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.
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 ,
Got it. Thank you, this all makes sense and is VERY helpful!
I suppose I’ll just go ahead and install the balance pot circuit that HP deleted. I’d be very curious to adjust for hum and see how it goes. Will a larger value pot have any side effects? Change the amp output sound much? If not, it would be good to have a larger value to allow for more hum cancelling adjustment, or at least that’s how i interpret it which could be incorrect.
As for the turning the single second filter (for screen grids) into a dual R/C > L/C filter in series, while retaining a total of 800 ohms to retain same voltage drop… what is the sonic advantage to using the inductor in second half vs a resistor? Inductor is just better at filtering?
I’ll also look into choosing values for creating a filter like the 200CD has… an L/C for before the very first rail. I’d love to nip this rail in the bud and get the cleanest DC possible to enter the unit.
 
I am very curious about cleaning up the power supply before first rail. I know that the output tube balancing adjustment should help and that it’s a standard solution for hum, but it would also be nice to have a power supply that is sending out great DC. Hooking up the PACO B-12 was inspiring. I would like to be able to use this custom power supply, built from the HP transformer and rectifier, for more than one project.

I looked into the Hammond 15x stuff and found something that might be really effective, which is the 159S.
https://www.mouser.com/ProductDetail/546-159S
Yeah, it’s 33 USD for a choke part, and it’s almost as large as the PT itself. But I’ve only spent about 60 USD on this project so far, and it sounds like a vintage amp that would cost thousands.

So using this calculator I entered 4 Henries, 100uF cap, 100mA (because the PACO B-12 shows that the unit draws a steady 100mA during idle), and 16.25 ohms per/Henry (because the 159S has a total impedance of 65 ohms), and the result is a ripple attenuation of 47.1dB, and voltage drop of only 6.5V. Sounds like a pretty great starting point before hitting any rail.

What’s the limit on capacitance at this spot? The first cap after rectifier would remain, and be 40uF. So the circuit would be 5Y3, 40uF to ground, 4H in series, then 100uF to ground. Is a 4H inductor enough separation from first cap to have a large cap, say 200uF, be after inductor?

I’m assuming that having a large cap after the inductor will offer a stiffer outer power performance, i.e. less sag. Is that how it works? Or does having an inductor involved mess with things and greatly change the whole performance of this output amp?
 
RCRCRC filters can be extremely effective, much more so than a single RC filter with the same voltage drop. For example three stages each of 1K and 100uF are a lot more effective that a single stage of 3K and 300uF. A gut called Scroggie worked this all out aeons ago.



Cheers

Ian
 
RCRCRC filters can be extremely effective, much more so than a single RC filter with the same voltage drop. For example three stages each of 1K and 100uF are a lot more effective that a single stage of 3K and 300uF. A gut called Scroggie worked this all out aeons ago.



Cheers

Ian

Excellent.
My concerns at this point are twofold:
1, the impact of whatever filtering method i use, on the power draw / availability I.e. the “sound” of the amp during louder music.
2, the health of the 1st cap / rectifier relationship. Say I put a RC filter after 1st 40uF cap. Say the R is 125 Ohms 10 Watts, and the C is determinable still. What is my limit? Can I go up to 240uF after only 125 ohms of barrier from the 1st 40uF cap? Or should I keep said RC cap lower? Like 100uF? Or just 40uF like 1st cap? And what is the reasoning behind this limitation?

I hooked up this exact setup and it already sounds quite good. 5Y4, 40uF to GND, 125 R, 240uF to ground. This provides main rail. Then rest is factory.. 800 R into 40uF which is screen filter.

With this filtering there is no need to increase the screen grid cap value like I had before because the original DC rail is already much cleaner.
 
LC filtering on the main HT supply is almost never seen in guitar amps ,
as you see it requires a choke as large as a the mains transformer ,
so the weight would be prohibitive ,
You do see it on some high end hifi amps though , the Williamson amplifier for instance , it also has excellent variable cathode bias scheme , which not only provides a means of balance but also to adjust the over all standing current , doing these things serves to reduce transformer distortions at low and moderate volumes , your amp will still distort when you push it , just a little sweeter than before .

