Greetings!

CHANGED: 4-11-2005
The OPT impedance is 12K.

This is my version of a "SPUD" amp.
Spud amps are single tube per channel and are proported to be the cleanest, most natural configuration.
Not many tubes can be used because a good degree of input sensitivity is required.
Favorite tubes to use vary considerably.
More sensitive tubes (usually pentodes) can be driven directly.
Less sensitive tubes use an input transformer for matching and signal boost.

One can argue that using an input transformer is cheating.
Since the transformer is technically an "amplifying device" it adds another layer of amplification to the "single tube" amp. In addition, the transformer adds its own character to the amp, defeating the idea of a "pure" single tube.

Now don't get me wrong, I am not saying that a Spud amp with an input transformer is a BAD thing. In fact there are several excellent amps out there that have been made this way. And that's the bottom line, isn't it? To get good sound.

My amp takes the position that good can be cheap. So that is the design objective. This amp was inspired by the many small output trannys (OPT) that have set free from old consoles; especially the European ones, that appear from time to time on eBay. Combine a pair of those with a used power transformer and you've got the building blocks for a small amp.

I had a pair of cosmetically challenged Telefunken OPTs lying around waiting for inspiration. I also had a pile of NOS 6197 tubes lying around, so it was a small leap to use these sensitive 7W pentodes for this amp.

The amp described here was actually built and currently resides on a breadboard. Eventually it will get to a chassis. So this amp represents not just an idea, but an actual physical unit. Also, during the design and construction a lot of variations were tried. So this amp represents near optimum operation (IMHO) for this tube.

First the output transformers:
I actually tried several. In fact, I need to retry some of them after the final circuit was decided. But the bottom line is you are going to need 10K plate impedance minimum. Anything from 10K to 13K seems to work well.  With this amp, the higher impedance seems to produce a better bass. (See note below.)

The OPTs I am running are about 15K:8R. these also have a considerable primary DC resistance (DCR) and that is a factor when making your power supply.

The operating point for the tubes was chosen to be 250V at about -3V bias. Self bias is used here, but with a small change, grid bias could also be used (better for the 10K OPT's). Since self bias is used in this version, bypass caps are required to maintain the input sensitivity. You can leave them out if your CD player (or other source) has enough drive voltage. If you omit them, you can use higher Z OPT's.

With a mere 3V bias, you can swing 6V peak to peak at the input, which translates to about 2V RMS. My CD player cannot drive this amp to clipping, which is a good thing! I get plenty of volume and never clip. Perfect for a pentode.

The secret to this tube is a stable screen voltage. (Most pentodes like a stable screen voltage.) So my power supply uses a true voltage divider for the screen supply. Another thing about this pentode (and others) is that some degree of isolation of the screens produces lower distortion. The kind of distortion that sounds like modulation ... very annoying.

To design a power supply for your 6197 amp, start at the RBL end and design back towards the transformer. I use 1K5 1W carbon comp's for the screen resistors. For a 150V screen then, I will need about 160V source. (This tube will allow up to 250V screen, but don't pop in a 6CL6 if you run it that way ... these tubes are NOT the same.) At 150V screen and 250V (255V actual) plate I got exactly 6.667 mA in each screen. So the screens are at about 1W standing disappation.

The values in my amp were chosen based on what I had available and the limitations of my power transformer (about 75mA). If you have a larger power transformer, you can adjust these values. So RBL was selected from stuff on hand. In fact, ALL the resistors in my power supply were selected based on actual measured values, which may be different from the indicated rated value.

So pick a value for RBL that will give you at least 7mA of bleeder current. The combination of bleeder current and screen idle current determines the value of RS. The voltage values (for my amp) were made exact as possible. So my 275V plate supply voltage takes into account the IR drop in the OPT primary. You should do the same. MY RS ended up being about 5200 ohms. I hand selected resistors until I got very close. Note that this resistor along with RBL will need to dissapate a lot of power.

So you measure the DCR of the OPT. You will draw about 25mA through this resistance. For a 250V plate, you then know what the plate supply voltage has to be. The difference between this voltage and the screen supply voltage (160V) is what you need to calculate the value of RS for your amp.

For my amp I used an old power tranny from the junk box. This tranny outputs 550V CT (275-0-275). It also has the 6.3V heater winding (you will need 1.5A rated). I use the 5Y3 tube as a rectifier and at the total current of about 75mA I get 290V at the input filter. Thus my primary plate supply is fed by a 220 ohm hand selected at 210 ohm resistor.

If you have a more powerful tranny, you can increase the bleeder current (or distribute the bleeder current) and may be able to get the supply set with "normal" value resistors. Also, you can use solid state rectification instead or a different rectifier tube. For a solid state full wave, you'll need about 210V per leg or about 215V into a bridge. There is also no reason not to use a voltage doubler if you want, and you'll need only 105V or so at 150mA or better. Whatever you choose, use the appropiate R to create the correct plate supply voltage and your screen voltage will fall into place.

In my version, I used a 47uf/450V filter at the power supply input. The plate supply is bypassed using a 120uf/400V unit. Since the amp is all class A, the 120uf is probably overkill. In the screen supply I opted for more overkill to try to maintain a solid supply in this area as per my initial design goal. Therefore, another 120uf cap is used here. The amp sounded the same using a 47uf cap here as well, but I think the bass was cleaner with the larger cap. 400V units can be used instead of 450V. Even 350V units can be used after the R although you'll be running near the peak recharge voltage. the screen cap can be a 350V unit.


Notes:
1. After I made the change in the screen circuit (added the 1K5's to each screen) there was a decrease in plate current which in turn caused a rise in all the voltages. The amp settled in at 263V plate (main B+ at the input to the filter rose to 295V) to ground at about 24mA or about 6.25W dissapation. The screen settled in at 160V source with 151V net at the screen to ground. Screen current actually went DOWN a tad. I replaced the original junk box 1000uf caps with some larger (physically) new caps of the same value. The change for all this in the sound is quite stunning. This jives with my pushpull version (see other pages) which "settled into" about the same range, but with a higher plate voltage. I like this tube a lot.

2. 6197. Nice looking power pentode about the size of a 6CG7. It has a 7.5W plate and a 2.5W screen. Max plate = 300V and max screen = 250V making this a versatile performer. Among its benefits is low distortion approaching 5% at rated output. Expect about 2W+ from this circuit.

3. I've used a lot of "vintage" and recycled parts in my amps, and one thing I have learned in a general way: the physically larger the cap for a given rating, the better it sounds. Now I am going to exclude the liquid dialectric (large can) caps from this statement; I'm talking about solid caps. So in this example (this amp) I replaced the tiny "matched pair" of modern bypass caps (1000uf/16V) that originally came off some circuit board, with some vintage 1000uf/16V caps that are about 4 times the volume. The change in sound was impressive as well as the deduction in distortion. I have found the same to be true of filter caps. For example 47uf/450V (the modern ones are pretty darn small) sound better and filter better when they are large. I have some that are about 1.25" long and about 0.5" in diameter, and the other ones that are 2" long and .75" dia sound and perform way better. FYI FWIW. So if you MUST bypass in the audio chain, use the best caps you can get but don't be silly and spend $30/ea. The Nichicon caps rated at 105-degrees all work great as to the Panasonics.

4. This amp is destined to sit on my bench to listen to while I work on other amps.