@Tubelab_com - And now I see why you gravitated to the 26HU5. It’s another OCTAL. Every time I go to buy compactron sockets they either are out, or put a 10 piece limit on it. You get octal sockets from anywhere, anytime, 50 at a time.
Thanks for that, wg. I’ve found a Lundahl that says it’s a 1300 ohm primary and rated for 360w and 430ma. That seems like it would do the trick, the current rating seems a bit low for that kind of power however.
I love the simplicity of the bass amp design, it would seem that all I’d really need to do is add two tubes to that and a 1650w and I’d be in the ballpark. Where did you set a bias point on yours? I would assume that it’d be close to a 6550 amount of current at 135w
I love the simplicity of the bass amp design, it would seem that all I’d really need to do is add two tubes to that and a 1650w and I’d be in the ballpark. Where did you set a bias point on yours? I would assume that it’d be close to a 6550 amount of current at 135w
Last edited:
I still keep coming back to the used Svt iron. It’s a higher primary Ra but it’s also very cheap used.
My fascination with TV sweep tubes started when I was making guitar amps as a kid in the 60's out of dead B&W TV set parts obtained from the local trash dump. Back then the 6BQ6 and 6DQ6 were the top dogs in the trash as color TV had not yet taken off and the few that had been sold were still worth keeping alive. In 1968 I got a part time job in a TV repair shop where I discovered the 6CB5, the 6CD6, the 6DQ5 and a few others. The 6CB5 became my favorite guitar amp tube easily making over 100 watts per pair. Somewhere in the early 70's I saw the biggest, fattest TV sweep tube I have seen in a GE color TV set. I had met the 6LW6. It is undoubtably the king of sweep tubes, but in the race to be the big dog, followed by the decline of tube TV and the chase to the bottom of the cost cutting barrel, there were at least 6 different versions of the 6LW6 in two different envelope diameters. Some of the monster fat bottles with huge heat radiating wings attached to the plate can sit at 60+ watts of dissipation all day long with no hint of redness. There are also several different versions of the 26 and 36LW6.
I had seen the 26HU5 in some TV sets and even collected a few. I was beating up on some 26LW6's in a test amp when I decided to stick some 26HU5's into the sockets to see what happened. Would you believe that despite the different published specs, a 26HU5 and a 26LW6 worked quite well together in the SAME channel of a push pull test amp and biased up nearly identically? I have only seen two different kinds of 26HU5's, GE made tubes, and Sylvania made tubes. I was in Orlando over the Christmas holidays in 2020 (I think) and went to see Stan at ESRC. He had lots of 26HU5's and said that nobody was buying. I grabbed a box full and told him that I would make arraignments to get many more once I got back home. Unfortunately, Stan passed before I got them.
I ended up with about 30 each of the main “Sylvania” and “Ge” varieties of the 26LW6. Syl in the small bottle, Ge in the large. That’s enough to build and maintain a couple of kilowatters. Got a handful of the 36V, and one good pair of the dual-plate 26V Ge variety. That pair will get used in something. At high voltage (900V) and moderate current I‘ll only need to run the screens at 100. Decent quantity of 21LG6, 21EX6, and 6HJ5 to play around in the 300-400 watt range. A few pair of 26DQ5 to keep that bass amp tubed up over the years, and maybe build a really nice “tweeter“ amp by refining the same circuit.
I set the bias at 26mA per 26DQ5. Sounds low, but it works very well. It’s the same current I run the 6550’s at in the monoblocks. You can run that with the full 700V and not over dissipate.
If you start to run out of drive with 3 pair and higher g2 (will require more swing) add the k-followers after the front end, and raise its supply to maybe 400V to get more out of that SN7. Mosfets are fine for the followers if you use them.
Probably just add the 3rd pair, and use the SVT iron. The power trafo will give you the 700 volts which the DQ5’s will take, and the OT is similar enough to the 1650W. To get the peak current up, raise g2 until you can get enough. 150, 160 volts maybe. Just use a small (100VA) trafo to generate the g2 and the 330 volts for the front end. 120-0-120 would be good - it will let you generate the two voltages needed and leave a little for regulation…. and it’s not some expensive “tube” component that they charge an arm and a leg for.
I set the bias at 26mA per 26DQ5. Sounds low, but it works very well. It’s the same current I run the 6550’s at in the monoblocks. You can run that with the full 700V and not over dissipate.
If you start to run out of drive with 3 pair and higher g2 (will require more swing) add the k-followers after the front end, and raise its supply to maybe 400V to get more out of that SN7. Mosfets are fine for the followers if you use them.
