Earthing and lead dress

[Mark]

I’m looking at an amp with noise issues. The noise is odd in that it sounds like mains hum but has a high frequency component of 170Khz to it.

The chief problem is it’s quite noisy on cathode bias mode, and it has been this way from manufacture, I have noticed that the noise does get quieter as the amp heats up which is strange.

IMG_8918.jpeg

As shown in the picture, I have added a 270pF cap across the plates of the P.I. stage and now both cathode bias and fixed bias have the same level of noise. The 270pF cap has addressed the high frequency component of the noise. The noise isn’t terrible but I think it’s louder than it should be. There was a wire for the external signal out (the signal comes from the input of the P.I. stage) which was running along side the heaters. I removed this wire and I will probably replace it with shielded cable.

I’m finding the earthing of the amp somewhat unusual.
The earth for the output jacks, the earth for the external signal out (signal from input to the P.I.) and the output transformer secondary are earthed at one star point. I don’t know if this will cause much drama.

IMG_8916.jpeg

The preamp section earthing is of some concern to me as the preamp earths are shared with power supply earths, and I have been told this isn’t a good practice.

IMG_8921.jpeg

IMG_8922.jpeg

As shown in the picture the preamp earth also has the preamp and P.I. filter cap earth, centre tap for the heaters, centre tap of the power transformer secondary, an earth for the bias supply, earth for the main supply 50uF+50uF caps, earth for the cathode bias cap and resistor.

It seems like a lot of stuff for one earth point especially when the preamp is earthed at that point.

Do you think I should separate the earths and where on the chassis is the best place for the new earth connection?

Thanks again for your assistance.

Getting my questions in before the new year.

[Raja_Kentut]

Hi Mark,

for a proven ground and earth wiring have a look here :

https://ampgarage.com/forum/download/fi ... &mode=view

All black wires are ground wires. There are three ground points :
1) left side the ground next to the stacked capacitors. All grounds of the power supply capacitors are connected here. This point is NOT connected to earth (chassis)

2) center of chassis below the large tube socket. All grounds of the powerstage tubes and PI are connected here. This point is NOT connected to earth (chassis)

3) right side below the two capacitors (which are the anodevoltage capacitors for the preamplifier stage) To this point all ground signal of the input signals are connected. Also the grounds of the input jacks. This point is NOT connected to earth (chassis)

From 1, 2, goes a wire to 3

The connection of ground and earth is next to the small black capacitor where you see „RÖ1“. The bolt that holds the contact strip is connecting ground to chassis. That is 3)

When you follow the thin black wires, you can see how the ground wires are laid. Next to the ground wires there are always thick red wires as close as possible to the black wires, they carry the anodevoltages for the amplifierstages. So always one black and one red to each stage. That is to minimize the loop area between forward- and return lines.

Sometimes you have a bad tube that causes oszillations. I had lastly one build where a 31kHz signal came out of the powerstage. I tried everything to fix that, no result. After changing one of the EL84 tubes everything was fine…

[Stevem]

How are you hearing 170Khz as you posted?

[TUBEDUDE]

I think mosquitoes hear in that range.

[bepone]

this looks some ceriatone or chinese amp kit.
of course speaker jack cannot be grounded to the chassis, but manufacturer should know where belongs that gnd point, not sell everything and after read solution on ampgarage

[bepone]

I think mosquitoes hear in that range.

artifacts are everywhere on hearing frequencies, it is not 1 oscillation freq. , it is summ and substract on kx main freq, k=0- any number

[Mark]

How are you hearing 170Khz as you posted?

“ The noise is odd in that it sounds like mains hum but has a high frequency component of 170Khz to it.”

I’m hearing mains hum and I see it on my the 170Khz on my CRO.

IMG_8912.jpeg

[Mark]

I think mosquitoes hear in that range.

It is now known that mosquitoes can hear sounds as far away as 10 meters (32 feet). They hear best in the frequency range between 150 to 500 hertz, which overlaps well with the frequencies of female mosquitoes in flight.
The mosquitoes’ frequency range for hearing also overlapps with human speech. The most energetic frequencies of an average human vowel are in the range of 150 to 900 hertz, so they technically should be able to hear people speaking. However, there is currently no evidence that they use this to locate and hone in on people. It is well known that mosquitos pick up sensory cues such as carbon dioxide, odors and warmth to locate people.

https://asknature.org/strategy/long-ran ... n-eardrum/#

[Mark]

Hi Mark,

for a proven ground and earth wiring have a look here :

https://ampgarage.com/forum/download/fi ... &mode=view

All black wires are ground wires. There are three ground points :
1) left side the ground next to the stacked capacitors. All grounds of the power supply capacitors are connected here. This point is NOT connected to earth (chassis)

2) center of chassis below the large tube socket. All grounds of the powerstage tubes and PI are connected here. This point is NOT connected to earth (chassis)

3) right side below the two capacitors (which are the anodevoltage capacitors for the preamplifier stage) To this point all ground signal of the input signals are connected. Also the grounds of the input jacks. This point is NOT connected to earth (chassis)

From 1, 2, goes a wire to 3

The connection of ground and earth is next to the small black capacitor where you see „RÖ1“. The bolt that holds the contact strip is connecting ground to chassis. That is 3)

When you follow the thin black wires, you can see how the ground wires are laid. Next to the ground wires there are always thick red wires as close as possible to the black wires, they carry the anodevoltages for the amplifierstages. So always one black and one red to each stage. That is to minimize the loop area between forward- and return lines.

