I've spent the last few days trying to solve distortion issues with the test amplifier, and today I can finally see a big light at the end of the tunnel. Going through the amp step-by-step, I discovered that the transformer coupled voltage amplifier was now no longer working at low distortion levels. When I added the output stage, I attempted to clean up the wiring and introduced some problems. After many hours of trouble-shooting with no success, I did the easy thing and choke loaded and cap coupled the voltage amplifier to the output stage instead. This fixed the voltage amp distortion issue. I will deal with the interstage transformer coupling issues later.
Unfortunately, that did not solve all my problems; distortion was still high from the output stage. After testing many theories and proving them all wrong, I finally stumbled onto the solution. I switched the primary connections around on the output transformer, and the distortion dropped considerably and did not sky rocket with more input.
So, at 1 watt out, THD is about 0.1%. At 13 watts out, THD is about 0.8%. The 2SK182ES is operating at Vds of 54V at 1.3A. The output transformer is configured for 16 ohm : 8 ohm.
The next step is to sort out my interstage coupled voltage amplifier. I think it's an issue of connecting the primary and secondary correctly, the same problem that I had with the output transformer.
Unfortunately, that did not solve all my problems; distortion was still high from the output stage. After testing many theories and proving them all wrong, I finally stumbled onto the solution. I switched the primary connections around on the output transformer, and the distortion dropped considerably and did not sky rocket with more input.
So, at 1 watt out, THD is about 0.1%. At 13 watts out, THD is about 0.8%. The 2SK182ES is operating at Vds of 54V at 1.3A. The output transformer is configured for 16 ohm : 8 ohm.
The next step is to sort out my interstage coupled voltage amplifier. I think it's an issue of connecting the primary and secondary correctly, the same problem that I had with the output transformer.
Attachments
I ran some more tests on the choke loaded/capacitor coupled voltage amp version today to establish a baseline to compare against when I get the interstage coupled version running. For full disclosure, the choke that I used is a Magnequest EXO-003 from my old toob daze. It is gapped for 60mA and I think the inductance is 40H; its DC resistance is about 300 ohms.
Power (8 ohm load) THD
1W 0.0096%
4W 0.20%
9W 0.30%
13W 0.80% (from yesterday)
16W 3.15%
I have also included an oscilloscope screen capture for 16W out showing the sine waves at the amp input (purple), input to 2SK182ES (yellow), and the output to the 8 ohm load(blue). The input is about 5Vpp, the output of the voltage amp is 53.6Vpp, the output to 8 ohm load is 32.0Vpp.
I also measured the distortion of the choke loaded 2SJ28 voltage amplifier at 60Vpp and it is 0.089%. So, it is well below the distortion of the output stage.
Power (8 ohm load) THD
1W 0.0096%
4W 0.20%
9W 0.30%
13W 0.80% (from yesterday)
16W 3.15%
I have also included an oscilloscope screen capture for 16W out showing the sine waves at the amp input (purple), input to 2SK182ES (yellow), and the output to the 8 ohm load(blue). The input is about 5Vpp, the output of the voltage amp is 53.6Vpp, the output to 8 ohm load is 32.0Vpp.
I also measured the distortion of the choke loaded 2SJ28 voltage amplifier at 60Vpp and it is 0.089%. So, it is well below the distortion of the output stage.
Attachments
It is a nice polite flat. Some amps go crazy when slammed to the rails. Looks like you turn-down a part-dB and it will be clean.
The flat only on one side suggests a different bias to compromise top and bottom flats, but 3rd usually sounds worse than 2nd.
I was big into single ended Directly Heated Triodes for a long time before I discovered SITs, thanks to this forum. Now, I'm totally converted. I still have some tube amps, but they're just sitting there.
At the risk of repeating myself, I believe the price of these large Tokins is still a good deal. Just check to see how much a 300B or 845 or 211 tube costs. And tubes wear out.
At the risk of repeating myself, I believe the price of these large Tokins is still a good deal. Just check to see how much a 300B or 845 or 211 tube costs. And tubes wear out.
At last! The coupling cap is gone. I finally got the interstage transformer wiring all sorted out and the 2SJ28-InterstageTransformer-2SK182ES-OutputTransformer amplifier is working.
I ran a couple of quick tests before shutting down for the night. The distortion performance is a little better than the choke loaded/cap coupled voltage amp version but otherwise, I didn't see any great differences. I ran the RightMark Audio Analyzer suite of tests, as well, mainly to see how the amp performs with respect to frequency response. I was happy to see that there were no glaring frequency response problems.
I ran a couple of quick tests before shutting down for the night. The distortion performance is a little better than the choke loaded/cap coupled voltage amp version but otherwise, I didn't see any great differences. I ran the RightMark Audio Analyzer suite of tests, as well, mainly to see how the amp performs with respect to frequency response. I was happy to see that there were no glaring frequency response problems.
Attachments
I made some more measurements of the test amp. Included are distortion measurements at various power outputs into 8 ohm resistor load. An oscilloscope capture of sine waves at 16 watts out is also included. The purple trace is signal in, yellow is output of voltage amp, and blue is output at 8 ohm load.
