The follower has now been working perfectly in my bedroom system for a couple of months. Very nice, with the obvious exception of the bass. Df=13 just does not cut it with those notoriously hard to please speakers.
Biamping looked like a must, so i decided to explore a crazy avenue: build an identical follower but using high Gm mosfets in order to obtain a bit more damping. Long term this is not going to be a viable solution, except perhaps in winter but it also satisfied my curiosity how a mosfet follower compares to a sit follower when used full range.
Distortion certainly went up but not that much, about 2-3 times. From ultralow it became lowish
The sound character did not change in any dramatic way, voices are still very nice but the sound is considerably more dull, less detailed and less spatious. -3db is still above 100kHz. Df about 50.
Biamping is clearly better in every way, but the bass, although more acceptable, still sucks. Despite my desire for keeping it SE the bass will clearly need to be PP and directly coupled.
First pic is of unfinished mosfet follower, second is of the biamping setup.
Biamping looked like a must, so i decided to explore a crazy avenue: build an identical follower but using high Gm mosfets in order to obtain a bit more damping. Long term this is not going to be a viable solution, except perhaps in winter but it also satisfied my curiosity how a mosfet follower compares to a sit follower when used full range.
Distortion certainly went up but not that much, about 2-3 times. From ultralow it became lowish
The sound character did not change in any dramatic way, voices are still very nice but the sound is considerably more dull, less detailed and less spatious. -3db is still above 100kHz. Df about 50.Biamping is clearly better in every way, but the bass, although more acceptable, still sucks. Despite my desire for keeping it SE the bass will clearly need to be PP and directly coupled.
First pic is of unfinished mosfet follower, second is of the biamping setup.
Attachments
Is that an Apple Mac Pro 300B amp??
No. It's a dual PS for a phono and line stage. 83s on top. Must weight about 40kgs. Apple knew what they were doing
I did some distortion measurements yesterday. It was a hot day with ambient temperature at 27+ degrees C and high humidity, so having 200W of heat pumping into the room was no fun. But as always, the diy was fun.
As usual, my rig comprised a Focusrite Scarlett 2i2 V2, a Victor (diyAudio member vicnic) 1kHz oscillator, a diy Akitika/Dan Joffe switchable attenuator, two 4 Ohm resistors in series mounted on a heat sink, and Room Equalization Wizard program.
The 1kHz oscillator outputs about 1Vrms maximum so I had to use my Luminaria preamp to provide voltage gain. I was able to get enough voltage from my Luminaria to get the amplifier to output 40W into 8 Ohms. It did not have enough voltage output to get the amplifier to clip.
I also measured the Luminaria's distortion at Vrms output of 3.15V and 20.0V, which correspond approximately to voltage that would get the amplifier to output 1W and 40W into 8 Ohms respectively. The Luminaria's measured distortion was quite close to the measured distortion of the Luminaria and amplifier combined. The results highly suggest that the amplifier's distortion is low and that a large portion of the Luminaria driven amplifier's measured distortion is due to the Luminaria's distortion. However, I don't know how summing of distortion of cascaded stages is calculated so I cannot determine the amplifier's contribution to the distortion. To manually measure the distortion of the amplifier by itself I need a low distortion gain stage, which I do not have right now.
Also as mentioned, I don't have a preamp/voltage gain stage that can output enough voltage to clip the amplifier, so power output at clipping will have to wait. However, I think 50W may be attainable.
I have attached the Luminaria/amplifier distortion measurements and that of the Luminaria alone for comparison:
Luminaria/THF-51S Mu Follower at 8 Ohm: 1W, 5W, 10W, 20W, 30W, 40W
Luminaria at 10k load: 3.15Vrms, 20.0Vrms
As usual, my rig comprised a Focusrite Scarlett 2i2 V2, a Victor (diyAudio member vicnic) 1kHz oscillator, a diy Akitika/Dan Joffe switchable attenuator, two 4 Ohm resistors in series mounted on a heat sink, and Room Equalization Wizard program.
The 1kHz oscillator outputs about 1Vrms maximum so I had to use my Luminaria preamp to provide voltage gain. I was able to get enough voltage from my Luminaria to get the amplifier to output 40W into 8 Ohms. It did not have enough voltage output to get the amplifier to clip.
I also measured the Luminaria's distortion at Vrms output of 3.15V and 20.0V, which correspond approximately to voltage that would get the amplifier to output 1W and 40W into 8 Ohms respectively. The Luminaria's measured distortion was quite close to the measured distortion of the Luminaria and amplifier combined. The results highly suggest that the amplifier's distortion is low and that a large portion of the Luminaria driven amplifier's measured distortion is due to the Luminaria's distortion. However, I don't know how summing of distortion of cascaded stages is calculated so I cannot determine the amplifier's contribution to the distortion. To manually measure the distortion of the amplifier by itself I need a low distortion gain stage, which I do not have right now.
Also as mentioned, I don't have a preamp/voltage gain stage that can output enough voltage to clip the amplifier, so power output at clipping will have to wait. However, I think 50W may be attainable.
