Increasing life of tube by reducing cathode current - will it work?

[AusTex64]

I'm working with a circuit that uses 6DJ8 family tubes in a White follower topology on the output. No output transformer. Problem is tube life is about 20 hours before emission starts to drop on the tube. Guessing that the tube is being run too close to edge of design. Am considering lowering cathode current/emission with the hope of extending tube life. Will it work?

Another question is which tube of the family might offer the best life? All are 6.3V heaters on a regulated DC supply, BTW.

Here is some data on my options:

6DJ8 - 365mA heater current, 1.8W anode dissipation per side.
https://frank.pocnet.net/sheets/191/6/6DE4.pdf

6922 - 300mA heater current, 1.5W on single side, 2W max on both side operation simultaneously. https://frank.pocnet.net/sheets/049/6/6922.pdf

7308 - 335mA heater current, 1.65W anode dissipation per side.
https://frank.pocnet.net/sheets/168/7/7308.pdf

I always thought that all these tubes were basically the same tube, like 12AX7 and 7025 are related. 7025 being a noise and vibration selected 12AX7. But all these tubes are obviously different in heater, cathode and plate construction, to offer different performance tube to tube.

Thanks!

 

[DaveP]

I always design my circuits so that the tubes run about 60% of max.

If you have the space, use single triodes or triode wired pentodes like a 6JC6A, there are a lot of under used TV tubes out there for a few bucks each.

Best
DaveP

 

[AusTex64]

Thanks Dave. That's an excellent suggestion.

I failed to mention this is an existing piece of gear that I have to work with, not a new design. So I'm kinda stuck with trying to make what's there work best. 

 

[PRR]

> life is about 20 hours before emission starts to drop

That's totally wrong. Raw emission should be 300+% of maximum rated current for several thousand hours.

Reducing heater power will make it worse.

Note that most tube testers WILL eat-up the cathode quickly. The "good/bad" indication flows MAXIMUM emission current in diode connection. Limit testing to a split-second, don't hold the button and burn the life out of it.

What cathode current are you actually running?

What heater-cathode voltage?

Golden-age tubes or new-made? (I guess 6922 and 7308 must be old-made.)

 

[AusTex64]

PRR, thank you very much for the reply. Here are some answers to your questions:

Cathode current - 10mA

Heater is approx. 6VDC

Cathode voltage is approx. 3.14 VDC on the lower half of the tube differential of 2.86 VDC

On the upper half:

Cathode voltage is 151 VDC

differential of 145VDC

Tubes are old, you are correct. Haven't had great luck finding modern tubes that sound as good and are made as well as old ones. The 7308 are Amperex and the 6922's are RCA branded Amperex (I think).

 

[ruffrecords]

Posting a schematic would be enormously helpful.

Cheers

Ian

 

[AusTex64]

Ian, thanks for your post. Agreed. Unfortunately I don't have a schematic that I'm at liberty to post.

 

[gyraf]

Something is clearly wrong design-wise.

You don't want to run the 6DJ8-family on much more than ca. 100V cathode-to-heater differential.

Also, this family was designed for 90V operation, with an ABSOLUTE MAXIMUM anode voltage of 125-130V (depending on manufacturer) - which you seem to exceed significantly.

If you need such, maybe try looking into the 12BH7 or even the Russian 6N23P types.

That said, MANY modern-production 6DJ8 type tubes shows exactly the early drop in emission you mention - have you tried with any other NOS types to check for this?

What sort of power do you dissipate in each triode? What is your supply voltage?

Last: What is the application - and why can't the design be changed to fit with reality? (my only guess would be in cloning something which was once "repaired" in wrong ways. Demeter?)..

Jakob E.

 

[merlin]

What sort of power do you dissipate in each triode? What is your supply voltage?

From his description you can see it is almost 1.5W per triode and 300V supply. That's right on the datasheet limit for dissipation (both triodes simultaneously).

What is the application

Probably a headphone amp.

 

[gyraf]

..but the 6DJ8-family are frame-grid types, specifically designed for operating nicely at moderate plate voltages - never intended for something like 300V...?

Jakob E.

 

[moamps]

..but the 6DJ8-family are frame-grid types, specifically designed for operating nicely at moderate plate voltages - never intended for something like 300V...?Jakob E.

