Help with Carl's fluxgate Gradiometer ckt.

[John Smith]

[John Smith]

Re: relative range of detectors ?

[John Smith]

Re: relative range of detectors ?

[John Clayton]

Help with Carl's fluxgate Gradiometer ckt.

I was an ET in the Navy and have been an electronic technician for over 18 years.

I built the FMX-1 mag from a kit purchased from Fat Quarters software for my cousin who is a Land Surveyor here in Kentucky.

I only have a Fluke Digital Multimeter to troubleshoot this non-functioning device with, but here is what I have found:

1.) The sound at the speaker is not consistent each time I turn on the mag power and wait ten seconds. Sometimes the VCO threshold knob has an effect, sometimes not.

2.) I carefully re-checked for solder bridges and loose connections since, in my experience, intermittent operation of most devices is usually caused by a loose connection, none found in this case.

3.) After powering on the circuit, I measured the voltage at the output lead, referenced to ground, of both FGM-3 sensors. They were the same @ 3.53 and 3.50 VDC. I know this is a gate signal around 100Khz or so.

4.) I measure 5.01 VDC at the pins of both sensors. The capacitor polarity was verified, and the LC filter is in place at the sensors.

5.) I measured the voltage output from the DAC and it is usually 2.503 VDC when powered up. I believe that means the circuit is maxed out and is seeing what corresponds to max difference between the sensors, if everything were working correctly.

6.) I measured the voltage of each of the 8 bits, reference to ground, between the SCL007 and the DAC. Each of them are always 1.290 VDC, which should be a TTL zero according to the Analog Devices data sheet for the DAC I received with the kit. These bit voltages do not change as I move one of the FGM-3 sensors or bring a magnet near one of them.

That last finding tells me two things:
a.) The SCL007 is faulty ( output byte does not change with a change in one of it's inputs ).

b.) The DAC is faulty (the output voltage does not reflect the input data byte, i.e. MAX voltage out with all zeroes in; the DAC data sheet truth table does not reflect negative logic.)

I continued to check other items...

7.) I removed the lead, at the circuit board, from the output pin of one of the sensors. I checked the bits of the SCL007 digital output, none of the bits went "High". I then shorted pins 17 & 18 of the SCL007 to simulate both sensors giving the same exact input to the SCL007. There was no change at the bit-level on the output of that chip.

Now, I believe the input sensitivity of my DMM is 20,000 ohms/volt. It operates in the 2volt range for this measurement when checking the digital bits, so I do not believe my multimeter is loading down the output of the SCL007, or am I wrong in this?

Sometimes the output of the DAC goes to about .7, (point-seven), VDC and I get a concurrent slow ticking at the speaker then, but the 8 bits never change.

I first replaced the lead I removed from the one sensor, then removed the lead from the other sensor. The sign-bit out of the SCL007 stayed zero in both cases.

Does anyone have any ideas?
Should I call Fat Quarters and ask for a new DAC and SCL007?

Like I said, I have been in electronics for quite some time and know how to be cautious about ESD when handling components. When the parts arrived from Fat Quarters, both of the suspected chips were in ESD DIP shipping tubes.

The MC34119 audio amp, the LM393 comparator and the 2 transistors, however, were stuck in plain old styrofoam. None of those components seem to be malfunctioning, but I wonder about the way they handle chips when I saw that.

This is the way the circuit has always acted, it has never worked properly.

Any help or suggestions would be greatly appreciated.

John Clayton
[email protected]

[Jan]

Re: Help with Carl's fluxgate Gradiometer ckt.

John,

Carl will no doubt help you as well, but perhaps I can give some ideas.
I did not build this design but one that only uses the sensors and SCL007 part.

First determine if the SCL007 or the DAC is the problem.
Remove the AD557, LM393 and MC34119 so the SCL007 is the only one left connected.
Then measure pin 2 (Sign bit) on the SCL007.
It should change level when you move a magnet from one sensor to the other.
If not, there is something wrong with the sensors or SCL007.
If the sign bits seems to work measure pin 13 (MSB) on the SCL007.
When you move the sensors this bit should pulse

If the SCL007 seems to work put the AD557 back in the socket.
Pin 16 on AD557 should change between 0-2.5V

If this seem to work... continue with LM393 etc.

It is very tricky to try to find the fault without a Scope.
Perhaps one of these cheap scopes on a PC will do ?

A few points on the sensor.

