Post by Owen Y on Sept 26, 2018 19:54:11 GMT 12
Currently I am using these 2A3 Push-pull power amps - which I built (in 2009) because: - I wanted more power than a SE 2A3/300B amp could produce - I wanted to avoid 300Bs (costly) & use 2A3s (which I had a few of). - I wanted better bass control & dynamics (better damping factor of a PP output). - I wanted an amp that would suit all types of music (which I found that an all-triode SE amp did not - although a pentode triode can do it better). - Push-pull also offers lower distortion ('distortion cancelling'). - I already had some nice Output transformers (vintage 25W Partridge 6k6:8). - I had read good things about a design built by 'Tossie' in Japan. However, I knew that I would have to accept some minor compromises: - PP would not have the absolute HF purity of SE. - A tube phase-splitter design 'could' have some degree of 'crossover' distortion. 
|
Post by Owen Y on Sept 26, 2018 19:59:15 GMT 12
The freq response looks like this....  - Bass was surprisingly good, only 0.4dB down at 20Hz as can be seen (the Partridges seem to be excell in this regard, the 2A3s would usually be loaded ~5k PP & 6k6 may sacrifice a little power but possibly lower distortion?). - HF rolls off a little quicker, 1.6dB down at 20kHz but not bad.
|
Post by Owen Y on Sept 26, 2018 21:53:46 GMT 12
The schematic: - Tossie's circuit from Valves' World, with my changes, partly to suit a 2 mono-chassis build, marked on. - I omitted the Negative Feedback circuit, after experimentation (I wasn't convinced by the sound).  (A correction - the 76 input tube cathode bypass cap has been restored.)
|
Post by colinf on Sept 26, 2018 22:22:58 GMT 12
Nice! You get about the same damping factor out of a PP amp as an SE amp, assuming there is no NFB on either. The driver stage can be made more symmetrical to get even lower distortion by swapping the 18k resistor for a constant current source (made up from a resistor in series with a constant current diode), and making the two anode resistors 33k.
AMR-iFi R&D
|
Post by Owen Y on Sept 27, 2018 11:07:37 GMT 12
Yes I guess that damping factor is similar between SE & PP when you have OPTs on both with primary imped to match the tube/s & sec imped to match loudspeaker. Previously I had a single 'full-range' driver lspkr & SE amp was ideal. (As Nelson Pass explained.) But eventually I devised more wide bandwidth lspkrs & wanted better bass control & dynamics. P-Pull balance - thanks for the CCS suggestion, which I should try (along with the Blumlein Garter cct), but life is short  , I am lazy & I have tried to minimise SS bits in my amps. However (thanks to Murray Dick among others) I have a plan involving more iron (& $$ unfortunately) - IT coupling
|
Post by Owen Y on Sept 27, 2018 13:35:26 GMT 12
Some further test results: Max power before clipping is around 7.6 Watts into 8 ohms (with no feedback). About the same as 300BSE but triple the 2.5Wpc power available from self-biased 2A3SE. (Tossie gets ~12 Watts (50% more) using higher dissipation Sovtek 2A3s.) With my 100+dB/W lspkrs, just 0.025W will generate over 85dB SPL at the listening seat 4m away. Input sensitivity is high at 340mV for max power out. 
|
Post by foveaux on Sept 27, 2018 18:46:44 GMT 12
Meant to say many moons ago - I admire your chamfered timber feature leading down to the valve mounts - so lovely seeing timber, rather than the usual silver/black, for audio gear 
"I see music as a lifetime affair." [Rory Gallagher]
"Free - I miss that band, but when I look back, we were very young" [Paul Rodgers]
850 posts
|
Post by Owen Y on Sept 27, 2018 21:40:44 GMT 12
Thanks foveaux - I can do woodwork easier than metalwork, & I've got a garage full of nice old timber. But sometimes you need metal for electrical shielding 
|
Post by colinf on Sept 27, 2018 21:45:04 GMT 12
Agree, the casework is lovely. I thought about an interstage transformer but that wouldn’t be the lazy way out!
AMR-iFi R&D
|
Post by colinf on Sept 27, 2018 22:00:53 GMT 12
Tha article about damping factor on full range drivers was very interesting. The fact that you removed the nfb of your push pull amp and the sound improved, I wonder if damping factor was involved in your sonic choice.
AMR-iFi R&D
|
Post by Owen Y on Sept 27, 2018 22:24:58 GMT 12
Square waves are revealing in a different way: 1kHz looks quite tidy, as it should, with a suggestion of LF roll-off on the tops.  100Hz shows more LF roll-off.  30Hz indicates significant LF weakness (but my signal generator is responsible for a good deal of that).  At 10kHz we see a hint of crossover distortion (discontinuity) & a touch HF ringing.  And at 20kHz things look a bit uglier, HF roll-off (vertical edge slope), the crossover discontinuity again & maybe some harmonic distortion beginning to show (on the tops). 
|
Post by colinf on Sept 27, 2018 22:39:09 GMT 12
What output power did you measure it at? Does crossover distortion disappear at lower power? Square wave testing into transformer coupled circuits usually isn’t very revealing.
AMR-iFi R&D
|
Post by Owen Y on Sept 28, 2018 11:58:26 GMT 12
Hi colinf - I think this test was taken at same 1W output. Although my ears tell me that this amp sounds very good, I wanted to check my amplification chain prior to some loudspeaker testing which I am gearing up for ( new thread), something I've been wanting to do for years - optimise my loudspeakers in-room, using measuring tools.
|
Post by Owen Y on Oct 4, 2018 21:47:14 GMT 12
Final measurements: Amplifier Gain is 23x ( 27.2dB) - with no feedback & using NOS RCA 2A3s (vs Sovtek 2A3s).
|
, I am lazy & I have tried to minimise SS bits in my amps.
