Now for the "Math Rock of the Day" question,

Actually, it isn't quite that simple. (Well, not to me, anyway...)

I have a practical problem that I'd like to solve using stuff that I have.

Namely, some UGN-3130 Hall-effect sensors - coupled with bits of wire, ferrites, what have you - to determine when current through a circuit is above or below 2A, +/- 10%.

According to the datasheet, these little jewels trip when gauss is 175, and release when it is -175, within a certain temp range and all that stuff.

So, what I'm thinking is that I can establish a -175 static gauss bias on one side of the Hall-effect, and let it get overpowered on the other side by the desired 2A cutoff point.

Sure, I could buy a more modern Hall-effect sensor that could do all that for me. What's the fun in that?

Trouble is, I go searching for calculators for this kind of thing, and usually bump into formulas that include integrals, which make my eyes immediately glaze over - I'm not a calculus type of guy.

I'm thinking there has to be a relatively simple way to calculate Gauss from a coil (whether air or iron core) and then use the distance formula (btw, nobody has described the units in the searches I've done; distance squared vs field strength) in order to have a bias Gauss coil/electromagnet on one side of the sensor to keep it turned off, or have it turned on via a 2A current through a coil.

I hope this all makes sense. If you really want a schematic, say so.

 

I guess you could calculate the position of the sensor, or get experimental and use one of these critters - http://www.allegromicro.com/en/Products/Categories/Sensors/linear.asp

I might experimentally position the sensor and observe the output until some comfortable level came out. Then use a comparator to signal 2 amps sensed.

 

That's the thing; I could get some linear Hall-effect sensors and use a comparator, or using a switching Hall-effect sensor (like I have) and use the hysteresis of the sensor vs the intensity of the field as the comparator.

Since it's -175 Gauss to turn off, and +175 Gauss to turn on, I'd like to convert that to a range of around 200mA (ballpark) difference in the current, to give, say, a range of 1.9A to 2.1A.

I picked up some small ferrites the other day, both rod and toroidal form. Perhaps I'll just fiddle with it and see if I can come up with anything. I really would prefer to know how to calculate Gauss as a function of number of turns... the calculators I tried did not really seem to make sense.

This one does close to what I want:
http://www.calctool.org/CALC/phys/electromagnetism/solenoid
but the B-field calculation seems a bit strange. If I made an air core flat spiral on a PCB that was very thin (say, 1oz copper) the calculator indicates gauss would be far stronger than a wound core even 1/8" long.

 

That linear output sensor might function as a probe to see how various windings produce external fields.

By the way, that + and - 175 Gauss figure seems to indicate that the current has to reverse. I don't think it will show a unipolar 200 ma difference.

 

What I had intended to do was to have a coil on one side of the Hall-effect sensor with a static current flowing through it, providing the negative Gauss with an offset to turn of the sensor output; basically around -3,325 Gauss, which would correspond to -1.9A, or 3,500-175 Gauss.

When the current sensor coil hits 3,675 Gauss, the net effect on the Hall-effect sensor will be -3,325+3,675=175, causing it to turn on.

In this way, the Hall-effect sensor provides the hysteresis. The adjustment is provided by the spacing between the coils, and the Hall-effect sensors' placement between the two coils.

Does that not sound like a feasible concept?

 

As luck would have it, as I was rummaging around in my inductor box, I discovered that I had a set of four high-current inductors with cylindrical magnets glued on the top of them... and I can trigger the HES on and off with it. I think perhaps a bit of tweaking on the number of turns, or maybe a parallel current path via resistors, and I'll have a viable sensor with hysteresis.

 

You can run some experiments with all that.

I have yet to mess with Hall effect sensors. I saw an article about someone using fragments of magnets glued to cockroach nymphs to track motion with 3 sensors. They were sensitive enough that he could see breathing movements.

 

I hadn't really messed with Hall-effect sensors before either.

I'd bought about a dozen of these UGN3130's in Philmore packaging 10 years or so ago, and they got lost in a move. Recently ran across them again, and while designing an experimental H-bridge, decided they were the easiest and lowest-loss way to monitor current.

Getting way OT here, but I started trying to use a sense resistor on the bottom of the H-bridge, common to both sides, and switching the high side to limit current. Forgot about the current recirculation through the high-side MOSFET body diode when switched off. Switching the high side was mandatory due to the use of N-ch MOSFETs all around, and leakage in the high-side drivers.

So, that meant the sense resistor had to be moved to the high side. However, that also meant building a high-side supply for comparators, and I felt that things were getting a bit too Rube Goldberg-ish. If I had an "over-the-top" comparator sitting around, or at least one that had R-R inputs, I might've gone that route; but no such luck.

So that's when I hit on the Hall-effect sensor idea.

I'm going to be trying them out for stepper motor feedback as well. I have some NEMA 23 size motors, and a few years back I picked up some surplus Hall wheels for next to nothing that have 200 teeth on them and 1/4" hubs - perfect for what I'm doing.

 

Sounds like an interesting project. Hall effect for high side sensing sounds easier than trying to live with mega common mode voltage.

I would have been a coward and used a linear output sensor with a window comparator, but you have the goodies already.

Lost in a move? I put my stuff in safe places. They do turn up on occasion, but only after I have done a work-around.

 

Sounds like an interesting project. Hall effect for high side sensing sounds easier than trying to live with mega common mode voltage.

Well, that's why I abandoned that approach pretty quickly.

I would have been a coward and used a linear output sensor with a window comparator, but you have the goodies already.

After some more experimentation, I think I'm going to have to go with a HES that's more sensitive.

In order to get these HE's to switch, I had to use a 2lb holding force neodymium magnet and get it within about 3/8" of the sensor to turn it on, then back off to nearly 1/2" to get it to turn off.

I'd need a pretty good-sized coil with lots of turns to generate that kind of gauss, even if it were butted right up against the sensor.

Roughly 6,100 Gauss had no effect whatsoever; 9 turns, two layers of #20 magnet wire wound on a spool 175 mils long on a cylindrical iron ferrite core; est. permeability of 200 (unverified). I'll have to do some more fiddling around to figure out just what the perm is of that core material.

Lost in a move? I put my stuff in safe places. They do turn up on occasion, but only after I have done a work-around.

You don't know my wife. She's a tinkerer's bane. Her idea of "tidying up" is to put a bunch of unrelated stuff in boxes that are labeled something completely different, and jam it in a closet someplace. We moved about 7 years ago, and many things wound up stuffed in boxes that I still have no idea where they are.