If you want to hear the sound of a tube amp with no choke try any modern Fender or Marshall amp , they have a resistive screen feed ,
When you drive the output stage it tends to clip a bit abruptly , its very clean right up to the point of clipping , but it overloads quite suddenly after that ,
In contrast an amp with a choke in the screen supply can supply a lot more punch on the peaks , were talking well into double digit distortion , never the less, if you do an impulse test (into a load resistor), the momentary peak power from the circuit using a choke is higher , and the overload characteristic changes for the better sonically .

Again theres no reason to try and rethink the masters work ,
just use the same cathode bias scheme HP came up with ,
You'll need a wirewound pot , I usually go with a 10 turn Spectrol 3W which allows ultra fine adjustment .

You probably wont need as large a cap after a choke to get adequate filtering ,
 
Well here is the schematic, all stitched together, and with the current mods as of this date.
That’s very interesting about the possibility of making the distortion even more natural and gradual than it already is right now. Without NFB, it’s pretty nice. And WITH NFB, it isn’t so nice. It is clean forever and then bam it’s a clipping point.
So this 201C amp is already better than that. But if it can be made to have even a smoother progression from clean to distortion, I’m all for it. So you’re saying just replacing R31 with an inductor + resistor that then totals 800 ohms? And keeping C12C there, and not adding another cap?
I’m planning to try a double R/C filter between power and 1st rail, instead of the single R/C filter. But i am concerned about the resistors being such low values like 50R each or something. But that’s what would be required to avoid too much voltage drop.

View attachment HP 201C Amp Mods.png
 
Last edited:
I know I'm late to the game, but did you check the orientation of the power and output transformers? These should be at 90° to each other (laminations). Also sometimes the covers get removed and forgotten, this will let the magnetic fields radiate farther (then weird science emerges). (Also, my old Marshall circa 1970 did have a choke. As did my old Twin Reverb from the same era.) (And why does that HP schematic have the full-wave-rectifier diodes reversed?)
 
Hi!

I wonder how it sounds connecting the speaker to pins 1 and 5 of output transformer....I'd probably bypass the output attenuator. Choke for power supply filtering is very effective ...I recommend putting it first, in place of the 125 ohm resistor. and leave R31 as is. Maybe need a low value resistor in front of or just after the choke...like 10 ohms...just have to try and give it a listen.

You can still use PACO power supply to investigate....for instance, power center tap of output transformer from PACO to see what if any hum is picked up in circuit from tube V4 or the +310 rail. Who knows but maybe the balance pot on output tube cathodes is totally cancelling the power supply ripply passing through output transformer primary....if so you can use a smaller choke and put it in place of R31.

DC filaments help get rid of/reduce hum sometimes as can a filament supply center tap. The HP201C is using a pseudo center tap for filaments. Some tubes have a failure mode in which they pickup more hum from ac filament than they did when they were new....I'm not sure what broke but I run into it frequently enough that I always check by subbing in a new tube (I'm referring to V4).

Do you have a variac? I always use a variac to bring up voltage slowly after I've made changes in the power supply. I use a smallish fuse in the variac....3 amps for testing line level stuff and 5 or 6 amps for power amps. I use fast blo for best protection since I'm bringing up the variac voltage slowly...so don't need slo blo fuse.

Lastly...the recording you posted sounds good...be careful not to lose that sound. I make recordings at various points of a mod making sure to keep levels the same for easy comparison later. I just record with my focusrite scarlett 2i2 cuz it's good enough....and I can afford to replace it if I blow it up which I have done. I keep a Hum Eliminator in front of the Scarlett except when I really need to have a the "instrument" impedance (a mere 47kohm) but even then I have a film capacitor in front of that...I think it's 10.0uf 600v....physically big but worth the protection. I have at times made recordig of the B+ rails. I have 100kohms on the output side of that capacitor and make the connection to the B+ rail before connecting to the Scarlett to protect it from the high voltage pulse that will come though the capacitor upon connection to the B+ rail. Note on the Hum Eliminator: good frequency response out to 50khz but low frequencies saturate is easily. Have to keep levels to -10db if I'm looking at frequency below 50 hz. I plan to buy a better transformer for this front end application. Non-saturation up to 15db should be findable.
 
