external supply I will want to maybe have a filter bank on board supply and also a cap at entry point on 201C, or something like that. Maybe just leave 201C as is how I have it, bring PT and rectifier external, give the rectifier its 40uF cap, and then decide upon some resistor value to put between power rig output and arrival point on 201C which is C12B (formerly the rectifier cap). It’ll drop some voltage, but I believe it can all be worked out.

I have no need for this to be convenient or portable. It’s just for a private home studio. I’m not ever a live musician. So a nicely built external power supply is fine by me. Safety concerns will be addressed whole heartedly though.
 
Ok, i checked out audio file...excellent! Sounds really good. Where did you connect signal...at C9 or the pot feeding it (R28?)? The point where you sub'd in the external ps is same point I would have sub'd.

Building or buying external power supply might be easiest way to go now that you know that works.

As for what might have been the problem: PT magnetic field into vacuum tubes could have been the prob but I've never seen that...many hi-end tube hi-fi amps have the output tubes located right next tot the PT and they're quiet as regards output hum. . PTs put out more field on two sides than the other two...I forget exactly but i think the field is lowest on the laminations sides.

I have often got rid of power supply related hum by moving the PT to it's own chassis and so I never knew for sure if problem was magnetic field radiating into components or eddy currents. I have often tried finding best orientation for power transformer in existing chassius and found that only works well occasionally...usually I can find a new orientation that reduces hum to about half of what it was, sometimes a bit more, but not enough.

Does the PACO power supply have ripple voltage on output? And if it does is it same more or less as ripple voltage of 201C..cuz if it is that suggest that in 201c, from regulator on the power supply circuit is not a problem and only the PT need be moved to it's own chassis. Of course if PACO is very low or no ripple supply then 201C rectifier circuit is still a suspect.

So if next you try remoting the power transformer and keeping the rectifier circuit the same then you'll know...rectifier circuit may cause ripple currents in chassis. With PT remoted, and if there is hum on output then it will be easier to try solution of single ground point, like on rectifier filter can tab.

Good that you have made that audio sample...it is now a good reference point to refer to when you make a solution of one kind or the other cuz it sometimes happens that things sound better with one hook up than another for who knows what reason.

So when it's all done and sounding great you have done this: Winning
 
Ok, i checked out audio file...excellent! Sounds really good. Where did you connect signal...at C9 or the pot feeding it (R28?)? The point where you sub'd in the external ps is same point I would have sub'd.

Building or buying external power supply might be easiest way to go now that you know that works.

As for what might have been the problem: PT magnetic field into vacuum tubes could have been the prob but I've never seen that...many hi-end tube hi-fi amps have the output tubes located right next tot the PT and they're quiet as regards output hum. . PTs put out more field on two sides than the other two...I forget exactly but i think the field is lowest on the laminations sides.

I have often got rid of power supply related hum by moving the PT to it's own chassis and so I never knew for sure if problem was magnetic field radiating into components or eddy currents. I have often tried finding best orientation for power transformer in existing chassius and found that only works well occasionally...usually I can find a new orientation that reduces hum to about half of what it was, sometimes a bit more, but not enough.

Does the PACO power supply have ripple voltage on output? And if it does is it same more or less as ripple voltage of 201C..cuz if it is that suggest that in 201c, from regulator on the power supply circuit is not a problem and only the PT need be moved to it's own chassis. Of course if PACO is very low or no ripple supply then 201C rectifier circuit is still a suspect.

So if next you try remoting the power transformer and keeping the rectifier circuit the same then you'll know...rectifier circuit may cause ripple currents in chassis. With PT remoted, and if there is hum on output then it will be easier to try solution of single ground point, like on rectifier filter can tab.

Good that you have made that audio sample...it is now a good reference point to refer to when you make a solution of one kind or the other cuz it sometimes happens that things sound better with one hook up than another for who knows what reason.

So when it's all done and sounding great you have done this: Winning

I’ve been skipping R28 because it’s a low value so I’ve been concerned about loading down whatever circuits that I’m feeding into it. So I’m just going straight into C9. I’m using the ground of R28 though, since it was clearly intended to feed audio to C9. Sounding great. But when i do implement a preamp into this unit, I’ll probably put in a different value pot there and use that as a second stage pot for preamp output to C9. 1st stage pot will be in middle of preamp stages.

