• Disclaimer: This Vendor's Forum is a paid-for commercial area. Unlike the rest of diyAudio, the Vendor has complete control of what may or may not be posted in this forum. If you wish to discuss technical matters outside the bounds of what is permitted by the Vendor, please use the non-commercial areas of diyAudio to do so.

Tubelab Universal Driver Board, 2015 version

The Universal Driver Board (UDB) was a concept that was underway in 2009 when the 6L6GC in AB2 thread appeared. The design concepts are discussed throughout that long thread. The circuit features two cascaded LTP's made from dual triode tubes. I have made several different versions of a similar design with 9 pin miniature, and octal tubes. The Octal version was used for Chrish's amp design, and for my test breadboard. Several different types of output tubes have been tested through my breadboard. The thread is here:

http://www.diyaudio.com/forums/tubes-valves/133034-6l6gc-ab2-amp.html?highlight=6L6GC+AB2

Shortly after Chrish's amp was done, it became evident that my engineering career was in jeopardy, and I had to devote most of my time to keeping it alive, so my amp never got finished. In 2014 my 41 year engineering career ended, so we packed up everything and left south Florida. The "6L6GC" thread fell into a 4 year sleep during that time.

It turned out to be a hasty exit and I had to sell, give away, or trash over half of my stuff. What was to be a long careful packing, labeling and moving job got compressed into 3 weeks, so things were essentially randomly packed. The breadboard built during that amp design effort has never been found, so I assumed that it was lost, and set out to build a new one.

I laid out a new board that used 9 pin tubes since there are far more different flavors to choose from. I had a batch of the boards made by a commercial PCB vendor, since they were to be the next Tubelab product. That was early in 2015. About this time my daughter returned home with 4 kids and a husband, which we are supporting, so Tubelab again got pushed back into the corner. That situation has still not been resolved, and often requires most of our time. In the exit from Motorola and south Florida, I also lost access to my web site. I am working on that as well, but that's also why there hasn't been any recent updates.

The batch of UDB's sat in their shrink wrap on the shelf, until we moved again into a bigger house. In late 2015 Chrish woke up the original thread, which prompted me to build a pair of my UDB's and a new breadboard. First seen in Post #630, the board, and new breadboard finally saw life in early 2016, post #651. There have been a few random tests since then, but nothing too serious, just 250 watts from some transmitting tubes.

I have ideas on building a few amps using these boards, and I have sold 4 of them, so it's time to put together some documentation. There will be more posts in this thread in the coming days with real data on test amps. The initial choices for output tubes is KT88's, and the amp has already been fired up on bench supplies.
 
Here is the schematic for the not so new board. Being a "universal" board requires a different parts list for each application, and some parts are not used in each application. A generic parts list is done, but it had NOT been optimized for any specific application. Specific versions are coming soon.
 

Attachments

  • LTP_driver_board.pdf
    24.6 KB · Views: 1,891
  • Univeral Drive Board generic BOM.pdf
    191.9 KB · Views: 1,109
Random Pictures of the KT88's cranking out 50 watts in triode from 450 volts. The audio analyzer reads 2.8% THD.

Off to take the grandkids to a Christmas show...one of them has a part in it. Maybe some more info later tonight or tomorrow.
 

Attachments

  • P1000322_x.jpg
    P1000322_x.jpg
    903.9 KB · Views: 4,448
  • P1000324_x.jpg
    P1000324_x.jpg
    191.8 KB · Views: 4,440
  • P1000325_x.jpg
    P1000325_x.jpg
    695.8 KB · Views: 4,253
Last edited:
Here is the complete generic BOM in zipped excel format. It contains the note that were chopped off in the PDF version.

Why would someone want to use this board, what can you build with it, and what else do you need?

This board is a single channel driver circuit for a push pull vacuum tube amplifier. I have used it (or it's predecessors) to make P-P amps from 20 WPC to 250 WPC. It will drive small tubes, the typical audio tubes (6L6GC, EL34,Kt88...), DHT's (45, 2A3's, 300B's), TV sweep tubes in conventional G1 drive, screen drive, and both grids (dual drive), large transmitting tubes like the 811A, 813, 211 845, 3D21, and well just about anything that can be used to make an audio amp.

To make a typical stereo amp, you would need a power supply with the necessary transformers, two driver boards, 4 output tube sockets, 2 output transformers and the usual chassis or cabinet.

The power supply needs to generate the heater voltages for the pair of tubes on each board (typically 6.3V at up to 600 mA each) and heater power for the output tubes. You will also need a negative voltage supply of at least enough voltage to fully cut off the output tubes being used at 2 X the B+ voltage being used, and a positive supply of at least 50 volts or so for conventional G1 drive. Screen drive systems use a lower negative voltage supply (about 50 volts), but somewhere in the 200+ volt range for the positive supply. You will also need a B+ voltage for the output tubes, which will also be used for the driver tubes.

