Sound Practices Magazine Onine!

Not Your Father's Dyna....
A poor man's SE Amp
Article By Dave McDonald and Matt Kamna
From Sound Practices Issue 10, Winter 1995

Dumpster diving Dyna fans take note!

How did I end up with a single-ended Dyna stereo 70? A little background will help to understand. As a kid I grew up with vacuum tubes — Knight Kit regenerative and superhet shortwave radios, 12W mono amplifiers, and later a Citation 30W integrated amp kit. There were solid state kits too, including an 80W stereo amp. I did a lot of tinkering in those early years, mostly making solid state radios, science projects, and the like.

A couple of years ago over at my friend Matt's house, I was exposed to SE WE212E amps flooding the room with light and sound. The 212E amps were too expensive for my budget. I also assisted in the creation of 2A3 and 300B amps. Listening proved that a low power amp could give that same wonderful sound.

I remember one Oregon Triode Society "DIY night" meeting where I got to hear many amps and speakers that were either modified or built from scratch. There were pentodes, triodes, tri-amp systems driving horns, etc. Those were exciting listening days!

I decided to pick up an inexpensive vacuum tube amp and start experimenting, so a Dyna ST-70 soon followed me home. Music listening was much more comfortable compared with my solid-state amp.

Then I checked the ST-70 on the test bench. The amp had a bit of a problem putting out full power. While troubleshooting the problem, we uncovered a design flaw that causes premature 7199 failure. Ironically, the "failed" 7199s sounded better than the new RCA tubes that I installed. My theory was that the unbalanced phase splitter drove the PP output stage toward single-ended operation. The old tubes were much smoother and warmer sounding. The person that sold me the amp heard the difference and agreed with me.

We tried operating the EL-34s in triode mode by disconnecting the screen from the output transformer and tying it to the plate through a 100 ohm resistor. The feedback cap from the transformer screen tap connection was left intact.

Then we tried running the amp with the main feedback loop open and the 7 kHz pole at the input pentode removed. Measured open loop performance was hopeless. The damping factor had dropped to 1.5. The new frequency response was down 3 dB at 42 Hz and 26 kHz. I didn’t even want to try to listen to my ported 2-way speakers with an amp with that low a damping factor.

I looked at commercial mods, but I decided that mod kits with excess gain were only going to make my amp sound more like a transistor amp. I spent a lot of time thinking about what I could build on this chassis that would be better than stock. I wanted a simple circuit that used the stock Dyna parts and sounded great. While tinkering, I continued to enjoy listening to my Dyna with "bad" 7199s, a JVC XL-V161 CD player, and some old 91 dB sensitivity KLH Model 33 loudspeakers.

Next I tried a variation of the triode mod mentioned above but with the global FB disconnected and feedback from the output transformer secondary back to the EL-34s installed to improve damping factor. I tied the 4 ohm taps to ground. Then I connected the 16 ohm taps to the EL-34 cathodes through 15.6 ohm resistors.

In this configuration, the driver circuit was simply not up to the task of driving the triode-connected EL-34s. I came to the conclusion that the 7199 drivers were hopeless without some heavy circuit changes. I feared that the circuitry was going to get more complex.

Another thing I didn’t like about the stock ST-70 was that the power transformer ran way too hot. While pondering how to relieve the heat problem, I had a flash of inspiration. Why not remove two EL-34s and go single-ended? The B+ and filament draw would be cut drastically.

Typically, single-ended amps require an air gapped output transformer to prevent core saturation under DC current. So I carefully disassembled the output transformer laminations and put then all back together with all the "E"s on one side, a .010 inch spacer in the gap, and all the "I"s on the other side of the spacer. I hooked up the single EL-34 in "100 ohm triode mode" with 3 dB feedback around the output tube and transformer.

Hum increased because there was no global feedback and no push-pull cancellation of B+ ripple in the output stage. A 100 uF capacitor was added in parallel to both the output and driver B+ supplies to reduce the ripple and make the unit much less objectionable to listen to. The selenium bias supply rectifier was also replaced with a fast, soft recovery diode to increase reliability. The bias supply output filter cap was increased in value to 330 uF.

At first, we tried to use the triode section of the 7199 as the driver. Using a pentode driver stage was not even considered, but the 7199’s low mu of 17 made it unsuitable for a simple single gain stage. We switched to a 12AT7 driver circuit used earlier for driving the really low gain 300Bs. One section of the dual triode was used per channel.

We sifted through seven EL34s to get the best results on the test bench. The old leaky paper coupling caps were replaced by some new MIT polystyrene jobs. In the process of making the modifications, some of the traces were cut off the old phenolic circuit board. There was no going back to the stock circuit. This was either going to work and sound good or not. Now we were ready to listen to a single-ended Dyna amp!

Initially we listened for a while on a highly modified three piece CD player and a set of tweaked Spica TC-50s. Then it was off to dinner and another fun D.I.Y. O.T.S. meeting. A group of a little over 30 people showed up for the event where the amp was  very well received. The system used had a slightly modified CD player and Lowther horn speakers with about 100-101 dB sensitivity.

Then it was time to try it at my house on the JVC player and KLH speakers. Later it was tried on the Spica using the JVC player. The KLHs played louder, but the Spicas were more controlled in the bass and had a different sound overall. Using the 84dB sensitivity Spicas required turning up the volume to hear very quiet sections of classical music. Then during loud sections of the music clipping was heard. When overdriven it sounded better than a solid state amp, but not as good as some other tube circuits.

We listened to a variety of music: jazz, classical, vocals and soundtracks. The midrange was smooth and it also had a good response at both the high and low end. It was easy to just sit back and listen. Several listeners commented on how different it sounded from the stock Dyna.

The power transformer runs much cooler now, and all of the original magnetics were utilized, even if they are modified. I have been enjoying listening to this new single-ended amp and expect to do so for some time to come.

Perhaps you will roll your own SE Dyna and maybe think of further refinements to try yourself. For example, we paralleled the EL-34s to increase the output power. In order to optimize power and distortion it was necessary to move the plate connections down to the screen tap on the output transformer primary winding. This yielded 16 watts at 5% THD and 14 watts at 1% THD.

ST-70 SE Measured Specs
10 Watts output @ 5% THD @ 1kHz
6 Watts output @ I% THD @ 1kHz
Voltage gain  17.5dB
Residual hum and noise 1.2 mV RMS
Damping factor 3.8
I KHz square wave response looked excellent
Freq. response - full power bandwidth -3dB @ 15Hz and 60 kHz