diyAudio (
-   Tubes / Valves (
-   -   Glowing Plates of Doom (

TV7 24th September 2012 04:41 AM

Glowing Plates of Doom
Now that I have your attention can I call upon the gurus to help me with a problem....:D

I just finished building a phono preamp head unit ( no power supply yet) and based the circuit around the Mcintosh MC2200 preamp (see attached), this circuit was almost exactly the same as on paper, the only differences are when I used more standard values of parts that I had available, I.E a 1K resistor in the first stage cathode instead of the 953 ohm....

All voltages are correct to within a couple of volts in the CF stage, the heaters are up at 100V from a decade of two 100K in series from the 200V rail. (CF at 135V) So plenty of headroom there....

The sound? well to my ears absolutely fantastic! who would have thought a feedback RIAA would work so well. But here's the problem.....

When running into my MC275 the plates on the KT88 run red hot. Even at low low volume. This surprised me because I am limiting the volume input with the amps pots and the output from the preamp is no louder than standard RCA jack levels. There is no DC on the output.

After a minute running hot I had a small flash over. This convinced me not to try again until the problem is resolved. Perhaps it is some sort of out of band frequency, sub or super sonic that is driving the amp crazy. I would have thought I would get some sort of indication but the sonics are sweet and non-distorted.

The HT is 270V from a 6l6 regulated bench top supply, (very clean) the heater supply for testing is a 12 Volt car battery.

I would have build SYs phono preamp but who can afford the parts? this preamp was made with all leftovers, including a chassis bought in Japan 10 years ago.... hence some odd looking bits and pieces....

Help me Obi-Wan :D

HJWeedon 24th September 2012 11:44 AM

Hi TV7

I may be the "guru" you like to get advice from.

Using standard 5% values for the components is in my opinion quite OK. The effects from variation of gain from tube to tube is probably much greater than the effects of the resistor tolerances. The simplest "fix" would probably be to replace the output cathode followers with 12AU7s they should not oscillate as willingly.

Without knowing which tubes are used, it is difficult to tell why something is oscillating, but you are probably correct in your assumption that something is oscillating at supersonic frequencies. My guess would be the output cathode followers in your preamp. I would put a 100 ohm resistor in series with the output cable. I would also put about a 1k resistor in series with each of the cathode follower grids. In other words a total of 4 places. the purist would perhaps say that the resistors would add something in the signal path, that is true, but the effect would be in the several tens of MHz range which is totally inaudible. In addition 220 ohm resistors in series with the anodes of the cathode followers would help as well. Since the schematic does not indicate the tube types it is difficult to tell why the stage is oscillating.

Before you embark on the modification trail you should borrow yourself an oscilloscope to verify that my supersonic assumption is correct. The oscillation may as well be your power amplifier as your preamp. I have seen power amplifiers oscillate with a specific length of input cable as well as cathode followers oscillate with cable loads.

Power tubes glowing red is very bad for the tubes especially power pentodes for audio use, the screen grids can melt in a flash without warning.

I hope this helps.

Hans J Weedon.

TheGimp 24th September 2012 12:56 PM

10K-47K would be more appropriate for small signal triodes as grid-stop resistors.

The only critical component values are those of the RIAA network. This is true whether in the feed-forward or feed back-path.

Note that R93 is part of the feedback network as well as C64 and R96. Although the latter two are less critical as their frequency is near 1Hz.

TV7 24th September 2012 01:02 PM

Thanks for the info Hans, the first amplification stage is a 12AX7, the cathode follower is 12AT7. After reading your very helpful post I will insert blocker resistors (is that the term?) in the signal path and the anodes for the CF. The KT88s are GEC NOS that I really shouldn't be messing with hope they made them to last.... And Gimp, yeah the RIAA is probably the only part of the circuit I was critical with, for the reasons you gave. Thanks again.:D

HJWeedon 24th September 2012 01:43 PM

Hi all.

I agree that resistors in the range of 10 to 47k as grid-dampers are more appropriate to quiet cathode followers from oscillating than 220 ohms but that usually all it takes. Perhaps the simplest thing would be to replace the 12AT7s with 12AU7s, pin for pin compatible and considerably less prone to oscillations. I would still add an inconsequential 100 ohm resistor in series with the output cable though.

Hans J Weedon.

PET240 24th September 2012 08:33 PM

Thanks Gents,

Very handy info!



TV7 24th September 2012 10:01 PM

Will try the solutions and the 12AU7 Hans, although how would a series signal resistor of 10K to 47K affect the CF bias?, possibly some. Maybe not at all? Thanks

tomchr 24th September 2012 10:50 PM

So you're saying that if you use this RIAA stage the plates of the output tubes glow red. But if you hook a different source to the power amp inputs, the output tubes do not glow red? Is the RIAA amp oscillating at RF by any chance?


TV7 25th September 2012 03:05 AM

Yes! dug the cro out of storage and low and behold a 1.8MHZ oscillator! would be a good radio transmitter if you discount the freq. shift ect... now to insert resistors and maybe change the at7s for au7s...... report back soon...

HJWeedon 25th September 2012 03:40 AM

Hi TV7

Glad you found the 1.8MHz oscillation. since the oscillation frequency is that low, we can discount the cable as the resonator, it is most probably the tubes and the feedback circuit that resonates. The 100 Ohm resistor in the output circuit should not do much. Changing to 12AU7 should not do it either, in my opinion the remedy would be to put 4.7k resistors in series with grids of the cathode followers. 4.7K does not change the bias of the cathode followers because normally the grid leakage current in a properly operating vacuum tube is in the low micro-amps.

I looked at the neat and tidy layout of your circuit. On second thought 1.8MHz oscillations could be caused by improper bypassing of the power-rails. A close by 10 to 47uF cap to chassis might do the trick. Use an electrolytic cap with sufficient voltage rating for bypassing the rails. The reason is that regular electrolytic capacitors are quite lossy at 1.8MHz and would dampen any resonances in the power wiring. You may have used film capacitors for bypassing the power rails. All film capacitors have very low losses at MF frequencies, and may form unwanted resonant circuits.

That is all for now. Please let us all know how you got rid of the oscillations, it will be a great learning for us all.

Hans J Weedon.

All times are GMT. The time now is 11:32 AM.

vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2017 DragonByte Technologies Ltd.
Copyright 1999-2017 diyAudio

Content Relevant URLs by vBSEO 3.3.2