You got to monitor it and wait for more and known quality jfets, I agree. Maybe it is not really a problem if it steady settles somewhere. Just trim the positive side after the neg settles and leave it there. Now if it continues sliding after some more hours, you better match a couple of jfets from the new ones you gonna get and put them in. That will show if it was the jfet. I am also wondering about actually setting a +/-V match using the trimmers for somewhat higher voltage after 10-20 mins as I wrote above. Are they really steady 0 when fully closed? Its a plausible question. There are minute moving parts in there, and the heat around them changes with time, as the shunts run.
The njfet is not near the heatsink at all. However, the 150R gate stopper is indeed lying across the MOSFET and the heatsink very close only on the negative side.
OK. I will try the trimpots.
The "sensing capacitor" is made up from a 1.1uF MKP || 100uF electrolytic. I am not too comfortable with that arrangement. It is much better to have the film capacitor only. I may try taking out the electrolytic and see how it sounds. If it sounds better, I may select a higher current 2SK170 to increase the impedance to allow a lower corner frequency (and I know you are against higher impedance due to damping factor, etc, but this increase is not really too much). Or perhaps use a different part like the BF244B that can give 15mA current? Would BF244B fit here? I have a dozen of them not used. Or I may try taking out the film cap, or I may increase the 100uF.
I will also play with different capacitance of the output capacitor and select the best value. I expect this will make some difference.
I think it is time for me to go for another 1 hour listening then try the various things. I guess I may be too lazy to do such mods while enjoying music. It may even take me a few days to find the result.
OK. I will try the trimpots.
The "sensing capacitor" is made up from a 1.1uF MKP || 100uF electrolytic. I am not too comfortable with that arrangement. It is much better to have the film capacitor only. I may try taking out the electrolytic and see how it sounds. If it sounds better, I may select a higher current 2SK170 to increase the impedance to allow a lower corner frequency (and I know you are against higher impedance due to damping factor, etc, but this increase is not really too much). Or perhaps use a different part like the BF244B that can give 15mA current? Would BF244B fit here? I have a dozen of them not used. Or I may try taking out the film cap, or I may increase the 100uF.
I will also play with different capacitance of the output capacitor and select the best value. I expect this will make some difference.
I think it is time for me to go for another 1 hour listening then try the various things. I guess I may be too lazy to do such mods while enjoying music. It may even take me a few days to find the result.
Since it basically works plus it sounds good to you, you have certainly succeeded.
Thinking that you firstly implemented a symmetrical V1.1 and gave us feedback plus you have shown a working layout, you also broke new testing ground. Thanks for the tests. Good contribution.
As for finding your favorite sound for various capacitors used, as for picking values and qualities, this is something that has to do a lot with you and the synergy in the system. Me I prefer around 470uF over the resistive Vref element, but its on another circuit and system. Listen to music. Thanks again. Whatever you arrive to, certainly people in the thread would like to know. Me I would like to see a picture, especially because it has the buffers and I know it will help anybody that will attempt them by looking at your layout, because the high Hfe 300mHz BJTs they can be tricky at a capacitive gate. Seen that with p2p when I added 2 buffers in positive V1's.
P.S. Looked at BF244B. Says pinch off starts low in its datasheet, if it is actually around -0.5V for the ones you pick it will work steady, else if -1V or more it will play. Also its an RF part, I hope it will behave for showing any instability tendencies. Using the 244 its something that you have to test in practice really.
Thinking that you firstly implemented a symmetrical V1.1 and gave us feedback plus you have shown a working layout, you also broke new testing ground. Thanks for the tests. Good contribution.As for finding your favorite sound for various capacitors used, as for picking values and qualities, this is something that has to do a lot with you and the synergy in the system. Me I prefer around 470uF over the resistive Vref element, but its on another circuit and system. Listen to music. Thanks again. Whatever you arrive to, certainly people in the thread would like to know. Me I would like to see a picture, especially because it has the buffers and I know it will help anybody that will attempt them by looking at your layout, because the high Hfe 300mHz BJTs they can be tricky at a capacitive gate. Seen that with p2p when I added 2 buffers in positive V1's.
P.S. Looked at BF244B. Says pinch off starts low in its datasheet, if it is actually around -0.5V for the ones you pick it will work steady, else if -1V or more it will play. Also its an RF part, I hope it will behave for showing any instability tendencies. Using the 244 its something that you have to test in practice really.
Just came back from more listening. Confirmed that it sounds better than other regs I have tried, or should I say, I prefer it over any others I tried. It sounds more "RIGHT" than all others. But I also believe that I still need to put in up to a couple of weeks time to optimize it as there are areas of improvements I can address. It has been a full day for me and I should retire from it now. I feel a bit burnt-out now and will approach it (tune it) slowly in the coming days and post the results in perhaps 2 weeks time, while preparing to build v2. But since I don't have the full range speakers to listen to because I don't have my line level XO/EQ yet and I only temporarily wired up a less than ideal passive XO and ran only the MTM part of the MTMWW+Sub, I am tempted to use the v1 / preamp to do driver /room final measurements, so that I can listen to my "glorious" full range speakers. Perhaps I could then do better tests on my regulators. I will see which way I will go.
