Banned
Joined 2002
Re: GTG
I'm little confused with this..
you can't blame (hehe "blame") sort of part used in CCS for sound resulted from more prominent parameters in one amp.....in any case-good CCS will not influence sound in any way....
besides-I can't recall that you use Jfet in CCS feeding input LTP ,I remember that you keeped IRF mosfet as CCS part.....did you changed that?
agree with that
looks that jfets are sorta gentler to use than mosfets......if we can talk generally...
BillWW said:
I'm little confused with this..
you can't blame (hehe "blame") sort of part used in CCS for sound resulted from more prominent parameters in one amp.....in any case-good CCS will not influence sound in any way....
besides-I can't recall that you use Jfet in CCS feeding input LTP ,I remember that you keeped IRF mosfet as CCS part.....did you changed that?
agree with that
looks that jfets are sorta gentler to use than mosfets......if we can talk generally...
Well, The input capacitance of the 9610 is 170 pf and thats pretty high compared even to the ZVP3310 device! Small JFETS certainly cannot be 170 pf......
No, still have to order the parts. Its VERY expensive to build them compared to through hole versions! I think the resistirs alone at Digi-Key were $75.00! I will have it ready the next time around though. We will compar it to my Mike W Mini-A.
I am still planning on making some changes to the MOSFET A-40 as you suggested.m We will compare again at the next Shootout in a couple of months.
Mark A. Gulbrandsen said:
for decent preamp (say-as yours with 12B4) and cables that difference in input capacitances can't have drastic influence on sound (even counting on Mr. Miller etc);
some other factors are in case , transfer functions etc ,but talking about these factors when we already know consequences is wasting of time......
jfet as current source
Choky,
I have tried all three (irf9610, zvp3310a, and j176 jfet) as a current source and will only use the j176 now, because of how much of an improvement it makes. I can only guess the linear nature of the jfet and also lower capacitance of the jfet makes a huge improvement to the detail of music.
I only directly swapped out the irf9610 (200 pF) vs zvp3310a (50 pF) and noticed a great big improvement in the highs, although the low frequencies were worse with the lower transconductanct zvp3310a.
I further tried the j176 as a current source and that sounded much better than the zvp3310a.
I also swapped out the 9.1 volt zener with an LED for a voltage reference and that was the icing on the cake.
Before I lost my left channel, it really started to sound like a triode tube amp more than any other solid state amp I have ever heard.
I do not know what is going on with the various current sources. You have previously stated it does not matter on the sound quality due to a current source is a current source, but not to my own ears which I go by.
Choky,
I have tried all three (irf9610, zvp3310a, and j176 jfet) as a current source and will only use the j176 now, because of how much of an improvement it makes. I can only guess the linear nature of the jfet and also lower capacitance of the jfet makes a huge improvement to the detail of music.
I only directly swapped out the irf9610 (200 pF) vs zvp3310a (50 pF) and noticed a great big improvement in the highs, although the low frequencies were worse with the lower transconductanct zvp3310a.
I further tried the j176 as a current source and that sounded much better than the zvp3310a.
I also swapped out the 9.1 volt zener with an LED for a voltage reference and that was the icing on the cake.
Before I lost my left channel, it really started to sound like a triode tube amp more than any other solid state amp I have ever heard.
I do not know what is going on with the various current sources. You have previously stated it does not matter on the sound quality due to a current source is a current source, but not to my own ears which I go by.
Banned
Joined 2002
Re: jfet as current source
you can claim contrary only if you are convinced that each time you set same amount of current with changed CCS active part;
If you are sure with that fact-I believe you,but that puzzles me;
if you are not sure-there is cause of sound difference
in any case-I'm glad that you make progress with better sounding amp
BillWW said:
I do not know what is going on with the various current sources. You have previously stated it does not matter on the sound quality due to a current source is a current source, but not to my own ears which I go by.
you can claim contrary only if you are convinced that each time you set same amount of current with changed CCS active part;
If you are sure with that fact-I believe you,but that puzzles me;
if you are not sure-there is cause of sound difference
in any case-I'm glad that you make progress with better sounding amp
Re: Re: jfet as current source
All I did was swap in the irf9610 vs zvp3310a with the previous Aleph 3 using 100K and 10k feedback resisters. In that comparison, the zvp3310a had less bass control, but much better detailed high frequencies.
