The "$250" Amp? - 6LU8 Compactron SE, UL

Wigwam Jones

Caesar non supra grammati
(post 5 has "ALL" the amp info figured out and listed by Kegger)

This thread is started at the behest of Kegger, whom I am currently working with (along with Squidward) to build a Compactron-based amplifier at low cost.

It all started a couple months ago, when I found a reference to an old article in Nuts-n-Volts magazine for a Compactron-based stereo amplifier. A 'compactron' is a simple enough idea - in the last days of tubes, GE had the bright idea to keep them alive a little bit longer to compete with transistors by placing the guts of two or more normal tubes inside one glass envelope. So a compactron is a twofer, or a threefer, etc.

As I understand it, a lot of these were used in TV's, as 'sweep tubes' (whatever that is) and as a result, they have a very linear characteristic curve (this is a good thing for music fidelity). They're also dirt cheap, because there were millions made, and because they don't get routinely poached by eBay shoppers looking for tubes for their crunchy guitar amps.

So I bought an old copy of the magazine, gave it to Squidward, and asked if he could help me build an amp. He was interested, got Kegger interested, and the next thing I knew, it was off to the races. Squid would call me every so often and tell me about some new idea that Kegger had to make the unit even more interesting, more powerful, more cool, more...well, you get the idea.

I had no problem with that - Squidward told me that he and Kegger were going to each build themselves one as well, and I thought cool, should be great.

After some time, though, I thought back to the original idea. A pair of 6T9 compactron tubes, a pair of cheap Edcor output transformers and one power transformer, and a handful of parts.

That was the original purpose of the 'other' thread. I wanted to see if anyone was like me - wanting to build the absolute cheapest, easiest to build amp possible, learn about amps along the way, and have some fun. I didn't mention this history in that thread because I didn't necessarily want the design to be a compactron-based tube, just something easy and cheap and bog-standard.

The problem is, as Kegger pointed out to me on the phone, that all the old hands are not going to want to build a crappy amp. They want to help, but they want to build amps that interest them. So that's kind of how that thread went to hell and scared off all the n00bs like myself.

Kegger said hell, let's just do the basic compactron amp, el cheapo style, but with room to mod it after it's done. He even started speaking Klingon to me about PakProtector's bizarre mathematical formulae - yeah, sounded like Greek to me. But he agreed to do the original amp I wanted - with the exception of a more 'interesting' tube, the 6LU8. I'll let him explain why it is better than the 6T9 that the original Nuts-n-Volts article called for (because I don't know).

http://www.nj7p.org/Tube4.php?tube=6LU8

It's like saying you are interested in new curtains and the next thing you know, someone's ripping up your driveway because if you're going to do new curtains, you definitely need a new floor, and if you're going to do that, you probably want to spruce up the paint, and if you're going to all that trouble, the yard should be resodded, and man, look at the cracks in that driveway. Guys, chill. All I wanted was new curtains.

So no harm, no foul, no blame.

But this thread is about the 6LU8 compactron amp for next to nothing in price, easy construction, and nothing else. Anybody wants to add that EL34's singled-ended to a borgwarner 4 speed transmission and a Dana 60 with locker would be the better way to go is going to get a pinch. Start your own bizarre notion thread, this one's mine.

We're going to do this week by week, with lots of photos and explanations and time for people who want to join in to source their parts and tools. That's because I'm broke alla time and I've got nothing for tools - not even a soldering iron. So slow, easy, cheap, and well-documented is what this is all about.

And there you go. Now Kegger can wax philosophical or something.
 
hey-Hey!!!
Don't want no Locker...they eat tires waaaaaay to fast. Torsen OTOH...:)

The 6LU8 is about a 12AT7 triode in the same envelope with a 6AV5 pentode. These two parts will make up the voltage amplifier section and power section respectively. First things first, download the 6LU8 sheet from Frank Philipse's site.
http://www.mif.pg.gda.pl/homepages/frank/sheets/123/6/6LU8.pdf

The pentode section "may" not work so well rigged Ultra-Linear. The high screen voltage would offer characteristics that may be hard to utilize with 40% tap location. It should work with a 60-70% U-L tap, but that gets us into custom OPT land. As a triode it has a mu of 6.5 and a plate Z of ~750 Ohms. That will work just fine. Now with a 5k plate load, a 40 mA idle current, and B+ of ~350 we'll want a solid g2 supply sittiing at ~100 V. That is easy with a 0C3 VR tube in a shunt regulator( tentative parts selection leaves us a bit of current to burn ).