As for pictures, it is late at night in Sydney now. I will do it in the coming days.
The picture of my speakers is already made so here it is. It does depend on the regulator to sing.
The tweeter is the very best of the top of the range Dynaudio's Esotar T330D, capable of 1000W transient with 93dB sensitivity used 2k and above. The Midranges are twin ScanSpeaker 6.5" revelators 18W8531G00, the best measured by Zaph, housed in Open back U-frame, used from 180Hz up. There are twin 10" Scanspeak revelators 26W8861T00 in seperate U-frames used from 60Hz up. The subwoofers are not in the picture. They are Peerless XLS 12" in monopoles.
The picture of my speakers is already made so here it is. It does depend on the regulator to sing.
The tweeter is the very best of the top of the range Dynaudio's Esotar T330D, capable of 1000W transient with 93dB sensitivity used 2k and above. The Midranges are twin ScanSpeaker 6.5" revelators 18W8531G00, the best measured by Zaph, housed in Open back U-frame, used from 180Hz up. There are twin 10" Scanspeak revelators 26W8861T00 in seperate U-frames used from 60Hz up. The subwoofers are not in the picture. They are Peerless XLS 12" in monopoles.
An externally hosted image should be here but it was not working when we last tested it.
I just want to be sure which one is v2.
I found yours, the original from Salas, the one for DAC... I checked from pg120..
We know that, has been talked here before. At low frequencies near DC they are complemenary enough for gm. Of course someone can use the Harris too, or 244
v2 is more of a topology. You can find the topology in post 1168
link to post 1168
Some resistors and the zeners have to be changed for your needs (voltage out, current). Post what you intend to power with it and we can help you with that.
hifinutnut, nice looking setup!
Last edited:
Logic level mosfets may work. For v1, something like this
NDP6020P Datasheet pdf - P-Channel Logic Level Enhancement Mode Field Effect Transistor - Fairchild Semiconductor
For v2, many choices of n-channel logic level mosfets.
NDP6020P Datasheet pdf - P-Channel Logic Level Enhancement Mode Field Effect Transistor - Fairchild Semiconductor
For v2, many choices of n-channel logic level mosfets.
I don't think it will hold all of its subjective signature if you go BJT though. If you care about such non objective stuff better go LL as Iko suggested.
Greg,
Oh, you are here. Have you got measurement equipment to measure my v1.1? If you do have, you can do us a great favour.
To be honest, I published my PCB layout so that the experts may pick up what I have done wrong. I would hesitate to publish it for others to use, as I never know if there are hidden oscillations or whatever there. Ears can not be fully trusted.
I have an Aussie EE friend who said he would give his scope and signal generator to me for permanent use, as he has quit his own EE business and am preparing to move to Thailand to start a new life. I have never used them before, and am not sure if I would go any further in this so I have not got them from him yet. I was hoping that I could complete my system without them.
By the way, you are welcome to come to my home to audition my final system once it is ready, and bring your stuff too. You heard my Optimos last time. After that, I found that I did something wrong in wiring the signal ground. I redid all ground work and now it sounds at the next level up. I think it is now a truly reference level amplifier.
Regards,
Bill
Last edited:
Most series regulators have inductive outputs. So film capacitors, or even low ESR electrolytic capacitors can not be used at the output.
What about v1 and v2? Are they inductive at the outputs? The recommended value of 470uF at the output suggests a reasonably low ESR. If it does not resonant, it means at least the output is far less inductive than most series regulators. The question is, how low the ESR can the regulator tolerate? Would it oscillate if you put a film capacitor there?
I know that the regulator provides a very low impedance at the output so a film capacitor may not be necessary. But that may be in theory only. Above some point, e.g. 200kHz, the impedance of the regulator has to rise again. This could make it inductive (but not necessarily so). Modern opamps often work till 10MHz and above. So local film capacitor bypass is generally recommended. In practice, I (subjectively) found such bypassing causes ringings that destroy the integrity of sound. RF can get into the circuit in many ways. Without bypassing, there is another set of problems. It has never been easy to work with opamps due to the wire resistance, wire inductance and parasitic capacitance, especially with some inductive outputs from voltage regulators.
I have been chasing in this game for a long time with series regulators without satisfactory results. What is your take?
What about v1 and v2? Are they inductive at the outputs? The recommended value of 470uF at the output suggests a reasonably low ESR. If it does not resonant, it means at least the output is far less inductive than most series regulators. The question is, how low the ESR can the regulator tolerate? Would it oscillate if you put a film capacitor there?
I know that the regulator provides a very low impedance at the output so a film capacitor may not be necessary. But that may be in theory only. Above some point, e.g. 200kHz, the impedance of the regulator has to rise again. This could make it inductive (but not necessarily so). Modern opamps often work till 10MHz and above. So local film capacitor bypass is generally recommended. In practice, I (subjectively) found such bypassing causes ringings that destroy the integrity of sound. RF can get into the circuit in many ways. Without bypassing, there is another set of problems. It has never been easy to work with opamps due to the wire resistance, wire inductance and parasitic capacitance, especially with some inductive outputs from voltage regulators.
I have been chasing in this game for a long time with series regulators without satisfactory results. What is your take?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.