I have a gutted but good transformers GT SE-40. The outputs are 1.9K and good for around 40W even though I think that’s optimistic. They have a cathode feedback winding off the secondaries as well as a UL Tap. The B+ is around 475-495V.
I’m soliciting for ideas for finals in a project using sweeps, or something different. I know I can use a pair of KT88 or three EL34, 6L6 per channel, but I’d rather have some fun.
Maybe three 6AV5 in UL per channel?
Thanks for any suggestions and I’m not picky about filament voltage. I can swap out the filament transformer.
I’m soliciting for ideas for finals in a project using sweeps, or something different. I know I can use a pair of KT88 or three EL34, 6L6 per channel, but I’d rather have some fun.
Maybe three 6AV5 in UL per channel?
Thanks for any suggestions and I’m not picky about filament voltage. I can swap out the filament transformer.
The maximum screen grid voltage rating is 175 volts for the 6AV5. Most will live at 250, many will live at 300 and slightly beyond, but none are going to live at 475 to 495 volts.
I assume that you meant 6FW5. GE and everybody else started out rating the 6AV5 at 11 watts of plate dissipation, since it was just the 6AV5GT stuffed in a bigger bulb. As TV's got bigger and needed more sweep power, the manufacturers just put bigger plates inside the 6AV6GA's bottle, but never changed the spec sheet. Most late vintage 6AV5GA's contained an 18 watts plate, but technically the 6AV5GA could contain ANY size plate from 11 watts to 18 watts. All 6AV5GT's had the 11 watt plate since that's all that fits inside the skinny bulb. This led to confusion and TV tech's sticking 11 watt tubes into sets that really needed the 18 watt tubes, so GE changed the number of the 18 watt fat bottled 6AV5GA's to 6FW5.
The 6FW5 is the same tube as the 6AV5GA, except that it will ALWAYS have the 18 watt plate.
6LW6's will slowly die in triode or UL at 400 volts. I fired 4 of them over a year running them in SE triode mode. They all go into a slow cook death at idle.
I have been looking into a 'LW6 PSE amp. It will run in pentode or a new mode that I am working on. Testing was underway with some cheap tubes that I can afford to blow up. I have 100 25DN6's that I got for 50 cents each. Again the spec shows 15 watts for the 'DN6, but most of them have 24 watt plates.
All of my tube work is temporarily on hold as I have taken on a contract engineering job to pay the bills.
Keep in mind that big sweep tubes will eat far more heater power than whatever came in that amp. You will probably need an auxiliary heater power supply.
Thanks!
I was planning on using the 26LW6 in Pentode with proper screen voltage. I almost always research your posts extensively before my sweep projects.
The 25DN6 is one I definitely may try since I have some and it’s relationship to the 6CD6 which has been used in PA amplification before be Dukane, etc. I’d think two per channel is optimal here, but maybe three with a higher bias.
The main thing I’d like to make sure is that I use at least half of the output capability of 35-40W.
I was planning on using the 26LW6 in Pentode with proper screen voltage. I almost always research your posts extensively before my sweep projects.
The 25DN6 is one I definitely may try since I have some and it’s relationship to the 6CD6 which has been used in PA amplification before be Dukane, etc. I’d think two per channel is optimal here, but maybe three with a higher bias.
The main thing I’d like to make sure is that I use at least half of the output capability of 35-40W.
The 6FW5 is not a bad choice either, but it would take a few to achieve 15 to 20 watts. I think I was getting 3 to 4 watts from a single tube in triode feeding a 5K OPT. I didn't try pentode back then (12 years ago).
I have a pair of Hammond 1628SEA OPT's. They have been in my collection since 2006. I have tried those OPT's in several amp designs over the years, but never found something that I liked well enough to actually build. The transformers have never been mounted.
I also bought a pair of the big Edcors at the same time, but the Hammonds got to me first, so I used them for amp development for all sorts of SE amps, but each time the Edcors just sounded better, so they got used. The Edcors are currently sitting in a SE amp that I built over 10 years ago. It was a 300B amp, then the 300B's got swapped for triode wired 307A's. Sometime before leaving Florida that amp got robbed for some parts, so the Edcors are idle and could be used in a new amp too.
Those big Hammonds are going to need to be driven from a low impedance source in order to sound decent, so pure pentode mode need not apply. The 50 cent output tubes (25DN6) and the 6CD6's for that matter DON'T LIKE more than 200 to 250 volts on the screens, so UL and triode are out of the question. What does that leave? It leaves pentode mode with some output tube plate to driver tube plate feedback, often called "Schade." See Petes big red board for an example.
Posted new P-P power amp design
If you read into that thread you know what I did with Pete's 18 WPC amp.
What I haven't done is explore SE operation with a similar design. I plan a SE design along those lines, and have created a somewhat novel way of making this work. It is a novel concept that hasn't been done before to my knowledge, but is totally unproven for SE operation where the idle dissipation is much higher. I have a push pull amp up and running cranking 50 WPC from some tiny sweep tubes. The SE breadboard is maybe half done and hasn't seen power yet. It simulates well, but it only exists in the computer at this time.
