Tweet
Jim,
I have had some rather interesting challenges with the magnetometer project
lately.
When I built the stand-alone unit, I used a lot of parts that I had on hand,
in particular a couple National LM2940 three terminal regulators that I had no
experience with. I installed them as per the National manual but now realize
that they were intermittently oscillating and occasionally shorting input to
output. This, I think, was the cause for the failure of the supervisor early
on. I was bringing the system up step by step and had no IC's installed;
however the supervisor was soldered down. I was using a bench supply that I
happened to have at 24v.The result was one fried supervisor.
The oscillations caused a lot of erratic readings that had me puzzled, the
machine would work OK for a hour or so, then start to give erratic readings, I
thought this was due to powerlines and the nearby radio station. I had hoped to
be able to give you some solid details about the paraffin sensor but did not
trust the numbers I was getting. I strongly suspect that the paraffin proton
source will not work. The numbers I was getting make me suspect that the uP was
measuring the resonant frequency of the toroid and the "bypass" capacitor as
they were far above any normal reading. Unless my back yard is directly over a
solid mass of iron! One never knows here in Kansas, our last home had several
lengths of old oil well casing buried close by!
Last weekend, the regulator really gave up with a sudden burst of power cycles.
By the time I realized what was going on and shut the thing off, everything
that had 5v going to it was dead, the silver mica capacitor I had across the
sensor had shorted and the LM394 in the preamp was open.
It's all fixed, no more LM2940's, instead old reliable LM317's, and a new cap
in the sensor. When you reach Arizona, you will find an order for an another
uP. Unfortunately; I am becoming too good a customer! I will use the repaired
stand-alone to test the amplifier assembles and sensors for the gradiometer.
I am in the planning state with the gradiometer, hope to start construction
around December. In particular, thinking over how I will build the digital
section. Either protoboard or wire wrap, I have not decided which yet. I will
be glad to share any problems that I come across. It is a surprise to learn
that I will most likely be the first to get one working, I thought I would be
one of the last. How long have you had the various articles on the net? I
estimate that I will have my gradiometer ready for turn on about February of
next year. A slight slow down due to the on line course I am taking. It has me
dusting off my calculus text! The course is very good, however, it strains the
Internet, if I look at it over my lunch, it will usually go through about 1/3
of the lecture, then lock up. The cure is heroic -I get up at 4:00 am and drive
down to the office and watch it. Makes me feel my age!
Some things I have learned with the stand-alone unit:
Your schematics have the squelch gate located with the intermediate amplifier.
I have built two different versions of your pre amp - squelch -intermediate
amp. using some cast aluminum IF strip enclosures that are mounted on the
protoboard, almost "plug-in". By chance, I built one squelch ckt. In the same
"chamber " as the preamp. This did not work; the result was a continuous
oscillation. The oscillation went away when I moved the squelch circuit into
the same area as the intermediate amp. As you showed.
My problems with the comparator oscillating were due mostly to a misswire and a
little bit due to poor lay-out. There is nothing wrong with your design, from
what I know of these things it should cause no problems if the builder does
his/her part.
I also built one assembly using the same LM394 for both amplifiers - I made it
work, but it took some doing to prevent oscillations.
If the shielded wires going from the relay to the preamp are not grounded at
the relay but left open, the preamp will oscillate (at least both mine did! I
opened the grounds to see if I could get away with one side open.)
I found the addition of a buffer made from one of the unused LM2904's op amps,
used to drive a headphone, was most useful to get an idea of what was going on,
especially when one is away from AC power for most test equipment..
I will build some sort of circuit board to mount the relays and associated
wires for the gradiometer. With the stand-alone unit, I soldered the various
wires directly to the relay - this does not take much abuse.
I mounted the relay in a home made can made of mu-metal. Using a current probe
as a sensor, there is a noticeable weakling of the field created by the relay
as compared to one with no shield. I have always heard that bending mu-metal
destroys it's shielding, but this seems to fly in the face of that. By the way,
I have a good collection of fairly small pieces of mu-metal and a couple Radar
and 'scope shields, should you have any call for this material, let me know, I
will never use it all and would be glad to give some of it away.
Your plans for circuit boards should eliminate most of these sorts of problems.
I suspect there are not many "out there" who build from scratch. Most of the
hobby articles I have seen usually have a PC board and the parts for sale, a
sort of modern HeathKit. Your article/ plans seem to me to be much like what I
would expect to find in QST.