I have updated the schematic to show the amplifier as tested.
I haven't listened to the amp yet and that is the next step.
I have updated the schematic to show the amplifier as tested.
I haven't listened to the amp yet and that is the next step.
Attachments
I did notice that, and repeated measurements produced the same results.
The power supply I built for testing maxes out at 60V, so I can't test any higher. I would have to build another power supply. I would need some higher voltage capacitors, also.
I don't know if I want more power out. More dissipation also means more heat to deal with. I'll have to think about this a bit.
The power supply I built for testing maxes out at 60V, so I can't test any higher. I would have to build another power supply. I would need some higher voltage capacitors, also.
I don't know if I want more power out. More dissipation also means more heat to deal with. I'll have to think about this a bit.
Today, I hooked up the amp to my test speaker and played some CDs with a portable all-in-one player. I listened for about 3 hours to various CDs.
I had remounted the 2SK182ES properly on a smaller heatsink with an insulator and thermal paste and added a small fan underneath the heatsink for extra cooling. During the whole time, the heatsink remained just warm to touch and the bias current remained stable at 1.29A.
My test speaker is a mish-mash with the high frequencies covered by an Altec 32 bent horn with an 804A compression driver, and the low frequencies handled by an unknown 12" speaker in a vented 2.5 cubic foot box. An old Altec crossover handles the frequency division. The low end is not great, but it's ok for tests.
The quality of the output of the boom-box is obviously not the greatest, but still the sound was enjoyable. The bass overall is a bit weak, but when needed, it was sufficient. I played a few CDs from the 70s and 80s, including Paul Simon's Graceland, Sade's Diamond Life, and Madonna's the Immaculate Collection. All these CDs have good hard hitting bass and percussion, and the system reproduced them surprisingly well. The impact was there, and even with just one speaker, the sound level was more than loud enough. Certainly, if I had a party, this system would be a hit, especially after some refreshments. I think the Altec horn loaded compression driver contributes a lot to the attack and the clean sound.
Overall, the output was clean and typical of the source follower sound. I have other SIT source follower amps including ones with 2SK180 and 2SJ28, so I am familiar with the sound. If I didn't know there were transformers in the chain, I would not have guessed it.
I may move this collection of parts upstairs and hook it up to my main system for more testing, but I'm leaning towards this design for the final build. It has enough power for my speakers, and the cooling requirements are not onerous.
I had remounted the 2SK182ES properly on a smaller heatsink with an insulator and thermal paste and added a small fan underneath the heatsink for extra cooling. During the whole time, the heatsink remained just warm to touch and the bias current remained stable at 1.29A.
My test speaker is a mish-mash with the high frequencies covered by an Altec 32 bent horn with an 804A compression driver, and the low frequencies handled by an unknown 12" speaker in a vented 2.5 cubic foot box. An old Altec crossover handles the frequency division. The low end is not great, but it's ok for tests.
The quality of the output of the boom-box is obviously not the greatest, but still the sound was enjoyable. The bass overall is a bit weak, but when needed, it was sufficient. I played a few CDs from the 70s and 80s, including Paul Simon's Graceland, Sade's Diamond Life, and Madonna's the Immaculate Collection. All these CDs have good hard hitting bass and percussion, and the system reproduced them surprisingly well. The impact was there, and even with just one speaker, the sound level was more than loud enough. Certainly, if I had a party, this system would be a hit, especially after some refreshments. I think the Altec horn loaded compression driver contributes a lot to the attack and the clean sound.
Overall, the output was clean and typical of the source follower sound. I have other SIT source follower amps including ones with 2SK180 and 2SJ28, so I am familiar with the sound. If I didn't know there were transformers in the chain, I would not have guessed it.
I may move this collection of parts upstairs and hook it up to my main system for more testing, but I'm leaning towards this design for the final build. It has enough power for my speakers, and the cooling requirements are not onerous.
Attachments
I am continuing my listening evaluation of my test amp. I know that my test speaker is not up to par, but I really don't want to take apart the amp and move it upstairs to my good speaker system. Ironically, I have an Altec Voice of the Theatre setup not being used but it is in semi-storage and I don't have passive cross-overs for them.
My next thought was to try to improve my test speaker. To do that, first, I needed to see what I've got. So, I pulled out my test microphone and measured the frequency response with REW. Well, it sure was revealing. First, there was a big suckout at the crossover frequency (800 Hz); speaker phasing problem. I fixed that and remeasured.
The next obvious problem was the high frequencies need to be padded down hugely. I am already using the maximum attenuation on the Altec crossover unit so no help there. Being lazy, I added series resistors to the horn driver to bring its output down. After several attempts, I managed to produce a reasonable frequency response. Not great, but heck of a lot better than before.
I listened again for about 3 hours this afternoon, and being a proud daddy, I have to say I like the sound. Obviously, the frequency balance is more natural. I could hear the tone fundamentals better whereas before, the harmonics would have been more prominent.
