Resistance Sensor Problem

What's wrong with measuring it between the ends you are holding it by?

If it's an incremental thing, i.e. etching it bit-by-bit and checking the diameter vs. position to control it better, you could do that with a higher precision total measurement, since the resistance of the rest doesn't change.

I wonder if capacitance is any help. A coax capacitor goes as the log of the ratio of diameters. Probably not so good against off-axis alignment though.

Tim

-- Deep Friar: a very philosophical monk. Website:

A difraction of light or ultrasound waves over the wire could provide pretty accurate non-contact measurement of the diameter.

Vladimir Vassilevsky DSP and Mixed Signal Design Consultant

If you used a non-conductive pulley to form a loop of wire after it had emerged from the bath, and arranged things so that wire leaving the loop was in electrical contact with the wire entering the loop, you could make this loop a shorted turn around an inductor.

The core of the inductor would ideally be a circular cross-section toriod, and the circular cross-section would be the axle of your non- conducting pulley.

In practice, two U-cores or two C-cores would form a perfectly acceptable substitute for a toriod.

You would use a fair number of turns on the drive/sense winding for the inductor so the voltage drop around the wire loop on the pulley isn't high enough to produce a destructive current in the wire you are looking at.

You will drive the coil from a stable AC voltage source, and monitor the in-phase AC current you need to drive through the coil to sustain this voltage. The assumption is that you won't need much current in the absence of the shorting turn, and that most of this this current is going to be in quadrature with the drive voltage and mainly determined by the inductance of the drive coil and the frequency of your AC sourvce. There will be some in-phase component in the absence of the shorted turn due to the resistance of the drive coil, but you should be able to pick a frequency where the inductive reactance is a lot bigger than the resistance.

When the shorted turn is in place, you will need to put more current through the drive coil to maaintain the same voltage across the coil, and the extra current will be current circulating around the shorted turn, divided by the turns ratio. The voltage across the shorted turn will be the voltage across the coil - again divided by the turns ratio

- so you are going to know the resistance around the current loop, unfortunately includng the resistance of the bit where the two ends of the loop rub against one another.

It might be good enoug.

-- Bill Sloman, Nijmegen

4-wire measurent with liquid contacts would be interesting.

It's already in a conductive liquid. Run it through four close-fitting conductive slots or sleves and do an AC 4-wire measurement. You might need some insulating or grounded guards, depending on the geometry. The AC measurement would untangle the resistance measurement from the

10 amp etching current.

But this would need some amount of circuitry.

John

ehsjr a écrit :

Use two more rolls and do a 4 wires resistance measurement.

sense sense roller roller | | v v _ _ wire__/0\O______________O/0\________wire ^ ^ | | current current roller roller

--
Thanks,
Fred.

This would be the right solution if the biggest part of the problem were a varying resistance between wire and wheel, but would not be able to deal with the wire bouncing fully off the wheel.

I think you're taking aim at the right problem, though. I'd also watch the dynamic behavior of the measurement, and look for measurements that were out of range -- i.e., times when the current goes to zero (indicating that one or the other of the current-injecting wheels is not making contact at all), and times when the voltage goes to zero (indicating that one or the other of the voltage reading wheels is not making contact).

If you make the measurements at a higher rate than you really need, you'll have the freedom to discard the ones that are bad, and still salvage the correct behavior of your control loop.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Do you need to implement control loops in software?
"Applied Control Theory for Embedded Systems" was written for you.
See details at http://www.wescottdesign.com/actfes/actfes.html

Is size the problem or resistance? If Size, use a Laser Micrometer. Many of those units give PID options to regulate things.

As for the R method, that would be simple with an offset roller system. The rollers are conductive using a material like brass/copper and would be wide. The idea is to wrap several turns of this wire around each roller. It feeds on one end and ends up spooling off the other end, this would be the pay off roller. The take up roller would be the same, entry to one side, wrapped a few times and pulls off the other side of the roller.. Doing this will give you plenty of contact area with no tension since you can use very light weighted rollers with a conductive surface and needle bearings. A slip ring connection can be made there.

