Noisy high-gain preamplifier

[jesusdecos]

Hello folks,
Sometime ago I posted a thread about a preamplifier that I built Building a mu-follower stage It is intended to amplify small-signal output from high-impedance transducers of musical instruments. It has a Baxandall tone stack to be more versatile. Here you can find a picture of the thing:
IMG_2289.JPG
Measurements show good agreement with theoretical calculations. OK, I found a problem in practice which is some crispy sounds listened when switching to on mode. When turning up/down the volume pot those sounds are increased. After having the preamplifier some time in on mode, let's say 20 minutes, the crispy sounds completely disappear, even when turning up/down the volume pot. Any idea about this behavior?
Jesús.

[jesusdecos]

Schematics attached.

[SY]

Whenever there are odd noises with ECC88, I always think "oscillation" first. You might try larger gridstoppers. 1-2k, perhaps. Make sure the body of the resistor is close to the tube pin. Then a cathode stopper, a few hundred ohms, again tight to the tube pins.

[richwalters]

SY's got it....The ECC88 will hard hog oscillate that will faithfully run on it's own accord around the 470 MC/s frequency range without it's owner being aware of this. I'v often wondered how such a tube could be designed, but its true life as a 1950-60's Colpitts TV tuner and or FM tuner head with it's ample inductance quenched self oscillation.
Your problem is caused by the audio signals being modulated by the RF oscillatory modes into a mini AM/FM mini transmitter and the preceding stage struggling to demodulate the mixed signals into a detector.
A voltmeter connected to a decoupled cathode (keeping things still) then slide the finger down the suspect tube envelope....any capacitive disruption to the oscillation will disturb the readings. A screening can will also indicate.
There must be alot of ECC88 designs out there that are oscillating and owners unaware that it is doing so. A cure is to decouple right on the socket tags and a ground tag very close to. The pic I've enclosed doesn't make for easy recognition of the components and layout as it often when the solution ends up as a bird's nest.The trouble is once a cure has been achieved, it's not uncommon for the wretched tube to shift it's oscillation to a frequency higher up. This one of the penalties of using RF tubes in hifi applications......the excellent bandwidth ft.
I've enclosed a pic of my standard solution; bit difficult to follow the underside of an ECC88 paralled cathode follower. The anode & grids are particulary responsive to decoupling and ferrite beads, the ECC88 WILL NOT WORK properly with lanky wiring schemes and a good quality digital scope with an FFT analyser makes things much easier. The solution like the 1960's FM tuners will most probably result in a "birds nest" with decoupling components jostled for the shortest position.

Another solution is find a tube which is tamer, but sonically the ECC88 is the best one gets.

richy

richy

[jesusdecos]

Hi again,
Thanks for the feedback, guys. My first suspiction was that the thing was oscillating. I couldn't see any evidence of oscillation in a RF analog spectrum analizer, up to about 1 GHz though. I made the measurements quickly and the spectrum analizer isn't among the best ones. I'll take some more precise measurements in a better microwave spectrum analizer that starts in 3 kHz up to some GHz. In oscilloscope measurements I noticed the effect of noises as a random AM modulation of the amplified tone. The 220 Ohm grid stoppers aren't tight to the tube pins, I must fix that. I may consider adding RF ferrite beads to grid stoppers as richwalters pointed out, as close as possible to socket pins. I hope that cope with those odd noises.
Now a couple of questions. Do I need to add grid stoppers to the cathode follower section or just to the common cathode section? I weren't sure about this and I included grid stoppers in all grids. Cathode followers are a must? How do I add those cathode stoppers? Unbypassed?

[richwalters]

Do I need to add grid stoppers to the cathode follower section or just to the common cathode section? I weren't sure about this and I included grid stoppers in all grids. Cathode followers are a must? How do I add those cathode stoppers? Unbypassed?
-----------------------------------
My circuit uses the cathode bootstrap to raise grid impedance so the likelyhood of oscillations is higher.

I ended up fitting ferrites then removing and seeing. This tube played me decoy tricks and between tube makes there wasn't much change. The decoy is; a capacitive probe connected to the circuit can dampen the oscillation and suddenly none to be seen. Since we are dealing with possible self oscillation RF energy, Use an non intrusive loop or a couple of turns as (in pic) to an analyser and spy around the circuit. One will pick analogue FM radio broadcast signals...one must differentiate. There isn't another way; just the good old hard analogue RF stuff.

Of course I might be barking up the wrong tree..could be something else...heater circuit troubles ?

richy

[Thatch_Ear]

I tested a quart bag of used 6DJ8s pulled out of a Tek scope for a buddy who was trying to get it working. I went over to use the best Hickok made which is the kind that uses punch cards. A 12AX7 and other tubes like that use 1 card and since it is a dual triode you press a button to get a reading on the second side. Makes sense.
A 6DJ8 takes 4 cards and the button because it is a dual, so there are 8 things that can go wrong with one of those. You use a tester that just tests the plate and grid chances are a lot of tubes that do not work correctly are going to pass with flying colors.
I don't recall any HiFi from the days when Tek scopes were built that had a 6DJ8 in it. Certainly nothing made in the US used them. They are easy to design around and they are worth the buck it cost to replace one.
All this gear that uses them is probably a passing fad. They were designed for industrial use that had nothing to do with audio at all.
It can be argued that many of the tubes used in audio were not designed strictly for audio use, in fact only the 12AY7 was designed with only audio in mind, but most of the tubes used in audio were used in radio and the ones that sounded best were used in amplification and built upon. The 6SN7 is a direct descendent of the 27 which was used as a power tube in radios in the 30s. The 12AU7 was supposed to replace it but isn't as good.
Look at what the tube dealers offer and most of what is in the market for 6DJ8s were made after 2000 and the rest were made for military use and not as replacements for anything having to do with signals, they went into test gear.
If you have a problem with a piece of gear that uses 6DJ8s the first thing to take a close look at is the 6DJ8.
You want a high gain preamp use little pentodes like the EF86 or a dual triode like a 12AX7 with one side driving the other. I think the problem with the 6DJ8 is that there are just so many things that can go wrong that a good percentage get past QC and into the market as good new tubes.
Buy NOS made for the military. JAN 7308s aught to last longer and have much better QC.

Thatch