I would also like to extract as much power as possible from my OPT's. The Edcors are rated for 25 watts, and the Hammonds are 30 watts. I want to be at or above those numbers. These transformers are 5K ohm so I will need over 600 volts to get there. The Hammonds have multiple secondary taps so operation as a 2500 ohm OPT is possible, but this may suck up too much magnetic headroom, only testing will tell.
Starting with the very basics, a class A amp can never achieve better than 50% efficiency, and getting close to 50% requires A2 operation. Some class A SE amps run far lower, My trioded 6550 amp eats 420 volts at 100 mA per channel with about 40 volts dropped in the cathode resistor for 38 watts input to get 7 watts out. That's about 18% efficiency. The same amp in UL makes 15 watts for almost 40% efficiency. I'm guessing a pentode or three with some feedback will be in the 40 to 45% range, so......
If I want 30 watts out, I'll need to burn 70 to 75 watts in the output tubes at idle. That requires 3 or 4 of the 24 watt sweep tubes or two or three 26LW6's or 26HU5's. I'll play with the cheap stuff first, and may use them if they work out. I'd rather save the LW6's for a BIG P-P amp.
You want around 20 watts, so you need to burn 45 to 50 watts in the output tubes. That's 2 or 3 tubes in the 24 watt class, or marginally one 26LW6, depending on the 'LW6. Some of the GE's have large heat radiator fins welded to the plate. They can eat 60 watts without meltdown. The Sylvania made tubes should stay around 40 watts.
Those numbers assume an optimal match between the tubes and OPT. Mismatch in either direction reduces the efficiency.
Yes, but the plate of the output tube will swing up to 2X B+ in normal operation, far higher if the amp is driven into clipping into a speaker as in a guitar amp. This means that the feedback resistor must eat about 1 KV. Don't use one wimpy resistor, it will EXPLODE (done that). Use two or three 1 or 2 watt resistors in series that add up to the required resistance value.
As you add in the feedback resistors the plate voltage on the input tube will go up, so you must increase the value of the plate load resistor on the input tube. My experiments wound up in the 150K to 470K range for the feedback resistor. Then tweak the plate load to get the same plate voltage as the open loop case.
The books, and many "experts" claim that this type of feedback will not work if the input tube is a triode, and I have found this to be true.....sometimes. It relies on a three way resistor divider to determine the amount of feedback, the feedback resistor, the plate load resistor on the input tube, and the plate resistance of the input tube itself. The plate resistance of a triode changes a lot with the signal making for a variable feedback effect. Sometimes this can work in your favor, many times it does not.
My experiments will use a pentode for the input tube for this reason, and to get enough gain to allow feedback. The 12AT7 will usually fall short when trying to drive a big sweep tube. It needs a lot more drive than the EL84 in the RH design.
The books, and many "experts" claim that this type of feedback will not work if the input tube is a triode, and I have found this to be true.....sometimes. It relies on a three way resistor divider to determine the amount of feedback, the feedback resistor, the plate load resistor on the input tube, and the plate resistance of the input tube itself. The plate resistance of a triode changes a lot with the signal making for a variable feedback effect. Sometimes this can work in your favor, many times it does not.
My experiments will use a pentode for the input tube for this reason, and to get enough gain to allow feedback. The 12AT7 will usually fall short when trying to drive a big sweep tube. It needs a lot more drive than the EL84 in the RH design.
Last edited:
My simulation uses the 6EJ7 because I have a model for it. There are several tubes with the same pinout and higher Gm that will plug into the same socket. I generally try a few of about 5 types that I have a lot of and see which one gives me sufficient gain and the lowest distortion.
6EJ7, 6JC6, 6JD6, 6HM6, 6KT6, and 6HT6.
The old standby 12BY7 is another good choice but their use in the Citation amps and several ham radio transceivers sucked up most of them making the rest expensive.....there are pin compatible tubes there too.
12BY7, 12GN7, 12HG7, 12HL7 and 11HM7.
In the octal world the transconductance is lower, but there are probably several tubes that will work.
The old standby is the 6SJ7 but the Gm is pretty low at 1650
The 6AC7 has a Gm of 9000. I have used it as a driver before, but it only comes in a metal can, no glass.
The 6AG7 has a higher Gm at 11,000 but again I have never seen a glass one.
TV sweep tubes generally don't like, or benefit from G1 being driven positive. I have not needed to drive them with followers, but I haven't tried making a class A sweep tube output stage in a long time.
George, are you sure? RES has a $5 listing for 6AC7GT and a datasheet for the GT is available. A listing does not mean actual stock on hand. A call is definitely in order.
Also, the GE datasheet for the metal can version has an explicit DC heating warning, when the type is used in audio service.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.