I have attached the Luminaria/amplifier distortion measurements and that of the Luminaria alone for comparison:
Luminaria/THF-51S Mu Follower at 8 Ohm: 1W, 5W, 10W, 20W, 30W, 40W
Luminaria at 10k load: 3.15Vrms, 20.0Vrms
Attachments
-
Luminaria Right Ch 20.0Vrms.jpg -
Luminaria Right Ch 3,15Vrms.jpg -
THF-51S Mu Follower Luminaria R Ch Vds 33,0V Iq 3,0A 40W.jpg -
THF-51S Mu Follower Luminaria R Ch Vds 33,0V Iq 3,0A 30W.jpg -
THF-51S Mu Follower Luminaria R Ch Vds 33,0V Iq 3,0A 20W.jpg -
THF-51S Mu Follower Luminaria R Ch Vds 33,0V Iq 3,0A 10W.jpg -
THF-51S Mu Follower Luminaria R Ch Vds 33,0V Iq 3,0A 5W.jpg -
THF-51S Mu Follower Luminaria R Ch Vds 33,0V Iq 3,0A 1W.jpg
I actually did 0.3R up and 0.2R down. I was going to do 0.25R up and 0.25R down but I ran out of room on the board. 0.2R up and down would probably have worked just as well. The 0.3R up and 0.2R down gives slightly higher distortion. The distortion is low though so the difference is small. I don't have a good reason for my choice - not an EE so I just slap components together and hope for the best.
I believe most of the distortion measured is from the Luminaria preamp. I think the Luminaria is a good match for a follower amplifier as the Luminaria has that nice single ended SIT/VFET harmonic distortion profile.
I believe most of the distortion measured is from the Luminaria preamp. I think the Luminaria is a good match for a follower amplifier as the Luminaria has that nice single ended SIT/VFET harmonic distortion profile.
I don't have a good reason for my choice - not an EE so I just slap components together and hope for the best.
My approach too
I bet the Luminaria really sings in this setup!
It's cooled off a bit, only about 25 degrees C so I decided to heat my room up a bit. Room heaters at 100W is the choked loaded 2SK180 follower for the left channel and at 200W is the mu-follower loaded THF-51S follower for the right channel. Additional heat provided by a Luminaria preamp and ACP+.
The ACP+ is used to amplify the signal from my TDA1541A dac. The dac has only a 75 Ohm I/V resistor converting a +/-2mA signal so maximum output is 0.15Vp.
All I can say is that I really like the sound. I haven't had the 2SK180 followers in my system for a while because of the heat, so my memory is weak, but as far as I can tell, this THF-51s follower amp sounds just like the 2SK180 follower. There is a clarity to the sound. Individual instruments and voices are distinct, and bass is good. There is a sense of realism. Now this is with the 2SK180 on one channel and the THF-51S on the other channel so it is not a 100 percent mu-follower THF-51S follower sound. So this is only a conditional opinion.
Now to build the left channel.
Attachments
Looking good Ben! Your description of 'clarity to the sound' is spot on with my (single) experience. I've consistently been surprised/delighted with the clarity and realism of the little Sony 2SJ28 follower – it beats anything else I've heard on that front. I like the sound a lot, which is why I'm especially excited to give the 'Lazy' Singing Bush a try.
Well done!!
Well done!!
The left channel is finished. It's up and running with no problems, and I am listening to the THF-51S Mu Follower in stereo. The sound is very similar to the choked loaded 2SK180 follower. I'm not sure whether I can tell the difference between the two. Perhaps this THF-51S Mu follower has a slightly fuller sound. Or perhaps not. It sounds really good. All I can do is repeat what I said in post #94. There is clarity and realism. I'm happy.
The left channel has lower distortion than the right channel, but the left channel is slightly noisier.
The left channel has lower distortion than the right channel, but the left channel is slightly noisier.
Attachments
Thanks.
I use one of those typical rectangular bridge rectifiers bolted to the chassis. I do the Quasimodo thing but that does not do anything to the low frequencies. For CLC I have 44mF-Hammond 159ZJ-44mF.
I try hard to minimize loop areas in wiring, which are twisted, and also minimize loop areas in the power supply and amplifier circuits. Keeping power supply components and AC power away from low level audio signals is important too. The chassis of the monoblocks are actually quite compact but I must have done everything right. I was quite pleased when I saw the measurements. And there is no noise audible from my speakers.
Since I have high sensitivity speakers, I try my best to keep the noise level low from the start, rather than having to trouble shoot and fix later.
I use one of those typical rectangular bridge rectifiers bolted to the chassis. I do the Quasimodo thing but that does not do anything to the low frequencies. For CLC I have 44mF-Hammond 159ZJ-44mF.
I try hard to minimize loop areas in wiring, which are twisted, and also minimize loop areas in the power supply and amplifier circuits. Keeping power supply components and AC power away from low level audio signals is important too. The chassis of the monoblocks are actually quite compact but I must have done everything right. I was quite pleased when I saw the measurements. And there is no noise audible from my speakers.
Since I have high sensitivity speakers, I try my best to keep the noise level low from the start, rather than having to trouble shoot and fix later.