Mount the sensors on a stable non-magnetic platform and make sure the sensors are aligned.
Do not even think about using them just lying on the workbench (the SCL007 will be max-ed out)

Make sure you are not working beside a strong magnetic field like a loudspeaker, these sensors are very sensitive and max out easily.
In my Mag a small slightly magnetic screwdriver is enough to max out the sensors.

Make sure that the connectors on the sensors point to the same direction, these sensors must be aligned in exactly the same way.
So the sensor connectors must point in one direction.


I hope this helps a bit.

Best regards, Jan.

[John Clayton]

Re: Help with Carl's fluxgate Gradiometer ckt.

>John,

>I did not build this design but one that only uses the sensors and SCL007 part.

>First determine if the SCL007 or the DAC is the problem.
>Remove the AD557, LM393 and MC34119 so the SCL007 is the only one left connected.
>Then measure pin 2 (Sign bit) on the SCL007.
>It should change level when you move a magnet from one sensor to the other.
>If not, there is something wrong with the sensors or SCL007.
>If the sign bits seems to work measure pin 13 (MSB) on the SCL007.
>When you move the sensors this bit should pulse

QUESTION: How fast is this pulsing? Is it something that would show on the 'analog' scale of my Fluke 77 Digital Multimeter?

>If the SCL007 seems to work put the AD557 back in the socket.
>Pin 16 on AD557 should change between 0-2.5V
QUESTION: Under what conditions should this change take place? I am finding it very difficult to get anything other than 2.5VDC. Perhaps if I removed both output leads from the sensors, then I would expect the voltage you mention to go to zero.... good idea, I'll try that.
>If this seem to work... continue with LM393 etc.

>It is very tricky to try to find the fault without a Scope.
>Perhaps one of these cheap scopes on a PC will do ?

>A few points on the sensor.

>Mount the sensors on a stable non-magnetic platform and make sure the sensors are aligned.

NOTE: I have them mounted in a very solid 1.25" Schedule 2 PVC pipe as mentioned in Chris' article. I followed his advice on hiding the alignment screws under the cap of the pipe. I used nylon screws. My application of his design is normally used in the vertical direction with one end of the pipe pointing down, close to the ground and the other end up towards the sky. The pipe is much shorter than Chris suggested. Mine is only 36" long, which should further reduce sensitivity, and that is desired. I only need to DETECT ferrous metals and their approximate location, not to accurately measure magnetic differences.

Both sensors are pointing in the same direction.

The adjustable sensor is mounted at the bottom of the pipe with it's leads pointing inwards towards the inside of the PVC pipe. That is the end closest to the ground during use. The other, fixed, sensor is mounted with it's leads pointing outwards, away from the inside of the pipe, so that both sensors (the ends without the leads) are facing in the same direction.

Would you agree that my test of disconnecting one sensor's output and shorting pins 17 & 18 of the SLC007 together should simulate a perfectly aligned PAIR of sensors and cause correct operation, (no audible output or zero VDC out of the DAC),?

>Do not even think about using them just lying on the workbench (the SCL007 will be max-ed out)
>
>Make sure you are not working beside a strong magnetic field like a loudspeaker, these sensors are very sensitive and max out easily.
>In my Mag a small slightly magnetic screwdriver is enough to max out the sensors.

>Make sure that the connectors on the sensors point to the same direction, these sensors must be aligned in exactly the same way.
>So the sensor connectors must point in one direction.

>
>I hope this helps a bit.

>Best regards, Jan.

Thank you for the help, Jan!

I got a call from Erich at Fat Quarters and after disscussing things with him, he decided to just send me the SCL007 and DAC chips. I must say that he was very, VERY courteous and understanding. GREAT customer service!

I will replace one at a time and see which is the problem.

You mentioned PC o'scopes; do you know of a good o'scope card I can run on my PC that is not too exspensive? I would think I can get by with a good 1Mhz scope card for now, or even 500 Khz.

Let me know if you do.

Thanks again!

John

[Carl]

Re: Help with Carl's fluxgate Gradiometer ckt.

Sorry I did not reply sooner, I'm on the road right now. Jan's suggestions are good ones. It sounds like the SCL output bits are not at the right level as they should be pretty close to either 0v or 5v.

It's terrific that Erich is sending some replacements to try. These types of circuits are difficilt to debug with only a DMM, but let me know if you still have problems and I'll be glad to help. Another builder sent me his board and sensors, which I got working in short order, and would be glad to do the same for you, no charge.

- Carl

[John Clayton]

Re: Help with Carl's fluxgate Gradiometer ckt.