I’m going to build an external supply using a 4”x6”x1.5” aluminum chassis. Will put the power transformer and rectifier and cap(s) in it. It’ll also of course receive earth. Then this earth will be brought to the 201C chassis along with the other leads. This effectively replaces the PACO B-12. I have some things to decide about what caps to keep/put where though.

I did measure a little bit of AC on the B-12 high rail output, however i used a DMM not a scope. I have a nice analog multimeter i will hook up soon to get a steadier read on things, but the DMM shows pretty low AC ripple at idle (when no audio is playing/happening) of around 10-15mV, and then when audio is passing the ripple wanders up and down around 30-150mV.

I suppose the worst case scenario is after i build the hook up the separated factory supply that I realize it’s still causing hum. That would be a bummer haha. But I really prefer making this power external no matter what. In the HP service manual, there is a statement in the instruction that clearly indicates that HP knew that there is always going to be hum coming from the power supply…

“””
3-8. LOWLEVELAPPLICATIONS.
3-9. To avoid excessive noise and hum at low output levels ( -40 dBm or less), it it good practice to attenuate the oscillator output -30 or -40 dB and use the amplitude control as an output vernier to obtain desired level.
“””

So, after all of this experimentation, I’m more and more getting the feeling that HP knew this power location was going to cause noise and that it would just be overcome by the super loud oscillator and smart gain staging. The idea of giving this output circuit the quietness it deserves, not what HP knew it was sacrificing, is pretty fun. I hope it works out.


Go Yuzuru Hanyu!!
 
I’ve been thinking more about how to substitute this nice PACO B-12 power supply by externalizing the 201C’s factory power circuit ( PT, rectifier, cap #1).
Knowing that the max limit the rectifier cap can be is 40uF, and knowing that there is benefit to having capacitance at both the power unit side of umbilical and the circuit side of umbilical… maybe a compromise would be to just divide the capacitor? I could put 20uF by rectifier in supply unit, have my 3 foot umbilical, and have another 20uF on main unit circuit which connects locally to output transformer. In parallel that’s 40uF.
Anybody ever try something like this?
 
I brought the factory supply out and used 3 foot long leads for earth, HV DC/ground, and 6.3/6.3.
I decided, for now, to keep things as low risk as possible and hook up a new first stage 40uF right at the rectifier which is with the transformer. Then, i snipped the local 40uF on the main unit that was previously this cap. And that’s where the DC gone in, after that deleted cap. So the circuit overall is the same as factory, but there is now a 3 foot lead of 340 volts between the rectifier cap and the main unit.

The deeper rumble, which I refer to as the “mains hum” because it sounds like the sound you hear when you bring any sensitive gear/pickups near a power transformer, has totally vanished as far as i can tell. So that’s great news, it’s really gone. This has to be a direct correlation with either the increased air space between transformer and audio circuit, or PT chassis earth connectivity distance from the main unit chassis.

There is still a 120Hz hum though, probably the same level as 240uF had accomplished before, but which i can now hear because the “mains hum” is gone. The 120Hz hum behaves similarly to before when incrementally adding things. But now i can audibly CONFIRM that going from 240uF to 340uF does nothing. The 120Hz remains the same. Before i could hear this because the mains hum covered it up. So i couldn’t tell if going to 340uF helped or not. But it doesn’t.

So what that means is that the remaining 120Hz hum is probably from PRIOR to this second rail. And if it’s on the first rail, it will be sent right to the output transformer primary center tap. This makes me want to attempt to clean up the first rail….

But unfortunately I’ll be messing with the rectifier > output transformer relationship if i add something between it. Is it possible to add a very low value resistor in series between the rectifier out/cap and the main unit entry point which goes right to the output transformer? Asking because maybe I could create a R/C filter directly after the first 40uF rectifier cap.

