45's in Stereomour = wow!

[Dr. Toobz]

I managed to snag a pair of NOS Type-45 tubes from the 30's or 40's (Slivertone brand, ST shape) for just $60 and made the few circuit changes to my new Stereomour to see what all the buzz was about. They sound great! The 45's have zero hum on my amp, even over headphones wired to the speaker terminals to assist in performing hum cancellation (as my ancient, cheap multimeter does not go low enough). When plugging the Sovteks back in, the sound got harder, the soundstage more forward, and the bass less dynamic. The 45's seem to do everything the 2A3's do, but better. Also, the lower output power is not a problem in an urban condo with wood floors

This does lead to a question: the Rp on the 2A3 is 800 ohm, and the 45's are somewhere around 1800 ohm. Since the primary resistance of the output transformers is not changed between tube types (e.g., remains at 4k) and therefore maintains the same step-down ratio, would not the secondary impedance at the speaker terminal be double that of the 2A3 for a given tap? (So, 8 ohms becomes 16, etc). I ask this because my Klipsch speakers seem to be more like 5 ohms versus 8, and sound very congested and bloated unless using a lower impedance tap than 8 ohms. I originally wired the 2a3 version of my Stereomour for 2 ohm outputs, as to obtain the best damping factor.

 

[Paul Birkeland]

Hello Angelo,

While it is true that the typical RCA operating point on the datasheet specifies an 800 Ohm plate impedance, this is one of many possible plate impedances depending on how you run the tube.

I will quote PJ from a post on another forum about 2A3 operating points:

"At 300v, 40mA, the reference load is about 5300 ohms

At 300v, 50mA, the reference load is about 4000 ohms

At 250v, 60mA, the reference load is about 2500 ohms"

The rule of thumb is to triple the plate impedance and use that as a loading resistance, so you can see your 800 Ohm number tripled gives 2400 Ohms under the 250V/60mA condition.  The 300V/50mA condition is approximately where the Steremour runs, and we could work backwards to guestimate that the plate impedance would be about 1.3K (knowing PJ, I would guess it's actually a little less than that). 

With all of that basically set in stone for the 2A3, the possible changes for running the 45 are plate to cathode voltage and current, so all that was necessary was to select an operating point that yielded roughly the same plate impedance, then adjust the B+ dropping resistors and cathode resistors (I made this sound easier than it is).

Now, how this all works with your speakers is another story.  A speaker that is 8 Ohms nominally might have an impedance dip that goes way lower, and in that case going to a lower transformer impedance is a great idea! 

-PB

 

[Paul Joppa]

Dr Toobs - no-feedback SETs have an output impedance that is usually 2-5 times lower than the nominal load impedance. Other than that, your analysis is correct - the 45, using the same transformer ratio, will have twice the source impedance of the 2A3.

PB - the "3 times plate resistance" rule is just one point on a continuum of good approaches. Plate resistance varies approximately with the cube root of current; the 2A3 is 800 ohms at 60mA, 850 ohms at 50mA, and 915 ohms at 40mA.

The primary impedance is usually optimized for maximum power output, which happens when clipping is symmetrical - limited by grid current on positive grid excursions, and by current cutoff at negative excursions. At this operating point, the output source impedance is the plate resistance times the impedance ration of the transformer, plus the resistive losses in the transformer windings. Assuming resistive losses at about 5%, that would be 0.4 ohms on the 8 ohm output. Figure the transformer primary impedance also includes that 5% as resistive, so a 4K transformer has an impedance ratio of 3800:8 if you ignore the resistance of the wire. The actual output impedance with a 2A3 at 50mA is then 0.4 + 850*8/3800, or 2.2 ohms. With a 45, it would be 4.0 ohms.

By using a lower-impedance secondary wiring, you will get a lower source impedance (better damping), but lower maximum output power as well. The distortion at that lower power is also lower, which somewhat compensates, but the comparison is complicated. Mostly what you hear is just the lower threshold of clipping. The clipping will be asymmetrical with an off-optimum load impedance, which emphasizes second harmonic over the third - a less unpleasant distortion - so the perception is even more complicated than the analysis.

 

[Dr. Toobz]

Thanks for the very detailed explanations! As with the SEX amp I built in 2008 or 2009, I will have to play around with different taps and see which one represents the best compromise with the 45 tube and Klipsch combo. The impedance switching boards look to be the next thing I'll buy from Bottlehead....

 

[Dr. Toobz]

Interestingly, the best sound seems to be with the standard 8 ohm taps, which makes me believe that there is more to the equation with my Klipsch speakers than just the damping factor. The 12" woofers in these things are sealed, not ported, and the -3dB point is about 50Hz. The sound on the 2 ohm taps is almost too lean and dry, or over-damped. While I was concerned about the bass sounding too "poofy" on the 8 ohm taps, it sounds tight and natural with the 45's. This is a different experience than when using the speakers with the SEX 2.0 and Specos, where lower output impedances (even via cathode feedback) sounded better. However, the SEX amp through the 8 ohm taps on the Specos is liable to have a higher THD than 45's through the much better Bottlehead OPT's, so maybe that's what I was hearing before.

In any case, I think I've finally settled on the perfect match for my Heresy speakers. Perhaps now I can actually listen to the music versus constantly analyzing the sound quality!

 

[Doc B]

45s tend to sound a little lighter weight on the bottom end, and that may have something to do with the perception of cleaner bass. They also have a very black background, which also helps the bass to sound cleaner.

 

[Dr. Toobz]

I have to say, I am more and more fascinated with this amp the more I play it! The 45's are just perfect: zero hum on my 99dB/W speakers, black background, tuneful bass, eerily-realistic mids, buttery-smooth highs. It's everything I loved about my SEX 2.0 amp, just more of it.

Interestingly, I was not very impressed with the Sovteks, finding them a bit "hifi" and sterile sounding. Running them in the amp as pseudo-45's didn't do it for me, either - compared to the 45's, they have a smaller soundstage, less bass, and are more "zingy" sounding where the 45's are smooth. If I had judged the amp on the 2A3's alone, I would have been happy, but the 45's have something special that I just "got" right away. I'm glad I decided to go through the trouble of finding a pair of NOS 45's, which ended up being less expensive then a pair of new 2A3's.

Maybe it's time for some meshplate 45's next? 

 

[azrockitman]

I'm building my first Stereomour soon.  Based on a few comments I've read, I tend to think I would like 45's better than the 2A3's.  Any opinion on whether I should first stick with building it with 2A3's and then altering later to 45's?  Or just start with 45?  I love building.  Not as fond of "undoing" and altering post production.  Thanks for any comments. 

 

[jimiclow]

In my case, I started with 45 configuration as I also don't like desoldering parts.
From the start, I also upgraded the coupling and parafeed caps, upgraded all electrolytics to Panasonic 105C, stock pot to Alps.
With the stock 45 configuration, the tubes are ran hard at >11 watts dissipation. I had to change the power supply resistors to drop the voltage. Right now it's running at 7.7 watts, producing around 1.5 watts/channel, still plenty for a quad of 97db Hoyt Bedfords and tubes last longer.

 

[azrockitman]

Thanks jimiclow!