A power supply PC board had been laid out, but not yet built and tested. The same power supply board can be used with the TSE board to improve its performance and reduce heat.

The only components needed around the output tubes is a small resistor in the cathode connection for easy bias measurements, and possibly a stopper resistor on the grid leads. These are usually mounted on the tube sockets, so a PCB may not be needed. Some universal output PC boards have been laid out, and proto versions are being tested with these driver boards. I have also used turret boards from AES for easy - swap output boards.

1st picture, 4D32's mounted on bare sockets, cranking out 250 watts.
2nd picture, a pair of $3 25L6's (like a 6W6GT) mounted on AES boards, making 40 watts on 400 volts.....yeah a bit over spec.
3rd picture, a shiny new board before I even used it.
 

Attachments

  • Univeral Drive Board BOM.zip
    8.8 KB · Views: 508
  • Universal Driver Board_x.jpg
    Universal Driver Board_x.jpg
    580.3 KB · Views: 1,133
  • 25L6_CrazyDrive_x.jpg
    25L6_CrazyDrive_x.jpg
    872.8 KB · Views: 3,969
  • 4D32_bigOPT_x.jpg
    4D32_bigOPT_x.jpg
    706 KB · Views: 4,124
Last edited:
I recently received two Universal Driver Boards from George and am hoping to build an amp to drive some vintage AR3a speakers. I've built 5 of George's boards and am looking for something with a little more power than the SPP which is currently driving those speakers.

Per George's suggestion I'm thinking of KT88's for power tubes.

I'm looking forward to this project.

thanks George.
 

Attachments

  • IMG_2241.jpg
    IMG_2241.jpg
    492.4 KB · Views: 1,122
I put together the parts list based on what parts were in the board that I had been using last, but that was two years ago. After playing with it last night I believe that it was last used for a screen drive, or dual drive amp.

Some of the resistor values that are on the current BOM will be changing. I have three of these boards built up, and they are all different. I think I will fire each one up tomorrow to see which parts do what.

I hope to have the list correctly filled in for a generic driver board that will work on 350 to 450 volts in a few days. That should cover most use cases.
 
The neon lamp does absolutely NOTHING (OK, it's a 1 pF cap)......once the tube heaters are hot.

When the board is first powered up the CCS will pull V1's cathode to the negative rail, which could be several hundred volts. Who knows where the heater windings are biased, but usually at some positive voltage, and V1 could be any one of several possible tubes. The neon bulb clamps the cathode voltage at about -75 volts during warm up so that the HK rating doesn't get seriously violated. If the negative supply is less than 75 volts the neon bulb can be safely omitted.

My usual application of this board develops +150 and -150 volts from a cheap isolation transformer like a Triad N-68X. This works for all output tubes except screen driven sweep tubes and big transmitting tubes, but puts the negative voltage in the blown HK range for many common input tubes.
 
I'm assuming there will be a way to build the driver board so that several different output tubes can be used without any changes to the driver board? For example an el34 or kt88? I realize the bias on the output tubes would need to be adjusted but other than that no changes? This is what I'm looking to do with it. 50wpc, triode, no feedback. Thanks. I'm hoping these boards sell very well for you, George! I know I want a pair along with the the power supply board once everything is sorted out.
 
I spent a few hours today tweaking out the board. I made some serious headway, and developed enough confidence in the boards to drop in some tubes that I have been saving for about 30 years.....Lets just say that I bought these for $15 each!

I had been chasing my tail every time I fired these up thinking that there was something wrong. The distortion would rise as the power was increased then drop a bit, then rise again. I dug up one of the boards from the 6L6GC in AB2 thread that uses 6SN7's, and it indeed does the same thing......I dug deeper into my "box of broken dreams" and found the original universal driver board from 2007, but it's pretty dead, so I didn't get to test it.

I separated the two stages, tested and tweaked each individually, then put them back together......Still happens. OK, I got it. You have a triode gain stage, it has a built in nonlinearity due to non constant Mu. You have an identical gain stage, which has an identical nonlinearity being fed by the first stage, but is operated out of phase with the first stage. Part of the nonlinearities will cancel each other. The cancellation will be almost perfect IF the stages are operated under IDENTICAL conditions, but they are not. The second stage sees a higher signal voltage, and runs at a higher voltage and current. This is why the cancellation is best at about 2/3 of maximum voltage swing. It is also why the cancellation is strongest when the same type (identical Sylvania 6CG7's) of tubes are used in each stage, and non existent when a 12AX7 is driving a 12BH7.

I agreed that the first test for this board would be with KT88's in triode, so that's where I went today. I started out with the same 20 year old pair of Chinese KT88's that I always use for experiments where I am not sure of the outcome. If one blows up, no big deal. I bought about 50 of these when I made guitar amps in the 80's. The quality was so bad that one sparked out so violently that the glass shattered. Others would red plate even when operated well within the ratings. I never trusted any of them beyond 35 watts per pair in pentode mode, and most were used in single ended Fender Champ clones. There were however 4 or 5 survivors that have followed me for all of these years without exploding. This pair has cranked out over 100 watts in pentode mode on 525 volts, so I figured that they should deal with a full on attack in triode on 450 volts, and they survived all testing.

