The fact that it oscillates at zero volume confirms that a grid stopper will cure it. At present your volume pot is acting as a stopper, except at zero when it is a short.
The ECC88 family has high gain up to UHF frequencies, so it sees any short piece of wire as a resonator. Combine that with the usual follower instability and you have problems. Add grid and cathode stoppers and all will be well. Personally, I wouldn't worry about carbon comp - you don't need especially low inductance as all you are trying to do is damp an oscillation by adding some resistance.
The ECC88 family has high gain up to UHF frequencies, so it sees any short piece of wire as a resonator. Combine that with the usual follower instability and you have problems. Add grid and cathode stoppers and all will be well. Personally, I wouldn't worry about carbon comp - you don't need especially low inductance as all you are trying to do is damp an oscillation by adding some resistance.
In a ECC88/6DJ8/E88CC/6922/6N23P/other "family members"-Cathode follower, a grid stopper is mandatory, and a cathode stopper is highly recommended - 22..47 Ohms will do nicely.
Paranoid people like me add small ferrite beads over any lead to the socket except for the shield connection, or even a small anode stopper (10..47 Ohms were mentioned as effective by Tim de Paravicini in a discussion I had with him a few years ago)...
Parasitic oscillation of cathode followers is very commonplace - sometimes the only noticeable effect is a seemingly microphonic stage, as the cathode follower has effectively become a RF condenser microphone. Sometimes one gets 'buzz', 'whistle' or increased distortion - effects vary...
Regards,
Michael
=->
Paranoid people like me add small ferrite beads over any lead to the socket except for the shield connection, or even a small anode stopper (10..47 Ohms were mentioned as effective by Tim de Paravicini in a discussion I had with him a few years ago)...
Parasitic oscillation of cathode followers is very commonplace - sometimes the only noticeable effect is a seemingly microphonic stage, as the cathode follower has effectively become a RF condenser microphone. Sometimes one gets 'buzz', 'whistle' or increased distortion - effects vary...
Regards,
Michael
=->
The heater current in EH6922 is 375mA, so they are not true E88CC clones. (Clearly EH wants us to think they are, or they wouldn't call them 6922, but rather 6DJ8 which is probably more likely what they are). I've got 100 of the EH6922s and haven't had one bad yet. I've got a handful of NOS types also, and honestly can't hear much difference. As mentioned, grid stoppers are a must.
Just started this reading this thread. Oscillation was going to be my 1st suggestion. I have a 6922 pre, that was originally for 12AU7. When it was converted no gridstoppers were added. There was a 1.6 MHz oscillation. Grid stoppers cleaned that up (and advanced the performance of the pre).
dave
Another difference between ECC88 and 6922 is that the 6922 (and the E88CC) can be microphonic. The 6922 wasn't meant for audio, so it doesn't have any anti-vibration mica springs (which the ECC88 has).
So the 6922/E88CC *may* be prone to shrill or emphasized high frequencies caused by microphony.
The best tube of the family is by far the E188CC, but it is expensive...
So the 6922/E88CC *may* be prone to shrill or emphasized high frequencies caused by microphony.
The best tube of the family is by far the E188CC, but it is expensive...
Have you actually measured the EH6922 at 375mA? The spec sheet says "not less than 285mA, not more than 335mA". Also, the EH6922 is a 6N23P-EV clone according to the manufacturer.
Maybe something has changed over time, but I have tested many 6N23P-EV, either in their original Saratov/Reflektor ¨MIL¨ incarnations, or as Sovtek or EH ¨6922¨ from various production runs (date codes from 1990ies to 2007), both with shiny and with grayed/blackened plates.
As a routine test, to avoid ´surprises´, I always measure the filament current at nominal filament voltage after two minutes of warmup.
For the 6N23P-EV the spread has been in the interval ´sciticen17´ mentioned in his post...
Regards,
Michael
=->
PS: (1) Interestingly, I had one batch of Russian-made (Novosibirsk 1980ies production 6N23P) ¨6DJ8/ECC88 Gold¨stamped by a well-known British rebrander, which did have filament currents in the spread 355..390 mA - in these, the actual plate construction did look markedly different from the well-known Saratovs...
PS: (2) As for DF96s post: The ECC88/6DJ8 and the PCC88/7DJ8 were originally designed for cascode stages used as VHF input stage in TV tuners, audio application had not even been an afterthought at the time of introduction.
For the E88CC/6922 the application as RF/IF Cascode or mixer stage was the foreseen prime application, but it was declared useable as well in phase inverters and in digital computer circuits as multivibrator or cathode follower...
As a routine test, to avoid ´surprises´, I always measure the filament current at nominal filament voltage after two minutes of warmup.
For the 6N23P-EV the spread has been in the interval ´sciticen17´ mentioned in his post...
Regards,
Michael
=->
PS: (1) Interestingly, I had one batch of Russian-made (Novosibirsk 1980ies production 6N23P) ¨6DJ8/ECC88 Gold¨stamped by a well-known British rebrander, which did have filament currents in the spread 355..390 mA - in these, the actual plate construction did look markedly different from the well-known Saratovs...
PS: (2) As for DF96s post: The ECC88/6DJ8 and the PCC88/7DJ8 were originally designed for cascode stages used as VHF input stage in TV tuners, audio application had not even been an afterthought at the time of introduction.
For the E88CC/6922 the application as RF/IF Cascode or mixer stage was the foreseen prime application, but it was declared useable as well in phase inverters and in digital computer circuits as multivibrator or cathode follower...
Yes I have. Does EH have a datasheet for the 6922? I've only found NOS-brand datasheets for this tube.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.