So possibly a choice between triode or pentode rigging for the power section. The triode-strapped power stage working with a 5k plate load will not need any negative feed back; the amp will run open loop. With the pentode connection we'll need some FB. The RH84 offers a very well designed FB circuit( somebody please follow up with a link to that schematic ). ( UL should be tested though )
cheers,
Douglas
 
Here are three thoughts on mains transformers. Please note that I've stuck to a more traditional approach, and avoided voltage doubler and other schemes. The voltage doubler might be able to be implemented cheaper than these options, but if you wanted a more traditional supply, these could be good options:

Antek AN-2T300: $35. Has good windings for high voltage, and some 6.3 and 12V taps for heat and other things. If one wanted to go cheap, you could use solid state diodes for rectification and the 6.3 to heat the 6LU8 tubes. For more cost (but more cool factor), one could use the 6.3 winding to heat a pair of damper diodes for rectification, and use a separate transformer for heat for the other tubes.

Allied 6K7VG: $47. Good for high voltage. Has a dedicated 5 volt winding for a tube rectifier. Has a nice 6.3 volt winding for the 6LU8 tubes. A decent "all-in-one" solution.

www.musicalpowersupplies.com (MPS) has a bunch of $33 power transformers. Unfortunately, the PT325 is out of stock, and it looks like one of the more interesting ones. The PT275 looks interesting as well, and could be used with solid state rectification to provide an all-in-one solution for both heat and HV. Or, like the Antek, you could use all of it's 6.3V for damper diode rectification, and get a separate heater transformer for the 6LU8 tubes.

One could use a 5V or 6.3V heater transformer with either the Antek or MPS PT275 transformers to use tube rectifiers. A 5V rectifier would probably be something like an octal 5U4 tube. 6.3V rectifier choices include the 9 pin 6CA4/EZ81, or a pair of compactron damper diodes (for maximum bling). The dampers are cool, and they are cheap tubes, but they will require a bit more heater current and a pair of slightly more expensive sockets.
 
The 6LU8 is about a 12AT7 triode in the same envelope with a 6AV5 pentode. These two parts will make up the voltage amplifier section and power section respectively. First things first, download the 6LU8 sheet from Frank Philipse's site.

The 6LU8 tube has two sections. One section is small, and we will use this as the "front end tube" of the amplifier. The other section of the 6LU8 is large, and we will use it as a power tube.

Here is a data sheet for the 6LU8:

http://tdsl.duncanamps.com/show.php?des=6LU8

The pentode section "may" not work so well rigged Ultra-Linear. The high screen voltage would offer characteristics that may be hard to utilize with 40% tap location. It should work with a 60-70% U-L tap, but that gets us into custom OPT land. As a triode it has a mu of 6.5 and a plate Z of ~750 Ohms. That will work just fine. Now with a 5k plate load, a 40 mA idle current, and B+ of ~350 we'll want a solid g2 supply sittiing at ~100 V. That is easy with a 0C3 VR tube in a shunt regulator( tentative parts selection leaves us a bit of current to burn ).

You can run the large power section of the 6LU8 in three ways: pentode, pseudo-triode, and Ultra-linear if you have an output transformer that supports it. Unfortunately, it would appear that Douglas thinks that the 6LU8 is not a good candidate for UL operation, unless the output transformer was custom made. That would blow the budget for this project, so it probably won't happen.

It would appear that Douglas thinks that with the addition of a Voltage Regulator tube (0C3), we could run the 6LU8 in "pentode mode". From my experience, this would give more power than triode mode, and may not sound radically different.

Triode mode operation would likely be the simplest scheme, but would yield the lowest power.

It is possible that one could implement the amplifier such that one could switch from one mode to the other, depending on final design choices.

So possibly a choice between triode or pentode rigging for the power section. The triode-strapped power stage working with a 5k plate load will not need any negative feed back; the amp will run open loop. With the pentode connection we'll need some FB. The RH84 offers a very well designed FB circuit( somebody please follow up with a link to that schematic ). ( UL should be tested though )

Here is a link to the RH84 schematic:

http://www.tubeaudio.8m.com/RH84/rh84.html

Note that it runs it's power tube, the EL84, in "pentode mode". It does not have the addition of a Voltage Regulator tube, as Douglas suggested, but rather it uses a less sophisticated method: A resistor from the High Voltage supply (B+ in tube lingo) to the second grid of the EL84 pentode (2.2K on the schematic).