I progressed to your new circuit using 220K ohm and 22.1K ohm and further tried comparing the irf9610, zvp3310a and j176 as current sources.
The only thing I can figure is in Nelson's amps, there must be some gain from the current sources and they affect the sound.
I for sure could hear the irf9610 had much better bass control vs the j176 and zvp3310a so transconductance does affect the sound and I also feel after extensive listening the capacitance in the current source does seem to affect the highs too. I would like to still eventually bring a jfet mini aleph, but am waiting for some parts to arrive before I finish that project.
Bill
choky said:
you can claim contrary only if you are convinced that each time you set same amount of current with changed CCS active part;
If you are sure with that fact-I believe you,but that puzzles me;
if you are not sure-there is cause of sound difference
in any case-I'm glad that you make progress with better sounding amp
All I did was swap in the irf9610 vs zvp3310a with the previous Aleph 3 using 100K and 10k feedback resisters. In that comparison, the zvp3310a had less bass control, but much better detailed high frequencies.
I progressed to your new circuit using 220K ohm and 22.1K ohm and further tried comparing the irf9610, zvp3310a and j176 as current sources.
The only thing I can figure is in Nelson's amps, there must be some gain from the current sources and they affect the sound.
I for sure could hear the irf9610 had much better bass control vs the j176 and zvp3310a so transconductance does affect the sound and I also feel after extensive listening the capacitance in the current source does seem to affect the highs too. I would like to still eventually bring a jfet mini aleph, but am waiting for some parts to arrive before I finish that project.
Bill
thanks to all
just a note of thanks to Luckylyndy, Billww and especially mark g
for a great afternoon of great sounds great company and delicous steaks .
even though i was I the hot seat with the misses
for delaying her birthday night out.
The magnavox amp spoken of earlier is the same circuit gabev
has been raving about and it is indeed surprizingly good.
the last filter cap section has benn increased to 470uf , and the coupling caps are now panasonic polypropylene types otherwise it is as torn from the console.
the a40 spoken of was entirely built from my "junkbox" and is under power supplyied with about 26 volt rails from a 320va sola
giving somewhere around 28-30 watts(8ohm) if i remmember correctly
all of the amps had a distinctivley different flavor and did many things right. the standout for me was the ksa-50 I didnt expect
the punch it delivered yet did not sound "hard"
oops sorry for the long winnded post
just a note of thanks to Luckylyndy, Billww and especially mark g
for a great afternoon of great sounds great company and delicous steaks .
even though i was I the hot seat with the misses
for delaying her birthday night out.
The magnavox amp spoken of earlier is the same circuit gabev
has been raving about and it is indeed surprizingly good.
the last filter cap section has benn increased to 470uf , and the coupling caps are now panasonic polypropylene types otherwise it is as torn from the console.
the a40 spoken of was entirely built from my "junkbox" and is under power supplyied with about 26 volt rails from a 320va sola
giving somewhere around 28-30 watts(8ohm) if i remmember correctly
all of the amps had a distinctivley different flavor and did many things right. the standout for me was the ksa-50 I didnt expect
the punch it delivered yet did not sound "hard"
oops sorry for the long winnded post
Re: jfet as current source
Hi Bill
I am also somewhat confused. The constant current source (CCS) is to get constant dc current. The first concern is to be whether CCS provides a perfect constant dc current or not. And, the second concern could be whether it works without noise or not.
I agree that CCS hardly affects the sound quality of the amp as much as we sense it.
