I'm designing an "overkill" amplifier and am looking to design a ultra high impedance, ultra low noise current source. Right now, it consists of a cascode current mirror using a THAT300/320 transistor array and two green LEDs to set a reference voltage.
My web research has found that use of LEDs to set the current source's reference voltage is about the best one can do to get low noise with a down-side of temperature dependence for actual current.
One question is this: what LEDs are suggested. I've seen references to "rectangular LEDs". Why should they be rectangular?
What would be a good PN choice in surface mount? Why?
My original design used a "filtered" 1N4xxx series "low noise" zener diode, but I found that using said part (without filtering) yielded a noise voltage of 80-100x worse than a junk-box T1-3/4 green led. Even with a "tapped" resistor filter, it was still worse than a LED.
After days of searching for info, I've come up dry. It's time for me to consult the experts!
Thanks,
John
My web research has found that use of LEDs to set the current source's reference voltage is about the best one can do to get low noise with a down-side of temperature dependence for actual current.
One question is this: what LEDs are suggested. I've seen references to "rectangular LEDs". Why should they be rectangular?
What would be a good PN choice in surface mount? Why?
My original design used a "filtered" 1N4xxx series "low noise" zener diode, but I found that using said part (without filtering) yielded a noise voltage of 80-100x worse than a junk-box T1-3/4 green led. Even with a "tapped" resistor filter, it was still worse than a LED.
After days of searching for info, I've come up dry. It's time for me to consult the experts!
Thanks,
John
You should be able to find some pointers in these threads:
http://www.diyaudio.com/forums/showthread.php?threadid=114813
http://www.diyaudio.com/forums/showthread.php?threadid=35821
http://www.diyaudio.com/forums/showthread.php?threadid=114813
http://www.diyaudio.com/forums/showthread.php?threadid=35821
jgedde said:
Simply wrong conception of engineering design - you say you want to maximize 2 parameters while using a topology that requires trading them off against several other unstated constraints
the way to find a acceptable balance is to look at the larger amplifier circuit context to determine which ccs parameters impact the rest of the circuits' performance the most
"the best" off-road truck tire isn't very good on a sprint bicycle
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.