Yes let me clarify my previous post as I checked a bunch of variations.
To summarize:
1) Pot connected -- external input jack connected/shorted. Silent on both ends, buzzes most in middle of the pot.
2) Pot connected -- external jack open. Silent on min, buzz progressively increases to max.
3) No input at all -- constant noise
Grid leak resistor is 1M but the pot is 50K and that should shunt it.
Either way I can experiment with 220K as that is the grid leak value that I used on a few Tubelab amps I've built that use the 12AT7.
Worth a shot, I have a few small value mica caps to try.
Interestingly, I was moving the scope probe around, and those spikes did not appear on the input terminal or the grid of the tube. They did appear though at any point in the feedback loop, i.e the 12AT7 plate, 100K feedback resistor, 220uF cap and on the 240R grid resistor.
50k pot fine. Silly question are the heaters floating or are they center tapped to ground. Does sound like heater coupling either on PCB or in valve itself. Note both halves of the valve are pinned differently making one side more susceptible.
Last edited:
Yes, good point on the heaters. If center tapped filament supply is available, then simply elevating heaters by connecting center tap of heater supply to cathode of EL84.
Yes heaters is one area I will need to be testing.
The heater winding is center-tapped and that is connected to the star ground (can be observed in the picture).
I want to try elevating the heaters but I don't have the correct resistors or cap so will place an order for those today.
I didn't try connecting the CT to the cathode, worth a shot.
As mentioned in the OP, when using a regulated heater supply this specific noise went away but it introduced some odd low-level hash.
There may be a different issue with the regulated heater supply because under load, I can only get it to output 5.7V even though it is rated at 5A. (total heater current is 1.8A AC or ~3.6 DC).
So I'll need to find a different 6V supply to try there.
The heater winding is center-tapped and that is connected to the star ground (can be observed in the picture).
I want to try elevating the heaters but I don't have the correct resistors or cap so will place an order for those today.
I didn't try connecting the CT to the cathode, worth a shot.
As mentioned in the OP, when using a regulated heater supply this specific noise went away but it introduced some odd low-level hash.
There may be a different issue with the regulated heater supply because under load, I can only get it to output 5.7V even though it is rated at 5A. (total heater current is 1.8A AC or ~3.6 DC).
So I'll need to find a different 6V supply to try there.
Yes the trace to pin 9 goes over the side with the issue...
Although the trace is on the opposite layer and 90deg to the other trace so I find it hard to believe that would be the issue.
The thing that confuses me here is that the issue goes away by either running heaters as DC or using the iPhone instead of the pot. 2 totally unrelated (seemingly) mitigations with the same outcome
I’ve had very good luck with elevating filaments with the output tube cathode, simple and works very well.
I would bet $$ that its simply that the heater on that channel is the culprit in your via layout.
This audio circuit, with the 1M Rg, has been built successfully with no hum by a ton of people.
The amp is very quiet for the most part, its just this slight buzz that is audible on headphones that brings to my awareness that there is some issue that should be resolved. If I put this in my main rig I doubt I would hear it.
I disconnected the heater CT from the chassis.. crazy buzz and noise.
Attaching it back to either the chassis or Pin3 of the output tube closer to the PT has the same effect.
Yes, it is the black wire going from the HP jack back to the star ground in the picture
It could be that there is a ground loop or an issue with shared ground between the 2 chanels. Is there no decoupling of the Zener on G2?
It only takes a couple of pf to couple the higher harmonics of the heater buzz onto the grid at 100k. Its not helped if the transformer generates the rectified HT as this get onto the heater windings. That left track of the input does cross over the heater. Add a few pf deliberately to see if the effect is the same. A buzz is usually capacitive coupling. Your spectrum shows the channels to be not dissimilar at LF but very different at HF.
Last edited:
The star ground is to level all ground but should never be used to draw current. It does here as the signal will flow from the PCB via the start ground to the outputs.
No decoupling as per the original schematic
Is there anything I can adjust to rule this out ?
The Zener should be decoupled by a small capacitor (my choice would be to add a big cap to ground here as the tube is in pentode mode and G2 is beeing a grid adding some amplification) The cables to the outputs should go directly from the Board and not via the star ground (3 ground cables from the board, 1 to the headphones connector and 2 to the speaker connectors). Sure 1 cable needs to be connected still to this point as reference.
Last edited:
Please note that there are output transformers on the output.
The primary is connected to the output tube plate and B+.
OPT secondaries (+) are wired to a DPDT (4ohm / 8ohm switch) and wired to the binding posts from there.
OPT secondaries COM are wired directly to the binding post and tied back to the star ground for safety.
The headphone out is wired from a 16ohm tap to L/R, with the sleeve tied back to same star ground
The primary is connected to the output tube plate and B+.
OPT secondaries (+) are wired to a DPDT (4ohm / 8ohm switch) and wired to the binding posts from there.
OPT secondaries COM are wired directly to the binding post and tied back to the star ground for safety.
The headphone out is wired from a 16ohm tap to L/R, with the sleeve tied back to same star ground
Ok, i understand. But i do not see the ground wire from the headphones output to the transformers. Do you use the earth on the mains connector? And is so is you preamp also connected to earth? In this case do not connect the earth direcly but use back to back diodes (safety).
Attachments
Last edited:
The first sentence is very interesting. It seems to confirm that the spiky noise is being coupled in from the heater supply, as everybody seems to assume already. Does your power supply have a common transformer for heaters and (rectified) B+ ? If so, then properly snubbing the B+ secondary winding will remove the source of the spikes.
For an excellent and revolutionary point of view, see Morgan Jones' article in _Linear Audio_ Vol.5 and Mark Johnson's in _Linear Audio_ Vol.10 .
YOS,
Chris