That’s a pretty great idea, about tossing a rail signal into an interface to record it and have a recording of the ripple/noise. With those protections of course. Yeah similarly I use my cheapest interface, a UA Volt1. Sounds terrific anyways.
Yeah I’ve been studiously taking notes of things as I record them, making notes in DAW for each recording about its power configuration. Mostly for hum performance as i change things. I also play music through the amp just to make sure it sounds as expected still. But I’ve focused on hum and compared all recorded options.. The schematic above, with external factory supply, is about equal in noise performance as when the PACO is hooked up. They are different noises though.. PACO has zero 120Hz hum but has a high buzz that sounds like grounding. The schematic above has a tiny about of 120Hz hum that sits at same volume as PACO’s buzz. So if you recall when I said in previous posts that PACO is dead silent during music playing, ie normal levels, well I can now say the same about this 40uF, 125R, 240uF external supply setup. Like the PACO, It is effectively so quiet that other things will have a higher noise floor, such as the old ribbon mic that I use for recording amps. I’m at a point where I’m splitting hairs now. The noise is kinda made irrelevant.

So, if the schematic is healthy for current draw through the 5Y3, particularly during power up inrush, then I might be done here sonically speaking. But I WOULD like to try dividing up the RC into two stages that retains the same voltage drop but lowers the 120Hz even more.. that would be the normal 1st cap 40uF, then a 62.5 R / 120 uF, and another 62.5 R / 120uF. (Obviously just a close practical value would be the R.) But I’m concerned about then only having 62.5 ohms between 1st 40uF and a 120uF.
So maybe I need to run some design software to check for inrush. Or maybe I can install a second ON switch for between 40uF and rest of circuit after.

As for chokes.. my concern there is changing the musical performance of the amp. Although, I like the idea mentioned a few posts ago about make the screen filter a choke rail so as to get a more natural progression of distortion. Anyways, with R/C stuff, both the original design and these newer R/C’s, I absolutely love the sag that the amp has during hard hits and loud musical moments. It is maybe the best part of the amp. It’s very noticeable and very awesome. It would be a shame to go through a bunch of hoops only to find out that adding chokes stiffens the amp too much, I guess by way of better regulating the rails.
 
Its as simple as this , ALL the great tube amps have choke fed screens ,Vox AC30 , Fender Bassman , Marshall , Hiwatt ,Boogie
Messr's H/P werent worried about hum or how the circuit behaved under overload ,
all they were concerned with was to output a continuous sine wave of defined amplitude ,as stable as possible ,

Definately do sepperate output and pre stage HT filters into two seperate lines !
You should bite back against noise a bit more ,
 
Obviously later H/P in the 200CD needed the entire HT filtered by CLC pre output transformer centre tap to get to the new cutting edge spec .
 
It looks like the Hammond 194B is the same choke as used in screen grid of the Bassman. I’ve always loved the breakup sound of a bassman. At 22 bucks, maybe it’s worth experimenting.

I ran some tests with the PSUD2 trying a few different R/C arrangements. It was very helpful to see exactly what the problem is with adding large caps at this stage, which is that that 5Y3’s spec’d 0.2 seconds of transient high current is violated and it ends up being around 0.3-0.5 seconds depending on R/C values. However all R/C configs I tried DO end up well below the 400mA operating current limit. So that’s good. But the only way to keep the startup inrush quick is to have smaller caps. Those smaller caps don’t filter very much ripple, but ENOUGH ripple for the original commercial purpose of the device.
So if I run the R/C filter that I have going, which is 40uF 1st cap then 125R/240uF, before the main rail, the inrush takes about 0.4 seconds go down under 400ma. The software gives the alert.
So I could run it like this and basically I think I’ll just burn through 5Y3’s a bit faster than usual. Not the end of the world maybe.

But back to the choke on screen grids.. it’s intriguing for the distortion so might give it a shot. I would love to have a gradual change of overdrive instead of a wall of change. Having no NFB is already in that direction!
 