Sometimes you have a bad tube that causes oszillations. I had lastly one build where a 31kHz signal came out of the powerstage. I tried everything to fix that, no result. After changing one of the EL84 tubes everything was fine…

I will check that out. Thank you for taking time to explain this to me.

[Mark]

I tried moving the cathode bias earth and the various centre tap earths to no avail.

Earthing isn’t my strong point generally and I find these type of faults hard to resolve.

IMG_8924.jpeg

[Mark]

I tried isolating the preamp/P.I. stage earth connections from the output stage earths and I didn’t hear any significant difference. I put a 270pF silver mica cap across the plates of the P.I. stage and that made a bit of difference, but nothing significant. I shorted out the 470K grid stopper to the P.I. stage as it would enable me to earth the input to the grid of the P.I. stage, I had dead silence when the input is earthed, when I turned the volume up the hum/noise reared its head, thus the issue has to be with the input to the P.I. stage (pin 1 of the 6SL7 valve.)
The other changes might have had an effect as I thought I tried earthing the input to the P.I. grid before, I could be wrong though.

IMG_8935.jpeg

IMG_8936.jpeg

[R.G.]

The amp is oscillating. Although it's possible that both are connected, the RF is likely to be a different problem than the hum.

Grounding ("earthing" - I'm in the USA, and it's hard to change terms, as you know from the opposite direction ) is easiest to understand by thinking in terms of all ground/earthing wires being resistors, not zero-resistance. Any current flowing in a resistor/ground wire causes a voltage across the wire/resistor. If you mistakenly connect some signal amplifying stage to the end of a wire carrying heavy currents, then the stage will amplify the voltage created by the current in the ground wire. The practice of grounding/earthing for low noise is to force currents to flow through their own ground wire, not a ground wire shared with some other amplifying stage.

The general idea is that the bigger the current, the more you need to return that current back to where it started from, and on its own ground wire.
Case #1: rectifier current return wire; connect the negative-side return to the power transformer ONLY to the negative side of the first filter cap, and have no other ground wire share any part of the wire going back to the transformer. The rectifier return wire carries the biggest current pulses in the amplifier, and sharing any part of its return/ground wire creates a hum that can't be gotten rid of any other way than proper wiring.
Case #2: output tube cathode current return "ground"; the cathode currents need their own wire to return to the negative side of the first filter cap. These wires carry a full-wave-rectified replica of the audio in current form, and sharing any part of this ground wire/resistor feeds this signal back into the stage that shares the ground wire. Connecting earlier amplifying stages to this ground wire for their own grounding/current return amounts to feedback, and this, by itself, can cause oscillation and hum.
Case #3: PI and speaker return; the OT secondary carries high currents to/from the speakers. For lowest noise, the speaker currents need to flow back to the OT secondary, and not through a shared grounding wire or the chassis. Nearly every amp uses a phone jack that's connected to the chassis for this purpose. Usually you get away with this, but sometimes you don't and you get hum and oscillation. The PI return ground should be connected to the negative terminal of the first filter cap on its own ground wire. Sharing the cathode ground of the output tubes means the PI is being fed a signal of the full-wave-rectified-audio currents times the output stage's ground wire resistance. Likewise, if the speaker return current shares a wire or path across the chassis with the PI ground current, the PI is being fed a "ground" signal from the speaker currents.

These are just the high points - and noted in order of high currents through ground wires. If you're thinking that this means that every single ground point ought to have its own ground wire, then, yes, you're right. That is "star grounding". No one has the patience or enough wire to do the full job. Nearly the full effect can be had by clustering all the grounds to one closely-connected bunch of stuff and using a separate wire for that cluster. Clustering the grounds to one preamp tube is an example that nearly always works well. The PI tube is another.

Star grounding can be shown to be hum and feedback free before wiring. Every other ground/earth wiring technique uses some form of intense thought or luck in combining ground currents that semi-cancel or cause issues below the detectable audio level.

[bepone]

I would never understand why the people use chassis for the electrical points and connections, current passing.. it is the worst option, full of resistance and possible oxides with time.
People should start to think about chassis is from plastic and never use it for audio, there will be 100 problems less.

Proper thinking is explained up, from @raja, insulated gnd points on capacitors negative terminals supplying dedicated stage, ie- daisy chain of local star grounding, and is not possible to have currents mixed like in chassis gnd option with random connecting like seen on many amplifiers, telling me that amp builders doesnt have a clue about what they are doing, just a c/p one from the other

[Mark]

Hi RG

Thanks for your reply.

