I guess the next person knocking at my door will be the UL police. Although I don't really need to, I always use a ying/yang diode for chassis to circuit ground isolation.
I agree this problem is some sort of ground loop issue. From what I am reading it is a non raspy 50Hz hum that is present only with an external grounded source is applied. It could be filament, but that should hum even without a source. While I am leaning heavily to the power supply as the issue, other possibilities exist.
I can't tell from your drawing, but were do you bring your cathode resistors back to? How do you tie your PS ground to the star point. Did you try a different GZ rectifier? How about a different input tube?
How about a picture of the wiring?
I agree this problem is some sort of ground loop issue. From what I am reading it is a non raspy 50Hz hum that is present only with an external grounded source is applied. It could be filament, but that should hum even without a source. While I am leaning heavily to the power supply as the issue, other possibilities exist.
I can't tell from your drawing, but were do you bring your cathode resistors back to? How do you tie your PS ground to the star point. Did you try a different GZ rectifier? How about a different input tube?
How about a picture of the wiring?
Last edited:
Hello people. Thank you VERY much for your suggestions so far, every post is appreciated.
Here is the current situation:
I installed another PCB myself into a similar amplifier (a stereo version, but only one board connected) and got the SAME problems. I tried my best to follow good grounding technique. I will try to list as many observations as I can recall right now
I originally designed and made these boards for a Radford STA15 I am restoring for someone, which is nearly complete. I have had it all repainted and any exposed metalwork chromed. It is a real labour of love.
Here is the current situation:
I installed another PCB myself into a similar amplifier (a stereo version, but only one board connected) and got the SAME problems. I tried my best to follow good grounding technique. I will try to list as many observations as I can recall right now
- With nothing on the input, the amp is as quiet as you'd expect.
- With a CD walkman connected, there is a SLIGHT low level non-raspy hum, but it is otherwise acceptably quiet.
- With a grounded source connected (an audiolab preamp fed with the CD walkman), there is an unpleasant raspy buzz, which at first seems to be an earth loop but I'm not so sure..
- An identical amp, connected to the same mains outlet (a Radford STA15) WITHOUT my new boards (originals) does not buzz, which is worrying.
- If you float the amp with my boards in, the buzz gets WORSE!?
- With a CD walkman connected directly to the amp (with a jack to phono converter), despite the quietness, if you touch the EARTH connection of the unconnected phono lead, it picks up a buzz???
I originally designed and made these boards for a Radford STA15 I am restoring for someone, which is nearly complete. I have had it all repainted and any exposed metalwork chromed. It is a real labour of love.
Beautiful work. Well worth saving.
Seems to me like you might have oscillation. Looks like you have grid stoppers on the output sockets. But the grid, plate and cathode resistors for the input tube are some distance from the socket pins. And they are partly carried by PCB traces, which will probably be more inductive than the fairly stout wire from board to socket.
I've done this type of earth lift in the past too. And with properly rated components, the amp common can be safely lifted from the chassis earth, even if not UL approved. However, I have found, in every case, that if I get all the common connections done properly, i.e., paying close attention to current flow, it's completely unnecessary. With systems that contain many interconnected components, the earth connections have to be starred or properly branched too.
Sheldon
Thank you 🙂
Possibly! I have scoped the anode though, I haven't noticed anything unusual. I ought to get it out again and have another poke around, but I have been summoned to do outdoors work today 🙄
I would have thought it was unlikely, since the dominant pole capacitor shunts the anode resistor for that stage and the roll-off starts at about 20Khz, although I am open to the possibility. The circuit does specify a 500pF cap and I'm using a 470pF since 500pF is apparently an uncommon value for the box type film caps I'm using (by the time I'd thought of simply putting two 1nF's in series I'd etched the boards 🙄)
Another weird thing I have observed is the raspy buzz while connected to the preamp changes whether or not the preamp is on. When it's on, it is rather raspy, and when it's off, less so.
It's like the earth on this amp is acting like another input, I really can't figure out how or why 😡
Hello bigwill, I do not want to side-track the thread, but would like to take a moment to reply to Mr. Wavebourn, so please excuse me for a few moments...
1.
I agree with you in principle. However I mention this because bigwill obviously wants to optimise the amplifier, so it is worth his time to consider all items associated with hum, etc., even if they are not directly part of his current solution. Considering these things, and having a list of them, will help bigwill with all future endeavours.
2.
It is in fact legal. Commercial amps do indeed have it that way. Also, if I change an amp at my home it is fully legal; I can even shoot it… shooting a man is illegal, but shooting an amp is OK !