The breadboard was tested with the Chinese KT88's in TRIODE mode. The B+ power supply was a Fluke 407D bench supply that I got on Ebay for $45 (non working). I fixed a broken switch with some JB weld epoxy, but other than that, it's all original including the 807's and bumble bee caps. The +/- 150 volt MOS (for the MOSfet PowerDrive) is a simple bridge rectifier and cap connected to a Triad N-68X transformer(seen to the left of the breadboard). An old HP 204C audio oscillator provides the 1KHz at about 0.12%THD, so it sets the lower limit of measurement. Clip lead interconnects and Walmart LED shop lights also raise the noise floor.

I took basic data on the test amp with a 3300 ohm load, and a 6600 ohm load. The same $16 guitar amp quality OPT (I bought 1000 pounds of these in the early 90's) was used for each case, I just changed the output tap. Bias for each output tube was set at 45 mA which is 20 watts idle dissipation (arbitrarily about 60% of where I run KT88's in my SSE amps 80 ma on 430 volts, 35 watts).

The "experimental feedback resistors" in the parts list were used, and worked great. There is a total of 240K on the driver board, and 300 K off board, so each path has a total of 540K working against a plate load resistor of 43K, for about 4 db of local feedback.

The results for a 3300 ohm OPT:

power THD B+ current output stage efficiency

0.1 W 0.14% 90 mA 0.02 %
0.5 W 0.19% 95 mA 1.1 %
1.0 W 0.25% 100 mA 2.2 %
2.0 W 0.34% 102 mA 4.4 %
5.0 W 0.58% 115 mA 9.7 %
10 W 0.99 % 140 mA 15.9 %
20 W 1.66 % 175 mA 25.4 %
50 W 1.76 % 260 mA 42.7 %
60 W 2.72 % 285 mA 46.8 %
70 W 5.22 % 305 mA 51.0 %

Maximum tube dissipation is reached at 60 watts output, and is 34.125 watts. It drops to 33.6 watts at 70 watts output. Maximum tube current is reached at maximum power (70 watts) and is 152 mA. Both of these are within the specs for KT88 and 6550.


The results for a 6600 ohm OPT:

power THD B+ current output stage efficiency

0.1 W 0.15% 90 mA 0.02 %
0.5 W 0.22% 92 mA 1.2 %
1.0 W 0.28% 94 mA 2.3 %
2.0 W 0.39% 96 mA 4.6 %
5.0 W 0.62% 104 mA 10.7%
10 W 0.94 % 118mA 18.8 %
20 W 1.44 % 143 mA 31.1 %
40 W 1.93 % 182 mA 48.8 %
50 W 5.01 % 199 mA 55.8 %

Maximum tube dissipation never reaches 23 watts. Maximum tube current is reached at maximum power (50 watts) and is 99.5 mA. Both of these are very conservative for KT88 and 6550.

Perhaps 5000 ohms would be a good compromise.......I just happen to have two excellent sets of 5K OPT's. That's for another test day.........

But wait, there's more!!!!!!

After seeing how well some crusty old Chinese KT88's worked, I plugged in a "matched pair" of Electro Harmonix KT88's. I tested these at a few random points and at 3300 ohms only. The distortion dropped slightly across the entire power range without any significant changes in efficiency. Perhaps the matching done by the Tube Store over 10 years ago helped.

Still more.......

I dug deep into my secret stash and pulled out a quad of NOS JAN GE 6550A's still in their military packaging with 10/83 dates. I got them in the late 80's. They have never seen power......these things are stupid money today! Were they still good? How good?

I did a quickie matching job....leave the bias pot where it was with the EH tubes, plug in each tube, give it 5 minutes warm up, then turn on the B+ and read the current. pick two pairs from 4 tubes.

Surprise! The expensive tubes really do work better than the others. The distortion at 1 watt was 0.157% for one pair and 0.160% for the other. This compares to 0.22% for the EH tubes, and 0.25% for the unbranded Chinese tubes.

The distortion was lower across the board for the GE. The next surprise comes at 70 watts. While the other tubes have ran out of steam and started to clip, the GE's still had a bit left. the distortion at 70 watts was 3.95% and 3.98% VS 4.54% for the EH and 5.22% for the Chinese tubes.

I changed a few resistor values to make these boards sing. I will update the BOM when things aren't changing every few minutes. I plan to make the second board like the one I tested today, and maybe make a few more......I got some big sweep tubes to test, and some 300B's and some 7403's and ......
 

Attachments

  • P1000334_x.jpg
    P1000334_x.jpg
    907.8 KB · Views: 1,205
  • P1000333_x.jpg
    P1000333_x.jpg
    782.8 KB · Views: 1,149