FWIW, I built an RH84, and it is an excellent sounding amplifier. I would not hesitate to use it either as-is, or as a jumping off point for further development. (I ran the RH84 at the Fest, so maybe some members heard it.)
 
More info, (updated with final building links)

(all of the following has been finalized)

I'm still shooting for UL, as the specs for the tube aren't quite complete and I Have a good
idea it will workout, if not triode wiring is easy enough. But with a 300v supply I think UL has
a solid chance of working well here. (Testing TubeLab did on a 6AV5 in UL was very inspiring)

Yes like WIGGY said ED(SQUIDWARD) did bring me this 6T9 article and it looked interesting.

But the true (I don't like that tube and can find something I like better) got the best of me. :)

Started looking through my stash of tubes when I came across the 6LU8 again got me going.
It's (to me anyway) basically a 12at7 triode and a lower powered version of an EL34 pentode. (In 1 bottle)

I said heck that looks pretty decent to me for a single ended amp and having a plate rating of
14 watts on the pentode an the Screen having a voltage rating of 300v I thought yah those UL
Single ended (Edcor) outputs could be perfect. So this 6LU8 Single Ended UL project was born.

Now I had a few of these, found more cheap ones and there's another tube with same pinout an
same basic structure but a different type pentode section that can be played around with as well,
the 6MF8 could be tweaked in to mess around with to, but the 6LU8 we will do our build for here.
(The 6LR8 is the 9pin equivalent to the 6LU8, they have been tested and work just fine here to)

I've attached a basic curcuit layout (that for the moment values are not set) in a cathode biased
setup, one can/could build a bias supply plus add fixed bias as well (I may go that route on mine).
EDIT: (The circuit is now tweaked and ready to go, it's cathode bias)

This will be the basic format, later I'm sure many tweaks will be done, 1 could be run the outputs
in triode and take the UL taps to power the front end ALA semi E-Linear like Douglas made famous.

(Like ED had mentioned) (Shade feedback)
Another is a RH84 type setup with it's feedback scheme in place, there will be quite a few different
ways to go with the amp that can be a lot of fun experimenting with and learning along the way to.

Some 300V supply and the Cheap/Inexpensive Edcor SE UL outputs are the things I am envisioning
will be the main stay for anyone doing the thread build approach, some may go solid state for there
rectification and some tubed, (I plan on EZ81/EZ80 or 6BY5 for mine in the tubed variety)

Between 2.5K-5K SE outputs should be able to work on this one. (suggest 5K if your buying them)
(There is no published specs for what were going to do, the above should work well, but 5k better)
EDIT (5K is the output transformer to use, 2.5k will work though but not optimum)

If some want to follow along with there parts bin to build from I'm sure that will be just fine here to.

Well as said, below is the basic schem and we can continue to plan the amplifier project from there.
Any and all help/suggestions are welcome (keep the scope of the project in mind please)

----------------------------------------------------------------------------------------------

(large caps 100-220uf will provide a clean an beefy supply, any value in between there will be fine)

Link to post showing solid state supply. (tweak the first cap size to adjust output voltage for 300v)
http://www.audiokarma.org/forums/showpost.php?p=1849499&postcount=61

Link to post showing tube rectif supply. (tweak the first cap size to adjust output voltage for 300v)
http://www.audiokarma.org/forums/showpost.php?p=2055368&postcount=308

Layout of what the supply looks like using a 6BY5 tube rectifier(EZ81 would be similar)
http://www.audiokarma.org/forums/attachment.php?attachmentid=288636&d=1311435754

Links to parts suppliers.
http://www.audiokarma.org/forums/showpost.php?p=2993123&postcount=709

Link to a parts layout/"bom"
http://www.audiokarma.org/forums/showpost.php?p=2991998&postcount=708

(Updated, Working Amplifier Schematic, as of 03-30-2011) (roughly 4.5w, and 1.5v sens)
5w cathode resistor is fine, 10w is 5x the draw, (750ohm a bit less current then 680ohm)
 

Attachments

  • 6LU8_PINOUT_MRKD.GIF
    6LU8_PINOUT_MRKD.GIF
    4.7 KB · Views: 1,533
  • 6LU8_FNL_Pins_Act_6_VLTS.GIF
    6LU8_FNL_Pins_Act_6_VLTS.GIF
    15.6 KB · Views: 1,710
Last edited:
Our high voltage power supply will require a filter section after the rectifier. This filter section will be comprised of capacitors, and resistors and/or chokes. Chokes are superior to resistors in this application, but chokes also cost a lot more.

