Best tubes for ~100w 1.9k p-p?

[m6tt]

I am building a big amp for a friend, and I have a 100w 1.9k push-pull transformer to start from. Power supply will be a 240v iso transformer in reverse (from 115) into a voltage doubling bridge for at least 480v+, screen supply can come from center tap if necessary. This is a limited budget project, where a quad of good quality el34 is probably too expensive. So far I have considered GU50 and soviet 6L6GB as the output tubes, as well as *av5. Top caps are probably to be avoided, although the chassis isn't designed yet, so interlocks of some sort could be an option. Any ideas? I'd like to be operating the tubes somewhat within specs, too Thanks, you folks are the best!

[SY]

AB pp-parallel 6L6GC should get you close to 100W at that B+ and load impedance.

100W is an ambitious target unless you've got a few other projects under your belt. Just remember, it's only 3dB louder than 50W, barely noticeable. If you build an amp to have graceful overload and recovery, it will sound more powerful than a higher power amp with more typical behavior.

[danFrank]

Hi,
You can use 3 pairs of 6L6 in push-pull (3 tubes per side of center tap on 1900ct push pull transformer). This will give you 100 watts easy with 480 volts for B+ and be a pretty good impedance match. You can feed the 6L6 screens with the 480 volts also.
You said 4 EL34 was too expensive, so I won't mention that, even though it's also a good choice.
I am assuming that this will be a musical instrument amp? You will need at least 500ma B+ for 6 6L6 tubes in push-pull.
Daniel

[m6tt]

75-120w is target, no firm figure. This isn't my first, not by a long shot, and should be easier than my stereo 75w which had dual OPTs on each channel and 16 tubes total (multiple inputs & mixers). This is a musical instrument/vocal PA with the goal of being very loud in a small club. In my experience, this is possible with smaller output power, but not with fidelity I'd enjoy for vocals (We'll assume standard speaker cabs with relatively poor efficiency, I'm sure 5w and a big horn would do it, but I'd like to preserve the ability to use whatever speaker cabs are around).

I'd prefer to use no more than 4 output tubes, to save chassis real-estate. There is going to be quite a complex preamp on the same chassis, so four is the limit...I do have two 6lw6, but they're not matched at all (at least on the top-cap less tester I used...I guess the screens aren't matched?). I think I'm saving those for a big SE instrument amp later.

I'm certainly considering screen drive or A2 operation, since the preamp will include a balanced CF output. This is where I don't have much experience, do these modes require higher primary impedances more like class B? Is there any hope of a pair of GU50 at 1.9k? Thanks as always...

[m6tt]

SY-I intend to use drivers dc-coupling into the power tubes, small-ish coupling caps elsewhere (72hz is really as low as I need here) and a small global feedback around PI, drivers & power tubes. That's pretty much all I know about graceful recovery, and overload. I could always ditch global entirely and layer on more local, I suppose. I tend to design without and do global last, and generally only to tame high frequency gremlins and tighten bass (i.e. not much feedback).

[Wavebourn]

GU-50 will need twice more of B+ for 100W output from a pair.

[tubelab.com]

AB pp-parallel 6L6GC should get you close to 100W at that B+ and load impedance.

The 6LW6 is one of my favorite tubes but it really wants more than 480 volts to really crank. It has been an observation that in order to sell any MI amplifier it better use tubes that are available at the usual places musicians go for supplies and servicing, unless you are planning to maintain it forever.

I am in the process of developing a driver board capable of feeding just about any tube pair, or pairs in just about any configuration. One of the first things that I did with this board was to attach a pair of $5 Chinese 6L6GC's. I set the power supply for 500 volts , 400 volts for the screen. The driver board uses mosfet followers to feed the output tubes all the grid current that they can eat! This allows A2 or AB2 operation, and more power.

Initial testing revealed 90 watts output with a 3300 ohm load. With 4 tubes and a 1900 ohm load I think you will get at least 150 watts.

Much of this was discussed in the following thread. My experiments start at post 127.

6L6GC AB2 Amp

[SY]

tubelab: The 6LW6 has a top cap?

SY-I intend to use drivers dc-coupling into the power tubes, small-ish coupling caps elsewhere (72hz is really as low as I need here) and a small global feedback around PI, drivers & power tubes. That's pretty much all I know about graceful recovery, and overload.

You know more than most. That's a good list.

[danFrank]

I have built 50 watt and 100 watt Marshall clones where the only difference is the number of output tubes, and I always thought the 100 watters were much louder and "punchier". With the 50's I would never feel it in my gut, but the 100's always hit my gut when power cords are played through them. I know there's only 3db difference between 50 watts and 100 so there's more to it than the 3 db's. Maybe peak current passing through the output tubes?

[m6tt]

Wavebourne-I feared as much, and the electrolytics will be expensive! I think I'll do a GU50 project when I have the time and some motor run film caps I have a big television PT that's 1800v ct, frankly it scares me .

SY-Thanks! I've both followed too many diyaudio threads (not enough, though!) and sat around with a soldering iron and a telecaster... I did a partially dc-coupled guitar amp and I had to buy new sponges to get the drool off the floor .

Tubelab-I was definitely leaning toward tubes that could be found in a small town music shop (such as my own) since this will be a gigging/touring amp. Definitely not going to a technician, so something they could find at the guitar center in just about wherever seems to be ideal. Do you think SN7 cathode followers can deliver the juice? Or do I need even lower output impedance (i.e. sand)?

Dan-I agree with the 50w/100w thing in music amps, and I think it has to do with peak current at clip. Generally regions well beyond fidelity. I've noticed a similar effect with tubes that have fatter cathodes...say 7ey6 vs. 6v6. It has a higher Rp, but at clip it seems to deliver more bass than 6v6, even when the circuit is adjusted somewhat. Is it more noticeable at clip or at the magic 1w?