I read your e-mail about the new sensor design. I am looking forward to your
revised spreadsheet, as this is one area of the gradiometer that I was still
considering. It would be nice to come up with a sensor that would fit inside
readily available PVC pipe. This is similar to the one I made from an old wire
spool; it's a shame I never got to the point where I could make s/n
measurements. I do believe this would be quite a challenge! I will do so once I
get the thing running again.
A couple problems I discovered way back on my first attempt:
If one attempts to use PVC pipe and end caps, the caps project over the
sides of the outer pipe - you end up with a lot of airspace. I tried to find
some flat PVC to make my own end caps of and was not successful at that time.
I also thought about cementing two end caps together then drilling a 1 or 2
inch hole through the two to fit a length of pipe and thus create a "toroid"
but gave it up. I had several articles that claimed that one liter was the
ideal volume, this scheme gave far less, probably about twice what there is in
one of those rock-a stack toys!
My eventual sensor was made of the wire spool with a Plexiglas "collar". I was
able to join the two with cement made from pieces of Plexiglas dissolved in
acetone. A major disadvantage is the relatively small inside "hole", 1in dia.
The final measurements are outside dia. of 4 in., inside dia. Of 1 in. and
height of 5 inches. I had one attempt leak from the stress caused by the
winding of the wires. The second was successful.
I found another article of yours on the net; it is the schematic of a more
sophisticated magnetometer that uses digital filtering. What is the status of
this?
About building circuit boards- sure wish I could be of help but I am still at
the tape and Mylar state. My amplifiers are laid out pretty much like the
schematic and work well.
This sort of thing has created an entire department at my company's main
facility, they use a very sopisicated computer program and the last I heard
were up to 10 layer boards.
Trivia:
Your comment about building boards for rocket use reminds me of yet another
Army story. There is a remote chance we "crossed paths" some time past. I
worked on the Redstone and Pershing missiles and participated in seven Redstone
firings at the White Sands test range. They were launched from a place known as
Oro Grande Range Camp. The Army, always trying for good public relations,
invited various schools to add high altitude experiments to these shots. One
had "stuff" from five different schools, two of which were Canadian. (The
Redstone had lots of empty space between the instrument package and the fuel
tank, about 4 ft dia. X 5 ft long). Who ever did the weight and balance forgot
about the additional weight. The result was when fired the thing rose about a
foot or so and stayed there until it burned off fuel. This was early in the
morning just as the daily dust storm blew in. The storm arrived early - the
Autopilot was really working to keep the thing upright and yet correct for the
wind, we could see the missile vibrating severely and start to drift down wind.
That area is covered with sand dunes held in place with a sage bush or two,
just high enough to catch the tail of the rocket. We were sure either the
safety officer or the dunes were going to destroy the thing and promptly left.
I am sure I ran the first two-minute mile. Neither occurred, the missile
cleared the dunes and went on a short flight of 50 miles. This would have been
in the mid 1960's, most of the details have gone due to my poor memory!
I hope the trip from Canada to Arizona goes well, I think it would be a nice
scenic, peaceful drive. (If you drive it).
All for now, pardon all the rambling
GwBower
--
On Fri, 13 Oct 2000 20:57:37 Jim Koehler wrote:
>The Proton Mag Forum
>
>Hi everyone.
>
> I've recently been corresponding with a fellow in France who is interested
in making a large toroidal sensor. We've been kicking ideas around and ha made
some suggestions which got me thinking of another somewhat toroidal shape for a
proton sensor which might even be better than a simple toroid. It is the shape
shown in the attached file. It has all the advantages of the toroid - good
noise rejection, insensitivity to orientation with respect to the magnetic
field. However, it is also easy to make .... AND, it gives a higher S/N!
>
> You could make a core like this by using two different diameters of
plastic tubing. Glue end plates on the two concentric sections of tubing (and
cut out the centres of each end, of course) and you have it. I'm quite excited
about this because I have done the theory for this shape and it looks like you
can get very high S/N - much higher than for a simple toroid (one with circular
cross-section) of the same radius. That is because S/N depends on the volume
of the fluid and it will obviously be much higher for this shape than for a
similarly sized toroid with circular cross-section.
>
> I am revising my spreadsheet to incorporate this shape and expect to have
this done in a week or so. I'm on my way down to Arizona for the winter next
week and will probably not get the revised version onto my Web site before I
get there.
>
> Jim
>
>//////////////
>
> Jim Koehler
> RR2, Site 292, C56
> Courtenay, BC, V9N-5M9
> Canada
>
> Email: [email protected]
>
>//////////////
>
>
>
>
>_________________________________________________ _____________________
I have had some rather interesting challenges with the magnetometer project
lately.