3.
As point 1 above.
4.
Layout and the attendant wiring of the components can introduce problems. Displaying the layout may help others to understand the issue, so that they may better provide advice.
OK, back to the thread...
Enjoy the sunshine. G.
bigwill,
It sounds like there is a loop formed by the input signal and corresponding ground conductors, which is getting a current induced in it by an AC field. There would be no circuit for this current unless something were connected to the input. The induced current could induce a hum voltage across the input impedance, across the grid resistor, and possibly also between the grid resistor's grounded end and the star ground point, as well as between the ends of the signal and ground conductors. (See Faraday's Law, or, Maxwell's Equations. A time-varying EM or M field will induce a time-varying current. [And a time-varying current will induce a time-varying field.])
If that is the problem, then you need to make sure that you run the ground and signal conductors as close together as possible (preferably twisted together, and preferably also shielded), ALL the way from the signal input connector to the grid input, or absolutely as much as possible. Then the signal's ground conductor should go directly to the grounded end of the grid resistor, while staying close to the resistor.
If unshielded (i.e. not a pair of wires inside a shielded cable with shield connected on one end only, to the chassis most likely), the twisted pair should be kept as far away as possible from all power, large signal, and dynamic current conductors, or, should only cross them at a right angle. And even shielded cable should be kept as far away as possible from changing magnetic fields.
The same loop-area-minimization technique should always be applied to EVERY pair of conductors, including input signal/ground, AC pairs, DC power/ground pairs, output signal/ground pairs, et al. Twist each pair together for as long as possible. And use shielded cable if possible. On PCBs, always try to run corresponding pairs' traces as close together as possible, for as long as possible.
A (somewhat) similar problem was solved per the above, in the thread at:
http://www.diyaudio.com/forums/tubes-valves/167617-help-ground-hum-4.html#post2204734
This article is very good:
http://www.diyaudio.com/forums/diya...udio-component-grounding-interconnection.html
Cheers,
Tom Gootee
It sounds like there is a loop formed by the input signal and corresponding ground conductors, which is getting a current induced in it by an AC field. There would be no circuit for this current unless something were connected to the input. The induced current could induce a hum voltage across the input impedance, across the grid resistor, and possibly also between the grid resistor's grounded end and the star ground point, as well as between the ends of the signal and ground conductors. (See Faraday's Law, or, Maxwell's Equations. A time-varying EM or M field will induce a time-varying current. [And a time-varying current will induce a time-varying field.])
If that is the problem, then you need to make sure that you run the ground and signal conductors as close together as possible (preferably twisted together, and preferably also shielded), ALL the way from the signal input connector to the grid input, or absolutely as much as possible. Then the signal's ground conductor should go directly to the grounded end of the grid resistor, while staying close to the resistor.
If unshielded (i.e. not a pair of wires inside a shielded cable with shield connected on one end only, to the chassis most likely), the twisted pair should be kept as far away as possible from all power, large signal, and dynamic current conductors, or, should only cross them at a right angle. And even shielded cable should be kept as far away as possible from changing magnetic fields.
The same loop-area-minimization technique should always be applied to EVERY pair of conductors, including input signal/ground, AC pairs, DC power/ground pairs, output signal/ground pairs, et al. Twist each pair together for as long as possible. And use shielded cable if possible. On PCBs, always try to run corresponding pairs' traces as close together as possible, for as long as possible.
A (somewhat) similar problem was solved per the above, in the thread at:
http://www.diyaudio.com/forums/tubes-valves/167617-help-ground-hum-4.html#post2204734
This article is very good:
http://www.diyaudio.com/forums/diya...udio-component-grounding-interconnection.html
Cheers,
Tom Gootee
Last edited:
Sure. But is a suggestion to do illegal things legal? That is the question.
I hate bureaucracy too. But...
I think you have a galvanic loop, i.e. adding the external ground priovides a conductive path for the noise to be imposed on the input of the amp (as opposed to an EMI or radiated noise issue). 50Hz is an indication that it is from the AC line or power supply primary side (assuming full wave rectification).
I see that your input RCA jacks are isolated from the chassis but you have two black wires connected to each. Earlier you say you are using a star ground. If one of the black wires goes to the star ground, where does the other one go? A star ground only works if all the wires truly connect at one point. In an amp like yours, there may be 20 wires or more. Every one of them except the middle of power supply filter chain needs to connect within a fraction of an inch of each other for best results.
I agree with Wavebourn. If you really and truly have a single point ground, a lot of the "EMI" issues go away. It's amazing how little shielding is needed at line levels and impedances we use in audio. Most of the noise problems are due to a galvanic circuit.