Glad you got use of Duncans PSUD ,
its a great little program , no real learning curve , just get stuck in and it works ,

Bear in mind your running a pair of 6V6's not 6L6's like the Bassman , so perhaps a slightly lower current choke will do , that might end up in a smaller size ,perhaps higher resistance component .
You might need to experiment to find a good quiet spot to place your choke ,relative to the output transformer , but you have plenty of real estate with the mains transformer gone ,
If you drive the output stage with a signal then attach the scope to the choke ,it forms a kind of search coil , you'll be able to find the optimum spot and orientation for minimum interaction .

Your observations on feedback are good , it tightens up the sound at lower volumes , but doesnt break up as nicely .
 
Here is a recording to demonstrate bringing the factory external, and also with an R/C filter for 120Hz. It starts with the original hum for reference.

https://www.dropbox.com/scl/fi/b9l3...LY.1.aif?rlkey=fr0tm1ih80y6f91rn12p9jdmh&dl=0

0-5 seconds: the unit as it was after just a recap, as the beginning of this thread. 201C input grounded. Clear hums at low level. This is both mains hum and 120Hz hum.
5-15 seconds: unit as it is now… factory PT/5Y3/40uF brought a few feet away to kill the mains hum, and with a 125R/240uF R/C filter before first rail to reduce the 120Hz hum. Input grounded. Like the 1st audio demo, i clapped my hands loudly right next to tubes a few times, then banged on the workbench a couple times. Just wanted to have evidence that there is audio.
15-20 seconds: UA 610 Solo is connected to input, with the 610’s output volume all the down. Pretty noisy unit even without any volume coming from its own circuit! Might someday address this, maybe by bringing power external haha.
20 seconds: guitar playing. In between phrases, all you’re really hearing is the noise of the 610 underneath it all. The small 120Hz hum in the 201C is a moot point when the 610 is connected.

Note, one of the best things about this setup is that there is such a low amount of guitar buzz itself, no matter which way you turn while playing. I assume the transformer on the output of the 610 is helping in this regard. For this reason I’m planning to put a transformer on the output (plate to line) of my preamp section that I build. We’ll see how that goes.

This more or less completes the goal of this thread. Thank you everyone for your help! I’ve learned a LOT from this, and feel more comfortable attacking other related issues in other equipment now, in particular in basic power supply design.

I may end up improve it one more time by putting in a double R/C since in running PSUD tests it seems that my current inrush would barely get worse than the current single R/C.

The intriguing thing that I’m trying to better understand is that perhaps the “alerts” in the PSUD are too rigid and my current inrush isn’t even a problem for the 5Y3 no matter if i use the single R/C or the double R/C….

The datasheet says the plates can handle up to 2 amps for 0.2 seconds. It also says the plates can handle up to 400mA operating current. I could be wrong, but i believe that the PSUD applies this as a hard corner that can’t be touched…. You need to get below 400mA at 0.2 seconds. It does in fact accomplish this in the factory design, thus no alert goes off. But both of my new versions trigger the alert.

But my inrush doesn’t ever reach 2amps, first of all. And second of all, it gets pretty close to being under 400mA by the 0.2 second mark. In both, it takes until around 0.3 seconds to get below 400mA. So, it’s not like I’m asking the 5Y3 to handle 2AMPS for 0.3 seconds. Rather I’m asking it to handle like 500mA, until 0.3 seconds. Or something like that. So I don’t think I’m really taxing the living daylights out of the 5Y3 during startup. And I’m certainly not over taxing it during operating state.

Attached are the three versions below. This is mainly to show the current inrush. The DC voltages aren’t perfect since it’s tough to recreate a whole functional estimate of the 201C unit/load. But if i were to zoom in, you’d see the main rail DC waveforms clean up successively. The 1st DC is a whopping 4000mV, so 4 VOLTS, peak to peak. This is what HP sent in to the 201C as its main rail, and is what the output transformer was seeing. Then, the single R/C filter brings it to around 200mV, which is (5-15 seconds in audio example above) which I could live with. Then the dual R/C filter brings it to around 100mV.

———————

1st is factory
2nd is single filter 125R/240uF (5-15 seconds in audio)
3rd is double filter 68R/120uF > 68R/120uF.

IMG_0303.jpeg

IMG_0304.jpeg

IMG_0305.jpeg
 
Last edited:
Back
Top