Carl,

Yours is a gracious and kind offer. I hope to get some type of O'scope emmulator card for my PC soon. I used to use those when I worked in the Calibration Lab before and programmed in HP-IB, (now IEEE-48. I would love to be able to verify that the pulses were, in fact coming from the sensors as your article describes. Getting a DC reading from these sensors is very far from proving them good or bad, but it seems very unlikely that BOTH could be bad. So, based on that, and since the two DC readings are so nearly identical, I must assume the sensors are OK. I would expect a fast-changing, 100khz signal from a 5 volt source to read as I described on a DC voltmeter. Erich seems to believe that the loading effect of my voltmeter on the digital bits from the SCL chip are more in the order of 1 Mohm or greater with my Fluke meter. I haven't seen the spec. sheets for this meter in quite some time, but I believe he is correct.

Jan's suggestions are fantasic! I thought I had isolated the problem effectively with my own tests, but actually removing the chips and testing again is a great idea and in this 'modular' type circuit are perfectly applicable.

Thank you for responding.

I will post the outcome on your board to this thread.

Regards,
John

[Jan]

Re: Help with Carl's fluxgate Gradiometer ckt.

John,

The answers to your question are not realy relevant anymore, but a question is a question is a question.

Question 1

Pin 13 is the MSB bit, so the bit with the highest value (decimal 12.
If every thing works OK this bit comes on when the value of the SCL007 is 128 or higher.
The SCL007 will try to balance the two sensors in the first 10 seconds. This bit should be 0 after 10 seconds.
If you move a small magnetic field to one of the sensors the bit should go on and if you remove the magnet again the bit should go off.
Thios should be measurable with your voltmeter.

Question 2

The AD557 gives an output from 0-2.5 volts max, depending on the 8 bit value off its input.
00000000 = 0V and 11111111 = 2.5V
So depending what the SCL007 outputs you will see a voltage between 0 and 2.5 volts.

Question 3
If you connect pin 17 to pin 18 on the SCL007 and use only 1 sensor to drive these pins I would think the SCL007 should output a stable 0 to all its output pins (after about 10 seconds)

Question 4
The problem is that I live in Holland so I have only European sites.
Just to give you an indications have a look at

http://www.velleman.be/index.asp?lan=1

I found also a kit that uses the soundcard of the PC to act as a scope
It looks like it is just what you want.... for free... :-))))
Have a look at http://www.vego.nl/14/07/01/14_07_01.htm

To download klik on the blue text "Klik hier !" under the header "Down-loaden?"

The webpage is in Dutch but the software is in English... so that is exactly what we want.
It uses the soundcard as a 2 channel 15kHz scope.
Ok... not very high frequency but at least you can have a look at the SCL007

Because my own old Tektronics scope is on the blink, no triggering anymore, I a was also looking into these PC scopes. (specially the first URL is gave you)
But to be honest.... I think I will go for the Tektronics TDS2002... have to save for it first though..

[Robert]

Re: Help with Carl's fluxgate Gradiometer ckt.

John,
I may have missed it, but the one thing I never saw mentioned was the startup procedure you have to perform every time you turn the gradiometer on. In the first 10 seconds you must rotate the gradiometer through its full range. If this is not done properly, it can certainly cause your readings to be off.

Good Luck,
Robert

[Malapais]

Building the Fluxgate Magnetometer (by Carl Morelnd)

Hello guys,

I am bran new to the world of Fluxgate. I have a Dalton Proton Mag and although it is simple to use and has great sensitivity, It is a real energy hog! Have to carry around a heavy motorcycle battery and it doesn't last very long. I was impressed with the article I read about Carl Morelands Fluxgate Mag that operates on a 9 V battery and ordered the FGM-3 kit from Fat Quarters. I just finished the circut board this evening and am ready to go get my PVC parts for the frame. I have some ideas that may or may not be good. I am hoping someone that has built one of these units can give me some advice before I begin.


I want to mount the circut board inside the 1 1/2" PVC instead of putting a Box in the middle of two pieces of PVC. I think it will be less likely to flex during use that way. Dalton got his circut board inside the PVC and his board is larger than the one I got in the Fat Quarters Kit so it should work.

Where I am stuck right now is how to determine how far to put my detectors appart. I am interested in as much depth penetration as I can get. There is no info provided with the kit that explains the pros and cons of spacing between the detectors. I am hoping to be able to see down to a depth of 10 feet. Is that reasonable for this FGM-3 type of gradiometer and if so what spacing between the detectors will I need to achieve that?

