Tips on getting low resistance connection to DUTs for measuring low ohms?

[cdev]

My DUT in this case is an old antenna tuner I got for cheap because the seller couldnt get it to tune right, the reason was that it has been lubed with a controversial graphite loaded grease - which is now black gunk.

Anyway, this is the deepest cleaning its gotten so far and what must be hundreds of q-tips and at least six old toothbrushes black gunk is still coming out. The critical connections in the bearings of the roller inductor and variable capacitor need to be squeaky clean. To evaluate performance I am measuring the ohms as I clean.

My old standby alligator clips arent doing it for me.

There is no way I am going to be able to do this right without better leads.  I think soldering the wires to the device is the best way to get a low ohms connection but the banana plugs I have available themselves don't appear to be solderable, they have screws. So I guess I am stuck with compression. At least the screws are gold.

I am using my UT61E as the meter- using relative mode.

There must be something else around here that I can use.

This is the first time Ive been concerned about such low resistances. In the case of RF circuits, though its worth it because only when the "Q" is super high do they really sing.

[Kleinstein]

For accurate low ohms measurements one would use a 4 wire connection. However if the meter does not support it, like most handheld meters, one would need an extra constant current source and use the meter only to read the voltage.

The resolution of the meter is limited anyway. So I would expect good banana plugs to be well good enough.

[edpalmer42]

It looks like that meter has a resolution of 0.1 ohms.  If you need more than that, set up a four-wire ohms arrangement.  Use a constant current source of some convenient value and measure the voltage drop across your DUT.  The connection resistance is irrelevant because you make your voltage measurement inside the current connection points.

I didn't see anything in the manual that said what the UT61E uses for a test current, but it could be 10 ma or less.  If you could use a 100 ma test current, a voltage reading of 1 mv would equal 0.01 ohms.

Make sure you don't put too much current through your DUT!

Ed

[bson]

If you have a decade resistance box with a suitable range and a suitable ammeter (coil or picoammeter) you can use a Wheatstone bridge. Not sure how well a DMM will work due to the burden (never tried, haven't done this since school), but a used dial coil ammeter can be had for peanuts.  You don't need it to be accurate even since all you need is to balance the bridge by centering the dial returning the dial to its resting position.

Edit: I'm confused clearly.  A voltmeter can be used in a Wheatstone bridge!

[Vtile]

Go with Kelvin aka 4-wire measurement. The beaty of it is that it cancels out the wire resistances ..like said above. Volt meter input impedance will produce error even in this setup, but it is such a fraction of non-existent with modern megaohm range meters that you should calculate the effect just for fun. Don't go with bridges they have all sort of traps because of non-linearities, wire effects and meter inputimpedance related errors. It is zero detection setup like an old mechanical scale only at its best at near balance/zero conditions.

With a quick calculation 100mA source would give 100uV per 1mOhm if I didn't make a silly mistake. This would be still shown by your meters 200mV range... edit I DID .. I'll go to sleep edit2. No I did not make a silly magnitude error, but I'll bury my head to pillow anyway.

[cdev]

I do have an old center reading D. Arsonval meter movement made by Simpson - in mint condition, (but its the more modern looking square plastic design) and the legend on it simply says "Tension".

 A Wheatstone bridge does really sound like the ideal thing to do with it.

Maybe with a pot and a digital readout of the resistance.

Thank you!

[Brumby]

Personally, I'd go with a 4-wire measurement - or a home-brew equivalent.

Since I don't have a 4-wire meter, I'd go with the constant current source flowing through a reference resistor and then the DUT.

Measurements requiring accuracy:
 * the reference resistor resistance (out of circuit)
 * the voltage across the reference resistor (in circuit, constant current flowing)
 * the voltage across the DUT (in circuit, constant current flowing)

Use Ohms Law to determine the resistance of the DUT - since the current through the two resistance is the same.

The actual value of the current is unimportant for the maths involved - but, as said above, be aware that it's not too much for the DUT ... AND that it is not enough to cause thermal variances in the resistors (either the reference or the DUT).

Edit:  Actually, you don't really need a constant current source for this.  Since this is a static topography, just make sure that the current doesn't vary between the two voltage measurements.  I'd check V_Rref, V_DUT, V_Rref, V_DUT, V_Rref - and if they are consistent, that will be perfect.

[edpalmer42]

Brumby and Kleinstein, we're overthinking this.

Cdev doesn't need to know milliohms.  All he needs is a relatively constant current at a value the DUT can take.  Use a lab power supply on constant current or maybe 30V - 50V with a big resistor in series.  Whatever. Put the current through the DUT and measure the voltage across the DUT.  Note the voltage, clean some more, see if the voltage has dropped.  Also note if the voltage is stable as the roller inductor is moved through its range.  Actual value or precision of current and voltage isn't important - it's more of a nice-to-have.

Ed

[Brumby]

That's a fair comment - but the reason I went with the approach I mentioned was this:  Just because the voltage doesn't drop, doesn't mean you have reached the best point possible.  It may simply mean that where you are at does not respond any better to the process you are using.

By having an actual resistance figure, you can better judge whether further efforts are warranted.

[TimFox]

If you don't need high accuracy, just high sensitivity (low ohms-value resolution), then you can use the most sensitive voltage scale on your meter with one pair of leads, and use a simple power supply plus power resistor to supply the current through the second pair of leads.  Since the voltage across the low resistance you seek is quite small, you need not worry about accuracy if you use a 1% power resistor and a few volts of DC (rating the resistor to handle the power with the voltage across it), measuring the voltage before proceeding.
For example, with a 200 mV full-scale and 0.1 mV resolution, if you apply 2.0 V through 20 ohms/1 watt (100 mA), your voltmeter resolution corresponds to 1  m and the full-scale corresponds to 20 m .  If the resistance be higher, then increase the resistor and/or voltage accordingly.

[cdev]

Hmm..

I have some temperature sensitive liquid crystal sheeting (basically poor mans thermography)

 I could use some current flow (limited via a power resistor and constant current) to get a better idea of the exact location of the resistance and its changes as the roller is rotated.

Unfortunately, today is not cold enough to make this work unless I crank up the single AC here. So i'd rather just wait a day or so - wait until say tomorrow morning - assuming its cooler.