A simple high voltage power supply filter could be described as a "CLC" filter. This means that the rectifier is followed by a Capacitor, then a choke (inductors are abbreviated as "L"), then a final Capacitor. If we use tube rectification, the first Cap needs be sized such that it is not so large as to stress out the rectifier. Most rectifiers have published specs as to what size cap they are able to handle. Most input caps need to be less than 40uf.

Douglas suggested a nice high quality oil cap of a reasonable size around 10uf or so. These "motor run" caps are nice in these applications because they can tolerate very high voltage, and a lot of "ripple current" (AC that hasn't been filtered to DC yet). Douglas suggested ASC caps from Allied, but I'll look around: I think I know of a few places that might be a better deal.

As inductors are superior to resistors for power supply filtering, and single ended amps require a high quality power supply, we really should try to use them in our implementation. Dave Newman (dnewma04) has suggested the RDC-100 "Economy Universal Filter Choke" from RadioDaze to me in the past:

http://www.radiodaze.com/RD-transformer01.htm

I showed this choke to Douglas, and he liked it a lot. He even suggested that after the first cap, each channel of amplification could get it's own choke and second filter capacitor. One might express this topology in shorthand as the following: C(LC)x2 where there is a single Cap after the rectifier, and then a split where each leg gets it's own Choke and final Cap.

The second Cap(s) in this power supply can be much larger than the first. They are shielded from the rectifier by the Chokes, so there is no danger of burning the rectifier out with a large value. One could put a cap from 50uf to 500uf in this location. A larger cap will provide more filtering, but in some audio circles people believe that the best solution might be to use "just enough capacitance, and no more". I am confident that we will find a good C2, or second capacitor choice.
 
On to the new thread. I'm still in.

And still just as stupid and still without any tools.

I have a small tool set, and I'm quite close. Drillpress that can cut large holes( and small ones ), and nearly anything else I can think of needing for this project...:)
cheers,
Douglas
 
I'm in on this too.. very interesting so far, and for some reason, I am understanding some of this.. but will be reading up on some other things.

yes, I want larger amps, and will have some eventually, but until then, this will help me learn about tubes and currents and such in the mean time.

Wished I wouldn't had done so much partying since HS.. I'm having to relearn all I forgot in my electronics class.. but it's looking fun and well worth the effort!
 
Wish I would have taken an electronics class.

I have a job, a mortgage, and a child. I cannot take an engineering course at my local community college because I lack the proper math prerequisites. So you know what? I'm taking a math class this term, as I have for the two terms before that. By next fall I'll have enough math under my belt to start learning engineering. It lasts for two years, then I transfer to another technical college...

Washtenaw is a community college in Ann Arbor. They offer an electronics course Tuesday and Thursday evenings.

We can't not try to learn this stuff. If we want to be knowledgable we have to be willing to learn.

Made in America needs to mean something again.
 
So it is a choice between triode or pentode rigging for the power section. The triode-strapped power stage working with a 5k plate load will not need any negative feed back; the amp will run open loop. With the pentode connection we'll need some FB. The RH84 offers a very well designed FB circuit( somebody please follow up with a link to that schematic ).

Again, the RH84 schematic is here:

http://www.tubeaudio.8m.com/RH84/rh84.html

Feedback is applied from one stage of an amplifier to another. It has the benefit of making an amplifier more stable, and as a form of error correction to make sure the signal output looks like the signal input, just amplified. It's a useful tool, and it's use is fine. It has sort of a knee-jerk bad reputation because of some designer's overuse, where it tended to produce sterile sounding amplifiers that measured very well.