To ensure I understand what you are saying, this is how I read your reply.

1/ The centre tap of HT winding is to be grounded separately with the earth connection of the first filter cap.

2/ The ground for the output valves is also to be connected to the HT centre tap and first filter cap.

3/Here is where I get a little confused. One side of the output transformer needs to be connected to the speaker ground/common, and not to ground itself.

The PI return ground should be connected to the negative terminal of the first filter cap on its own ground wire. Sharing the cathode ground of the output tubes means the PI is being fed a signal of the full-wave-rectified-audio currents times the output stage's ground wire resistance.

I read this as the P.I. should be grounded with the first filter cap which is with the HT centre tap and the Output stage cathode earth. I would have thought you meant to the negative side of the filter cap smoothing the P.I. supply?

The other thing is the chassis is rather small. Do you have suggestions as to the best spots on the chassis for the ground points?

IMG_8938.jpeg

I agree that not using the chassis for grounding offers more control over where the current is going, though I would also have used an aluminium chassis to stop current from the transformers getting into the equation. I suppose I’m looking for a relatively simple fix.

Thanks again for your help all.

[R.G.]

1/ The centre tap of HT winding is to be grounded separately with the earth connection of the first filter cap.

Yes. Put another way, the CT of the power transformer HT winding must connect to the negative terminal of the first filter cap, and not to anything else, especially not to the chassis or to any other "ground/earth" wires. The power transformer HT CT is not really a ground wire in the way we usually think of. Rather, it conducts the ... large... pulses of current that the rectifiers fill the first filter cap with, but on the negative side. As amp owners and builders, we really don't want those pulses going anywhere but into and directly out of the first filter capacitor.

2/ The ground for the output valves is also to be connected to the HT centre tap and first filter cap.

Yes, but I would state that as "a wire carrying the cathode currents for the output valves connects to the first filter capacitor negative terminal." The HT centre tap happens to also connect to that point, but it is for other reasons. The connection of output valve cathodes to the HT centre tap is a byproduct of them both needing to connect to the first filter cap negative terminal. Because these wires carry large currents, it is important that no other "grounds" connect to these wires on their way back to the One True Ground.

3/Here is where I get a little confused. One side of the output transformer needs to be connected to the speaker ground/common, and not to ground itself.

The PI return ground should be connected to the negative terminal of the first filter cap on its own ground wire. Sharing the cathode ground of the output tubes means the PI is being fed a signal of the full-wave-rectified-audio currents times the output stage's ground wire resistance.

I read this as the P.I. should be grounded with the first filter cap which is with the HT centre tap and the Output stage cathode earth. I would have thought you meant to the negative side of the filter cap smoothing the P.I. supply?

Grounding issues can be very confusing. You're in good company.
We (technical-inclined humans) have by theory and practice decided to use one and only one point as the reference voltage against which all other voltages are measured. For many reasons, this point is the negative side of the first filter capacitor in nearly all guitar amps. Ideally, EVERY SINGLE CIRCUIT GROUND should have its own, separate ground wire leading to this point. That's why they call it star grounding. The HT centre tap should connect there, on its own dedicated wire. The output valve cathodes should connect there, with their own private wire; the PI tail current should also return there on their own private ground wire. The bypass caps feeding the PI and preamp sections should really have their own ground wire leading back to the star ground point. The filter cap smoothing the PI supply is intended to do two different things. One is to "short circuit" or bypass any mains/rectifier hum currents back to the star ground point (the negative of the first filter cap) so these currents don't flow in the PI circuit, causing hum. The other thing is to provide an AC short circuit/bypass for the wobbles put on the power supply by the audio-frequency currents pulled by the other stages so these currents don't get amplified by the PI. In either case, the bypass cap does a better job if its negative returns to the star ground point instead of the negative side of the PI circuit.

Speaker return current is another confusing point. The speaker winding of the OT is truly floating. All speaker currents leave that winding, travel through the speakers, and return to that winding. The speaker output can be tied to the chassis "ground" by using a grounded-bushing speaker jack and no currents flow on the chassis. That's OK. Things get complicated when you try to take feedback from the speaker output back into the PI; now currents can flow from the speaker output winding to the rest of the circuit. Using feedback from the secondary, it is really best to float the speaker ground from chassis, run a ground wire from the speaker return wires to the one true ground, and then run the feedback resistors/etc. to the PI. This ensures that speaker currents stay off any conductors that reference audio ground in the amplifying circuits.

The hard part in all of this is getting used to thinking where the currents are flowing, not what the voltages are. A current flowing in any real wire causes a voltage drop by the wire resistance. That makes an offset voltage across the wire, and that offset voltage can be amplified unintentionally. You can force the currents to flow in single wires to the One True Ground point, and that prevents its offset voltage from being amplified.