Lazy Singing Bush

[Zen Mod]

I forgot how long I have these pcbs; heck, even added (just for giggles) one small cap, just to have a reason to make new gerbers
Anyway, N-MOS/Schaded N-MOS/SIT source follower, loaded with P-MOS in Mu-Follower arrangement, blatant shameless flipped Xerox of Papa's BAF 2015 amp Mu part

Mosfets are IXYS hockey pucks
SIT can be THF51, 2SK180, 2SK182

Single rail supply, GND up, neg 60Vdc down (remember that GND is State of Mind - I did use +/-30Vdc PSU from my T-Bed 2 accordingly wired, without using center rail)

Iq 3A2

Necessary pictorial info about "main" pcb


(late) edit on Jan 25. 2024. - change of Mu resistors values, as per Ben Mah's suggestion - post #17, https://www.diyaudio.com/community/threads/lazy-singing-bush.388164/post-7069779

in short - 0R2 at SIT side, 0R3 at puck side

 

[Zen Mod]

Necessary pictorial info about Mu (load) pcb

 

[Zen Mod]

now, proof is in Da Pudding

pictures of pure MOS iteration

 

[Zen Mod]

graphs for 8R load

NB that with my measurement setup (EMU1212 with funny blackbox interface) I can't put more juice in amp than 10W@8R; frankly, I don't care for more

for all iterations, I can declare it as very fine SE amp in range of 20W, with wild burst of power/headroom - up to 50W

go figure

 

[Zen Mod]

Pure MOS, graphs for 4R load

 

[Zen Mod]

MOS Schade

R16 changed to 750R, installed R17

there is another (extreme) combo of said resistors values, R16 1K1 , R17 10K, giving more SIT-ish transfer function, but that's too much for my liking

 

[Zen Mod]

MOS Schade, 8R load graphs

 

[Zen Mod]

MOS Schade, 4R load graphs

 

[Zen Mod]

OK, time for THF51

remove MOS, mount SIT, change ZD3, rotate gate zener

 

[Zen Mod]

SIT, 8R load

 

[Zen Mod]

SIT, 4R load

 

[Zen Mod]

Few notes:

- THD Spectra can be also altered changing ratio of two resistor groups in Mu Follower - as I did accommodating various SITs in origin Singing Bush; now, we can see from THD graphs that in all 3 iterations main output part somehow is preferring lower loading; that means - if we change Mu for lower gain, (increasing value near Source , decreasing value of other resistor group), main output part is seeing lower impedance
though, how far to go with that is strictly personal taste tyhing, thus I didn't go further than contemplation

-Power Up/Power Down thump in speaker - no more than gentle pull or push of speaker cone; even Lowther will cope with that without any consequence; applies for all 3 shown iterations of amp

- will try with 2SK2087C, most likely with decreased rail to 50V and Iq in range of 2A; maybe even trying IRFP9140 instead of IXYS P-Mos; today or, more likely, tomorrow

- yes, Choked Singing Bush and Warbler pcbs are with Fedex

 

[von Ah]

Fascinating stuff, ZM. Well done. I expect you are enjoying this.

 

[Zen Mod]

I'm waiting for Ben Mah's comment/comparison with his amp(s) - he's liking these very much

 

[Botte]

Looks similar to my build. I’m sure yours holds the Bias more steady. My amp drifts up till it hits thermal equilibrium.

Here is a graph with 2.3A bias at 65v. Varying Vds changes the distortion. The 4ohm runs out early due to the lower bias.

thanks for bringing it to the group. How about a standalone frontend to drive it?

 

[Ben Mah]

Comparisons - my power supply puts out a bit more than 60V, probably up to 65V or so depending on the time of day as the wall AC voltage varies by up to 5 or 6V. I run my amps at 3.0A and the THF-51S Vds at 33V . I take my speaker output from between 0.3R at the Mosfet and 0.2R at the THF-51S.

With a clean 1kHz input signal and output to 8R, one channel had 0.15% THD and the other channel had 0.066% THD, 2nd harmonic dominant. It maxed out at approximately 50W into 8R, THD was bit less than 2%. I was using an cobbled together KP926B voltage gain stage for the 50W measurements and an ACP+ to boost an oscillator signal for the 1W measurements. However I listen to music with my Sony VFET 2SK79 preamp/gain stage and the THD is approximately 5% at 50W. Previous I had used a Luminaria preamp for listening to music and the THD was similar to the 2SK79.

So the performance of my amps is similar to the Lazy Singing Bush at 1W. Zen Mod's 1kHz signal seemed to be at its limit at 10W into 8R and 20W into 4R as the higher order harmonic distortions appear very high. That is probably not the amplifier's distortion profile but a reflection of the signal source. Testing of a high power follower amplifier can be tough to do without a clean signal. Otherwise the measured distortion is that of the signal source and not of the follower amplifier.

My Mu Follower Follower THF-51S Amp

 

[Zen Mod]

Tnx Ben

Knowing that you already did your homework regarding Mu resistor groups value, I didn't followed it further

So, I believe it can be carved in stone as 5*1R as upper group and 5*1R5 as lower ( puck source side) group

 

[Zen Mod]

Now, as I promised - play with 2SK2087C

rail decreased to -50Vdc
Iq decreased to 2A; did that prior to reading Ben's post, so for my tests I did stay at 1:1 ratio of resistor groups , decreasing them to 0R33 (3*1R in parallel); easy to recalculate to 3:2 ratio

schematics, edited to show differences ( said Mu resistors, ZD3 in biasing circuit, upper resistor in IRF cascode)

 

[Zen Mod]

first tests with IXYS puck in Mu

 

[Zen Mod]

same, 8R load graphs