Coil magnetic field measurement

[Zollie]

Hi, I would like to have some suggestions for solenoid valve magnetic field measurement.
Some specifications: sensor diameter must be smaller than 14mm but <6mm preferred.
Accuracy: min. 0,1 % (max. field about 100 mT)
Frequency max. 7 KHz
NMR sensors are too big, fluxgates haven't got the bandwidth and hall elements and GMR sensors cannot fulfill the accuracy criteria. SQUID and SERF do not even come in question.

Anyone has got any ideas?
Thank you!

 

[VEBill]

Here's a lead I just found. I didn't dig all the way down... ...I'll let you investigate further to see if they have anything suitable. Good luck.

Senis Transducer Brochure (pdf)

 

[IRstuff]

Unless you rework the problem, you've presented an impossible situation. I would revisit the accuracy requirement, particularly w.r.t. to the Hall and GMR sensors and look to either loosening the accuracy requirement or doing a direct calibration of the sensor or both

TTFN
faq731-376

 

[VEBill]

I wonder if the Watt Balance might provide some ideas.

 

[Zollie]

Thanks for the answers!
@ VE1BLL: Senis seems ok, I am looking for a cheaper alternative. I found this:

http://www.hoeben.com/productdata/Datasheet-HE144-20120601.pdf

As far as I see it, Watt Balance is for mass measurements.
@IRstuff: The problem really has high criteria, they might be lowered if there was no solution.

 

[VEBill]

The Watt Balance also involves extremely high precision calibration (or perhaps calculation) of solenoids and their magnetic fields. Perhaps there's some advice in the related literature; just in case you run into a road blocks.

 

[IRstuff]

A Watt Balance approach potentially could be taken, but I've only seen that approach on radiometers, and even then, the absolute accuracy was only around 0.5%. Additionally, with a 0.1% measurement tolerance, you'd need submicron positioning accuracy, particularly the closer you get to the poles.

Your cited Hall effect device appears to net out at about 0.5%, when you account for all the parasitic effects, like temperature, Joule heating, nonlinearity, etc.

TTFN
faq731-376

 

[Zollie]

It is important, that my coil is about 15 * diag=10mm, so watt balance should come out of question.

Which cited Hall Sensor are you referring to? Please note, that the temperature coefficient of the offset can be mostly compensated by the HE 144 Sensor.

 

[VEBill]

The 'Watt Balance' has been in the physics news recently because it is one of the technological approaches being considered to replace the prototype kg mass held in that Paris vault. Wiki: "As of 2011, the accuracy record is held by the U.S. NIST, with a relative uncertainty of 3.6 × 10⁻⁸ ..."

They use clever techniques to eliminate unknowns. It's those clever techniques that I was intending to point to with my mention of the topic. Basically, that 10^-8 makes "0.1%" look pretty easy in comparison.

 

[IRstuff]

Why is there such restriction on a solenoid valve?

TTFN
faq731-376

 

[Zollie]

I see the point on watt balance and regarding accuracy, it surely can be used. It's just, that it would be easier to buy a sensor, rather than constructing one from scratch.

Well, the solenoid drives an analog electromechanical valve with really small movements.

 

[IRstuff]

How small? People have built 1 nL dispensers without going to this level of complexity:

http://www.nordson.com/en-us/divisi...ts Pumps Valves/DS-DJ-9000-NordsonASYMTEK.pdf

TTFN
faq731-376

 

[Zollie]

The movements are about 100 micrometers.
As a summary my thoughts are:

Now I would like to choose between 2 options. There is the Hoeben Sensor, that I have linked before. This sensor is mounted on a circuit. The accuracy is a bit higher than 0,1%, as ohmic offset voltage / temperature can be corrected.(There is a manual for it on the website.) There are also the parasitic effects, IRstuff mentioned before.

The other option is a calibrated probe with a total accuracy of 0,1%. The link:

http://www.senis.ch/docs/default-so...agnetic-transducer-datasheet_rev.pdf?sfvrsn=0

Here I do not have to build a circuit, therefore the total accuracy may be better as in case 1.

 

[Skogsgurra]

Before you get too involved in balances and such things: Do you need to follow the 7 kHz flux with high fidelity? Or is an RMS value good enough?

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[Zollie]

RMS should be good enough. It's just that because of the electronics most sensors have a low pass behaviour.