Archives

All posts by George

I get far more email than I am currently able to answer. Considerable email asks “how do I order a board”? I have updated the PC board ordering page in an effort to reduce these emails. I get a few emails asking technical questions about the PC boards. These will all be answered, but lately the I can only respond to Tubelab email on weekends. Tubelab email access is forbidden at work, and lately I have been at work until late almost every weekday. For now I can still check in to the diyAudio vacuum tube forum at lunch and after 5PM from work. There is nearly always an active thread about Tubelab amplifiers under construction. The people are friendly and there are plenty of people who have already finished their Tubelab amps if I am not available. Anyone can read posts, but you need to register to post questions or download anything. I have started compiling a FAQ page based on previous email and forum posts. It will be added to this web site as time permits.

Another Recession

I am building this page on the road somewhere in western Pennsylvania. Sherri and I are staying at her mothers house for a short while. We should be returning to Florida in a week or so.

The communications industry has been hit hard by the economic downturn with hundreds of layoffs in the plant where I work. I am usually at work 10 to 12 hours a day lately doing my best to stay employed. Unfortunately no job is totally secure, and the expenses associated with Sherri's mom have forced a careful optimization of Tubelab's expenses. So for now, Tubelab is operating with a fraction of my time, none of Sherri's, and limited funds. The budget for new projects is very limited so except for small parts all development projects in 2009 will come from my "junk box" which fortunately is well stocked. There is no budget for transformers or other costly parts.

When the economy started to slow down last year, sales of PC boards dwindled too. This forced the delay in the launch of the Tubelab SPP. Fortunately it is nearly ready. For now future products must wait until the current new project (SPP) pays for itself.

There are two P-P designs currently under development. The “Small” one, now called the SPP, and the "Big" one, now called the Universal PP driver. The Small one has progressed to the "third prototype PC board" stage. It is currently going through my rigorous testing, and I have been using it for several experiments. I plan to take this design through to a production PC board, but it will not be ready until the 2nd quarter of 2009.

The circuitry is not all that unusual. I Have already been informed that it bears a resemblance to SY's "red light district" amplifier, and Morgan Jones "Bevois Valley" amp. The truth is that it is about as simple of a P-P amp that I could come up with that yielded good performance. All three of these amps look like descendants of the Williamson to me.

Why have I done "yet another version of this common design?" I have breadboarded countless variations of small P-P amps. Some used the 6AQ5 tube, and some used the 6BQ5 / EL84 tube. Every amp was evaluated for performance, simplicity, and most important, the ability to be assembled by builders of varying skill levels and work the first time. One of the main requirements was for a stable design that didn't require any component tweaking to make it work. This design worked out to be the best. I chose the 6BQ5 / EL84 tube mostly because that is what was requested in the email that I received. The 6AQ5 is cheaper but produces slightly less power, and a slightly different sound. I may produce a 6AQ5 version of this board if there are enough requests for it.

As usual the board can be wired for triode, UL and pentode mode. Provision for global negative feedback is available. It can be used if needed, and it is needed in pentode mode.

There have been the usual delays due to "trying to keep my day job" activities and Sherri being away for much of the year, but the most recent delays are for two good reasons. I got some $1 output tubes that may work in this amplifier, so I had to try them out. It turns out that a PC board change was required. During the development of the original board I built a hybrid amplifier that used a tube - mosfet "Darlington" connection for the output. Again a PC board change was needed. The changes have been implemented and a prototype board made. The board works as expected with 6BQ5's but I have not tested the other combinations yet.

I have included some photos below.

The first board, it hasn't seen any electricity yet.

A careful look will reveal my first problem with this board. The driver tube on the right is installed. The driver tube on the left, won't fit in the socket! The holes for pins 2 and 3 are smaller than the pins on the tube. Next time I'll look more carefully at the sockets before installing them.

It is cranking some tunes here. The flying resistors are the GNFB resistors. I have not yet decided on an optimum value.

I added some much bigger transformers for much bigger sound.

Board on test. Note the distortion number of 0.346%

SY convinced me to try this. The cathode resistors and bypass capacitors have been removed and replaced with LED's. In this case three special LED's were mounted in series on a heat sink material.The amp does indeed sound fantastic with LED's in the cathodes of the output tubes. I am looking for an off the shelf LED that will work here. It may be included in the design.

It can light up the whole room too!

Tubelab SPP Boards

The SPP boards are in house. This one has been kicking out the tunes for a few weeks now.

It has also been the subject of some extensive experimentation. In fact I decided to build another one, and then three more were assembled so that I could shoot all of the pictures for the manual.

All of the pictures have been taken for the manual and I am currently putting it all together on my laptop while travelling. I have discovered a few photos than must be redone once I return home. We have ordered the parts for a small parts kit which will consist of everything needed to populate the PC board. Both should be ready within two weeks of my return to Florida.

The next step in the “not quite so Simple P-P” is a P-P Driver board. The design goals are simple, it needs to be capable of driving a pair of tubes to full power in P-P. It should be capable of driving anything from a pair of 6V6's or 6L6 type tubes in conventional grid drive up or cranking a pair of sweep tubes to near meltdown in screen drive or even cathode follower mode. This board must be able to drop in to the original 300Beast amplifier, replacing its driver board while offering equal or better sound quality.

OK, writing down some lofty goals on paper (or in the computer) is easy. Building the circuit and getting it to work is not. This project has been alive since before there was a Tubelab web site. In many ways it has evolved from my 300Beast amplifier. It incorporates a lot of the 300Beast design and many techniques that I have learned since the 300Beast was built.