The feedback in the RH84 design is the 100K resistor in the middle of the schematic (Rfb). It is taking a small bit of the output of the EL84 tube and funneling it to the 12AT7 stage before it. This is a very short feedback loop, from one stage to the stage just preceding it. Some amplifiers have long feedback loops that take signal from the speaker side of the output transformer, and loop it to the very first input stage. Some designers favor shorter feedback loops for sonic reasons. I believe it would be safe to say that the RH84 offers a short feedback loop, perhaps a good compromise between the need for feedback, and the desire to limit it's negative effects.

It is possible to "triode strap" a pentode, as Douglas mentions. To do this, you place a resistor (100 ohm, maybe?) from G2 (or second grid) of the pentode to the plate. This makes the second grid mimic the plate. The resistor is there to prevent ocillations, as the two grids experience almost, but not quite the same conditions, and need some sort of damping between them. A mechanical analogy might be to think of a piece of rubber used to relieve sympathetic vibration between two parts. In this triode strapped mode, feedback may not be needed. If you triode strapped an RH84, you remove the 100K Rfb resistor, and instead tie G1 to G2 via a 100ohm (or so) resistor. This would cut the power of the amp in half (from probably 5WPC to 2.5WPC or so), and sonics would be different.

To run the tube as a pentode, you provide a certain set voltage to G2 instead of tying it to G1. In the RH84 schematic, G2 is connected to the high voltage power supply via a 2.2K resistor. This is a very simple way of dropping the voltage to a set level and applying it to G2. Douglas has talked about adding a "voltage regulator" tube, this would be a more sophisticated and stable way to provide a set voltage. The resistor may allow for fluctuations in G2 if there are fluctuations in the power supply itself, while the regulated scheme will have much more stable characteristics. The design works well in pentode mode with the simple resistor, a regulated supply would certainly be considered an add-on tweak. You will need feedback to stabilize the amplifier in pentode mode, no matter how you implement it, however.

Ultra-linear mode would be implemented by leaving Rfb in place, and strapping G2 to the Ultra-linear tap of an output transformer, assuming it has one.

It is possible to switch between all three of these modes on the same amp, I believe. The simplest option would be to use a Double Pole Double Throw (DPDT) switch to chose between two of them. My knee-jerk reaction would be to switch between pentode mode and UL mode, if UL mode ended up working OK. Both modes would benefit from the feedback resistor, and both would give slightly higher power than triode mode. All three should be attempted experimentally, and perhaps the best two should be chosen as options. It would be possible to do switch between all three modes, and remove the feedback resistor during triode mode, but that would require more complicated switch logic, and more expensive parts. A cheap and easy solution would still probably include a mode switch with two modes, and no optional removal of the feedback resistor.
 
Last edited:
Hey, you've got my interest. If it's simple enough for me to understand & I can get parts w/o tracking them down from some difficult source, then I'd like to give this a try. The only electronics I ever built was an integrated amp kit back in the 60's. So I figure I know nothing, NOTHING! But I'm willing to learn.:yes:

Cheers, Jim
 
Hey, you've got my interest. If it's simple enough for me to understand & I can get parts w/o tracking them down from some difficult source, then I'd like to give this a try. The only electronics I ever built was an integrated amp kit back in the 60's. So I figure I know nothing, NOTHING! But I'm willing to learn.:yes:

Cheers, Jim

That's what I'm talking 'bout. :thmbsp:
 
It is possible to "triode strap" a pentode, as Douglas mentions. To do this, you place a resistor (100 ohm, maybe?) from G2 (or second grid) of the pentode to G1 (the first grid). This makes the second grid mimic the first grid. The resistor is there to prevent ocillations, as the two grids experience almost, but not quite the same conditions, and need some sort of damping between them. A mechanical analogy might be to think of a piece of rubber used to relieve sympathetic vibration between two parts. In this triode strapped mode, feedback may not be needed. If you triode strapped an RH84, you remove the 100K Rfb resistor, and instead tie G1 to G2 via a 100ohm (or so) resistor. This would cut the power of the amp in half (from probably 5WPC to 2.5WPC or so), and sonics would be different.

You mention one way of triode rigging a pentode. It is quite different form the g2-plate version. See the RCA amps running a type 46. The intended triode version for this amp is the g2-plate version, and the resultant mu is listed as mu g1-g2 in the data sheet( it is 6.5 ).

The g2-g1 strapping can be examined in the sheet for type 3D21WB from Tung-Sol, note the radically different plate curves between the two arrangements.
cheers,
Douglas
 
Back
Top Bottom