Progress has been slow, but there has been some progress. Actually, it's been one step backwards and two steps forward. I wasn't happy with the finish on the wood bases, so I got my belt sander out, resanded the surfaces and applied new stain and oil finish. I'm much happier with the result. Should have put more effort into it the first time; I should know better than that.
Next up was finalizing how everything was to sit on the top plate and figure out how to mount the fans under the heatsinks. Worked that out and then needed to source 5 1/2 inch long bolts to attach the heatsinks to the top plate. Finally found some 8-32 threaded rod that I cut to size for the job.
I'm going to mount the 2SJ28 VFET directly on the top plate and I've been going back and forth on how to protect myself from electrocution by touching TO-3 case which will be at about 180V. I had all sorts of artsy ideas but settled on an small aluminum plate clamped to the top of the VFET (with insulator). The nice thing about this is that it is easy to do, adds heatsinking surface, and doesn't look too bad.
The other issue I was concerned about was the source resistors for the 2SK182ES, I need about 4.4 ohms at approximately 1.8A which is almost 15 watts. I initially bought through hole resistors, but I became concerned with how to dissipate the heat in the small space below the top plate. So, after much thought, I decided that the answer was chassis mounted power resistors. Of course, that meant another order to Digikey.
Oh, and then I thought that since the output transformers have 4, 8, and 16 ohm taps, maybe it would be great if I had easy access to the 8 and 16 ohm taps. I do have a pair of Quad 57 ESLs. That meant the double pole speaker terminals that I had were out, and Digikey got more of my money.
So, all the holes have been drilled on the top plates, unless I have forgotten something. Here are some pictures showing how the finished amps will look. The good thing about the actual amp circuit is that there are not many parts so wiring shouldn't take too long. That could be a week, two weeks, or a month ...
Next up was finalizing how everything was to sit on the top plate and figure out how to mount the fans under the heatsinks. Worked that out and then needed to source 5 1/2 inch long bolts to attach the heatsinks to the top plate. Finally found some 8-32 threaded rod that I cut to size for the job.
I'm going to mount the 2SJ28 VFET directly on the top plate and I've been going back and forth on how to protect myself from electrocution by touching TO-3 case which will be at about 180V. I had all sorts of artsy ideas but settled on an small aluminum plate clamped to the top of the VFET (with insulator). The nice thing about this is that it is easy to do, adds heatsinking surface, and doesn't look too bad.
The other issue I was concerned about was the source resistors for the 2SK182ES, I need about 4.4 ohms at approximately 1.8A which is almost 15 watts. I initially bought through hole resistors, but I became concerned with how to dissipate the heat in the small space below the top plate. So, after much thought, I decided that the answer was chassis mounted power resistors. Of course, that meant another order to Digikey.
Oh, and then I thought that since the output transformers have 4, 8, and 16 ohm taps, maybe it would be great if I had easy access to the 8 and 16 ohm taps. I do have a pair of Quad 57 ESLs. That meant the double pole speaker terminals that I had were out, and Digikey got more of my money.
So, all the holes have been drilled on the top plates, unless I have forgotten something. Here are some pictures showing how the finished amps will look. The good thing about the actual amp circuit is that there are not many parts so wiring shouldn't take too long. That could be a week, two weeks, or a month ...
Attachments
I managed to get some stuff done so it's time for another progress report. I've been slowly working out all the little details of the build and I think there's not much heavy thinking left.
There are not many parts in the actual amplifier circuit. Options for mounting the resistors/capacitors include lug terminal strips (like old tube amps), printed circuit board, and perf board. I didn't really seriously consider terminal strips, but considered pcbs. However, I didn't want to wait the 2 to 3 weeks for the manufacture, so I went to the perf board option. Not many parts, anyways.
For a while now, I've been using LTspice to do perf board circuit layout. I can move components around and route the wiring to produce a good layout while trying to minimize loop areas. As a bonus, once I come up with a layout, I can run the circuit analysis to verify that there are no circuit errors in the layout. So, I have completed building the perfboard circuits for both channels and wired up the 2SJ28 voltage amplifier for the left channel.
I loaded the interstage transformer secondary with a 47kOhm resistor and performed some tests. The gain is just under 10 times and the distortion test results are as expected.
Next, I built the 2SK182ES cooling fan power supply filter. The fan will be fed by the 60V output stage power supply. The concern, as pointed out by ZM, is that the commutation noise from the fan motor will be fed back into the amp power supply. I had tested a simple RC filter with the fan connected to a regulated lab power supply, and the noise did not seem to be a problem. However, I was not sure how the CLC amp power supply would fare with the fan connected.
