6146A/B cathode bias not recommended?

[6267]

I saw this noted on the 6146a and 6146b data sheet–is it because class A will put it way to far past max dissipation? runaway issues? I figure the 6146 tube is a fairly linear one, probably even when overdriven, so I thought a cathode-biased stage would give the sound a bit more color than it would have otherwise. I'd be running two in push-pull, I don't want more than 40-60w, so I figure class A might be the way. I have a 500v unloaded B+ supply, and I can make up a regulated -150 max bias circuit if I must, but I have always liked the cathode biased sound a bit more. Filament current is not an issue. Is this tube not recommended for this application? Anyone tried doing something like this?

[Synaptic Amps]

Cathode bias is never recommended for high powered output stages like these because it's too inefficient and the consequences of cathode resistor failure would be catestrophic threatening not only your tubes but also your transformers if not much worse. If you really wanted to do it you'd have to use some nice aluminum housed wirewound cathode resistors attached right to the chassis for good heat dissipation, preferably one for each tube. I probably wouldn't want to use anything less than 25W rated. The 50W ones are only a dollar or so more from Mouser and only require a little more space. Does about 400 ohms per cathode seem about right to you? Of course there's no reason why you can't have your amp wired class A fixed bias and avoid these safety issues all together.

[6267]

I suppose that'd give me more B+ to play with, but I do like the cathode biased sound. I suppose the bolt on resistors should be safe enough, I'm not going to run the tubes at full power–although it would be cool to use an 0A3 or something to make an adjustable bias supply with builtin nightlight .

[Steve Conner]

6146s are rated for use with a separate screen supply of around 150-250V. I don't see any fundamental reason not to cathode bias them, but I can think of practical reasons not to.

The 6146B is a 35 watt (some sources quote 27w) plate dissipation tube, and that's an ICAS rating, not even continuous. So you won't get anywhere near "40-60W" from a pair in Class A. Therefore, you should use them in Class-AB if you want that much power, and fixed bias is best for Class-AB.

Also, it needs a lot of bias voltage. Even if you used a 200V screen supply, the datasheet says -50V grid bias for 25mA idle current. You may run a lot more than 25mA to get near Class-A, but on the other hand, you'll probably have more than 200V on your screens, which will increase the bias voltage requirement again. All of those 50-ish volts need to appear across your cathode resistor, wasting a lot of power. Compare that to EL84s which only need something like 5-10V.

[Bruce / Mission Amps]

6146s are rated for use with a separate screen supply of around 150-250V. I don't see any fundamental reason not to cathode bias them, but I can think of practical reasons not to.

The 6146B is a 35 watt (some sources quote 27w) plate dissipation tube, and that's an ICAS rating, not even continuous. So you won't get anywhere near "40-60W" from a pair in Class A. Therefore, you should use them in Class-AB if you want that much power, and fixed bias is best for Class-AB.

Also, it needs a lot of bias voltage. Even if you used a 200V screen supply, the datasheet says -50V grid bias for 25mA idle current. You may run a lot more than 25mA to get near Class-A, but on the other hand, you'll probably have more than 200V on your screens, which will increase the bias voltage requirement again. All of those 50-ish volts need to appear across your cathode resistor, wasting a lot of power. Compare that to EL84s which only need something like 5-10V.

Right. A pair of push pull EL84s in most amps with average B+ rails (cathode or fixed bias) will only need about 24-34vac PP signal across their grids to turn them totally inside out.
6146 tubes are very frequency sensitive.
In smallish <1.8MHz RF decks, I've run lots and lots of parallel pairs of 6146B tubes in class B with outputs of over 130 watts... at +600vdc. At 30MHz they might only make 85 watts in Class B.
At 150MHz in Class C FM RF decks they still will make a bunch of power but are down around 60 watts.
Seems like at audio freqs, a class AB pair should make at least 80 watts output though... Actually I wouldn't be surprised to see more then that if you allow some 10% to 15% distortion.

But, here's the rub, except for old Ampeg SVTs, I don't have any experience with them in audio circuits.
By the way, those amps have a separate and stiff, screen supply winding and screen node.

