Shared tubes and noise

[Colossal]

I just finished up on a cathode biased 2x6L6 amp very similar to the old Gibson EH150. It has a wonderful sound and is dead quiet. There are two channels: Instrument and Microphone. The Instrument is basically the clean channel and the Microphone is cascaded for a mild, bluesy but killer sounding overdrive.

I've attached a stripped down version of the preamp for your reference. V1, V2, and V3 are 6SQ7 octals and V4 is a 6N7 cathodyne PI. You will notice that V2 and V3 share a common anode and cathode bias resistor/bypass cap, essentially a common anode mixer arrangement.

I noticed something that I'd like to find a solution to for future builds: When plugged into the Clean channel (V3), if the volume on the overdrive channel is turned up, a little gain hiss/buzz is added to the background. It's only really noticeable when the OD volume is about 7-10 but I'd like to be rid of it if possible. Would splitting the shared cathode into two separately biased cathodes, say 2k each (from 1k shared) be a solution to this or is this just a consequence of the shared load resistor??

Thanks for any suggestions, in advance.

[tubeswell]

I think its more like the 'inert' o/d channel stage in front will be producing some signal, like shot noise or somesuch (maybe from random electrons in the first stage hitting the plate and bouncing around?, or maybe from that stage's plate resistor picking up some micro wobble in the HT and feeding it into the next stage), from just being 'on', and when you have the o/d vol up, then the grid of the 2nd stage that is connected to that first o/d stage would surely be picking this up and amplifying it? Interested to hear what others think.

[es345]

Looking to your example just one question: Is the overdrive input grounded if nothing is plugged in?

[Colossal]

I think its more like the 'inert' o/d channel stage in front will be producing some signal, like shot noise or somesuch (maybe from random electrons in the first stage hitting the plate and bouncing around?, or maybe from that stage's plate resistor picking up some micro wobble in the HT and feeding it into the next stage), from just being 'on', and when you have the o/d vol up, then the grid of the 2nd stage that is connected to that first o/d stage would surely be picking this up and amplifying it? Interested to hear what others think.

Yes, this is exactly what I think is happening (shot noise) since the added noise is not a nasty hum or anything, just an increased ambience/hiss/fuzz. It's only apparent when the OD volume is up to around 7-10 but it is there and if recording, the clean should be as clean and quiet as possible.

Looking to your example just one question: Is the overdrive input grounded if nothing is plugged in?

Yes, it actually is. I should have just posted the whole schematic. There are actually two inputs on each channel (at the request of the client) and the high gain jack on the OD is grounded, so both jacks are grounded (via the sleeve switch) when nothing is plugged in.

Thanks guys. I'm still wondering if splitting the shared cathode resistor into two separate ones (2k with 22uF bypass) will take the OD channels shot/other noise out of the picture. It would still be a common anode mixer though.

[Structo]

I don't know much but could you add some mixing resistors where the two channels combine?

[Colossal]

I don't know much but could you add some mixing resistors where the two channels combine?

The only issue I see with it is that both tubes are coupled via a common load resistor and share the same tone caps/pot

[es345]

what happens if you pull out v1? If there is no longer any hiss then its V1 via the grid to V2 as already mentioned.
I think separate cathode resistors will not help. The cathodes of V2 and V3 are grounded for audio anyway via the cap (25 uf).
There are inputs on the market with 2 embedded isolated switches.
Using that you can create a grounding of the grid of V2. e.g. with a VTL5C1 optical isolator if nothing is pugged in the OD channel.

[Colossal]

Unfortunately (for me) the amp is with its new owner so cannot try pulling V1 out but I did try that before the amp went away and yes, there was no sound when V1 was out so your idea about the noise from V1 via V2's grid is likely correct.

