Ghost Notes From Marshall, Bassman, Etc.

[Daver]

OK, maybe it's me. I've played several Marshall or clone amps as well as the tweed Bassman style amp. All seem to exhibit ghost notes to one extent or another when cranked and using the neck pickup. I know grounding and filtering can be a factor. Is there something inherent in the design?

I recently built my own take on the JTM45/Bassman and I get the same thing. I have used a star ground scheme and upped the filtering to 47uf for the PT plates and screens, 22uf for the PI and cathode follower stage and then added another, separate 22uf filter node for the first preamp tube stage. The preamp cathode bypass caps have been reduced in size. All this has helped, but the ghosting is still audible.

So is it this style of amp, something I'm missing or am I just sensitive to it?
Thanks.
Dave

[Diablo]

According to Aiken, ghost notes are a result of inadequate power supply filtering. There is ripple at high current draw and this makes the ghost notes.

What are you using for a power transformer, choke, and main filter capacitors in your JTM45?

Sag

[Daver]

Diablo,
As I mentioned above, I have 47uf filter caps each for the power tube plates and screens. I'm using a 3 Henry choke and my power transformer has 250ma of current available. I've heard this in other amps of this style, so I'm wondering what the common factor is here. Thanks!
Dave

[Chuck H]

It's the ground scheme. There are even flaws with star grounding IMHE.

Don't ground the preamp filters in the same place as the power tube filters and PT. A two or three star ground system works better.

Ground power tube filters, PT, OT, cathodes and bias supply on one star close to or right at the place where the AC ground is. At mid chassis, on another star, ground the presence, PI circuit and remaining filters. On a third star near the input make all preamp grounds except filters. ALL GROUNDS MUST HAVE THEIR OWN LEAD. No daisy chaining. Forget about grounding the backs of the pots and leave them to the chassis connection.

You must use low impedance (old school is low ESR) caps.

Try to locate star grounds and layout components so you can use the shortest leads conveniently. This applies to grounds in high gainers as much as AC lines. Not that the JTM45 in a high gainer per se, but most players like to crank them so...

Do this and ghosting will be minimized. So much so that you won't have to think about it.

Chuck

[Daver]

Chuck,
That's an interesting way of doing the grounding. I would have thought you would want to have the filters for the preamp input stages grounded at the same point as the preamp cathodes. I would also have thought the PI and Presence would get grounded at the same point as the output section. At any rate, I'll give your method a try and report back. I suppose the other amps I've noticed this problem in also had grounding "problems" as well. Thanks!
Dave

[pdf64]

If all these amps are being used with vintage style speakers with lightweight cones, you may be hearing cone cry. This is cone resonances which are not harmonically related to the input, in the same way that ripple on the B+ isn't with ghost notes.
Try these amps through a heavy duty speaker like an EVM12L to rule this out. Peter.

[Daver]

Peter,
It doesn't sound like cone cry. I've tried other speakers and used an attenuator to lower the volume and it still happens. Other amps into the same speakers don't exhibit the ghosting. Thanks.
Dave

[Chuck H]

Originally Posted by Daver

I would have thought you would want to have the filters for the preamp input stages grounded at the same point as the preamp cathodes. I would also have thought the PI and Presence would get grounded at the same point as the output section.

Absolutely. And I've done it that way too as it is "correct". But because of the fraction of resistance on the chassis I find that locating the preamp filter grounds between the preamp and the main chassis ground actually works better. I put the PI grounds there for the same reason. The presence is gounded there only out of convenience and hasn't been a problem. It's important to remember that an equivalent length of wire still has more resistance than the same distance on the chassis. So some of my grounding scheme is arranged to keep the leads from being too long. Often ignored, but long ground leads are very bad about having a small amount of voltage at their head and this is can be a source of ghosting, hum and channel bleed. If you locate grounds strategically to avoid crosstalk and keep short leads to avoid resistance, half the battle is won. Filters with alot of equivalent series resistance leave too much power supply noise riding on the preamp plate circuits. So good filters are also very important too. This is my experience anyway.

Chuck

[Daver]

Chuck,
OK, I rewired my grounds as you indicated. The ghosting did get better, but was still there. So I changed the grounding scheme again. I put the grounds for the PI/Presence and its filter at the same point as the power tube grounds. I tied the grounds for the second stage/cathode follower, its filter, the volume controls and the tone controls together at a point in the middle of the chassis. I then have the input stages and their filter grounded at the input jack. That made it a little better still. It's still there, but you need to listen for it. I'm using Tech Cap 500V axial electrolytics for all the filters except the input filter which is a 450V Nichicon radial. I've used these caps before without problems. So have I gone as far as I can? Thanks for the help!
Dave

[Chuck H]

Originally Posted by Daver

So have I gone as far as I can?

Probably. Remember I said "Do this and ghosting will be minimized. So much so that you won't have to think about it." These amps do seem to ghost a little no matter what. You can even hear it in professional recordings (some Angus solo's come to mind). Good work on locating the better ground points. That's the game. Try to "see" the ground path and assume there IS signal on it. Then locate the grounds to minimise crosstalk while still keeping leads as short as possible.

There may be one other thing that can help, but it's only a hunch...

I can't do this with my 1959 right now because it's in dire need of a full overhaul (6 big dual cans and a full set of power tubes, at least). Not in my budget right now. But since your amp is in good working order it would seem like a good place to try it...

