low voltage tubes (6GM8 and 12DL8)

[fdegrove]

Hi,

The mu (14) should provide mroe than enough voltage gain,

Yes, but not under the working conditions you describe here.

Oh, in case it is important- the tube is running on +/-12V rails like the BUF.

Which one exactly? The 6GM8 or the 12DL8?

Sorry, but I'm more than a little puzzled on how you hooked the whole lot up....

Can you redraw the circuit diagram?

Cheers,

[mig-ru]

What are you running with this setup, a speaker? A schem of the final layout and PSU would help.

[needtubes]

Hi all-

Attached is the circuit I used for testing. I think all the values are the same, but I substituted some at times. The heater was undervoltage, which might have been a problem, but I don't see why one volt would make all the difference in the world.

Tube and buffer run off of +/-12VDC. Heater voltage was 5V. The overall response was very weak and somewhat fuzzy. Any ideas as to why?

Thanks

[bournville]

Ideally tubes should be run well within 5-7% of the rated heater voltage for best performance - running a 6.3 volt tube at 5 volts in way outside that, and you could not expect it to perform well in those conditions.

[fdegrove]

Hi,

Any ideas as to why?

Other than the heater voltage, may I suggest to try the following:

Instead of running the tube from a bipolar supply, ground the cathode resistor and decouple it with at least 100µF/25VDC cap.

You should notice an increase in gain already...if more gain is required try lowering the value of that cathode R and see if it helps.

The ECC86/6GM8 is an odd tube that should be used with the curves in mind so that you stay away from the knee.

ECC86

Cheers,

[PRR]

Needtubes-

I suspect you know, but others here may not:

The 12DL8 space charge tetrode and its ilk are NOT normal tubes.

Any tube will run with 12V on its plate, at very-very low current.

What these space-charge tetrodes do is run the full 12V into a grid (G1) near the cathode. That SUCKS the electrons off.

Most of them go to the G1. But a few miss G1 and end up loitering in the space just outside G1.

This thin fog of loose electrons is equivalent to a cathode. Virtual Cathode is the name of the effect, and you can get it in most tetrodes and pentodes between G2 and P.

In the 12V space-charge tubes, this virtual cathode is sucked on by the Plate, but flow is resisted by control voltage on G2 (signal grid).

So you have a triode with a very funny cathode that takes both a heater and a large DC bias current.

G1 has no resistor, just wired to the battery. Likewise the cathode really needs to be grounded.

Running a 12DL8 with G1 resistor and cathode resistor, you miss the benefits of space-charge design. What you have is a pretty crappy triode, easily beaten even at low voltage by many "normal" triodes.

For "plate power efficiency", these tubes are not very good. Most of the B+ current is the yank-out current from G1, which does not contribute to signal output. However the total power in heater and G1 is less than the heater power in an oversized tube that could handle the signal current with normal cathode action. Even so, we put 6 Watts in to get 0.040 Watts out. The only way this makes sense is in a car, where 12V is free and HV is expensive. In 1957, they could do the whole RF/IF strip of a car radio with small pentodes working 12V, and the audio power with a huge Delco power germanium eating 12V. But they could not get the 10mW-100mW needed to drive the audio output from any normal tube working on 12V, and a second transistor cost more than these whacky space-charge tubes.

These tubes are freaks and are specfically designed to be used a certain way.

Set it up just like the book. Cathode grounded or a very small resistor (10Ω ). G1 to +12V directly or through 10Ω. G2 takes signal and is biased with a 2Meg resistor to ground. Plate current is claimed to be 40mA at 12.6V, plate resistance is then 480Ω. If you want to resistor-load it, just 100Ω will tend to cause a 100*0.040= 4 volt drop, plate voltage falls to 12-4= 8 volts, a big change. We won't get 40mA with any decent resistor load (this tube is really meant for transformer coupling!). The Gm then will be somewhat less than 15,000µMho, be conservative and say 12,000µMho. 1/Gm is 80Ω, plate load is 100||480Ω=83Ω, voltage gain is about unity. Current gain may be huge for very-small signals, though because we have to bias the control grid to the edge of grid current to get any useful plate current, for large signals the current gain may be more like 10.

[EC8010]

Thank you, PRR, for taking the time to type a most interesting post.

[fdegrove]

Hi,

Thank you, PRR, for taking the time to type a most interesting post.

I second that...absolutely baffling.

Need to tell the friends to stop calling me a walking library, I'm completely stunned.

Cheers,

[Brett]

Excellent post PRR.

The 6GM8's are an interesting little tube for a preamp. Welborne has a kit using them in a DRD configuration run off SLA's. I have the batteries, the tubes are cheap, and I have some nice trannies to do a PP version....
But no time.

[mig-ru]

Hmm... I've been wanting to tinker with low voltage tubes for a while (no expensive transformers to buy). Is there a list of or a site dedicated to low voltage tubes? Also, PRR thanks for explaining this dillema so well, I really enjoyed reading your post.