Grid stopper question

[g-man]

Deleted

[sluckey]

#3 should be fine.

[HotBluePlates]

Would having a resistor pair mounted to the sockets (for hi-lo function) and a separate grid stopper on the tube socket work? See picture 3. I have never seen it done that way so I'm wondering if it causes a problem? 

There's nothing wrong with #3 except that it uses 3 resistors where 2 would get the job done. But if you want add grid stopper resistance and don't want to do it by swapping resistors, you could do it the way #3 shows.

BTW, you wouldn't reduce the values of the resistors on the jacks in order to add a resistor at the socket (I'm assuming that's the reason the jack resistors are 43kΩ).

... my 18 Watt trem channel already has the resistors on the jacks and a single shielded cable running to the tube socket, but it still picks up RF noise. ...

How did you determine it was RF noise, and related to the input jacks?

[g-man]

Deleted

[HotBluePlates]

I did reduce those from 68K to 43K thinking that it would be better with a 10K in series. I have read that going lower on grid stoppers is better for lower noise ...

Yeah, I think the logic goes something like "resistors contribute noise, and very high resistances contribute more noise, so reduce resistance to reduce noise." The problem is you're trying to use a grid stopper to reduce high-frequency response as the way of killing RF noise, and the value of resistance determines how that approach works. Reducing the total resistance from 68kΩ to 53kΩ (43kΩ + 10kΩ) actually reduced your high frequency attenuation.

The formula for capacitive reactance is:


"C" will wind up being the Miller Capacitance of the 1st tube stage; we could calculate that, but let's assume it will be ~120pF, or 0.00000000012F
Let's use 20,000 Hz for "f" which is the upper limit of human hearing.

If you calculate "Xc" using those numbers, you get ~66kΩ. An RC circuit's response will be down 3dB when circuit-R equals Xc. If you swap the positions of "f" and "Xc" in the formula, you can take your known-R and find the -3dB frequency. For 53kΩ and 120pF, the circuit is -3dB at ~25kHz, so a little less high frequency attenuation.

... I have read that going lower on grid stoppers is better for lower noise ...

Since the circuit's effectiveness is dependent on the resistance value, a better option for reducing noise is to use a lower-noise type of resistor. Like metal film instead of carbon comp. But this will just cut hiss, if hiss is even an issue due to the grid stopper.

How did you determine it was RF noise, and related to the input jacks?

It's radio station signals that are being picked up, but I don't know for sure if it is due to the jacks. I recently re-did my ground scheme in that amp to get other noises minimized, but the radio stations are still picked up. I don't get those stations picked up on my other amps using the same guitars, cables, location, etc. so it is specific to this amp.

What's the amp? Do you have any pictures of it?

You have a 3-wire power cord, with the ground wire bolted to the chassis? Do you have any kind of ground-lift scheme? Did you do anything to try to isolate any circuit elements (jacks, pots, etc) from the chassis? (I'm not recommending that as a fix, just trying to find out what you have now)

What is your present ground scheme? How are filters caps, and the rest of the circuit in general, grounded? What did you change when you re-did the grounding?

Broadly speaking, to make an A.M receiver you need a diode detector and some amplification. An over-driven tube grid can act as the diode. But so can non-obvious wiring errors. If you have a 220kΩ or higher resistor handy, try swapping that in as a temporary grid stopper and see if you can kill the radio stations. If not, you probably have some other error or cause of the noise.

[terminalgs]

It's radio station signals that are being picked up, but I don't know for sure if it is due to the jacks. I recently re-did my ground scheme in that amp to get other noises minimized, but the radio stations are still picked up.

Ground the grid on that first input jack, preferably with a circuit ground as near to the tube as possible.  turn amp ON, and turn volume UP.  If the RF interference goes away, you are on the right track with your V1 grid troubleshooting.  If you still get RF interference, then the circuit is picking up RF elsewhere in the circuit.

[jojokeo]

You can try this for RF protection rather than messing with grid stoppers: use a .01uF cap w/ short leads grounding near the input jacks.
 
*I should mention also that the insulated input jacks (which are recommended for all modern amplifier designs) does not get grounded at this point as you can see? The pot's casings do get grounded as well as the shielded part of the wiring and serve as the shield for the signals. Signal ground along with the input jack's ground (two black wires in this case) do not connect along any of these various points. They get grounded at their respective ground point of their filter cap.

[g-man]

Deleted

[Ed_Chambley]

You can try this for RF protection rather than messing with grid stoppers: use a .01uF cap w/ short leads grounding near the input jacks.