When I built the stand-alone unit, I used a lot of parts that I had on hand,
in particular a couple National LM2940 three terminal regulators that I had no
experience with. I installed them as per the National manual but now realize
that they were intermittently oscillating and occasionally shorting input to
output. This, I think, was the cause for the failure of the supervisor early
on. I was bringing the system up step by step and had no IC's installed;
however the supervisor was soldered down. I was using a bench supply that I
happened to have at 24v.The result was one fried supervisor.
The oscillations caused a lot of erratic readings that had me puzzled, the
machine would work OK for a hour or so, then start to give erratic readings, I
thought this was due to powerlines and the nearby radio station. I had hoped to
be able to give you some solid details about the paraffin sensor but did not
trust the numbers I was getting. I strongly suspect that the paraffin proton
source will not work. The numbers I was getting make me suspect that the uP was
measuring the resonant frequency of the toroid and the "bypass" capacitor as
they were far above any normal reading. Unless my back yard is directly over a
solid mass of iron! One never knows here in Kansas, our last home had several
lengths of old oil well casing buried close by!
Last weekend, the regulator really gave up with a sudden burst of power cycles.
By the time I realized what was going on and shut the thing off, everything
that had 5v going to it was dead, the silver mica capacitor I had across the
sensor had shorted and the LM394 in the preamp was open.
It's all fixed, no more LM2940's, instead old reliable LM317's, and a new cap
in the sensor. When you reach Arizona, you will find an order for an another
uP. Unfortunately; I am becoming too good a customer! I will use the repaired
stand-alone to test the amplifier assembles and sensors for the gradiometer.
I am in the planning state with the gradiometer, hope to start construction
around December. In particular, thinking over how I will build the digital
section. Either protoboard or wire wrap, I have not decided which yet. I will
be glad to share any problems that I come across. It is a surprise to learn
that I will most likely be the first to get one working, I thought I would be
one of the last. How long have you had the various articles on the net? I
estimate that I will have my gradiometer ready for turn on about February of
next year. A slight slow down due to the on line course I am taking. It has me
dusting off my calculus text! The course is very good, however, it strains the
Internet, if I look at it over my lunch, it will usually go through about 1/3
of the lecture, then lock up. The cure is heroic -I get up at 4:00 am and drive
down to the office and watch it. Makes me feel my age!
Some things I have learned with the stand-alone unit:
Your schematics have the squelch gate located with the intermediate amplifier.
I have built two different versions of your pre amp - squelch -intermediate
amp. using some cast aluminum IF strip enclosures that are mounted on the
protoboard, almost "plug-in". By chance, I built one squelch ckt. In the same
"chamber " as the preamp. This did not work; the result was a continuous
oscillation. The oscillation went away when I moved the squelch circuit into
the same area as the intermediate amp. As you showed.
My problems with the comparator oscillating were due mostly to a misswire and a
little bit due to poor lay-out. There is nothing wrong with your design, from
what I know of these things it should cause no problems if the builder does
his/her part.
I also built one assembly using the same LM394 for both amplifiers - I made it
work, but it took some doing to prevent oscillations.
If the shielded wires going from the relay to the preamp are not grounded at
the relay but left open, the preamp will oscillate (at least both mine did! I
opened the grounds to see if I could get away with one side open.)
I found the addition of a buffer made from one of the unused LM2904's op amps,
used to drive a headphone, was most useful to get an idea of what was going on,
especially when one is away from AC power for most test equipment..
I will build some sort of circuit board to mount the relays and associated
wires for the gradiometer. With the stand-alone unit, I soldered the various
wires directly to the relay - this does not take much abuse.
I mounted the relay in a home made can made of mu-metal. Using a current probe
as a sensor, there is a noticeable weakling of the field created by the relay
as compared to one with no shield. I have always heard that bending mu-metal
destroys it's shielding, but this seems to fly in the face of that. By the way,
I have a good collection of fairly small pieces of mu-metal and a couple Radar
and 'scope shields, should you have any call for this material, let me know, I
will never use it all and would be glad to give some of it away.
Your plans for circuit boards should eliminate most of these sorts of problems.
I suspect there are not many "out there" who build from scratch. Most of the
hobby articles I have seen usually have a PC board and the parts for sale, a
sort of modern HeathKit. Your article/ plans seem to me to be much like what I
would expect to find in QST.