Of course, this is a very subjective thing, I admit to having a strong bias and I have a lot of skin in this. But I'm ready to put my money where my mouth and buy parts to build this amp.
My next thought was to try to improve my test speaker. To do that, first, I needed to see what I've got. So, I pulled out my test microphone and measured the frequency response with REW. Well, it sure was revealing. First, there was a big suckout at the crossover frequency (800 Hz); speaker phasing problem. I fixed that and remeasured.
The next obvious problem was the high frequencies need to be padded down hugely. I am already using the maximum attenuation on the Altec crossover unit so no help there. Being lazy, I added series resistors to the horn driver to bring its output down. After several attempts, I managed to produce a reasonable frequency response. Not great, but heck of a lot better than before.
I listened again for about 3 hours this afternoon, and being a proud daddy, I have to say I like the sound. Obviously, the frequency balance is more natural. I could hear the tone fundamentals better whereas before, the harmonics would have been more prominent.
Of course, this is a very subjective thing, I admit to having a strong bias and I have a lot of skin in this. But I'm ready to put my money where my mouth and buy parts to build this amp.
Attachments
I read Mark Johnson's post (#133) yesterday in the thread "BAF 2022 SIT AMPS" about grounding the drain of SITs in source follower amps to reduce noise and that reminded me that although I had done that in my previous two SIT source follower amps (2SJ28 and 2SK180), I had not done this for this amp. Well, this provided a perfect opportunity to compare the two grounding schemes.
So, first I modeled the grounded drain scheme in LTSpice to make sure that I did it right. Good thing I did, because I had neglected to switch the ground of the interstage transformer secondary and the bias was way out to lunch.
After I got that all sorted out, I changed the grounding in the 2SK182ES output stage of the test amp and ran a quick FFT test. And of course, Mark is right. Thank you, Mark, for the reminder.
I compared this revised scheme to measurements I did the other day, so I really should go back to the original scheme and retest it again. However, all the FFTs from the other day showed the same level of noise, so I think the results stand.
So, first I modeled the grounded drain scheme in LTSpice to make sure that I did it right. Good thing I did, because I had neglected to switch the ground of the interstage transformer secondary and the bias was way out to lunch.
After I got that all sorted out, I changed the grounding in the 2SK182ES output stage of the test amp and ran a quick FFT test. And of course, Mark is right. Thank you, Mark, for the reminder.
I compared this revised scheme to measurements I did the other day, so I really should go back to the original scheme and retest it again. However, all the FFTs from the other day showed the same level of noise, so I think the results stand.
Attachments
I've been ordering parts for the build and they are coming in slowly, bit-by-bit. In the meantime, I decided that I needed to learn how to make circuit boards. I have resisted long enough, and given the prices for perf boards (and other parts) these days, pcbs are a bargain. So, I downloaded and installed KiCad on my computer and got to work.
Now, this amplifier doesn't have many parts, so there are only a few areas that a PCB would be worth the effort. Most of the parts are in the power supply filters, so that's where I started. I managed to put together a CLC filter pcb for the output stage after going through a few steps (and stumbles). Learning KiCad wasn't too bad, but I found that I didn't have a footprint for the 5 terminal electrolytic capacitors, so that took a awhile to make one myself.
I liked the idea of seeing a 3D model of the board, and of course, I didn't have a model for the 50mm dia. x 94mm tall capacitors either. After searching high and low for a pre-made one, I finally decided that it would be easier to make one up myself. So, I installed FreeCAD, and with some help from the Internet, I made a crude cylinder to approximate the shape of a capacitor. Should have done that in the first place; it took hardly any time at all.
My output stage power supply is negative voltage out with choke filtering. The choke will be chassis mounted with leads going to the pcb. A bleeder resistor and LED indicator are included in the pcb. I chose to use a ground plane for the positive voltage ground.
This is my first pcb design so it is likely to have things that may be questionable, so any comments are appreciated.
Now, this amplifier doesn't have many parts, so there are only a few areas that a PCB would be worth the effort. Most of the parts are in the power supply filters, so that's where I started. I managed to put together a CLC filter pcb for the output stage after going through a few steps (and stumbles). Learning KiCad wasn't too bad, but I found that I didn't have a footprint for the 5 terminal electrolytic capacitors, so that took a awhile to make one myself.
I liked the idea of seeing a 3D model of the board, and of course, I didn't have a model for the 50mm dia. x 94mm tall capacitors either. After searching high and low for a pre-made one, I finally decided that it would be easier to make one up myself. So, I installed FreeCAD, and with some help from the Internet, I made a crude cylinder to approximate the shape of a capacitor. Should have done that in the first place; it took hardly any time at all.
My output stage power supply is negative voltage out with choke filtering. The choke will be chassis mounted with leads going to the pcb. A bleeder resistor and LED indicator are included in the pcb. I chose to use a ground plane for the positive voltage ground.
This is my first pcb design so it is likely to have things that may be questionable, so any comments are appreciated.