I am also curious about what kind of power switch to use, Toggle switches have steel in them. What is recommended? Is a volume control necessary? I want max volume all the time, Can I use a fixed resistor instead of a variable pot? I shorted out the R-6 connection since I will only be using a speaker output.

Can anyone share with me their own experiences with one of these units. ie: How far apart are your detectors? At that spacing... what is the sensitivity? How many feet away can you detect the presence of something like a miner's pick, or a shovel?

The only experience I have had is with the Dalton Proton Mag. As an example, with that I can detect a 5 ft. piece of barbed wire 10 feet away, but it is so cumbersome that it is impracticle for using in rough terrain, and then there is that huge/heavy battery problem.

Thanks,

Malapais <Jim>

[Carl-NC]

I want to mount the circut board inside the 1 1/2" PVC instead of putting a Box in the middle of two pieces of PVC. I think it will be less likely to flex during use that way. Dalton got his circut board inside the PVC and his board is larger than the one I got in the Fat Quarters Kit so it should work.

You can do that, just makes assembly more difficult. I don't think it will affect flexing... mounting a control box requires 2 very small screw holes, plus about a 3/8-inch hole for sensor wiring.

Where I am stuck right now is how to determine how far to put my detectors appart. I am interested in as much depth penetration as I can get.

My prototype was 5-feet. I tried 10-feet using larger PVC, but even that flexed and caused false signals. That's one drawback with fluxgate vs. PPM... fluxgate sensors have to remain perfectly aligned. Someone told me (possibly on this forum) that they used aluminum angle stock to stiffen their tube.

There is no info provided with the kit that explains the pros and cons of spacing between the detectors. I am hoping to be able to see down to a depth of 10 feet. Is that reasonable for this FGM-3 type of gradiometer and if so what spacing between the detectors will I need to achieve that?

Depends on target size. My 5-foot tube would detect the car at 20-25 feet I believe.

I am also curious about what kind of power switch to use, Toggle switches have steel in them. What is recommended?

Since it's a differential mag, when you align the sensors it will null out any steel in the circuitry.

Is a volume control necessary? I want max volume all the time, Can I use a fixed resistor instead of a variable pot? I shorted out the R-6 connection since I will only be using a speaker output.

Fixed volume will work.

The only experience I have had is with the Dalton Proton Mag. As an example, with that I can detect a 5 ft. piece of barbed wire 10 feet away, but it is so cumbersome that it is impracticle for using in rough terrain, and then there is that huge/heavy battery problem.

The fluxgate won't be that sensitive. That's another drawback with fluxgate.

- Carl

[Unregistered]

PPM detection

With my Geometrics G-856 proton mag, I can detect a .40 Glock at 20 feet. That's why I haven't switched to a fluxgate. I like the idea of a constant output rather than a charge/wait senario but I need the sensitivity and I can't afford a Cesium :-(

[Malapais]

Completed My Project

Thank You very much for the itemized response Carl.

Based on your replies I decided not to re-invent the wheel and just build my Fluxgate Mag according to the instructions provided. I put the electronics in the radio shack box recommended (instead of putting them inside the PVC like I was thinking) and used a pot/switch combo for the On/Off and volume control.

The only modifications I made were the addition of an earphone jack beside the Vol control and a couple more PVC fittings at the detector ends of the pipe so I wouldn't have to shave the coupling down to fit inside the 1 1/2" PVC pipe.

I made the last wire connection to the detectors about 11 PM last night and by 11:15 I was out in the moonlight tracking water pipes in my yard! I have not alligned the adjustable detector yet (I needed daylight for that) but, I must have it pretty close because there was only a slight freq. change when I rotated the instrument.

I really like the constant output of the Fluxgate compared to the charge and pulse mode of the Proton Mag.

It is a very nice design Carl! I look forward to getting the detectors alligned today and getting to work on some sensitivity tests.

For those of you thinking about building Carl's Fluxgate Mag. from the Fat Quarters kit, I highly recommend it. The only problem I had through the entire project was that one of the leads broke off one of my 56uHy inductors that had to be installed at the detectors and I could not find an exact replacement locally. A quick call to Erich at Fat Quarters and I was pleased to learn that I could use 100uHy inductors (which were available locally) in place of the ones supplied with the kit and I was right back on track.

Thanks Carl and Erich, from my preliminary experience just playing with the instrument in the moonlight last night I can already tell that this is going to be a great addition to my Treasure Hunting aresonal of equipment.


Jim Hatt <Malapais>

[Carl-NC]

Glad you got it working! Yeah, the inductor value is not at all critical. Just a voltage supply filter.