I read somewhere that the hu5 was essentially an lw6 without the cooling fins on the plate but I’m not sure how true that is.
Do you have a preference for a particular tube? I was considering 26lw6 or 26hu5 myself since I have a local source but I imagine it really doesn’t matter at this size of tube. 6kg6 would also work I assume, the cheap Russian tubes are pretty tantalizing
Last edited:
Tried it, wasn’t a fan. I do small amps usually but I can’t see how this is really much different, just bigger.
If the LW6 and HU5 are biasing the same, it’s probably “use whichever one you can reliably get more of”. That is, getting more of the SAME construction. There are many LW6 variants. I was lucky to get quantities of BOTH main variants. The problem with the KG6/EL509 is the gaddam magnoval base. Sockets are going to be problematic. When you do find 12 pin compactrons, they usually are good and there are no issues. Magnoval and novar, not so much. Neither one is done well these days and they are not interchangeable. Octals are the way to go if you can manage it.
With the SVT iron, you could probably get away with two pair of the full size (40 watt sweeps). The next size down (*DQ5 and similar, 24 to 28 watt) requires three pair. The SCREEN voltage will require adjustment. For this, you look at the plate curves. You need 1.2 amps, plus a little margin. Not too much margin, because any excess results in excess screen dissipation and NO more power.
The 26DQ5 was suggested because I get them for seven and a half bucks. Was five at one time. The curves look just like a 6550, same gm curve, same shape of the kink, same Rp, same g1 bias IF the screen is ajusted to give the same max current in the 300-400 mA range. So does the original prototype, the 6CB5. But those are harder to get and I have only one pair the same. Outside that current range, I’m sure it diverges but that’s where 6550’s are used when making amps like this. The 21LG6 is another attractive tube for this application, and so is the 21HJ5 which is CAPLESS !! But their turn on characteristics are more abrupt. In a high feedback design (like these) it may not matter.
That’s a great point, I completely forgot about the bases. That’s what I get coming from octal world though. I’ll get a handful of whatever he’s got more of, that look identical. Either the lw6 or hu5 look plenty beefy enough for 75 watts each(hopefully). Now the parts ordering begins.
Rated dissipation on those is 40 watts. Long term average. The max dissipation is B+^2/(2.5*Ra-a), and that’s somehwere around 2/3 power sine wave where it hits the max. The B+ is what it sags to at load, not the unloaded value - and that helps a lot. And it never stays there for long statistically. Everywhere else is lower, unless the quiescent is the limit. And I doubt you’d need to run them all the way up at the 57 mA, because I can get away with half that and still not have crossover distortion. If in doubt use six instead of four - and you could drop the g2 below 100V - and that low you could clip the $#1+ out of it and not see a glowing screen. You just pay for that with more heater power, and you get better efficiency otherwise. The TUBES aren’t the real limiting factor on power, or the cost driver. The TRANSFORMERS are. No matter how many tubes you parallel, or how badly you overload them, you get so many watts out of a particular set of iron. The B+ drops by X amount at Y volt-amps, and you get Z watts undistorted output out of a given OPT, assuming it’s the right load for the power trafo.
I’d already changed my mind after sitting on it a while lol, going with 6 for peace of mind and reliability. I can always throw more current at them in the process, I personally believe with Svt iron there’s plenty left on the table. Not that I need more, but knowing that I can have more if I want.
Go cheap and try a dual primary 300Va 35 volt torroid with 400v B+ .
This will test the sound and power
35 hz to 8k
This will test the sound and power
35 hz to 8k
I think I have to build another to test the output toroid idea I’m very intrigued by that. I wonder how cheap I could get it
A complete set of Antek iron would be about $300. 8T800 for the OPT, 1400 ohm a-a at 8 ohm or 700 ohm at 4 ohm load. Two 4T430 or 4T360 paralleled for B+ (600 or 500V, depending on whether you design for 8 or 4 ohm load). Use all four 6.3’s in series for your power tube heaters - 25.2 volts is close enough. You can have delayed B+/heater like I do on the monoblocks. 1T150 for the screen (200) front end (400), bias, and front end heaters. Cap couple the bias rectifier to make the negative voltage. A couple of mosfet regulators (with cap multipliers) and you can dial both voltages to exactly what you need, and get rid of all hum and noise at the same time. Nice for your preamp stages.