I read your e-mail about the new sensor design. I am looking forward to your
revised spreadsheet, as this is one area of the gradiometer that I was still
considering. It would be nice to come up with a sensor that would fit inside
readily available PVC pipe. This is similar to the one I made from an old wire
spool; it's a shame I never got to the point where I could make s/n
measurements. I do believe this would be quite a challenge! I will do so once I
get the thing running again.
A couple problems I discovered way back on my first attempt:
If one attempts to use PVC pipe and end caps, the caps project over the
sides of the outer pipe - you end up with a lot of airspace. I tried to find
some flat PVC to make my own end caps of and was not successful at that time.
I also thought about cementing two end caps together then drilling a 1 or 2
inch hole through the two to fit a length of pipe and thus create a "toroid"
but gave it up. I had several articles that claimed that one liter was the
ideal volume, this scheme gave far less, probably about twice what there is in
one of those rock-a stack toys!
My eventual sensor was made of the wire spool with a Plexiglas "collar". I was
able to join the two with cement made from pieces of Plexiglas dissolved in
acetone. A major disadvantage is the relatively small inside "hole", 1in dia.
The final measurements are outside dia. of 4 in., inside dia. Of 1 in. and
height of 5 inches. I had one attempt leak from the stress caused by the
winding of the wires. The second was successful.
I found another article of yours on the net; it is the schematic of a more
sophisticated magnetometer that uses digital filtering. What is the status of
this?
About building circuit boards- sure wish I could be of help but I am still at
the tape and Mylar state. My amplifiers are laid out pretty much like the
schematic and work well.
This sort of thing has created an entire department at my company's main
facility, they use a very sopisicated computer program and the last I heard
were up to 10 layer boards.
Trivia:
Your comment about building boards for rocket use reminds me of yet another
Army story. There is a remote chance we "crossed paths" some time past. I
worked on the Redstone and Pershing missiles and participated in seven Redstone
firings at the White Sands test range. They were launched from a place known as
Oro Grande Range Camp. The Army, always trying for good public relations,
invited various schools to add high altitude experiments to these shots. One
had "stuff" from five different schools, two of which were Canadian. (The
Redstone had lots of empty space between the instrument package and the fuel
tank, about 4 ft dia. X 5 ft long). Who ever did the weight and balance forgot
about the additional weight. The result was when fired the thing rose about a
foot or so and stayed there until it burned off fuel. This was early in the
morning just as the daily dust storm blew in. The storm arrived early - the
Autopilot was really working to keep the thing upright and yet correct for the
wind, we could see the missile vibrating severely and start to drift down wind.
That area is covered with sand dunes held in place with a sage bush or two,
just high enough to catch the tail of the rocket. We were sure either the
safety officer or the dunes were going to destroy the thing and promptly left.
I am sure I ran the first two-minute mile. Neither occurred, the missile
cleared the dunes and went on a short flight of 50 miles. This would have been
in the mid 1960's, most of the details have gone due to my poor memory!
I hope the trip from Canada to Arizona goes well, I think it would be a nice
scenic, peaceful drive. (If you drive it).
All for now, pardon all the rambling
GwBower
--
On Fri, 13 Oct 2000 20:57:37 Jim Koehler wrote:
>The Proton Mag Forum
>
>Hi everyone.
>
> I've recently been corresponding with a fellow in France who is interested
in making a large toroidal sensor. We've been kicking ideas around and ha made
some suggestions which got me thinking of another somewhat toroidal shape for a
proton sensor which might even be better than a simple toroid. It is the shape
shown in the attached file. It has all the advantages of the toroid - good
noise rejection, insensitivity to orientation with respect to the magnetic
field. However, it is also easy to make .... AND, it gives a higher S/N!
>
> You could make a core like this by using two different diameters of
plastic tubing. Glue end plates on the two concentric sections of tubing (and
cut out the centres of each end, of course) and you have it. I'm quite excited
about this because I have done the theory for this shape and it looks like you
can get very high S/N - much higher than for a simple toroid (one with circular
cross-section) of the same radius. That is because S/N depends on the volume
of the fluid and it will obviously be much higher for this shape than for a
similarly sized toroid with circular cross-section.
>
> I am revising my spreadsheet to incorporate this shape and expect to have
this done in a week or so. I'm on my way down to Arizona for the winter next
week and will probably not get the revised version onto my Web site before I
get there.
>
> Jim
>
>//////////////
>
> Jim Koehler
> RR2, Site 292, C56
> Courtenay, BC, V9N-5M9
> Canada
>
> Email: [email protected]
>
>//////////////
>
>
>
>
>_________________________________________________ _____________________