Regards
BillWW said:
Hi Bill
I am also somewhat confused. The constant current source (CCS) is to get constant dc current. The first concern is to be whether CCS provides a perfect constant dc current or not. And, the second concern could be whether it works without noise or not.
I agree that CCS hardly affects the sound quality of the amp as much as we sense it.
Regards
to BillWW:
I hope that you'll not be mad at me if I attach your sketch here....
I'm little confused with values there.......
I don't see any problem with using J176 as CCS part,but:
If we presume that you have at least 10 mA through first Diff pair J176 ,to achieve ~ 4Vdc at 412ohms resistor,then I have some doubts about value of 221E+50ohms pot above CCS fet......
if I=Vled+Vgs/R
we know that I must be at least 2 x 10mA
then R=Vled+Vgs/I = 2V2+0V3/0,02 = 125 ohms ,roughly...........
something starving here?
maybe you just need to find sweet spot for each part used (mosfet or jfet ) but without starving anything........
next time try to measure all values in all nodes ,to be sure that you will have enough parameters for later thinkin' and rethinkin'.....
hehe,food for thoughts .
I hope that you'll not be mad at me if I attach your sketch here....
I'm little confused with values there.......
I don't see any problem with using J176 as CCS part,but:
If we presume that you have at least 10 mA through first Diff pair J176 ,to achieve ~ 4Vdc at 412ohms resistor,then I have some doubts
about value of 221E+50ohms pot above CCS fet......if I=Vled+Vgs/R
we know that I must be at least 2 x 10mA
then R=Vled+Vgs/I = 2V2+0V3/0,02 = 125 ohms ,roughly...........
something starving here?
maybe you just need to find sweet spot for each part used (mosfet or jfet ) but without starving anything........
next time try to measure all values in all nodes ,to be sure that you will have enough parameters for later thinkin' and rethinkin'.....
hehe,food for thoughts .
Attachments
Choky,
We did feel bad not to hear Bill's newest incarnation of the Aleph 3 with jfets. I don't know what he was fiddling with at the last moment to cause the meltdown. He is claiming to have it up and running next week, but it sounds worse than that.
I wanted to compliment MikeW for such a beautiful job on the A-40 Mosfet. Nice cube design, a la Aleph 5, and the teak wood looked just marvelous! It looked the farthest thing from DIY, or should I say closer to one of Peter Daniel's creations.
Lyndon
We did feel bad not to hear Bill's newest incarnation of the Aleph 3 with jfets. I don't know what he was fiddling with at the last moment to cause the meltdown. He is claiming to have it up and running next week, but it sounds worse than that.
I wanted to compliment MikeW for such a beautiful job on the A-40 Mosfet. Nice cube design, a la Aleph 5, and the teak wood looked just marvelous! It looked the farthest thing from DIY, or should I say closer to one of Peter Daniel's creations.
Lyndon
choky said:to BillWW:
I don't see any problem with using J176 as CCS part,but:
If we presume that you have at least 10 mA through first Diff pair J176 ,to achieve ~ 4Vdc at 412ohms resistor,then I have some doubts
if I=Vled+Vgs/R
we know that I must be at least 2 x 10mA
then R=Vled+Vgs/I = 2V2+0V3/0,02 = 125 ohms ,roughly...........
something starving here?
maybe you just need to find sweet spot for each part used (mosfet or jfet ) but without starving anything........
next time try to measure all values in all nodes ,to be sure that you will have enough parameters for later thinkin' and rethinkin'.....
hehe,food for thoughts .
Choky, You are correct. I had to play with the values to give room for adjustment of the dc offset.
I did not eventually use a 221 ohm plus 50 ohm pot. I just stuck in a 500 ohm pot, but forgot to show that in my above layout that I e-mailed you.
I did initially run into a problem adjusting the dc offset with the higher 221 ohm resister plus a 50 ohm pot, so that was why I later just placed a 500 ohm pot to allow for using all mosfets such as the irf9610 and zvp3310a as a current source. The various fets deviated so much that 500 ohms seemed to be a better range for all of the mosfets and jfet I tried.
