Working Prototype of a Kelvin (4-Wire) Milliohm Meter

RE: Working Prototype of a Kelvin (4-Wire) Milliohm Meter

fmilburn — Fri, 12 Oct 2018 05:54:41 GMT

I am going to be away from the computer for a couple of days so here is an update if anyone following cares to comment.

[View:/resized-image/__size/620x469/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-bf0480ae-37a4-44dd-8ddd-dcb9d28e8fce/8562.contentimage_5F00_188715.jpg:620:469]

Summary of changes:

Added 4.7 V LDO - it can be bypassed if not required
Added tantalum footprint R9 if low pass filter desired
Added on / off switch
Added power indication LED to panel
Added op amp acting as comparator to panel to indicate out of range past 4 V - Shabaz: does this look right?
Added power headers in case panel meter or other stuff needs power
Added various test points
Added mounting holes

I am thinking that it would be possible to take two boards and populate only the parts needed and then connect through the test points to extend range.

Preliminary BOM:

[View:/resized-image/__size/620x600/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-bf0480ae-37a4-44dd-8ddd-dcb9d28e8fce/8176.contentimage_5F00_188716.jpg:620:600]

I found everything on Newark except the LDO Regulator - could not find a fixed value 4.7V LDO so may have to change. Everything else should be there except the panel meter which I am still waiting on from Aliexpress. Will provide jacks and send it to the multimeter if the panel meter doesn't work out - or maybe should provide jacks in any case.

RE: Working Prototype of a Kelvin (4-Wire) Milliohm Meter

shabaz — Thu, 11 Oct 2018 15:15:34 GMT

Hi Frank,

By the way regarding Q2 in the schematic, it looks like the part is moved but the net connections remained the same maybe (there is a line through Q2). Just wanted to highlight it, in case it messes up the final board, although you'll probably spot it anyway once you start on the PCB design.

[View:/resized-image/__size/432x306/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-bf0480ae-37a4-44dd-8ddd-dcb9d28e8fce/contentimage_5F00_188714.png:432:306]

RE: Working Prototype of a Kelvin (4-Wire) Milliohm Meter

shabaz — Thu, 11 Oct 2018 05:20:59 GMT

I got curious if autoranging could be done with logic gates : )

I'm not sure if this would work, it was just some late-night brainstorming : ). But maybe it is too complex, since a SPDT switch will work as well!

[View:/resized-image/__size/878x542/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-bf0480ae-37a4-44dd-8ddd-dcb9d28e8fce/contentimage_5F00_188713.png:878:542]

RE: Working Prototype of a Kelvin (4-Wire) Milliohm Meter

shabaz — Thu, 11 Oct 2018 03:30:39 GMT

Hi Frank,

This is great to see, it it excellent work on the schematics and the prototype! I hope you're planning to stick this as a Project14 entry : ) it definitely looks like a handy piece of test instrumentation, especially since you're getting successful sub-milliohm accuracy.

Considering this is still using prototype resistors with high tempco, the results are really promising. I'd like to build it sometime too, I still need to get the op-amp and trimmer, and some resistors.

I think it will speed up measurements once lower tempco resistors are used too.

I was wondering why the accuracy dropped off slightly right at the bottom, and could it be the probes? From the photo, I couldn't tell, are the wires going to each clip separately? Sometimes they maybe solder them together on one side, which is fine if it is because that might explain the small delta at the lowest measurements.

Incidentally, although maybe it is not necessary, there is already half of a low-pass filter in the circuit for free : ) It just needs a capacitor (tant or electrolytic, maybe 1uF) from the bottom of R1 to ground. It could help if there is any noise on the supply line.

Regarding different ranges, I can't think of an easy way with this arrangement. Maybe it is a bit wasteful (but the parts are cheap), the current source could be repeated, but for 1mA, and have a SPDT switch to select either of the ranges. It could maybe (would need to be tested) be done with MOSFETs too, i.e. no need to use a SPDT if you wanted to do it electronically somehow, e.g. for future auto-ranging hehe. For that, the two current sources could be permanently connected together at the top, but the ground connection of each current source would instead be connected to ground via a N-ch MOSFET each. I don't know what the leakage is though, maybe it would need to be selected more carefully than a jellybean MOSFET, for this to work.

Anyway, congrats on the progress, and can't wait to see the PCB : )

RE: Working Prototype of a Kelvin (4-Wire) Milliohm Meter

jw0752 — Thu, 11 Oct 2018 02:45:45 GMT

Hi Frank,

You have made an excellent approach to this project. I think the way the expected results correspond to the actual is very encouraging. For the fun of it the other night and inspired by your project I found myself trying to measure the resistance in some copper wire I had in the shop. I also have some brass wire which I have used for shunts in the past as it has a higher ohms / length. It is a real challenge and I salute your excellent results. The use of the box for a temporary enclosure worked well to keep things positioned and stable.

John