Six power tubes at most $150. $18 apiece for the LW6’s (even less for the cheaper options!). Can’t build a 350 watt amp, even a solid state one, any cheaper. Well, yes you can, because I paid less than $100 for my 480V industrial control trafos. But at $144 the Antek option isn’t bad. I did this on the monoblocks and the only thing I paid thru the nose on was the 1650W’s. Now the 6550’s would have been more, but they were $35 a pop at the time for Tungsol. In any case, now or back then, the iron still costs more than the tubes. Using an extra pair just doesn’t seem that much of a burden.
When I get around to it (could still be years off), I’m doing this with the 8T800’s and a batch of 21LG6 - before tacking the kilowatters. I’m in the middle of a new home build right now and just recently had a setback. I’m facing foot surgery this Tuesday. In the middle of moving my warehouse, I had a winch lift collapse on me, sending me and the 100 pound horn loaded top cap straight down 7 feet. Broke the heel bone so I need a few screws put in. It will be after the first of the year before work on the new place resumes. The builder will get the fences put up in the meantime, and I’ll get the next load of rock put in on the driveway.
Six power tubes at most $150. $18 apiece for the LW6’s (even less for the cheaper options!). Can’t build a 350 watt amp, even a solid state one, any cheaper. Well, yes you can, because I paid less than $100 for my 480V industrial control trafos. But at $144 the Antek option isn’t bad. I did this on the monoblocks and the only thing I paid thru the nose on was the 1650W’s. Now the 6550’s would have been more, but they were $35 a pop at the time for Tungsol. In any case, now or back then, the iron still costs more than the tubes. Using an extra pair just doesn’t seem that much of a burden.
When I get around to it (could still be years off), I’m doing this with the 8T800’s and a batch of 21LG6 - before tacking the kilowatters. I’m in the middle of a new home build right now and just recently had a setback. I’m facing foot surgery this Tuesday. In the middle of moving my warehouse, I had a winch lift collapse on me, sending me and the 100 pound horn loaded top cap straight down 7 feet. Broke the heel bone so I need a few screws put in. It will be after the first of the year before work on the new place resumes. The builder will get the fences put up in the meantime, and I’ll get the next load of rock put in on the driveway.
A 300 VA toroidal power transformer will not handle 300 watts of audio at 41.2 Hz (Low E on a bass guitar with standard tuning). Saturation will set in well below 300 watts. With a B+ in the 600 volt range the OPT will see over 2400 volts peak to peak, or about 850 volts RMS of audio voltage. This is at 41.2 Hz, so more iron and inductance is needed.
The big Antek 8T800 might work, but I have never tried it and likely never will since it is a $105 experiment. What I have done was to stack two hamfest sourced "600 VA medical grade isolation transformers" on top of each other, using all 8 windings as the primary and winding a new secondary through the stacked pair. The primaries need to be connected so that the DC flux cancels in both transformers. There are several possible combinations, each will have differing high frequency response due to stray capacitance and inductance.
I have gone a different route for B+ in my killowatt amp. I have 13 of these wired in series. Note that you must remove the little capacitor from the output to the chassis ground if you don't like watching them explode. It's the little brown cap on the upper right next to the end of the heat sink in the picture. I got them back when they were $12 each a couple of years ago.
https://www.mpja.com/48-Volt-Power-Supply-3A-Cosel-LDA150W-48/productinfo/34713+PS/
I will be watching this with great interest, making something out of cheap and used parts is a favorite pastime of mine. Hope all goes well with the surgery, I’m unlikely to get my amp done any sooner than you. Being a professional musician tends to keep one busy
I may just try the control transformer and antek toroid from the outset. It really doesn’t save me much money from used Svt iron but my curiosity tells me I gotta know what that sounds like
It’s got to WORK. It’s designed for 1600 volts RMS to be impressed on the “primary”, and you’ll be running about half that. It should have enough inductance to work down to 25 Hz. The only question is how high does it go? With the high turns per volt, it may roll off sooner than one of their little ones running with only 240 volts RMS across their “primary”. But when their 50 VA units are used that way they sound fantastic. This is a nice little zero feedback PP class A I made. No measurements yet but pleasant to listen to. As long as there is some method of ensuring equal bias in the two sides for DC balance. With the TV sweeps I’d for sure be using fixed individual bias and set them carefully by hand. Experiments I’ve done even running a push pull pair open loop pentode getting the full 50 watts kept it in balance enough if it was balanced at quiescent.
Ive already bought a pair each of the 8T800 and 15T950. That expense is over for me. Just waiting on deck like George’s drool-over Plitron OPTs.
Ive already bought a pair each of the 8T800 and 15T950. That expense is over for me. Just waiting on deck like George’s drool-over Plitron OPTs.