How does the cap work?  It looks as if you have it connected to the pot?

[jojokeo]

It connects to the input jack's ground. Any RF or signal picked up via the guitar &/or chord riding toward the input gets shunted to ground via the cap right at the input jack.

[g-man]

Deleted

[jojokeo]

That is correct. Either Aiken or Merlin mentions it in their writings and where I got it from several years ago.

[Ed_Chambley]

Ed, the cap JoJo is referring to is the green one near the chassis (not the white ones near the pots).

That clears things up a lot.  Thanks.

[g-man]

Deleted

[HotBluePlates]

I eventually found it was related to the footswitch. Unplug the footswitch and the radio signal was gone. Plug in a shorting jack instead of the footswitch and it was still quiet.

Are you using a shielded cable for your footswitch?

[g-man]

Deleted

[jojokeo]

OK, I think I got this worked out. The Normal channel did not have grid stoppers and picked up some minor RFI. I added a 10k to each grid and a 0.01 cap at the jack ground and that channel is quiet. The trem channel picked up much more RFI even though it has the 68k pair on the jacks. I eventually found it was related to the footswitch. Unplug the footswitch and the radio signal was gone. Plug in a shorting jack instead of the footswitch and it was still quiet.

I mentioned I recently re-did my grounds. My original grounding was pretty random, and the amp has always been more noisy than I like. So I separated the preamp grounds and ran them to a single point. Helped a lot, but it may have increased the RF noise somehow? As recommended on the Mojo layout, I built it with a shielded cable running from the footswitch to V3 pin 7, and that shield was now grounded with the preamp signal ground. I decided to change the regular wire running between the footswitch and speed pot to a shielded cable, and at the same time I grounded both shields together to the chassis (no longer tied to the preamp ground). After a couple days testing it at different times of day I haven't picked up any more radio. Is it likely that the shielded cable on the footswitch being grounded with the preamp signals was enough to pick up radio noise? Or maybe the non-shielded wire that ran to the speed pot was picking it up? I'm really not sure but so far it seems OK.

First off, glad you got things quiet. Secondly, the grounding is just as important as the circuitry. Grounding IS the other half of the circuit. Currents don't disappear into a black hole (as Merlin describes). If they are drawn from a positive end of a power source, they have to find a way back to the negative end via ground. There should only be one circuit ground to the chassis for this reason and how ground loops are created. Circuit ground may form at several different places before leading to a common/star ground point though and non-audio ground should not be connected to audio grounds. Rather it should connect back to the "spur" as it's called which is a short wire to the reservoir ground point. If confused, read and re-read Merlin's grounding article. It's virtually impossible to take all of it in in a single reading. This helps explain the noise which you speak of (being random - and is of even more importance when an amp has several channels and/or high gain too).

That long footswitch cable is like having one long grid wire. Grid wires are ultra sensitive high impedance places where short wire lengths and/or shielding is of great importance and which acts as an antennae. The cap to ground I mentioned helps with any signal trying to ride along the shielding of either a footswitch (like you're using) and also regular guitar cables when on the input. The signal wire itself is where the grid stoppers do their job here.

[HotBluePlates]

The switch that came with the kit is not shielded. Just three insulated wires in a larger insulation layer.

That's the culprit.

[g-man]

Deleted

[jojokeo]

Makes me wonder why the kit comes with an unshielded switch if that can be the noise source.

Not a fan of kits in general myself. Even the plans for the layout & schematic I have issues with. Good thing your experience could get you through this as most newbies could've easily had numerous problems from what I can see.

[HotBluePlates]

... I'm going to make a shielded switch. ... Makes me wonder why the kit comes with an unshielded switch if that can be the noise source.

Not the switch; you need a shielded cable running out to the footswitch.
 
The footswitch itself winds up automatically shielded if you use a metal housing (and have some continuity from cable shield to the metal housing). The footswitch jack gets shielded as much as it needs by the chassis itself.
 
Look at an old Fender Reverb/Trem footswitch. The housing is metal with a plastic base. Within the cable is a single conductor for the trem (all it does is connect the wire to ground or not, determing trem oscillator operation but is not a potential noise pick-up point). The conductor for reverb is a shielded wire because it connects to a tube grid. To kill reverb, that circuit connects the hot to the shield (which is grounded at the footswitch jack). So the wire shielding in that case protects the grid wire from picking up noise and also provides part of the switching functionality.

[g-man]

Deleted

[HotBluePlates]

I see! Oops!

I envisioned a little box around the backside of the switch, behind the panel...