So, I added some inductors to the filter and put it all on a perf board. With the amp power supply feeding this filter and fan, oscilloscope measurements do not show any appreciable noise feeding back to the 60V power supply. Of course, the true test will be when the output stage is operational.
I've got four 2SK182ES SITs, so I just need to curve trace them to find a similar pair and then complete the output stage. Basically, hook up the SIT to the interstage transformer, perf board, source resistors, and output transformer. And hope the fan power filter works.
There are not many parts in the actual amplifier circuit. Options for mounting the resistors/capacitors include lug terminal strips (like old tube amps), printed circuit board, and perf board. I didn't really seriously consider terminal strips, but considered pcbs. However, I didn't want to wait the 2 to 3 weeks for the manufacture, so I went to the perf board option. Not many parts, anyways.
For a while now, I've been using LTspice to do perf board circuit layout. I can move components around and route the wiring to produce a good layout while trying to minimize loop areas. As a bonus, once I come up with a layout, I can run the circuit analysis to verify that there are no circuit errors in the layout. So, I have completed building the perfboard circuits for both channels and wired up the 2SJ28 voltage amplifier for the left channel.
I loaded the interstage transformer secondary with a 47kOhm resistor and performed some tests. The gain is just under 10 times and the distortion test results are as expected.
Next, I built the 2SK182ES cooling fan power supply filter. The fan will be fed by the 60V output stage power supply. The concern, as pointed out by ZM, is that the commutation noise from the fan motor will be fed back into the amp power supply. I had tested a simple RC filter with the fan connected to a regulated lab power supply, and the noise did not seem to be a problem. However, I was not sure how the CLC amp power supply would fare with the fan connected.
So, I added some inductors to the filter and put it all on a perf board. With the amp power supply feeding this filter and fan, oscilloscope measurements do not show any appreciable noise feeding back to the 60V power supply. Of course, the true test will be when the output stage is operational.
I've got four 2SK182ES SITs, so I just need to curve trace them to find a similar pair and then complete the output stage. Basically, hook up the SIT to the interstage transformer, perf board, source resistors, and output transformer. And hope the fan power filter works.
Attachments
Fired it up this morning and all the electrons generally flowing as planned. I made some distortion measurements at 1W into 8 ohms, with and without the cooling fan in operation. The fan does add some noise in the low frequencies, but the levels are fairly low. I'm not going to worry about it for now.
Attachments
I tried different bias settings for the 2SK182ES and decided to go with one that allowed for greater power output at the expense of more distortion at low power output. The final operating point for the 2SK182ES is 51 Vds at 1.46A. The cooling fan has kept the heatsink to a temperature that I can tolerate for quite a while with my finger.
I had a short listen with a test speaker and all appears well. No weird noises, other than the music.
I had a short listen with a test speaker and all appears well. No weird noises, other than the music.
Attachments
I finished the right channel amplifier yesterday, and after checking everything today, I fired it up. Actually, no fire, which is a good thing. I did bring it up it stages with a Variac while monitoring the voltages and it went smoothly; no surprises.
I ran some tests with an 8 ohm resistor load and the results are very similar to the left channel. Distortion is a little lower and noise a little higher, but not big differences.
Next up, is to mount the amps to their wood bases and take some glamour shots.
I ran some tests with an 8 ohm resistor load and the results are very similar to the left channel. Distortion is a little lower and noise a little higher, but not big differences.
Next up, is to mount the amps to their wood bases and take some glamour shots.
Attachments
I finally got around to playing some music through the new amps. I spent the last couple of days reorganizing my system so I can easily mix and match amplifiers. Since I biamp my JBL speakers, it is a little more complicated than if I had speakers with passive crossovers.
I got a new rack for the amps. It's actually a garage storage shelf, on sale at Canadian Tire. Big and ugly, but it was cheap and it is strong. Now my short interconnect cables were, well, too short. So, I took time out to build four sets of 6 feet long interconnects. Luckily, I had some Mogami microphone cable and some decent RCA plugs laying around.
Oh, and assembling the rack was an exercise in itself. Not quite as hard as assembling a barbecue or a piece of Ikea furniture, but I was sweating (but not swearing) by the time I finished.
And before I forget, I managed to release some magic smoke a couple of days ago when I had buttoned everything up and was about to perform some final test measurements. The amps were warming up and I was connecting the test gear, when I heard this loud hissing sound accompanied by this smell. Not good. I hit the amp power switch and prepared for the worst.
I thought it was a resistor frying to a crisp. I took the bottom panel off the problem amp and all the resistors were intact. I could see any obvious damage so I got my dim bulb tester out and powered the amp up. No problem. Powered up without the DBT. No problem. But I knew that something had failed.
