[Erik M]

Compadres!

Regardless if using magnatraction or not you will handle magnets.

Which mean using a Gauss Reader of some kind.

The most basic one would be one hall effect sensor, four penlight batteries and a multimeter.
The sensor give a volt reading from the middle of about 5V supplied.
Ie with no flux (which I will call it, regardless if this is "correct") the reading 2.5V, then up towards 5V for one pole and down towards 0V for the other one.
Problem is... When reading magnatraction there's fractions of millimetres and sub-millivolts to read. And they depend on where that middle voltage is. So a reading on your voltmeter of 3.495 might be the same as 3.792 on your friends.

So I have had a circuit laying around for some five years, newer gone to more than a bunch of PCB's etched and the one prototype.
But now I got at it again. A variable gauss meter (with dynamic(?) static(?) relativity).
Variable in three ways; south vs north, high sensitivity, low sensitivity.
It shows what pole is dominant and can be adjusted for intervals suitable for you current need.

And the schematic (given under the principle of GNU licensing) would look like this:
(nB! This is not a working schematic, values and gates are not tested for function etc.)

Uploaded with ImageShack.us

 

[marctownsend]

Thanks for sharing that erik, do you actually know the values for the components or where they can be found? Looks potentially very interesting.

 

[RichD]

Some years ago I saw an article on building a gauss meter using a Hall sensor, a voltage regulator and a volt meter. If you only wanted to read the relative magnet strength you could use a cheap uncalibrated sensor. If you wanted actual gauss readings you would have to sprimg for a far more expensive calibrated sensor. I already had a good gauss meter, but I did save screen grabs of the article if anyone is interested.

 

[Spiritlodger]

Rich, is this the article you are referring to Build your own Gaussmeter : I have constructed one of these, and used it to good effect matching magnets for slot cars.

The circuit I prefer for building a hall effect based gaussmeter was published in "Everyday Practical Electronics, Volume 28, No.12, December 1999 - Magnetic Field Detector, by Robert Penfold". Although this uses an analog meter, I find it is easier to use. The article also has comprehensive build instrutions, and a veroboard layout which makes building one very easy. This listing on ebay is for the correct issue: Ebay Item number: 110710050423

BTW ... is that Erik M from Sweden ?

 

[Spiritlodger]

I had a little rummage on the internerd ... found this: TheBackShed.com - GaussMeter

 

[Spiritlodger]

The ultimate solution would be to use a microcontroller hooked up to an lcd or similar display. One of the guys at DHORC has one of these ... very cool.

The following article shows how this can be done: Using Hall Effect Sensors with the Arduino

Here is another take on this theme, this time from an electric guitar forum: Gaussmeter project

 

[Erik M]

There's a problem using the hall effect sensor / voltmeter, you don't get readings where they are needed. You have to be able to calibrate where to gauge what you are looking at.

If you only want to measure a magnet's strength the above works perfectly fine, as does the VGR above.

If you want to measure, what would be the area of interest around here, magnet versus magnet versus placement in chassis then you have to find a way to calibrate the output.

And the used hall effect sensor in this case calibrates itself. And has a very nice resolution at that. At 5mV/G I can detect south/north of the ordinary Ø8x5mm neodym from over ten centimetres away. And that's the least sensitive of that range.

And it's, basically, microcontroller hooked-up...

That is, I can most probably find an IC, as in computer chip, that is more compact than using three LM339's as I do here. But I'm comfortable with these, more complex IC's will come later on for sure. But not right now.

Price of components of the above?
Shopping list:
[qoute]
1,54 1 1,93 SEK £0.18 0,22 € L-7104SRC-D LED hyper red 3mm
7,59 1 9,49 SEK £0.90 1,08 € EL204-15/T2C3 LED 3mm whit
17,56 1 21,95 SEK £2.08 2,49 € A1326LUA-T Hall effect sensor SIP3
14,55 1 18,19 SEK £1.73 2,06 € KB-2785YW LED light bar yellow 10x20mm
1,54 1 1,93 SEK £0.18 0,22 € SIL-net, 10 kOhm ±2 %
2,05 2 5,12 SEK £0.49 0,58 € SIL-net, 220 Ohm ±2 %
4,15 3 15,56 SEK £1.48 1,77 € LM339N comparator DIL14
29,29 3 110,- SEK £10.43 12,46 € Rotary switch HEX FR01-KR16P-ST-075A
0,42 4 2,11 SEK £0.20 0,24 € BC557 Transistor PNP TO92 45V 0,1A
0,19 22 5,14 SEK £0.49 0,58 € Metal film resistor 10 kOhm 0,4W 1%
~Sum: 192 SEK £19 22 €
[/quote]
On top of that PCB etc.
I could most probably cobble together a DYI kit with a dedicated PCB for perhaps 20% above that.
The main cost, as can be seen, are the three rotary HEX switches. But to get variable read intervals that are the same from VGR to VGR and setting to setting and enough of them too they are needed.
Using only one, to push apart south and north, ie narrowing the top interval, could work however.
With two you can adjust where you want the interval. And as I read this myself I come to realise two is probably enough.
Each setting [1 ... 16] will have full documentation.
So the material cost would end up at 155SEK, £16 and 18€.
With enough interested, add 20% and a dedicated PCB and it could be something worth doing.