It's white cathode follower or mu follower or srpp....Each triode's plate voltage is about +Vcc/2.
Anyway, the maximum dissipation per tube is 2W (both systems), so the tube  is obviously overloaded (by the plate voltage, Vkf voltage and dissipation).

 

[AusTex64]

Thank you guys for sharing your knowledge and insights.

It's a microphone preamp. 5751 for gain and 6DJ8/6922/7308 for White follower output. Was originally spec'd for 7308/E188cc. No output transformer. It is the cleanest, wide frequency response, hi fi yet not edgy, open sounding mic pre I've ever heard, hands down. I own RCA BC-2B, Neve 1073 and Neve 5012, Hardy M-1 and have used many others like Boulder, API, etc.  Otherworldly THD numbers, like .005% at +20dB at 1kHz - on a tube mic pre!!!. But it eats White follower tubes fast. Experimenting with different cathode resistors, which is assumed will raise THD numbers. Trying to find a happy medium.

Gyraf, we're using NOS tubes, Mostly Amperex 6922. But I've heard the JJ 6922 doesn't suck, and in a White follower shouldn't affect the sound so much as the 5751 and input transformer. Have not tested JJ 6922 yet. Your comments about early death are not encouraging, though. Your suggesting of 12BH7 like in LA-2A is not a bad idea either. 6BK7b has been suggested as a more robust possible substitute for the 6922. Will probably add THD, but after 20 hours THD rises with 6922 too. Have not tried them yet. Trying to make 6DJ8, 6922 or 7308 work, as that's what is in the original and it sounds so good I'm afraid to mess with the recipe too much.

 

[ruffrecords]

I have found the  EH 6922 tubes to be very robust. I have had poor results with JJ 6922s. The most robust 6922 type of tube I have come across are the Russian 6N23P. You may have to select them for noise but they are hard to kill.

Cheers

Ian

 

[AusTex64]

Thanks Ian! For some reason I have been suspect of Russian tubes. "Hard to kill" sounds quite appealing. Will have to get some of those 6N23P's and give 'em a try.

 

[gyraf]

For setups like the Aikido Cascode microphone preamp and the like, you could probably have good results with the ECC99 type tube for output driver...

Jakob E.

 

[CJ]

GE  5 star tubes are supposed to be good for 10,000 hours, no?

10,000/20 = you should be getting 500 times the life out of the tube that you get now,

that means circuit redesign or different tube choice,

so just changing brands won't make the nut,

how did they make the horizontal output tubes in a 25 inch Zenith be able to be left on 24/7 by blackout drinkers or TV crazy people and not burn out for 10 years that is what i want to know, recliner chair, carton of Marlboro red's, and the TV guide and i'm good to go,  You Asked For It, Rescue 8, Sky King, Twighlight Zone, then the test pattern with the weird Indian, can still hear that 15,750, wtf, over? 

check the hi-tek effects>

 

[My3gger]

I find JJ 6922 tubes failing very fast too. They really don't like a bit high anode voltage and dissipation.
Try  EH's version of E88CC, or Russian ones. Seems like Ian and i have the same experiences with this type.

 

[ruffrecords]

For setups like the Aikido Cascode microphone preamp and the like, you could probably have good results with the ECC99 type tube for output driver...

Jakob E.

The ECC99 is a handy tube. Much greater dissiaption than the 6922, similar ra so similar drive capability. It  has lower gm and a mu of 22 but for a White follwer this does not matter. Same pinout as 6922 but heater current is 800mA.  I used it in a headphones amp design.  It is capable of 400mW output in class A.

Cheers

Ian

 

[AusTex64]

Hi Ian,

I can only find the JJ ECC99 online. Is that what you've used in your headphone design?

It is a very robust looking tube. With 400V max on the plates, will likely hold up a lot longer to the 300V currently on the 6922's. Gotta make sure the unit's heater supply can stand up to 800mA current, but otherwise it just might be the trick. Thanks for that suggestion!

 

[ruffrecords]

Hi Ian,

I can only find the JJ ECC99 online. Is that what you've used in your headphone design?

It is a very robust looking tube. With 400V max on the plates, will likely hold up a lot longer to the 300V currently on the 6922's. Gotta make sure the unit's heater supply can stand up to 800mA current, but otherwise it just might be the trick. Thanks for that suggestion!

Yes that's the one. It is pretty bullet proof.

Cheers

Ian