As far as separating the cathode resistors goes as a possible solution, I would separately bypass them as well with their own caps so each grid would have its own path to ground but would still share the same anode. My thinking was that any noise coming from V1 while V3 is getting signal would not be transferred, but that might be flawed thinking since the anode resistor is still common in each case (hence the *name* common anode mixer )

[Merlinb]

My thinking was that any noise coming from V1 while V3 is getting signal would not be transferred,

Sadly no, separating the bias circuits would make no difference. The problem is simple; any noise coming from V1 is amplified by V2, and ends up mixing directly with the clean signal. The solution is to eliminate the noise in the first place, which of course depends on what sort of noise it is.

It it's heater hum then you could use a humdinger, or even a DC heater for V1.

If it's hiss then use metal film resistors around V1, and reduce the grid stopper to perhaps 10k. Also, remove the potential divider which comes after the OD pot and instead split the V1 anode resistor into two parts and take the output from the junction of the two. This will reduce resistor noise, while retaining the same level of attenuation that you already have.

[Colossal]

My thinking was that any noise coming from V1 while V3 is getting signal would not be transferred,

Sadly no, separating the bias circuits would make no difference. The problem is simple; any noise coming from V1 is amplified by V2, and ends up mixing directly with the clean signal. The solution is to eliminate the noise in the first place, which of course depends on what sort of noise it is.

It it's heater hum then you could use a humdinger, or even a DC heater for V1.

If it's hiss then use metal film resistors around V1, and reduce the grid stopper to perhaps 10k. Also, remove the potential divider which comes after the OD pot and instead split the V1 anode resistor into two parts and take the output from the junction of the two. This will reduce resistor noise, while retaining the same level of attenuation that you already have.

Merlin, thank you for the excellent comments, I appreciate it. First, the amp is all octals. It is remarkably quiet and I mean <i>silent</i>. You can have it on, turn up the guitar volume and walk away and never know it is idling! The heaters are elevated to DC by attaching the heater CT to the output cathode bypass cap and are seeing 41VDC. I used IRC GS-3 3W metal glaze resistors for all plate loads (all are 100k). The amp has a wonderful tone and is very warm, natural, and rich sounding.

The voltage divider is 1M (top) and about 150k on the bottom (bottom is actually adjustable via a trimmer to find the sweet spot for overdrive). I stayed very true to the original schematic and the only modification is the addition of that voltage divider. The only time the noise gets injected into the clean channel is when the overdrive (Microphone) volume is up high, around 7-10...that's when it's apparent. It's not a buzz or hum, just a little soft zzzzzzzzzzzzzz sort of sound, very subtle, but is enough to be noticeable and make me wish it were gone. I figured it was a consequence of the mixer and the 1M resistor in front of V2 at the voltage divider (you talk about changing voltage divider ratios in your book for less noise).

So you would recommend a split load and removing the voltage divider...the plate load that is there now is 100k. I should of course add a grid stopper at V2 in the absence of the divider. What values would you suggest for the split load? Grid stopper for V2 around 10k?

Thanks for the help on this.

[Merlinb]

It is remarkably quiet and I mean <i>silent</i>.

Except "with the overdrive (Microphone) volume up high, around 7-10...that's when it's apparent" it's not silent any more when considering only the OD channel, presumably?

What values would you suggest for the split load? Grid stopper for V2 around 10k?

Well since the existing divider is 1M-150k, I would suggest splitting the anode load into 87k-13k, which is the ratio but adds up to 100k (the larger one goes to the anode, the smaller one to the HT of course). Or you could try a similar ratio using standard values, say 100k and 22k, and the smaller one could again be made a trimpot, to serve the same function as you have now.

You don't necessarily need a grid stopper on V2 since there is still plenty of resistance between the anode of V1 and the grid of V2.

[Colossal]

It is remarkably quiet and I mean <i>silent</i>.

Except "with the overdrive (Microphone) volume up high, around 7-10...that's when it's apparent" it's not silent any more when considering only the OD channel, presumably?

Yes, technically I only noticed this issue when the player was playing clean and I happened to be messing around with the controls as he was playing. The pickup in his guitar added more buzz than is noticeable than when the OD channel is maxed and playing clean but it irks me nonetheless. I just want to remove it on principle but also so that if the player is mic'ing the amp while playing clean he is getting the cleanest, quietest sound. If you turn the OD volume to 5 while on the clean channel there is no added noise, only around 7-10 does it add this artifact to the clean side.

