EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: CopperCone on June 11, 2017, 05:26:45 am
-
I have a remote microphone. It is powered. It is meant for PCB use. It runs off 5VDD. It is just a op-amp buffer on the mic with no gain.
It has 3 connections, gnd, signal, and power. I have ferrite and small value capacitance glued to the sensor, (signal is shielded with a EMI filter.. the one component types).
Ferrite bead -> capacitor (nF) -> power (this is LC)
emi filter (bead type) -> signal (this is C-L-C)
This goes out to a tantalum capacitor (uF) that is about 2 inches away. (which is between power and GND). Between the microphone and the tantalum I twisted GND and signal together and coated it with epoxy. I will paint it with conductive epoxy next. This is done with 30 gauge wire.
From the capacitor, should I *using large wire*
1) branch it off, so there is two ground wires (signal ground and power ground, with two twisted pairs for the signal/signal ground and power/power ground
2) leave it as 3 wires? (twisting signal and ground together)
I think option 1 is wrong, but I am not sure.
also,
I plan to connect the epoxy paint to the cable shield, and connect that to the circuit ground at the receiver.
No matter what I do, there will be a fork in the road to a 4 wire split (since there is a power supply and a amplifier), but should it be as short as possible or as long as possible? I think the cable may be as long as two meters.
It seems by option 2 I am doubling my ground conductor length. So it can pickup more noise. But then you can put a common mode filter on signal. Especially if two separate boxes for PSU and receiver/filter are used (both floating). But then its making some kind of differential ground impedance and I think thats bad.
Passive sensors are much easier to comprehend...
Also, with option 1, would I just use a bead choke with 3 wires coiled in it at the earliest opportunity as close to the capacitor as possible?
So,
1) tantalum capacitor -> common mode choke for all 3 conductors (ring torroid) -> shielded 3 wire cable -> split as close to the two instruments as possible (maybe I will put a 5V daughter board in the amplifier)
2) tantalum capacitor -> ground split ->
gnd1 + signal -> common mode choke -> 2 meters cable -> common mode choke (lcl or clc filter maybe?) -> amplifier
gnd 2 + power -> common mode choke -> 2 meters cable -> common mode choke (? filter with big and small caps) -> power supply
But then imbalance in the chokes impedance causes weird effects?
the signal is AC.. .so I don't know what the ground offset will do (since with option 1 you basically short the two instrument grounds at the floating instruments, with option 2 you have a long wire connecting the instruments. with chokes in between).
This also applies to connecting a breadboard to multiple instruments.. I wanted to get as much gain as possible from the mic for a quiet signal..
Perhaps with option 2 would work if the instruments are shorted together using a grounding strap (like a low impedance cable shield) directly soldered to the PCBs where the reference rail is ? (rather then shorting at the signal input). I think this is optimal (but annoying since you need to modify the instruments to have a ground strap connection to their circuit ground..) or am i missing something?
Been a long day