I laid out this prototype PC board in a hospital waiting room in Pittsburgh during the Christmas (2008) holidays. I started building it when I got back to Florida. When I completed it and powered it up I found that I had swapped the top and bottom layers when making the board, resulting in a mirror image of what I laid out. After nearly tossing it in the trash, I realized that all I needed to do was solder a set of tube sockets on the back of the board. The board is alive and working now. More details are on the Universal Driver Board page.

Octal driver board

I have received numerous requests for an amplifier using all octal tubes. Octal tubes use up more board space than miniature tubes and are usually more expensive, so I have not done a completely octal amplifier yet. I decided that a complete P-P amplifier would not fit on a single PC board if it used octal tubes, so an octal tube amp must be broken up into blocks. While working on the Universal Driver Board a thread developed on the diyAudio forum dealing with designing an amplifier using 6L6GC's in AB2. The thread originator desired to use 6SN7's for the driver tubes. I was assisting in a "paper design" effort by modifying a Universal Driver Board for this project. I got a little carried away and 350 forum posts later I had a working a prototype amplifier that used all octal tubes.

See the Octal Driver Board page for more details.

Spud SE

I have developed another future product, the Spud SE. This is a "really simple SE". It started out as an email conversation about "Spud Amps" and evolved into this. There was very little interest in this project until lately. In the past two months there have been a few questions about the availability of this board, unfortunately I don't have the time to work on it right now, and I am not sure that it would ever pay back the initial PC board design costs in today's economy. This one is on hold for now.

MiniTron

I got a lot of email about the MiniTron. Some called it cool, some thought it was criminal, most was positive though. Either way, the MiniTron and its derivatives will have to wait until I have more time. Circuit Cellar asked me to write an article about it for their magazine. That article is in the current issue of Circuit Cellar (Oct. 2009). I don't have the time right now to do any further development on the software. The MiniTron now sits, lonely, all alone in the closet, waiting for some attention.

The MiniTron amplifier was designed for entry into the Circuit Cellar / Microchip design contest. In a sea of microprocessor powered IC based gizmos this vacuum tube amplifier WON ONE OF THE CATEGORY PRIZES!!! It is featured in the Oct 2009 issue of Circuit Cellar magazine, and is currently on their web site.I will be posting much more information about this design in the coming months. You don't care anything about an amplifier that has more chips than tubes? I don't blame you, but this project has taught me a lot about the distortion producing mechanisms in vacuum tube circuits. I am using this information to design some radically new vacuum tube circuitry, that is silicon free, as well as some unique hybrid designs. Unfortunately the events outlined below have severely slowed down the progress. The amplifier you see above was completed in early October, and hasn't been touched since.

Progress has resumed slowly with the design of an all tube cathode follower amp incorporating some of the technology developed during the development of the MiniTron. A new page for this design will be created soon. One component of the MiniTron is the cathode follower amplifier, which still has its own page. The complete contest submission including schematics, source code, and design details can be found here:

http://www.circuitcellar.com/microchip2007/winners/MT2209.html

Tubelab SSE update

There have been a few questions related to the construction of amps. Many builders request an exact parts list with part numbers and transformer recommendations for the “standard configurations” It is impossible to cover all of the possible combinations, but I have added specific transformer and configuration information to the Tubes and Applications page. There have been numerous inquiries about a small parts kit for this board. I have decided to offer a kit containing all of the parts to populate the PC board. Most of the parts have been ordered and I should have everything some time in November.

Tubelab SE update

There have been far more questions about the Tubelab SE. The parts list has been modified to reflect the changes outlined below.

If you have a Tubelab SE and it is working OK, leave it alone. If you have any instability, or have experienced blown mosfets, make these changes. If you are starting a new build, or in the process of building one, make the following changes.

Increase the grid stopper resistors on the 5842 (R31 and R32) from 1K to 4.7K ohm. These resistors should be carbon composition if at all possible. The 5842 will oscillate if given the opportunity. The higher stopper resistor value will help. If you are using a volume pot on the board, make sure that the case of the pot is grounded. The cable from the board to the input jacks should be connected to ground on the PC board, and connected to ground at the input jacks. All ground connections to the chassis should be made at the input jacks. See the wiring diagrams for the Simple SE for details. The oscillation occurs in the 100KHz range which is inaudible, but may be heard as a hum or buzz. It can also cause the output tube current to rise unexpectedly.

R14 and R25 set the operating current for the mosfets. The parts list shows a value of 20K ohms. This is good for boards that operate from 275 to 325 volts of B+. Boards that operate from 325 to 360 volts should use 30K ohms. Boards that operate from 360 volts to 400 volts should use 36K ohms. This will reduce the dissipation at higher voltages. The B+ voltage may be considerably higher than normal during the time that the tubes are warming up. This is especially true with low cost power transformers. Voltages of 450 volts and higher have been seen, which combined with high mosfet current can lead to blown mosfets on power up.

C6 should be rated for the full B+ voltage, and the expected rise on power up. This means a 350 volt cap is OK for 45 amps that run at 275 to 325 volts. A 400 volt cap should be used if your B+ is from 325 to 360 volts, and a 450 volt cap is needed for boards that run from 360 to 400 volts. It is unlikely that a 100uF cap that will fit the PC board can be found in the higher voltages. It is OK to use an 82 or 68 uF cap.