At the time of settling with the j176, 412 ohms seemed to drive the output mosfets better into lower impedence loads. When it was at 470 ohms, it did not drive my low 2 ohm load very well in the bass. So, roughly 408 to 412 ohms seemed to give me better drive current into 2 ohms with my heavy biasing of my amp. Previously, with the preliminary 470 ohms seemed to make the bass sound very weak without a lot of authority. After realizing that, I replaced the 221 ohm resister and 50 ohm pot with just the 500 ohm pot and then I could adjust the dc offset fine.
Also, those considering an LED for a voltage reference, I found about 2 to 4 volts to be roughly the right value for the jfet. I have not completly finished that part of my experiment yet, before destroying my left channel.
Initially with the 9.1 Volt zener where the above photo shows now a LED, the Zener would not regulate to the 9.1 volts with the j176 jfet current source in place.
That was also another reason why I choose to stick in the LED to still regulate at a lower voltage value, while the LED offers about 10 times less noise then the zener from what I have studied on diyaudio. It is a lot better than a LM... voltage regulator according to what I have studied and it substitutes nicely in place of the zener for those converting their Alephs.
I have also found just a higher value resister for a voltage source sounded inferior to the LED for some reason. Just listening with my ears though. I was very happy with the LED compared to a resister.
You would be surprised how nice the Aleph jfet can sound with the right current source. The jfets for the differential LTP alone did not sound that nice. With the jfet current source, it made a big improvement and especially with the LED as a voltage reference. I still would like to see if bypassing the LED would improve things too like with the zener, but need to get my amp up and running on both channels.
I recall my amp acting like it is now one time previously and suspect one of my output mosfets may be bad. I will have to check them out and that is my prime suspect at this time.
PS, Choky, on your babelfish, I noticed you placed the loading resisters in a different place than the original Aleph's do on their mosfet versions. Is this because a jfet is different than a mosfet? In other words should I remove my 221 ohm loading resisters on the j176 LTP and just jump a wire in place of these previous resisters? I left my 221 ohm loading resisters in place for now, since I did not see any harm in it, in case I wanted to convert back to the zvp3310a.
Bill
BillWW said:
I think that you think on gate stoppers-they are here for preventing oscillations (gate acting as aerial
)you can leave it here without any harm....
regarding sort of CCS,you remember that I use two bjts in CCS....no problem with any voltage (hehe-ready even for mini Babelfish without any modification)....and I think that sort is even less noisier than Jfet CCS.......but-anyway -good enough .......
PS maybe you can try ,when you repair that channel,connecting second Jfet as I do-to mosfet source.......difference or not-you'll hear ...
Aleph 3 jfet amp
Hey guys,
I got the Aleph 3 up and running last night with the jfet front end including jfet current source.
I still need to replace the zener with the LED for a voltage reference. I will have to do that tonight. I am very happy to have my beautiful sounding amp back. I missed it like a good friend.
I am eager to try the LEDs red ones, since I understand the lower brightness ones have very low noise. The green ones that I had were still better than the 9.1 volt zener.
So, I guess Saturday if everyone wants to hear the jfet Aleph 3 it should be available as long as Mark is home.
Regards, all
Btw, I came home with an extra power cord. Is anyone missing one? Dave?
Bill
Hey guys,
I got the Aleph 3 up and running last night with the jfet front end including jfet current source.
I still need to replace the zener with the LED for a voltage reference. I will have to do that tonight. I am very happy to have my beautiful sounding amp back. I missed it like a good friend.
I am eager to try the LEDs red ones, since I understand the lower brightness ones have very low noise. The green ones that I had were still better than the 9.1 volt zener.
So, I guess Saturday if everyone wants to hear the jfet Aleph 3 it should be available as long as Mark is home.
Regards, all
Btw, I came home with an extra power cord. Is anyone missing one? Dave?
Bill
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.