If you wanted too, you could put a tension motor on the second roller with a small catenary arm on a optical pot to control the tension between the two rollers.. I've found that a tension only with no feed back works ok in most cases at high speeds.

Our plant does have a machine that can handle small wire like that and it uses plastic beam supported open rollers with hard surface on them. The wire gets past through an induction heater. These rollers spin so easy that a little breeze that you can't even feel will turn them with no item wrap on them.. They're on graphite bushings and one roller has a metal sleeve on it with a slip ring. I suppose you could do it with both rollers.

Laser Micrometer would be my choice. It depends upon what dia the OP is dealing with. Another thing to watch out for is temperature effects of various metals. The final dia when its out of the chemical bath would be smaller, if of course, the bath is a higher temperature.

Cheers

We have a process done at work only in the lab for testing the DCR of conductors and that is to clamp some weights to the ends and drop the leads into a mercury tub with specific lengths.. The tubs are the connectors for the leads..

Out on the floor they use kelvin connections at testing, but this of course can not be done on a moving line..

Jamie..

I'd probably do it by looking at the optical diffraction, if the environment is clean. Or possibly by measuring its stress vs strain curve--put a constant tension on the takeup spool, a constant side load on the guide pulley, and measure the deflection of the guide pulley's mount. The elongation at constant tension will go as 1/(cross sectional area).

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal
ElectroOptical Innovations
55 Orchard Rd
Briarcliff Manor NY 10510
845-480-2058

email: hobbs (atsign) electrooptical (period) net
http://electrooptical.net

how about skip rollers anse use polished rods or wires in this layout instead to insure constant contact. connect each set of two or three rods on each end and measure resistance across them

wire -------o/o\_________________o/O\-------- wire

as long as you have tension, the wire will touch the 'o's which are cross sections of rods or ticker wires or whatever.

Why not call somebody who's already making wire and ask them?

Good Luck! Rich

Which leaves you with a well controlled diameter, but at the mercy of the resistivity of the material (and possibly roughness of the etch) for the resistivity of the resulting wire.

It's probably not a problem, but...

--
http://www.wescottdesign.com

And I wonder who would be in here that works on, designs and modifies such equipment.. ;) Among other things.

Jamie.

Thank You to all who replied.

Lots of replies. Nice :-)

Ed

Oh, it CAN be done; I've seen fine wire guided so that it contacted the (positive) meniscus of a mercury droplet. Any glassblower can build you a good four-wire set of Kelvin contacts for a horizontal moving wire.

Yup, that's true. On the other hand, essentially all manufacturing process control is based on correlating what you want to manage with what you can measure reliably and affordably. Optical and sliding-contact sensors are tough to keep working in most factory conditions. (My experience is mostly with Class 1 clean rooms, and it's tough enough even there.)

The method I posted won't actually work without some changes--the constant tension roller will keep the lateral deflection constant as well. You'd have to measure the guitar-string resonance instead--for constant tension, that gives you the mass per unit length.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal
ElectroOptical Innovations
55 Orchard Rd
Briarcliff Manor NY 10510
845-480-2058

email: hobbs (atsign) electrooptical (period) net
http://electrooptical.net

That sounds promising. Run the wire through a magnetic field, drive a current through it, pick up any oscillation of the wire in a non-magnetic manner (optical?), then add feedback to create an oscillator where the frequency of the wire oscillation at a constant tension is proportional to the mass of the wire. Alternatively, drive the wire mechanically (similar to bowing a violin string) and measure the oscillation with a magnetic pickup.

I'd probably try making an oscillator using a guitar pickup directly.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal
ElectroOptical Innovations
55 Orchard Rd
Briarcliff Manor NY 10510
845-480-2058

email: hobbs (atsign) electrooptical (period) net
http://electrooptical.net