Since ALL Marshall's ghost a little and ALL Vox's with a cathode follower do too, I suspect the cathode follower. It does tie the B+ rail to the signal chain without any resistance. So I'm thinking that since the ESR on your average filter cap is going to be somewhere between a 20 and 150 ohms (double this if your main caps are in series like most Marshall's) that's alot of ripple to tie directly to the signal chain. What if you put a smallish resistor, like 1k, on top of the cathode follower (instead of the B+ being connected directly to the plate it would feed through this 1k 1W resistor). I don't think it would audibly effect the amps performance or drop any significant volts. But it would create a significant voltage divider as seen by the 60hz on the B+ rail possibly reducing ghosting to 1/3 or 1/5 what it was. It's worth a try anyway. I'd use a 1W rated resistor for this as it's feeding both halves of the triode. Even though there won't be much draw across a 1k resistor it's best to play it safe.

I'd be very interested in how it works out.

Chuck

[Daver]

Chuck,
Good idea. I just tried it and it didn't change anything. Bummer. I then clipped in another 22uf cap at this new node for giggles, but that didn't change anything either. I think I may try eliminating the cathode follower altogether and see what happens. That'll take a bit more work. I'll report back. Thanks!
Dave

[Chuck H]

Oh well. Now we know.

Since you have it minimized now, try not to obsess about it or it will eat you up and make it impossible to play freely in your head. Angus doesn't seem to be hampered by such concearns , for example.

Chuck

[daz]

Chuck, can changing the ground scheme have tonal affects too, or is it only responsible for minimalizing noise and ghosting etc?

[Chuck H]

It's definitely most responsible for controlling noise, hum and ghosting.

But consider this. Since it affects ghosting there is obviously some 60hz signal riding on the ground leads, right. So the amps audio signal can ride there too. Any coupling via the ground leads can and does create small NFB and PFB loops in the signal chain. So in that respect tone is affected. Usually the amount of amp signal on the ground leads is so small that this phenomenon can be neglected. But sure, it can make a difference. The average player probably wouldn't notice unless it was grossley bad, as in bad filter caps with very high impedance or really sloppy ground lead dress. But if your super finicky you could get a little nutty trying to figure the best signal ground lead arrangement to avoid unwanted coupling. Since there is always coupling you'd first have to determine what was unwanted. So you can see where this line of tweaking could really get out of hand.

Chuck

[daz]

Thanks. What do you think about the way i did it....One single ground for the entire amp which is close to the power input and PT. The preamp and input are all connected to a heavy solid core bare wire that runs the length of the turret board and all the preamp grounds including the filters are connected to it, then the end of it at the PI end goes to the one common ground. Does that seem ok? It's a pretty darn quiet amp if i do say, and i never noticed any ghosting. But this has me wondering about how different ways to do this could affect the tone in possibly a good way.

[Chuck H]

Provided the buss wire you used is a nice big heavy one, what you have is not dissimilar from the way I do it. But don't pin it on me to be the authority on this. In fact, in this very post a better scheme than the one I suggested was discovered. Sometimes it's individual to the layout. I don't use a buss, I use the chassis. The chassis can be your "buss" if you follow the idea that distance = resistance no matter how you slice it. And since chassis are usually a nice heavy gauge they have lower resistance than buss wire. But when you consider the resistance difference between the chassis and a fat copper buss wire it becomes moot. Only if you were to use a small gauge buss wire would it matter. Then it would be as if you just daisy chained your grounds. With standard 22 or 24 gauge lead wire that's a bad idea.

Chuck

[daz]

thanks. I think i used 18 gauge if i recall.

[Chuck H]

I was actually thinking larger than that. But if your amp is quiet it doesn't matter.

Also, I use aluminum chassis (because they're easier for me to punch and cut) which has no electromagnetic properties. If I built an amp on steel I would probably use a buss because EMF from the transformers is supposed to be able to linger on the chassis. In that case I would think that a buss connected to the chassis ground would be the quietest arrangement.

Daz, how much ghosting do you have in your build? A little, a typical amount or a lot?

Chuck

[daz]

To be honest i never noticed it. If what you said is right, then i guess it's there but i can't hear it. then again i don't usually crank it way up either. Loudest i ever play it is about bar gig volume, which when i played out was about 3-4 on a JCM 800 master for example. Thats fairly loud, but you said a amp thats good in this regard will still have it but only noticeable when cranked. So maybe i'd hear it if it was on 10. I've had it there just a few times early on in the development but i didn't notice it then. Maybe if i had played it like that longer, i dunno. And amazingly mine has small filters ! I had big ones for a couple weeks recently but went back to the 32/32uf can and 22uf preamp nodes because i started realizing i didn't like that stiffness.

As to the chassis, mine is aluminum for the same reason. But i wonder about the sonic difference and whether it's really a determent or not to use steel. I have a 6V6 build that started life as a 18 watt and i re-did the preamp exacly like my EL34 amp, but it soumnds world different. It has a much more traditional marshall high gain preamp tone. My EL34 amp has it's own sound...marshallish, but with it's own unique character. The 6V6 amp has a much more traditional marshall tone, and i really like it and wonder is it could be due to the chassis. I've tried the JJ 6v6 tubes in my EL34 amp before and it still sounds basically like it does with el34. There *could be* something to the aluminum vs steel debate don't you think?

[Daver]

Chuck,
I reconfigured the Cathode Follower as a paralleled stage using the same plate and cathode R values. The ghosting once again is a tiny bit better. I think I'll "think Angus" and ignore the little bit that's still there. I do like the tone better without the CF, but it's a tad thinner. I'll need to juggle the tone stack or cathode bypass cap and see what I get. Thanks for all the input!
Dave