• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Protection for DC-coupled cathodyne stage?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hello all! I have a Morgan Jones “Bevois Valley” amp which seems to destroy the E88CC valves in the input stage every hundred or so hours of service life.

Symptoms are pops (particularly on startup) and hissing, usually followed by quiet/distorted sound in the affected channel. Measuring between heater and cathode of a failed valve with a megger, the insulation appears intact (well above 20M at 250V applied).

I’ve tried higher-value grid stoppers right on the pin (10k), and replacing the valve sockets, but to no avail. Recently, I’ve elevated the heater supply to 30V or so via a divider off the B+ (330k in the top leg, 33k in the bottom decoupled with 10uF 63V electrolytic). This seemed to fend off the hiss, but I still got pops on startup. I fitted 1N4007 between grid and cathode of the cathodyne sections to protect against possible high positive grid bias and flashover on startup. It’s been a week, and no more pops :) I’m just waiting on some more valves (Reflector 6N23P) to arrive, and then I’ll see if it’s fixed once and for all.

Just thought I’d share my experience, as it’s been a long-standing issue for me. Has anyone run into similar issues? What were your fixes?
 
A look at the Philips 6922 data sheet suggests you've been exceeding the heater to cathode potential limits, before biasing off B+. A 30 V. bias seems inadequate and 90 V. is what I'd shoot for.

There are several "flavors" of Saratov made tubes tubes you can use. The best is the 6H23Π-EB (6n23p-ev), which is both tough and decent sounding. FWIW, the ElectroHarmonix (EH) 6922 is 6H23Π-EB.
 

PRR

Member
Joined 2003
Paid Member
A look at the Philips 6922 data sheet suggests you've been exceeding the heater to cathode potential limits...

I read it as 150V cathode positive of heater.

But at start-up, V1a is cold, V1a plate and V1b grid-cathode go to full 285V for some seconds. {EDIT--- no, only 91V due to 47K:22K action.)

I'd be thinking a 150V Zener on V1a plate. No effect once hot, but limits the start-up surge.
 

Attachments

  • BevoisValley.gif
    BevoisValley.gif
    11.4 KB · Views: 294

45

Member
Joined 2008
Yes excessive cathode to heater differential, despite what it says in the datasheet.
This is a common problem with other circuits like SRPP, mu-follower and cascade using E88CC or ECC88. Best way to do it is build a stereo amp and use shared tubes between the two channels: one tube serves as input and one tube as splitter (each for both channels). Each tube has its own heater supply at the right DC voltage. The PCL82 is much better choice for that circuit as is if no triode-sharing between the two channels thanks to 200V limit for Vkf (that was necessary for series operation in televisions). Still getting around 12W power with 300V anode voltage so no noticeable difference respect to 15W....
The diode is necessary or not depending on how the anode supply kicks in. Generally it is recommendable.
 
This is the exact reason I insist on STAND-BY switches or B+ delay circuits.... I get bombarded with forum beatings that the tubes don't have cathode stripping issues... It's not about cathode stripping...it's about conducting current when B+ is applied to avoid this heater to cathode arc-over....
Another method to avoid this is to simply DC "float" the heater winding and use a bloody big bypass cap on the Center-Tap to ground....
 
I agree that startup would be more of a strain than Vgk.

It could be some detail you haven't described. Are you using the Maida regulator as MJ is? Did you build it with a big resistor divider to draw enough 10-15mA so the regulator is always regulating? If not, the regulator will allow full B+ through before the tubes draw current.
 
110v can be made by 82k resistor from285v and 2x1N5263b zeners , thus avoids caps . Heath w5m is in same problem ,the filament supply is left floating.
Excess of cathode-to-heater voltage for power tubes is a bad idea. Two resistors and a small value filter cap that's all you need to target a more suitable reference voltage. One small value electrolytic cap is cheaper than 2 zeners around here. I can't see any valid reason to avoid the cap. Nothing to do with signal. Done many times and works FLAWLESSLY.
 
Thanks for the suggestions. I took some measurements today to see where the cathodyne is biased, and it’s not as I expected. My Maida regulator is delivering the specified 285 V, but bias is much too high at around -1.0 V. I suspect it’s well into grid current! Playing around with load lines, theory seems to agree with measurement.

I’m thinking of reducing the input stage cathode bias resistor from 830 to 270 ohms, in an attempt to lower the cathodyne’s bias voltage. Otherwise, maybe I should just consider converting the cathodyne to be AC-coupled with cathode bias? Getting the cathodyne bias correct using DC-coupling is proving to be tricky!
 
Disabled Account
Joined 2013
I’m thinking of reducing the input stage cathode bias resistor from 830 to 270 ohms, in an attempt to lower the cathodyne’s bias voltage.


The better and surer way as I gathered is to bias the cathodyne stage first. Find a suitable bias for it for the highest head room and lowest distortion. Then adjust cathode resistor R9 the input stage so the plate of V6 is same as this U1 grid voltage, this voltage should stay the same after the connection is made between the two.

If I have to change IT to cathodyne stage I need to use a 13k resistor in plate instead if IT, but I need to re-bias this U1 stage for optimal performance. If you know simulation then it would save one some times to bread board it out. The sim params shown is quite close to actual thing.
 

Attachments

  • Allied IT stage.png
    Allied IT stage.png
    39.8 KB · Views: 138
The better and surer way as I gathered is to bias the cathodyne stage first. Find a suitable bias for it for the highest head room and lowest distortion. Then adjust cathode resistor R9 the input stage so the plate of V6 is same as this U1 grid voltage, this voltage should stay the same after the connection is made between the two.

Thank you, this is close to the approach I was going to take. I drew a new cathode load-line for the input stage so as to intersect the anode load-line at about Va=90V (my desired Vg for the cathodyne). Then, I calculated dV/dI of that cathode load-line to get a resistance of about 240ohms. The downside is that, by changing this cathode resistor, I alter the bias of the input stage, but hopefully not enough to enter grid current with signal applied. It also alters the proportion of gNFB applied.

I've just found the article for the ImPasse preamp, which uses another method to bias the cathodyne - a voltage divider off the input stage anode. Possibilities... :)
 
I use a 30sec delay circuit - at power-on, the filament voltage comes on immediately, but there's a delay before the AC is applied to the rectifier diodes.

This means that the moment the tubes see HT they're already conducting and voltages won't surge.

The other option is the arc-protection diode as described by Blencowe in the Cathodyne chapter of "Designing High-Fidelity Tube Preamps"
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.