Anyone experienced the sound of DC Cathode Follower valve degradation?

[bordonbert]

Hi.

Thanks to the efforts of Nickb in his JVM Cathode Follower thread we have a fair bit of insight into the problems that valves in a DC Cathode Follower (DCCF) role experience. He has pointed out with supporting evidence that the usual stress issues quoted are most likely not the real problem and he has gone a long way towards describing what that real issue could well be. That said, it doesn't tell us how the issue, whatever it may be, will begin to affect our playing use of the amp in terms of sound. Ultimately the valve will fail and should be fairly "noticeable" then but in the early stages what changes would you first hear?

Does anyone have any history of living with an amp with declining sound which turned out to be the DCCF and was cleared with only a replacement of that valve? I'm currently a user of an actual Marshall JVM205H, the amp which kicked off Nickb's exploration, (and a Hughes & Kettner GM36 occasionally), and it has a problem with a dullness and lack of sparkle drifting in after 30-45mins of playing. At startup it sounds fine, lots of top and the treble control working as it should. Then when the amp heats through thoroughly that top end slips off to a dullish quality, not distorted in any way, where the Treble and Presence controls appear to do less and need to be maxed out to partially compensate. It isn't desperate but it is noticeable. The issue is the same for all voices of both channels. Each one has its own dedicated gain control/tone control setup and uses a different selection of valves for its gain stages.

The amp had a full valve swap about 6 weeks ago and it was sounding great but the background to that DCCF issue in Nickb's thread has shown that that valve can go downhill pretty quickly given the stress. I've just swapped out the DCCF valve for a new Shuguang 12AX7B as per the advice that Chinese valves are more resilient in this area but at last night's gig it seemed to make no difference. That said I wanted others' knowledge as to the sound of a DCCF valve gradually going down. It's a follower so won't necessarily sound like any old gain stage as they go.

Of course other parameters could be the cause. I use the amp with a '60s Marshall 4x12". I now have this set up with a solid wood divider down the centre and switchable 2x12" sides. These are currently a pair of Celestion G12M Greenbacks on one side and a pair of Celestion G12H Heritages (one original '60s + one modern) on the other. These both sound fine in both separate and series/parallel modes, with the G12Hs maybe having a bit more top end and breaking up a little earlier than the G12Ms. I know they can be a bit laid back but I'm not looking for an in your face raw raucous sound as I play only classic rock, (real classic rock, late 60s through 70s, think Humble Pie/Free/Bad Company), so I am looking for a strong overdriven sound but certainly not getting close to metallic.

Guitars could be a factor too. I use a number of them for quick change different tunings. (As an aside, controversial though it may be, I love the Tronical system and could use that but it takes a little fairly quiet time and a bunch of guys around you who will give you that, and onstage string breakage is always a worry though it has never seemed a bad issue to me). I use either a PRS Bernie Marsden or a Gibson LP Traditional as standard, a Gibson LP Junior for open E and a Gibson LP Special DC for open G (guess who!). Nothing outrageous there. These are fed through my own buffer system, (I'm ex-electronic design), with a tiny amp in the guitar jack plug feeding a preamp at the pedalboard end. It is absolutely unrestricted in our audio range even with a little overkill, (10Hz-15kHz with RF reduction), and has headroom up to 2V pk-pk. It also does not use standard voltage amp techniques, it's a current based system immune to the usual noise and has proven reliable for years during which I didn't have this problem. The problem is the same even with this removed and a standard length cable in place.

Any experience you can offer?

[R.G.]

As one more extension to the MOSFET Heresies, I could point out that MOSFETs do not have heaters, and they make a dandy-fine DC follower.

See post here and articles on MOSFET Heresies at geofex.com.

[g1]

In my experience they don't 'gradually go down', they just fail.
I don't think your issue is related to the cathode follower 'syndrome'.

[nickb]

"Loss of sparkle", which I take to mean of of treble isn't something I'd expect from the CF dying.

Try this, plug in and play for 30 secs. Now walk away, leaving it on. Return after 45 mins and try again. Have you really lost HF or could it be a perception issue? Let's suppose you come down on the side of amp problem then I would set it up with a 2Khz sine tone running it through. Measure the output 2 min after switch on and without touching any controls repeat after 45 mins. If there is a difference then you can start the process of isolating it.

[bordonbert]

I must admit a drop in treble isn't what I would have expected either Nick but all of my real experience has been with solid state. The valve side, while my basic theory may be fairly sound and I pick it up easily once it is pointed out, is not so broad in the day to day stuff for issues like these. I always feel it is better to get a bit of backup from others with a more comprehensive background. I had been getting by on the suggestion you made , that it was just perception. It's even an argument I make regularly myself in the great "tone cap" and "sound of valves" debate. However, I can't keep fooling myself for ever and I am now having to investigate properly to be able to sleep at nights once again.

