DIY B15N - Noise Issue

[ndawson]

I did try the metal cap/cover to chassis ground a while back with no luck. Thanks for that tip.

I also checked the those 5.6k cathode resistors and they were solid to the first filter cap negative, lowest ohms, 0. I didn't check the B+ to ground end of the cathode resistors or shunt the grids to the ground end of the cathode resistor of the tube yet.

I've decided to go through and re-flow all the solder connections and double/triple check the component connections with the schematic. Just want to make sure things are or aren't what they seem. I'm also looking at all of the wiring/heaters/etc..

I'm also going to readjust my grounding and place a heavy duty/tooth grinding ground connection point at the chassis near the first filter cap negative. I'll use this as the only ground to chassis and isolate the input jack from the chassis. I'll also add the shunt cap to ground at the input jack to eliminate any potential rf noise.

I'll look at the specs of my scope.

[ndawson]

Rechecked and reflowed all of the connections. Moved the ground to chassis connection near the first filter cap negative and isolated the input jack from the chassis. Still noisy. Just wanted to do all of this before I continued the troubleshooting.

I was able to shunt the preamp tube grid and it's corresponding cathode resistor ground and doing so stopped the buzzing/noise. I did this for each grid/cathode section of the input tube with the same results. Any ideas here?

As I'm truly a novice at troubleshooting a tube amplifier and using a scope, what would be some tests I could do with the scope that could potentially find the source of the noise/hum? I've managed to use the scopes onboard signal generator and connect to the amp and get signal going. I've got the probe and ground clip ready to go. That's about it.

Here are the specs for my (cheap) digital scope:

Sampling rate：5MSa/s
Bandwidth：1MHz
Vertical sensitivity：10mV/Div-10V/Div
Time base range：500ns-20S
Voltage range：X1:士40V (Vpp:80V); X10: 士400V (Vpp :800V)
Trigger method：Auto / Normal / Single
Coupling method：AC/DC
DDS Signal generator
Frequency: 0-10KHz
Duty cycle: 0-100% (rectangular and sawtooth waves)
Amplitude: 0.1-3.3V
14 Waveforms—Sine wave, rectangular wave, sawtooth wave, half wave, full wave, step wave, anti step wave, noise wave, exponential rise, exponential drop, DC signal, multi tone, Sink pulse, Lorentz wave.

[ndawson]

I'm trying to understand what the shunt is doing here. I've been reading online that shorting the grid to cathode ground would bypass or short out that section of the tube. If that's the case, then when I shunt the input of the preamp tube and the noise disappears, it would mean that the problem lies before the grid? Before the input is the jack, 120k and 5.6m input resistors. Or maybe I've got this all backwards.

[Stevem]

For a test I would jump out those 120k resistors and just for S & G jump out the complete tone stack to the next gain stage.

[didit]

Hello --

I'm trying to understand what the shunt is doing here. I've been reading online that shorting the grid to cathode ground would bypass or short out that section of the tube. If that's the case, then when I shunt the input of the preamp tube and the noise disappears, it would mean that the problem lies before the grid? Before the input is the jack, 120k and 5.6m input resistors. Or maybe I've got this all backwards.

That seems fairly accurate -- before, or right at, the grid.

If I were building from scratch from a schematic the layout would be different. Specifically, both of those resistors would have been on/around that tube socket rather than around the input jack. However, if presented with the amp as-is this would be well down on a remediation list so would not start by rebuilding this "correctly".

Worth a go. Would also tend to toward using somewhat smaller values for both -- 47K and 2.2M. The grid stopper will work best if soldered directly on the socket grid tag, and it's straightforward to solder its other end to the core wire of that coax.

Best ..

[R.G.]

I'm trying to understand what the shunt is doing here. I've been reading online that shorting the grid to cathode ground would bypass or short out that section of the tube. If that's the case, then when I shunt the input of the preamp tube and the noise disappears, it would mean that the problem lies before the grid? Before the input is the jack, 120k and 5.6m input resistors. Or maybe I've got this all backwards.

Shunting the grid to the ground end of the cathode resistor does bypass or short out that section of the tube. There is a bit more to it, though. What is really going on in the triode self bias circuit (this is one of those) is that the tube stays at the correct DC voltages on all pins, but any input signal is "shorted out" by the near-zero resistance of the wire. This loads down any signal being fed into the grid, so the AC gain is dropped to about zero.

This is different from changing things so the tube is biased to cutoff (no plate current) or near-saturation. In your circuit, this tells you that the signal is probably not being fed into the cathode by sneaky ground currents, because the gain for that path is not affected much by shunting the grid. So the grid is involved. Wiring before the grid or feedback to the grid is the likely culprit.

120K and 5.6M are way big grid resistors compared to what most amps use, and putting them on the input jack is odd. If it were me doing this, I would try what didit suggests - lower resistors and right up on the tube socket, not out at the jack. It also would not hurt to run a coax/shielded cable from the jack to carry the signal to the grid. You might be able to just use a twisted pair from jack to tube.

The scope specs are useful. Having an 800V range on 10:1 scales is great! Just be careful not to probe plate voltages or high voltage supplies with the probe set to 1:1. Get familar with where the switch that chooses 1:1 versus 10:1 is and use it.

The point of the scope is to see if the circuit is amplifying/making ultrasonic or RF. 1MHz is good for audio, but radio-station RF is so high in frequency that this scope can't "see" it. Still very useful. Even 20MHz scopes can get blind to a lot of RF.

[ndawson]

It seems the straight wire to the grid and the 2.2m resistor have the most dramatic sound effect. I also noticed that having the jack floating away from the chassis has also reduced the noise. It is such a short distance, maybe one inch, from the tube socket to the jack that I didn't use coax cable. Just hard wired the resistor and short wire. Also, the ground on the jack is temporarily clipped onto the chassis with the green alligator in the attached photo. Now I have the star ground point connected to the chassis and this additional ground point connected to the chassis. There is still a slight buzz, but well below the level it was at originally and not noticeable at all when playing the bass through the cabinet.

IMG-2035.JPG

[ndawson]

Very little noise at this point, if any. Settled on a 1.2k grid resistor. I also changed to the 2.2m resistor on the input grid to ground as recommended. I’m going to run the amp through its paces tomorrow after reinstalling in the head enclosure.

Thank you for the suggestions and patience. I really appreciate all the knowledge and support here.


Nate

IMG_2038.jpeg

IMG_2037.jpeg

[R.G.]

YAAAY!!! Congratulations!

[didit]

Good hearing it's resolved and you're no longer hearing noise, now just making some.

..

[SoulFetish]

Nice work!

[rayrayman]

Jumping into this thread while chasing an issue in my own build...

Can you explain in more detail ideas about how the changes you made solved the problem? Was the original resistor just noisy or was it something you did with the ground locations?