[6267]

Would a VR tube be a stiff enough regulator, say 0d3 (i think 150v)? I've heard from a couple places that you really want to be careful about the screen supply, and also about blocking RF into the tube. I heard the ampegs were running the screens way over voltage. At this point I'm almost considering a fixed bias supply, regulated screen supply, and see what I can get out of the 6lw6s I have sitting around. I know that their emissions are a bit different though, and I worry about the transformer with that kind of current + an imbalance. Hmm...so many projects so little time Thanks for the info, it sounds like it just would have taken a very hefty resistor and added a few dollars to the power bill!

[Steve Conner]

The Penta Labs 6146B datasheet suggests about 120W of audio from a push-pull pair, running off 750V B+. This probably uses pretty cold bias, so I imagine the distortion would be high and not terribly musical. Musically terrible, maybe...

Like Bruce said, these are designed as RF amplifier tubes. So they will happily oscillate at RF frequencies. If you do want to play with them, I'd recommend stopper resistors on the grid, screen, and plate. It might seem like overkill, but parasitics can be hard to track down, so my philosophy is to stomp them as hard as you can from the start!

They are great RF tubes, too. My first ham rig came with one that had a hole burnt clean through the plate, and it still worked!

[6267]

sounds like a good summer project then–I'm actually looking for less power, not more, so I was hoping to run them at high current and only 400-500 volts. The chassis I'm using had an 0A3 regulator on it once, and I haven't stripped the socket yet, so I'm definitely considering some RF tubes with a regulated screen supply. Fixed bias sounds like it might be better though, so I'll have to figure out a nice little bias supply. It's going to be an amplifier head, so I'm not worried about plate caps too much, as I'm not going to be digging around in the back of it.

[Enzo]

In my formative electronics years back in the 1950s, I thought of the 6146B an audio modulator output tube. Seem to recall circuits using it for that in QST and the RAH.

[kg]

have you got a lot of these tubes laying around?

i don't know that i would design an amp around them seeing as you're probably going to have limited source options.

[Steve Conner]

That's a good point, but it could still be a fun learning experience. If you build it with a little thought now, you can always convert it to run 6550s if you ever run out of 6146s.

There are still lots of NOS 6146 around, and they don't fetch crazy prices because practically no-one uses them for audio. I guess the original Ampeg SVTs have been converted to run 6550s by now.

There's also the 6883, which is just a 6146 with a 12V heater. You can probably get these NOS even cheaper, because the heater voltage is wrong for Grandpa's boat anchor ham rig. (I have three of them and never found a use for them. Maybe I ought to try to get more for my SVT clone project.)

If you used a heater transformer with two identical windings, you could series them to run 6883s or parallel them for 6146s.

[traynor]

6146 tubes

I'm a little late to this thread but I have lots of 6146's used and NOS if anyone needs any. I'm looking into building a small SVT with them, just cruising for info. I have an SVT here now in for repair.

Brent

[jrfrond]

The only practical reason for NOT cathode-biasing them would be if space constraints prevented you from using a 50W chassis-mount heat-sinked cathode resistor. I have been building amps with cathode-biased 6550's for several years, and neither the bias system nor the use of a garden-variety, unregulated power supply has ever proven to be an issue. You have a lot of latitude with beefy tubes like the 6550, 6146, KT88, etc. On a pair of 6550's, I get 70W clean on-the-nose with a 250?/50W cathode bias resistor and 1K/5W screen grid resistors. I suspect you will get very close to that with 6146's.

For the record, cathode-bias 6550's sound WONDERFUL! Your 6146's should come close. Just make sure that you mount the bypass cap AWAY from the resistor, which should be bolted to the chassis with a coating of heat sink grease to maximize thermal conductivity. It WILL get hot, and it can cause premature failure of the cap. I mount it off on a circuit board.

I just want to add that cathode-bias does NOT automatically make it Class A. You can bias the tubes A, AB1 or AB2, or even C if you wanted to, by selecting the correct cathode resistor.

[6267]

I've still been considering this since the chassis is very large & steel and has a big 12.6v supply. Would leaving them unbypassed perhaps reduce their RF oscillation tendency? I don't necessarily want all the watts I can get, since I'd have to get a different OPT for this setup anyway. I'd probably run the screens with an 0d3 regulator or something, maybe zeners since it sounds like the SVTs liked to blow due to screen voltage getting way high. I have ten octal sockets and five noval, so getting the drive voltage to run anything won't be too hard...I have some 6lw6s, too, but the required OPT is going to be $$$. Since they don't match perfectly I'd rather use those for a 20-30w SE project (more expensive transformers, so probably not soon, but a super champ does sound tempting )