There is also a point about venues and crowds to be considered. I mostly play smallish clubs and pubs and a brief soundcheck may be all we get when the place is fairly empty. An hour or two later the room has filled with people, (you would hope!), and their absorbent quality comes into effect. That is a contributing factor I'm sure. I can't really test this like for like at home with the much smaller room. Oh, and the neighbours! You're right, I really should try to set up something like the simple tests you pointed out, they would set my mind at rest one way or the other.

I know there will be a number of simple Go->Nogo occurrences of this G1, I have always hoped it would be as simple as that, but I wondered if there was another way they can dropout gradually based on Nick's info re their degradation due to plain overcurrent. I can't really envisage what would be the eventual effect of gradually reducing amplification capability in a valve when it is set up with 100% feedback. The DCCF is a pretty different beast in its operation compared to the usual gain stages employed. That 100% feedback can cover a multitude of sins but up to what point? And also Nickb's findings that the valves can actually be recovered is one I really would not have expected.

As to MOSFETs RG, I would be with you in your praise of the solid state which always gets such an unfair rap with one exception. There is no MOSFET equivalent that I know of to the forward biased grid current method of using the DCCF to generate even order distortion which is - nice! (Diode/zener chain in parallel with the preceding anode resistor?) And both of my working amps, Marshall and Hughes & Kettner, do use this method as a way of sweetening up. But certainly, I'm no proponent of the "solid state is always a bad thing for tone" bit of "common knowledge". I always use my own in jack plug buffer + preamp for everything I do live now, (commercial versions are always an expensive compromise to me). It removes all problems with loading the guitars as that can be done in the plug and/or in the guitar and it gives a consistent base sound no matter what happens between guitar and amp. This has no effect on sound, it's neutral and absolutely silent, in any way that I can detect. Or anyone else on either side of the argument either, at least until some of them hear it has those tone sucking spawn of Satan in there messing with the signal, as we all know they must!

[nickb]

.... it has those tone sucking spawn of Satan in there messing with the signal...

I have to confess I had completely overlooked that as a possible explanation

[pdf64]

Of course there's the hearing compression effect, but I have a suspicion that in heavy use, as the whole magnet assembly heats up, the tone a speaker produces may change subtly, especially those with vintage type unventilated magnets (eg Greenback), moreso those also in sealed cabs. Dunno if it's purely the copper resistance changing or what, my hypothesis could well be nonsense, as the tone from close miced recording seems not to exhibit the effect

[bordonbert]

I have to confess I had completely overlooked that as a possible explanation

We-e-e-ell Nick, I'm sure this is a underground issue on this forum the same as it is on all of the others and I don't want to drag it out of the mud at the bottom of the pond but I always believe in speaking what I know to be true from my own work and investigations. That is what I hear all too often on the issue of solid state mixed in with vacuum. It ranges from, "you can't put an opamp into the signal chain without imposing the solid state signature on it" quoted by people who have cobbled together 53 different distortion pedals (12 of them worked, almost) all based on their favourite boutique type at £200 a pop which is itself based on a good old Big Muff but populated with obscure NOS Mongolese trannies and diodes from the '60s in the "choice" places at £15 a time, to, "it has solid state controlling the reverb switching? It's polluted, they've ruined it!" from people who have never handled a component in their lives but they talk to a few "gurus" who do.

The Hughes & Kettner GM36 I use is an awesome amp for the right music. It is a German High Gain Metal monster, (I don't play metal). It uses a simple two 12AX7 preamp and an all valve power amp and augments that with a neat little solid state clean buffer/TS type overdrive stage on the input and solid state active stages around the Gain, Tone and Volume controls. Each of those solid state blocks is very well designed. I have modelled each of them in Spice at length for analysis so I know them pretty well. They have almost a hifi like "lack of effect" on any signal passing through them unless the effect is intended. Opamps are seamless and invisible if used within their design parameters. They add nothing and take nothing away and for guitar audio use there are no magic types, serious analysis genuinely shows that even the venerable TL071 class is inaudible amongst much more expensive types. It's not a demanding field for that aspect.

It's not the classic path and that path is still a great one but it is not the obscenity it is made out to be. Inaudible SS stages do the mundane jobs well and remove interactions making it easy for the designers and valve stages give the tone. When that info concerning the GM36 came out into the world the amp which so many had raved over became a veritable plague dog for a huge proportion of players as it carried the horrendous virus of siliconitis. So many missed the fact that it was the same amp they already loved and it was only the knowledge that it had opamps and transistors in there which had changed their minds so it now sounded bad.

If any component has an audible effect on a signal then that audible effect can nowadays be identified and analysed on the bench, test equipment is that good. Any claimed audible effect can be proven to exist or not with blind listening tests. If you can genuinely hear it tests will prove that, if it genuinely exists we can measure it, if we can't measure it you genuinely can't hear it! (As you can sense, that level of fundamentalism ticks me off. Maybe to the point where I have become a fun-DAMN-entalist myself? If it gets too bad then just let me know by hitting me over the head with an output transformer! )

Anyhoo, I digress and I proselytise. Back to the ward nurse...