EEVblog Electronics Community Forum
Electronics => Beginners => Topic started by: XRFlight on November 13, 2023, 10:27:12 pm
-
Hello,
Trying to design and build a couple insturments as a way to learn before jumping onto bigger projects, as well as give myself some useful tools. Right now I am working on a picoammeter. Ideally with a picoammeter you want to have dual supplies to work with, currently I am using an op-amp to create a ground half way between the positive and negitive sides of a linear regular's output. I want this ground would serve to provide the driven gaurd to the sensitive input of the picoammeter, but also thinking about connecting it to chassis ground. From what I have read, it doesn't matter much what side of the power rails you connect to chassis when using the chassis as a faraday cage, so long as you connect something you will reduce noise. What stumps me is this, why not use gaurd? All the professional picoammeters/electrometers use triax cables with input in the inside, gaurd on the middle layer, and sheild on the outer layer. If the sheild could be held at gaurd potential however, then it would seem like the expense of the triax cable could be removed. Is there some sort of noise that will be induced by connecting the op-amp driven ground to the chassis or some other factor that I am missing here?
I can find all sorts of information about grounding to chassis, dual supplies, and guarding, however there seems to be little information about combinations of the three.
-
Not sure exactly to what you are trying to achieve, however triaxial signal cables intermediate braid are often used to implement "powershielding"; "power guarding"; or "ring-fencing", of low level inputs, bootstrapping them to increase the effective impedance of the input device, and also negate any electrical leakage in the insulating materials in the signal path.
Connecting to ground, or some fixed potential isn't going to achieve that powershielding/bootstrap effect.
If its solely to provide signal shielding then yes your system could well work, and all I can say is try it!
-
My understanding is that the guarding works by minizing the voltage differential between the signal and the next conductor current flowing to or from the signal would encounter across a surface or through an insulator, therefore minimizing the leakage current. My question is why not set sheild to the signal potential (a transimpedance amplifier (TIA) picoammeter keeps the signal potential near constant), so that the sheild can also be the guard?
-
Another way is to try it both ways. When I designed an AFE I was prepared for many predictable scenarios, I designed defenses, expect the leakage current inside a CMOS switch.
So my suggestion is try, fail fast and iterate.
-
Thanks, I will see if I can configure the circuit so I can ground the chassis from muliple places (only one at a time) and see where it goes.
-
You can make a Triax cable if you can pull out the center conductor and insulating layer of a larger coax, then put a smaller coax thru the empty bigger shield.Also look up this on US ebay
NEW Belden 9222 50 ohm Triax cable Sold in 10’ Increments $1.89 Per Foot RG 58/U
-
Thanks everyone for the replies. As things always go, after looking for an answer for a couple weeks and finding nothing, I stumbled upon the answer while looking for other information. It appears the reason that triax is used in commercial electrometers is that the ammeter/coulombmeter is floated relative to the chassis and other circuitry. The guard would be connected to an RF shield inside the case, which I would assume to prevent noise due to the lack of connection to chassis ground, while the outer shield would be connected to the chassis. It appears common practice is to ground the guard to chassis if the test inputs are at ground potential.
Extrapolating from there, I assume that it is likely desirable for the shield to connect to the chassis at the point of entry, but that care must be taken to ensure that doing so does not result in a ground loop or currents from elsewhere in the device flowing through the analog circuitry. For a simple battery powered picoammeter that contains no digital circuitry or other inputs/outputs, it would appear to me that standard coax would be at no significant disadvantage.
Of course now debating using triax connectors elsewhere for other reasons. I will post the design in the projects forum once it is ready.
-
Here is the HP Application Note for GUARDING: Hopefully quite instructive
http://hparchive.com/Application_Notes/HP-AN-123.pdf (http://hparchive.com/Application_Notes/HP-AN-123.pdf)