Going to put 45mA Mosfet output stage on same rails with telescopic cascode input stage also having its common base bias just resistor sourced? Too much interaction IMHO. Needs some rails insulation scheme.
For sure inserting CCS's in the two stages is the natural next step...for modeling the intial simple schematic resistors work well (could also insert a CCS symbol)
Yes, this can be bettered but i think the concept basically works just fine. First i had extrem simplicity in mind and now we see the limitations.
Why is that crasy..??? I do this on an IV-stage, must say the i sounds guite wounder full...ohh...I remember one LA-audio MC head-amp...That thing was quite wonderful......
a
The pre pre could be regulated some down to further reduce noise, as the front J-fets only sees the app 10V under the cascode... (like the front of the pearl 2)
Anyway My MAC...quit on me...so boards will start when i get the Poor (The mother of bad words) thing back....!!!!
a
The pre pre could be regulated some down to further reduce noise, as the front J-fets only sees the app 10V under the cascode... (like the front of the pearl 2)
Anyway My MAC...quit on me...so boards will start when i get the Poor (The mother of bad words) thing back....!!!!
Last edited:
I had the LA-Audio pre-prre and never got it to work without incredible microphony. I rebuild it with Jürgen Ultee. We used the Arthus Loesch circuit of a Fet-Tube cascode and that workes fine until today. I still have it. I will make a photo.
As to the Pre-Pre can we agree on plus-minus 24V ? That is really decent my friend. plus-minus 48V for the second stage whould be enough too because 60V DC is not really safe any more when you get your fingers on it ( OK, just ). Running the input stage on less voltage whould allow us to filter the PSU down from the second stage supply. I think Salas is on the mark that current draw in the second stage should not influence the delicate Pre-Pre. You already agreed that filtering down to the Pre-Pre is ok. I have a schematic ready in my head. I will post it later today and i also have an idea for a alternative transimpdance input that circumvents the unlinear input capacitance of the Fets and should be even faster at the expense of loading the cartridge somewhat. It will work well though with typical low impedance movings coils like the Lyras, Shelter, Zyx, My Sonic Lab, Transfiguration etc.
As to the Pre-Pre can we agree on plus-minus 24V ? That is really decent my friend. plus-minus 48V for the second stage whould be enough too because 60V DC is not really safe any more when you get your fingers on it ( OK, just ). Running the input stage on less voltage whould allow us to filter the PSU down from the second stage supply. I think Salas is on the mark that current draw in the second stage should not influence the delicate Pre-Pre. You already agreed that filtering down to the Pre-Pre is ok. I have a schematic ready in my head. I will post it later today and i also have an idea for a alternative transimpdance input that circumvents the unlinear input capacitance of the Fets and should be even faster at the expense of loading the cartridge somewhat. It will work well though with typical low impedance movings coils like the Lyras, Shelter, Zyx, My Sonic Lab, Transfiguration etc.
Lets get serious. Here is a high PSSR version of the input stage. No Leds, no elcaps, no conventional RC filtering. Bias is done with an adjustable micropower voltage reference. It has an extremely low dynamic impedance of 1 Ohm, runs on less then 0.1mA, has a tempco of 150PPM. The way i have connected it gives even lower noise. The 100kOhm resistor gives enough exess current to run the bases of the cascodes. Hfe of the cascode transistors should be around 200 or more. I use 2SA1015 and 2SC1815. i have degenerated the Fets a bit less and have doubled them. In MiiBs design the 20Ohm resistors that are AC wise in parallel worsen the SN of a ca.5 Ohm cart like the Lyras already by 4.5dB. Now we have 4 x 10 Ohm in parallel, giving 2.5 Ohm so we win 3dB. The doubling of the Fets also gives an SN advantage of 3dB and doubles Gm. I think another P-Channel has to disappear from the market but it is worthwhile. I can also tell you why plus-minus 28V is optimal. Doing much more then 5V over the Fets increases noise because of raising gate current and heating, doing more then 10V over the bipolars comes at the risk of avalanche noise. Making the collector voltage higher then 14V in this case does not give a dynamic range gain eather because to run them on half the PSU voltage is already optimal. If we keep the 60V supply for the second stage ( i whould go for plus-minus 48V if i had a say here ) we have to burn 32V. Fortunately the input stage does not draw more then 15mA so a regulator that can stand 500mW is ok. I have made a design like that, that works like a cap multiplier and a voltage regulator in one go but maybe MiiB has a better idea. I have also shown a transimpedance version. Yes, folks, that works with Fets too and i have shown an AC coupled version in Jan Diddens Linear 0. There i used high value, low voltage Polymer caps. Here, voila, is the DC coupled variety. I had that idea after Andy visited me and we had that delicious steak he paid for me. "The torture never stops" Frank Zappa.
Clever You....Looks really good...With the last version...Is there not a risk of frying the coils in case of drift..???
NO MAC...I HAVE.....🙁
NO MAC...I HAVE.....🙁
Last edited:
I never fryed a coil. Even my most badly offsetted transimpedance BJT stages where not able to destroy my rather expensive Titan i. For the faint at heart there is the Transconductance version.
For sure offset with Jfets will be low....but anyway....
Talking of Lyra...Anyone in the fora heard from Jonathan Car...He has though my visits grown to be a friend, but...seen or hear nothing since the tragedy in Japan...
Talking of Lyra...Anyone in the fora heard from Jonathan Car...He has though my visits grown to be a friend, but...seen or hear nothing since the tragedy in Japan...
Jonathan is in design mood. That has the consequence that he is even hiding for Stig. I skype with Stig regulary and he said "Tokio is nearly back to normal". I whould leave the country but that is against the japaneese culture of "Insisting, Resisting, Persisting". An amasing, admirable but somewhat ancient society.
Is stabilizing the cascode's base voltage enough to give the full PSRR? What lurks on R2, R6 from the rails, is not affecting any more?
Well, the cascode has some stabilizing purpose because the voltage at the Fets is constant. R2, R6 can be substituted with current sources. Yes, if the PSU has a flaw, the resistors can not supress this. It will simply downmodulate. Popa has shown in Linear Audio Volume 1 that a current source improves PSSR ( humm ) but does not reduce the noise ( hiss ). How a cascode can be made that is referenced to ground and has no seperate voltage reference i have shown in my FPS in Linear Audio Volume 0 cleaning the baby and not throwing it out of the water the same time. This little workaround did not stir any interest though as far as i can tell from the comunication i got. My PSU that i have not shown so far solves that problem to a certain degree too. I could go on and on. This is all managable with ever increasing complexity. The question is what we want this time. It started out with minimum complexity plus no use of obsolete components and now we are back into state of the art, high tech solutions plus some NOS thrown in for good measure. If an fine with that as long as do not bore the contributors and readers.
We want clean supply for no extra complexity. Since there is a clean rail, and good cascode base supply is established, we are OK short of folded cascode. Its good we discuss why and how non the less IMHO.
I think now that the circuit is fine. Not too expensive, not too complex and stiil performs well at least in simulation. This micropower references are not expensive at all despite the exotic name. The 10uF to ground can be made from 2 x 4.7uF Wima MKT-2. They are inexpensive too and measured under 10mOhm ESR at 100kHz. I build the 12V battery supply version and was happy with the measurement so it works in praxis and the latest changes will make it only better. Lets look at an equally simple and good PSU. I think that could be fun.
You may like what juma liked for his BF862 preamp. juma's BF862 pre. You also see how it mirrors there.
I attach the non Zener one.
I attach the non Zener one.
Attachments
