Wow, it almost looks like some Ghz RF receiver with that cooper wave guides, I'm wondering what noise floors this beast is hitting for all this screening taking place there, have to find this thread.... I didn't had to go to such extremes as my noise floor is bearly -70, -90dB on 50Hz and harmonics sensitive frq area. I traced every cm of wiring and now have 100% confidence that whats left is from cart, if I put foil on it, everything is clean. I refuse to post my cookie box shielded proto pic versus this peace of art 
So if what is left can't be taken care off by running balanced in stage I'm afraid it looses the whole point. Now I could be dumb or blind or both this evening, but I'm having second thoughts about this affair. It seems that the only reason why it works in proto is finite resistance of sk170 current sources. If I cascode them, everything starts to swing widely. Current just has no return path so it circulates between mirrors. Sim's do confirm this. So we have to reference inputs to ground, but then we have 2 noise sources that isn't shunted with cart resistance. I guess they would not add precisely by sqr root because of balanced CMRR but for sure they will lift up the noise floor by some amount. Have to check it in proto, and that I'm afraid would have to wait for tomorrow.
So if what is left can't be taken care off by running balanced in stage I'm afraid it looses the whole point. Now I could be dumb or blind or both this evening, but I'm having second thoughts about this affair. It seems that the only reason why it works in proto is finite resistance of sk170 current sources. If I cascode them, everything starts to swing widely. Current just has no return path so it circulates between mirrors. Sim's do confirm this. So we have to reference inputs to ground, but then we have 2 noise sources that isn't shunted with cart resistance. I guess they would not add precisely by sqr root because of balanced CMRR but for sure they will lift up the noise floor by some amount. Have to check it in proto, and that I'm afraid would have to wait for tomorrow.Attachments
Good you stashed, because right now almost everything listed as J74 is as real as Milli Vanilli.
Hi there,
sad news from the Milling front.
Due to a suspected serious Heart-attack,
my "Milling-man" is off the Job for at least
three month.
So i have to see what i can do as a alternative.
Keep you informed...
@Joachim,
this Motor-thingie is very interesting too.
Because i've build two variations for AC-motors.
the first was this one;
SinCos - die Motorsteuerung
After a bigger GB i've still got some sets in stock.
The second;
Roehren-und-Hoeren :: Thema anzeigen - Luxusnetzteil für Synchronmotoren mal anders
They are almost the same concept except serial/usb connectivity.
Sorry, both links are german.
sad news from the Milling front.
Due to a suspected serious Heart-attack,
my "Milling-man" is off the Job for at least
three month.
So i have to see what i can do as a alternative.
Keep you informed...
@Joachim,
this Motor-thingie is very interesting too.
Because i've build two variations for AC-motors.
the first was this one;
SinCos - die Motorsteuerung
After a bigger GB i've still got some sets in stock.
The second;
Roehren-und-Hoeren :: Thema anzeigen - Luxusnetzteil für Synchronmotoren mal anders
They are almost the same concept except serial/usb connectivity.
Sorry, both links are german.
Yes, returning the collectors to the opposite supplies gives a diamond buffer. That will work better, sure. The circuit shown is based on an idea of Baxandall who found that base and collector of a BJt can be on the same potential. It was used by QUAD and later others but with feedback.
I have a better idea that works fine. I have already build that input stage and presented it on this thread under the name "Hulk". I think i have that stage still lying around somewhere in perfect working condition. I will put it in the system tonight. As far as i remember, it sounded pretty damn good. Martina and i also used it on ETF2010 to good advantage.
I have a better idea that works fine. I have already build that input stage and presented it on this thread under the name "Hulk". I think i have that stage still lying around somewhere in perfect working condition. I will put it in the system tonight. As far as i remember, it sounded pretty damn good. Martina and i also used it on ETF2010 to good advantage.