Resist the temptation to strip it all out and start again. As you have discovered, just throwing "fixes" at the problem simply leads to more confusion and frustration. Putting a "ground loop breaker" inthe wrong place can make the problem worse. You really need to think through how 50Hz Ac gets connected to your amp input. Where's the unintended circuit?
The fact that hum increases when you connect an input cable is a clue. I would start by connecting a clip lead to the RCA jack ground and the other end to various ground points like the chassis and the AC input ground, etc. to see what happens. I agree that AC is somehow getting to the input. That seems to be the place to focus, and the 2 black wires make me suspicious.
Cheers,
Michael
I see that your input RCA jacks are isolated from the chassis but you have two black wires connected to each. Earlier you say you are using a star ground. If one of the black wires goes to the star ground, where does the other one go? A star ground only works if all the wires truly connect at one point. In an amp like yours, there may be 20 wires or more. Every one of them except the middle of power supply filter chain needs to connect within a fraction of an inch of each other for best results.
I agree with Wavebourn. If you really and truly have a single point ground, a lot of the "EMI" issues go away. It's amazing how little shielding is needed at line levels and impedances we use in audio. Most of the noise problems are due to a galvanic circuit.
Resist the temptation to strip it all out and start again. As you have discovered, just throwing "fixes" at the problem simply leads to more confusion and frustration. Putting a "ground loop breaker" inthe wrong place can make the problem worse. You really need to think through how 50Hz Ac gets connected to your amp input. Where's the unintended circuit?
The fact that hum increases when you connect an input cable is a clue. I would start by connecting a clip lead to the RCA jack ground and the other end to various ground points like the chassis and the AC input ground, etc. to see what happens. I agree that AC is somehow getting to the input. That seems to be the place to focus, and the 2 black wires make me suspicious.
Cheers,
Michael
Last edited:
Those two black wires are to simply earth the runners that the valves are mounted on (they are physically separate from the main chassis), the other wire is the negative from the output transformer, the runner is just picking its earth up along the way:
OPT- [wire] Valve runner [wire] Earth
OPT- [wire] Valve runner [wire] Earth
I THINK I have actually solved the problem. I think it was that damn 56 ohm resistor all along. 
Despite bypassing it numerous times before, what I did this time seems to have done the trick.
I shorted it out with a twisted clip lead on each end, and while there is still a slight ground loop like buzz, it is still there with the other amp (despite being seemingly quiet in this configuration before). It is probably just the residual natural ground loop between the source and the amp.
I don't know why this is suddenly working better than it did before. IIRC, when I had the other amp connected to the preamp before it was quiet, but now they are behaving the same, which is a relief.
Grounding the scope to the amp and probing the shield of an RCA cable connected to the preamp reveals AC noise, despite there being no loop. There is an AC difference between the earths of the two components, despite there being no loop and being connected to the same place!
The 56r allowed this voltage to be developed across it with little attenuation, so basically the preamp and amp were still bobbing up and down with respect to each other. Shorting this resistor tied them together and seems to have cured this
The resistor I thought would stop this sort of thing from happening seems to have hurt things a lot more
I am surprised at this behaviour to be honest, but I have learned a great deal in this troubleshooting process!
I will try floating the amplifier next and in theory it should be silent when connected to the preamp. Obviously I won't be leaving it like this 🙂
Despite bypassing it numerous times before, what I did this time seems to have done the trick.
I shorted it out with a twisted clip lead on each end, and while there is still a slight ground loop like buzz, it is still there with the other amp (despite being seemingly quiet in this configuration before). It is probably just the residual natural ground loop between the source and the amp.
I don't know why this is suddenly working better than it did before. IIRC, when I had the other amp connected to the preamp before it was quiet, but now they are behaving the same, which is a relief.
Grounding the scope to the amp and probing the shield of an RCA cable connected to the preamp reveals AC noise, despite there being no loop. There is an AC difference between the earths of the two components, despite there being no loop and being connected to the same place!
The 56r allowed this voltage to be developed across it with little attenuation, so basically the preamp and amp were still bobbing up and down with respect to each other. Shorting this resistor tied them together and seems to have cured this
The resistor I thought would stop this sort of thing from happening seems to have hurt things a lot more
I am surprised at this behaviour to be honest, but I have learned a great deal in this troubleshooting process!
I will try floating the amplifier next and in theory it should be silent when connected to the preamp. Obviously I won't be leaving it like this 🙂
- Status
- Not open for further replies.