What values would you suggest for the split load? Grid stopper for V2 around 10k?

Well since the existing divider is 1M-150k, I would suggest splitting the anode load into 87k-13k, which is the ratio but adds up to 100k (the larger one goes to the anode, the smaller one to the HT of course). Or you could try a similar ratio using standard values, say 100k and 22k, and the smaller one could again be made a trimpot, to serve the same function as you have now.

You don't necessarily need a grid stopper on V2 since there is still plenty of resistance between the anode of V1 and the grid of V2.

Thanks Merlin, this is very helpful. Yes, I see that the ratio works out the same; very cool. I think I will try the latter solution of 100k/trimmer as dialing in the trimmer pot (56k fixed resistor in series with 100k) in the existing configuration, I found the OD to be most pleasing with the higher (combined) value, so this equates to about 80-100k on the trimmer. Going lower (towards the base value of 1M/56k) the OD gets blatty and unpleasant. For the split load, I'll add the trimmer in front of HT as you suggest and go with 100k for the anode side. This is good as getting rid of the voltage divider will take the amp back to the original schematic anyway which I am trying to stay true to.

[Cliff Schecht]

Leaving the two stages always connected like this is going to have some issues. The noise in your build is just a side effect of budget amp type designs. While it does simplify the design and building of that section of the amplifier, having that stage always connected is going to interject noise and have other crosstalk problems. The easiest way to fix this is to just remove the channel from the signal path with a switch. You could perhaps figure out a simple channel switching configuration to take the OD channel out of the signal path when the clean channel is on.

Or you could get more clever and do the exact same circuit with one less tube. Coming up with an option to simply switch out the first stage of the OD channel will have the same sound as the configuration you posted (assuming you adjust the cathode and plate resistors accordingly). I build a lot of vintage inspired amps but use modern knowledge and materials to my advantage. There is nothing wrong with updating a circuit to fix problems while achieving a similar or identical sound.

[Colossal]

Leaving the two stages always connected like this is going to have some issues. The noise in your build is just a side effect of budget amp type designs. While it does simplify the design and building of that section of the amplifier, having that stage always connected is going to interject noise and have other crosstalk problems. The easiest way to fix this is to just remove the channel from the signal path with a switch. You could perhaps figure out a simple channel switching configuration to take the OD channel out of the signal path when the clean channel is on.

Or you could get more clever and do the exact same circuit with one less tube. Coming up with an option to simply switch out the first stage of the OD channel will have the same sound as the configuration you posted (assuming you adjust the cathode and plate resistors accordingly). I build a lot of vintage inspired amps but use modern knowledge and materials to my advantage. There is nothing wrong with updating a circuit to fix problems while achieving a similar or identical sound.

Thanks for the ideas Cliff. This project was for a jazz purist who was after the real EH150 sound and I did not take (m)any liberties with the design, only a few here and there. The idea was to capture the tone as closely as possible and according to the owner, it does that very well (sigh of relief). He has played a real EH150 and was after *that* tone. Second, he asked for two inputs per channel (high and low, not shown on the example schematic) so a student could play along. So, to have that functionality, there is some compromise. In truth, the added noise is very slight and only when the OD channel volume is way up and the amp is idling with no input is it noticeable. It is remarkably quiet otherwise. He will simply turn it down when playing or recording using the Clean channel.

I plan on building more of these and they are actually fantastic sounding amps. For v2.0, I have been considering a number of upgrades/changes to make it more modern while retaining the original sound. I like the philosophy of more modernized methods as we are not limited as in days of old.

That is an interesting idea about just taking the first stage completely out of the picture for the Clean sound as that is essentially what is happening anyway. The compromise would be to just use one input jack, but it would be pretty simple to use a switch to route the signal around V1 and switching in the clean volume pot at the same time.

Thanks,
Dave

[Cliff Schecht]

/me in a preachers voice

"Can I get a gutshot?!?"