Anyways that’s where I’m at. It is an improvement over factory, using only factory parts, because the mains hum is gone. I haven’t tried things in a musical context yet but it’s probably a pretty decent hum level. But it’s nothing compared to the PACO B-12 providing the power. I’d love to just mimic that thing’s circuit, but the circuit is very complex. It seems it is buffered and/or regulated by a pair of 6L6’s. I suppose a last resort is I could buy a smaller regulated tube power supply and dedicate it to this amplifier. But i would MUCH prefer to create a healthier DC situation using this factory HP supply.

Also, i tried 3 different 5Y3’s now. All perform exactly the same and come from working environments. I also tried a 5AR4 since it’s listed on chassis and in schematic, but that thing created a much much louder and new kind of buzz, and also boosted the first rail to 390V instead of 340V.
 
You need to be able to adjust the standing current across transformer half primaries ,
that will make sure the 120hz component nulls better ,
Throwing resistance/capacitance filters at it will reduce the noise a bit , but its much more productive to tackle the unbalance at this stage .

A choke in the screen supply can also help reduce noise , it will , in combination with what ever value of capacitor you use after it ,form a much more effective filter than the RC screen grid arrangement you have now .
 
Last edited:
You need to be able to adjust the standing current across transformer half primaries ,
that will make sure the 120hz component nulls better ,
Throwing resistance/capacitance filters at it will reduce the noise a bit , but its much more productive to tackle the unbalance at this stage .

A choke in the screen supply can also help reduce noise , it will , in combination with what ever value of capacitor you use after it ,form a much more effective filter than the RC screen grid arrangement you have now .
So, balancing meaning installing the pot R48, and its R47 to ground, which were deleted by HP, so that the cathode biases of output tubes can be balanced so as to phase cancel the 120Hz? I might do that. I wonder what their reason for deleting it was though, and installing a fixed resistance to ground that is shared by both cathodes. Would the sound change much by putting the balancing pot back in?
I experimented with a couple RC filter changes and yes the hum can be pretty well contained. Will probably try a few more combinations and see what works best to juggle voltage drop and hum reduction. My concern was that the first rail having resistors before it wouid be a big issue for the output transformer, but checking around other amp designs it seems somewhat common to have an RC filter, or even two, prior to the first rail.

I’d like to learn more about the choke option though. What would you suggest where?

The interesting thing in all of this is just appreciating how good of a high voltage DC signal is provided by the PACO B-12 supply. I guess there’s no need to filter much if the DC going in is nearly perfect already. When I hooked up that supply, I didn’t even need to use the high 240uF at the second stage. Everything was just totally quiet. Would be nice to be able to achieve that with new filtering. Maybe chokes are the way?
 
Tubes change slightly over the course of their life ,
Having it adjustable means it requires re-setting periodically ,
if its not set correctly theres a chance you could end up with more unbalance than the simple unadjustable scheme .
Fortunately is very easy to adjust by ear for minimum hum , so it doesnt have to be the big bother its made out to be .
HP's trimming procedure for the variable cathodes probably serves a different purpose entirely ,
it will be used to match the standing current dynamically , in otherwords while the unit is driving a signal ,
In your case you only need to match the currents so the 120hz better cancels in the output transformer ,without the presence of a drive signal .

I'd be tempted to try a small choke in the screen grid feed ,
the choke will have some internal DC resistance maybe a couple of hundred ohms ,
subtract the dc ohms of the choke from 800 , replace the 800 ohms resistor with that value , place the choke with another capacitor following it after the RC filter ,

Lets say the choke is 250 ohms , 800-250=550 ,

You've moved the mains transformer , rectifier and first filter cap off to a remote chassis ?
That leaves you with plenty of real estate on the amp chassis for a choke or two,

Just like you can an add an extra RC filter to the supply before the output transformer centre tap , likewise you can do an LC filter , thats the arrangement they ended up with on the HP200CD.

Check out the hammond 15x series chokes and try choosing a suitable component ,
remember to consider size ,current handling and inductance value .

Ive found with tube circuits with on board transformer based supplies -90db hum is around the best you can expect , with a remote PSU figures of -110 to -120db are possible ,
 
Back
Top