 

[Erik M]

Hm, I'll try to explain this better. But first I got to walk the dog. And rest.

Meanwhile at another establishment, namely at The Little Thief and over an excellent "English Breakfast"...

Indeed I am Roy. I even felt preeetty sure that was you under that nick.
Thanks for remembering me that fondly. At least I hope it is fondly!

But the reference is buried and no shovel in sight (including the one in your garden shed, it went astral last night...).

This picture from my "not quuuiiite graveyard" would satisfy any doubts:
Especially note what might be discernible on the rear bumbers of the left Marcos.
The #22 car is another sure give away.

I can still see before me a seriously lit up face when given a mug o'JAVA at five(?) in the morning. Don't remember exactly who it was, but the girlfriend was a sight for sore eyes for sure.

Uploaded with ImageShack.us

 

[Erik M]

marctownsend ~ I totally misread you. Value as in ohm and such, not value as in phenning.

I'll come back about that part. Together with the correct schematic.

 

[RichD]

Spiritlodger, yes that is the article. Since things like that don't always stick around I did screen grabs of the pages and saved those. Since the site still exists I bookmarked it this time. I have had a good gaussmeter (magnetometer) for some time and I have found that getting meaningful readings of magnets is not always straight forward. Magnets to not have symmetrical fields and those fields can be influenced by nearby magnets and by other materials. Different people can get different readings from the same magnet even if they use the same meter. If you are able to get the magnet by itself and fish around until you find the highest value you would have a number that might or might not reflect what the magnet would do in a slot car.

 

[Erik M]

You're definitely nailing a couple of the problems using a gauss-reader RichD.
And the mentioned and shown examples are as good as they are. And if one can work within these various limitations it is a good tool.

What I'm trying to accomplish is a bit more.
If you hook a gauss sensor to a voltmeter you'll notice that it "bottoms out" and that it thereby is really hard to get any numbers worth trying to use.

Yes, you can read a magnet, as it is. Especially if you grab your knowledge in geometry and physics. Ie a magnet has the same strength all around it (in broad terms gentlemen, in broad terms) and it's spread out so that it is the same area vs flux rating all over.
And this mean we got a dependable algorithm to think from.

The magnetic field is (more or less) a sphere. and has the same strength all around. But spread thin.
Let's say we got a reading of 3V at 25mm (with no reading voltage is 2.5V) and the sensitivity is 5mV/G.
3-2.5=0.5 » 0.5/0.005 » 100 "gauss". If we are anywhere near knowing anything this reading of 100 is spread out over (4×25²×pi~) 7'900mm² and a second reading at 15mm should then be (100×7'900÷2'800{=4×15²×pi}=) 282 "gauss per area of".

With a Ø8×5mm magnet this further ought to mean it has (100×7'900÷101{=2×4²×pi}=) 7'900G.
And here there's taken that a magnet's force start from the areas of poles (hence "2×" above) and "internal" magnetism is dis-regarded.
Problem is we can't do that since we know that a Ø8×5mm is stronger than a Ø8×3mm. The easiest way is to just add the length, as ×L and thus go from "force by area" to "force by volume".
As can be seen one can look at this in multiple ways, and that is also part of the understanding of magnetic fields, they're quite illusive and hard to define.

Anyhow...

For the purpose of slotcars and racing with them we don't really need the G of a magnet.
What we need to know is how two magnets compare and also how much flux we got at a certain point and thus the needed change in distance to the magnet.

And by this I mean that we...
One) ...want to match magnets to make better working motors.
Two) ...want to know how much magnatraction is generated from a magnet (one magnet at a time only).

In both cases we don't need to know any specific number, we need to be able to read high/low and some steps around it.
But as the magnets we use vary so widely in flux (ferrite vs neo) which leads to extreme distances to read the same interval between them we need to be able to calibrate our readers.
So we need to be able to set the sensitivity. It also helps if we can define what to look for. One for pairing to see if two magnets match and one for magnatraction to see what force we get from the same magnet at different height, and then place another magnet at a height that will give the same force.

Yes, it is the same in a way. But to be an effective tool we need to be able to set interval and resolution. And to be able to take a reading from one reader to another we need to have distinct settings.
And as I read this myself I come to wonder if just one rotary is actually enough. It might be. And that would be so much less expensive.

Am I thinking straight?

 

[Erik M]

To much text to press through I guess...?

Anyhow. What it boils down to is to make the (basically) exponential change in flux into a linear reading.
That way you can get a probable guess at Gauss for what shim you need to use to get within race specs etc.

And the same for when measuring magnets towards matching. One coarse to group them and then a finer to match them.

For magnatraction reading the group of racers can then decide on a specific plate to measure with (I got a copper-free fibreglass PCB for this). One spec of plate. 'Cause if you have to add shims etc to get the reader at the right distance you open up for confusion.
But if you can say "(MCR) VGR, (setting) 14, 6 (lights tops)" everyone would know what spec is in effect.