The good thing is there are not many parts in each amp. That left the electrolytic capacitors to check. Sure enough, one of the two parallel source resistor bypass capacitors had a bulge in its top. On further investigation, it turns out there was a bad solder joint at the capacitor connection to the junction of the source resistor and output transformer primary. The capacitor had only one electrode to connected to 60V positive ground and the other end, which should have been at about 53 volts or so, was at 0V so the 35V capacitor was subjected to about 60V. Luckily, the capacitor failed open instead of shorted so the SIT retained the full bias voltage.
So, this afternoon was the first listening session with the new amps in my biamped system. I had been using my two channel 2SK82VA/2SJ28 CCS source follower amp for the top end (900 Hz crossover) and my 2SJ28VA/2SK180 choke loaded source follower amps for the low end. I decided to try the new amps out driving the top end first. The replaced Sony VFET amp has about twice the distortion of the new amps at 1W output and about half the maximum power output.
I've built a few VFET/SIT amps now and I have found that all of them sound good and the differences are subtle. It takes extended listening to finally become aware of any differences. These new amps are no different. I could not hear any obvious degradation of sound with the new amps in place. That was a good thing.
I think the overall sound may have been more robust than before. It seems fuller. I would have thought the lower distortion would have the opposite effect. However, the increased reserve power can't be bad. The interstage and output transformers would have some effect on the sound too, I would think. All in all, I found the sound very enjoyable. I like the live feel that the single-ended SIT amps combined with big JBL speakers produce. It's like being in a bar listening to a band play. The sound envelopes me and I don't have to sit in one place to enjoy the music. I guess I'm not a big sound stage person.
I have included a couple of pictures showing my VFET/SIT amp stack. From top to bottom they are:
I got a new rack for the amps. It's actually a garage storage shelf, on sale at Canadian Tire. Big and ugly, but it was cheap and it is strong. Now my short interconnect cables were, well, too short. So, I took time out to build four sets of 6 feet long interconnects. Luckily, I had some Mogami microphone cable and some decent RCA plugs laying around.
Oh, and assembling the rack was an exercise in itself. Not quite as hard as assembling a barbecue or a piece of Ikea furniture, but I was sweating (but not swearing) by the time I finished.
And before I forget, I managed to release some magic smoke a couple of days ago when I had buttoned everything up and was about to perform some final test measurements. The amps were warming up and I was connecting the test gear, when I heard this loud hissing sound accompanied by this smell. Not good. I hit the amp power switch and prepared for the worst.
I thought it was a resistor frying to a crisp. I took the bottom panel off the problem amp and all the resistors were intact. I could see any obvious damage so I got my dim bulb tester out and powered the amp up. No problem. Powered up without the DBT. No problem. But I knew that something had failed.
The good thing is there are not many parts in each amp. That left the electrolytic capacitors to check. Sure enough, one of the two parallel source resistor bypass capacitors had a bulge in its top. On further investigation, it turns out there was a bad solder joint at the capacitor connection to the junction of the source resistor and output transformer primary. The capacitor had only one electrode to connected to 60V positive ground and the other end, which should have been at about 53 volts or so, was at 0V so the 35V capacitor was subjected to about 60V. Luckily, the capacitor failed open instead of shorted so the SIT retained the full bias voltage.
So, this afternoon was the first listening session with the new amps in my biamped system. I had been using my two channel 2SK82VA/2SJ28 CCS source follower amp for the top end (900 Hz crossover) and my 2SJ28VA/2SK180 choke loaded source follower amps for the low end. I decided to try the new amps out driving the top end first. The replaced Sony VFET amp has about twice the distortion of the new amps at 1W output and about half the maximum power output.
I've built a few VFET/SIT amps now and I have found that all of them sound good and the differences are subtle. It takes extended listening to finally become aware of any differences. These new amps are no different. I could not hear any obvious degradation of sound with the new amps in place. That was a good thing.
I think the overall sound may have been more robust than before. It seems fuller. I would have thought the lower distortion would have the opposite effect. However, the increased reserve power can't be bad. The interstage and output transformers would have some effect on the sound too, I would think. All in all, I found the sound very enjoyable. I like the live feel that the single-ended SIT amps combined with big JBL speakers produce. It's like being in a bar listening to a band play. The sound envelopes me and I don't have to sit in one place to enjoy the music. I guess I'm not a big sound stage person.
I have included a couple of pictures showing my VFET/SIT amp stack. From top to bottom they are:
- 2SJ28 choke loaded voltage amplifier with choke loaded 2SK180 source follower output stage
- the new 2SJ28 voltage amplifier interstage coupled to output transformer/resistor source follower output stage
- BAF 2015 common source THF-51S CCS loaded output monoblocks on the left; 2SK82 choke loaded voltage amplifier with 2SJ28 CCS loaded source follower output stage on the right
Attachments
