Electronics > Projects, Designs, and Technical Stuff
Uber low noise current source suggestions
<< < (4/5) > >>
Kleinstein:
10 nA noise for some 500 mA of current is 0,02 ppm level.  No way one get this with normal room temperature electronics. If possible at all this would be the range where superconducting devices are needed. So think about  a Josephson junction reference squid amplifiers and NMR based current reference.

I kind of doubt the requirements are really such a stable current, more like a very stable light output or a very stable temperature. But than it would probably better to directly control the intensity of temperature, but no the current. Current would than be only a control input.
tszaboo:
So, you design a current source for ~1ma, make sure it is low noise.
You know, JFETs, low noise opamps and stuff like that.
Get a good contract from your customer.
Connect 650 of it in parallel, and let statistics take care of the noise.
Maybe you will need larger number and lower current.
Marco:

--- Quote from: Kleinstein on July 04, 2019, 03:57:50 pm ---No way one get this with normal room temperature electronics.

--- End quote ---
I'm not sure if that's true as far as thermal noise is concerned.

Bigger problem is that you could only ever deliver the current noise through a resistor ... through a laser diode you are shot noise limited to ~7 nA RMS, so ~50 nA peak to peak. Throwing it in liquid helium won't help you.
ejeffrey:

--- Quote from: Marco on July 04, 2019, 05:24:59 pm ---
--- Quote from: Kleinstein on July 04, 2019, 03:57:50 pm ---No way one get this with normal room temperature electronics.

--- End quote ---
I'm not sure if that's true as far as thermal noise is concerned.

Bigger problem is that you could only ever deliver the current noise through a resistor ... through a laser diode you are shot noise limited to ~7 nA RMS, so ~50 nA peak to peak. Throwing it in liquid helium won't help you.

--- End quote ---

That's not true.  Laser diode driven by a low noise current source can easily be below the shot noise limit.  Think about it: if you have a sub shot noise current through a resistor and the resistor is in series with a diode, where is the noise current going to go?

Shot noise applies to a diode that is voltage biased.

At high frequency you will see shot noise because at high enough frequency the diode capacitance is low impedance.  But up to 100 khz is no problem.

Achieving the specified noise parameters is going to be difficult but doesn't seem impossible.
Marco:

--- Quote from: ejeffrey on July 04, 2019, 07:04:34 pm ---Shot noise applies to a diode that is voltage biased.

--- End quote ---

Oops, guess you're right.

Well in that case 20 current sources consisting OPAx189s and MOSFETs with 400 Ohm sense resistors should just about get you there in theory. Only 50 Euros of opamps.

PS. I guess if <0.1 Hz noise doesn't matter so much you could also just use precision BJT opamps, though you can't really reduce the number of current sources much if you want to keep the supply headroom the same (ie. ~10V) because the thermal noise of the sense resistors starts dominating.
Navigation
Message Index
Next page
Previous page
There was an error while thanking
Thanking...

Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod