Epe pic magnetometer

[Samy]

NEW DATELOGGER SD CART

Hello
here is a new data logger with Sd Card.

http://www.fmk-braunschweig.de/

is unfortunately in German
may be one or the other to rebuild it.
Projekt Datenlogger II
http://www.fmk-braunschweig.de/Proje...lOGGER_v26.zip

Datenlogger ist eine Weiterentwicklung des ersten Types. Es wurde eine andere
Druckzelle (MS5540B) verwendet. Dadurch besteht die Möglichkeit den barometrischen
Druck und die Temperatur auszulesen. Außerdem hat sich dadurch auch der
Stromverbrauch reduziert. Das Gerät könnte jetzt auch z.B. als Wetterstation eingesetzt
werden.
Der Datenlogger ist ein batteriebetriebenes Messgerät, mit dem 13 unterschiedliche
Messwerte über einen gewissen Zeitraum eingelesen werden können. Die eingelesenen
Messdaten werden auf einer handelsüblichen SD-Karte gesichert und können alternativ
auch über die serielle Schnittstelle ausgelesen werden. Die Auswertung der Messdaten
kann dann komfortabel auf einen PC mit Kartenleser erfolgen. Für den PC gibt es auch ein
grafisches Auswerteprogramm.
Das Projekt 'Datenlogger' ist für Modellflieger und andere Interessenten gedacht, die
elektronische Kenntnisse besitzen. Außerdem sollten sie schon einmal Erfahrung mit
SMD-Bauteilen gemacht haben. Sollten diese Punkte nicht zutreffen, so bleibt als letzter
Ausweg ein Bekannter der diese Kenntnisse besitzt.
Die Steuerung des Datenloggers übernimmt ein Mikrocontroller vom Typ AVR AtMega16.
Der Mikrocontroller wird direkt von der Batteriespannung versorgt. Des weiteren ist eine
Echtzeituhr vom Typ X1226 vorhanden und ein Absolut-Drucksensor vom Typ MS5540B.
Die Messdaten des Datenloggers können auf einer SD-Karte gesichert werden. Ein
Anschluss für eine serielle Schnittstelle ermöglicht, die Daten direkt zu betrachten.
Außerdem können über diese Schnittstelle Befehle gesendet und empfangen werden.
Informationen über diese Befehle können dem ASM-Programm entnommen werden.
Es werden 13 Messwerte verarbeitet. Das Datum und die Uhrzeit werden vom Echtzeit-
Uhren-IC X1226 erzeugt. Diese Messwerte werden zur genauen Bestimmung der
Messung benötigt. Ein Messwert zeigt die Batteriespannung der externen Batterie und
somit auch der Versorgungsspannung des Mikrocontrollers an. Die Messwerte für den
Servo-Impuls und der Servospannung bzw. Empfängerspannung werden über einen
Eingang, der mit einen Servokabel ausgerüstet ist, eingelesen. Über den Impulseingang
eines 3pol. Steckers können externe Impulse eingelesen werden. Hier kann z.B. ein
externer Drehzahlmesser angeschlossen werden. Die Flughöhe misst ein Absolut-
Drucksensor vom Typ MS5540B. Die Referenzhöhe wird vor jeder Messung gesichert.
Des weiteren können noch vier externe Spannungen im Bereich von 0 – 2Volt ausgewertet
werden. Der Drucksensor stellt auch noch den absoluten barometrischen Druck und die
Temperatur zur Verfügung.
Der Datenlogger ist ausgelegt für den Betrieb mit einer LiPo-Zelle. Das heißt, er muss in
einem Spannungsbereich vom 2,7 – 4,2 Volt arbeiten. Die untere Spannungsgrenze wird
durch Software überwacht und der Datenlogger schaltet unter 3 Volt ab, da sonst die
Backup-Batterie der Echtzeituhr die Versorgung des Uhren-IC's übernehmen würde. Die
vom LD-Spannungsregler erzeugten 3,3 Volt versorgen alle Bauteile auf der Platine.
Die auf der SD-Karte gesicherten Dateien sind so formatiert, dass sie ohne Probleme als
CSV-Textdatei in eine Tabellenkalkulation eingelesen werden können. Dort können die
Daten weiter verarbeitet werden.
Jede Messung erzeugt eine eigene Datei mit fortlaufender Nummer. Die erste Datei hat
den Namen „DATA0000.TXT“ die zweite den Namen „DATA0001.TXT usw. Die Dateien
werden immer ins Hauptverzeichnis (Ordner) der SD-Karte geschrieben.
Unterverzeichnisse (Unterordner) können mit dem Datenlogger nicht erzeugt werden. Das
Löschen der Dateien und das Formatieren der SD-Karte muss auf dem PC erfolgen.
Mit dem Taster auf der Stirnseite kann der Datenlogger eingeschaltet werden. Dabei
leuchtet die LED ca. 1 Sekunde auf. Wenn der Datenlogger im Betrieb ist, kann ein
Tastendruck die Messung starten. Die LED leuchtet dabei 2x auf. Ist die Messung
eingeschaltet, so kann die Messung durch einen weiteren Tastendruck beendet werden.
Auch hierbei leuchtet die LED 2x auf. Zum Ausschalten des Datenloggers muss die Taste
mehr als 5 Sekunden bei ausgeschalteter Messung gedrückt werden. Danach leuchtet die
LED wieder 1 Sekunde auf und der Datenlogger schaltet sich aus. Diese
Ausschaltfunktion wird auch bei Unterspannung der Batterie eingeleitet.
Wird die SD-Karte eingesteckt oder herausgezogen, so wird diese Funktion auch durch
ein kurzes Leuchten der LED angezeigt. Ist eine Messung gestartet, leuchtet die LED
durch einen kurzen Lichtblitz bei jedem Messintervall auf. So kann die korrekte Funktion
des Datenloggers überprüft werden.
Für die Betreiber von Modellraketen ist noch eine Höhenauslösung implementiert. Der
Logger misst die maximale erreichte Höhe und löst beim Unterschreiten der Höhe ein
Signal aus.
Mit den Befehlen der seriellen Schnittstelle können verschiedene Parameter gesetzt
werden. Zum Beispiel kann das Datum und die Uhrzeit gesetzt und gelesen werden. Der
Messintervall kann zwischen 1 und 6500 Sekunden eingestellt und die Messung gestartet
oder gestoppt werden. Der Befehl 'INFO' gibt alle gültigen Befehle seriell aus. Bei der
Übergabe von Parametern muss sich zwischen dem Befehl und den Parametern ein
SPACE befinden.
Der Aufbau erfolgt auf einer doppelseitigen durchkontaktierten Platine. Eine selbst geätzte
Platine ist auch möglich, da die Durchkontaktierungen so gelegt wurden, dass sie kein
Bauteil behindern. Falls noch doppelseitige Platinen vorhanden sind, können diese auch
beim Autor bezogen werden.
Das Schaltbild, das Platinen-Layout, die Stückliste und das aktuelle kompilierte
Assembler-Programm sind in der 'WB_LOGGER2_xx.ZIP'-Datei enthalten. Zur
Programmierung des AVR-AtMega16L gibt es im Internet genügend Informationen.
Technische Daten:
Platinengröße : 60mm Länge, 33mm Breite
Gewicht : 18 Gramm
Spannungsversorgung : 3 Volt – 6 Volt
Stromaufnahme : ca. 8 mA
Mikrocontroller AtMega16L ADC, Impulse
Uhrenmodul X1226 mit Backup-Batterie (1620), Datum und Uhrzeit
Drucksensor MS5540B Höhe von 0 - 6000m
Druck, Temperatur
Anschlüsse:
2pol. Batterieanschluss, 3 – 6 Volt
3pol. Für RC-Impulsbreite und Empfängerspannung 0 – 6 Volt
4pol. Molex-Stecker Ser. Schnittstelle, 38400 Bd, 8N1
Zusätzlich Versorgungsspannung
3pol. Molex-Stecker Impulseingang, 0 – 65535 Impulse/s
Zusätzlich Versorgungsspannung
Bei aktivierter Höhenauslösung : Schaltausgang
SD-Stecker Steckplatz für eine SD-Karte
6pol. (2,56mm-Raster) Stecker für 4 weitere Analogeingänge von 0 – 2,0 Volt
Zusätzlich Versorgungsspannung
6pol. (2mm-Raster) Programmierstecker für den Mikrocontroller
Taster Zur Steuerung des Datenloggers
2x LED Funktionsanzeige
Messstellen: Datum TT:MM:JJ
Zeit SS.MM.SS
Batteriespannung 0 - 4,096 Volt
RC-Impuls 80 - 248 0= Kein Impuls
RC-Spannung 0 - 6,144 Volt
Höhe 0 - 6000m
Impulse 0 - 65536 I/s ( U/min )
Spannung Kanal1 0 - 2,048 Volt
Spannung Kanal2 0 - 2,048 Volt
Spannung Kanal3 0 - 2,048 Volt ( 0 – 32 Volt )
Spannung Kanal4 0 - 2,048 Volt ( 0 – 80 Amp. )
Barometrischer Druck

Temperatur

[Samy]

Hello
here is a new data logger with Sd Card.

http://www.fmk-braunschweig.de/

[stephane.mg]

I confirm, avg is the best.

Concerning the epe, the fgm-3 specs told to use either a coil or make "extreme" measurements to find the constant for each sensor, to be able to correct them and make them react exactly the same, in case of using them in a gradiometer configuration.
What would be much more simpler to me, is to modifiy the pic program, to make it find the constants by himself :
Align the sensors, put them e-w, take measurement, turn left 90deg, take measurement, turn 180 right, take measurement, etc...
So the constants could be calculated by the program itself, but this means to reprogram the pic, and I dont have the material.
Also, I've read on the fgm-3 specs that a gradiometer is less sensible that a magnetometer to locate buried objects. The range of a single sensor is proportionnal to inverse / square the distance. So the range of two sensors is proportionnal to inverse double square the distance, making the device less sensible that a magnetometer.
So I decided to stop invest time on this gradiometer, I've learnt a lot of things anyway, but I'm seriously needing a precise device, so I buy the ppm sold by Willy.
Thanks for all the people that helped me here.

Best regards.

steph

[ivconic]

You are a very stubborn man. You always insist on not using Anti-Virus software. If you must use Windows then download AVG - it's free and works great! http://www.avg.com/product-avg-anti-virus-free-edition

Otherwise use Ubuntu Linux. Much faster, stable and doesn't need Anti-Virus software.

Huh! I guess you are right! I am old donkey. Dealing with pc's for 2 decades and have same old (stubborn) habits. Oftenly it bangs my had! I have few pc's at home. But my own pc (my real own pc - not to my kids) is old Siemens PII (450MHz). It is my most favorite "toy" i ever had! My best pal! My everything! But it is to slow for modern antivirus. Any antivirus installed slows it down so hard that i can not work at all.
Why am i so tied to that old machine? Because it is tested many times so far. Motherboard is hi quality, all the ports are tested with various programmers and everything is working perfect. I am dealing a lot with peripherals and most of those are handmades - i need accurate signal levels. So i tested few motherboards in the past and i was not happy until got this one! Eversince than, everything i could imagine simply works fine with this pc?!?
Unlike modern PIV's. I tested signal levels at PIV my kids have - horror!
So...kind a inert to collect enough money and buy real PIV, some real brand name. 99% of todays pc's are low quality, junk made in...you know where!
Year by year i delay that decission - finally to buy new machine for me.
So...pretty soon i must do something about it. Until than i will most probably suffer from simillar problems.
I do backups periodically. But some of my works forgot to backup! I lost shots and few schematics and pages of translations i worked on for nights and nights!
Well....now is to late for me to be smart!

[epitopios]

data logging software

Here is data loging software

Hi strujias , the summer is over and everybody back to his hobby, I saw the data logging software , but I think something is missing in the software programme , like .bin or .hex
What do you think ? did you work this software ?
friendly epitopios

[Qiaozhi]

I am very sorry cose i dont have those shots right now to post them here. Recenlty i got problems with my pc, i founded 92 trojan (userinit..or something) copies spread on my hard disk and i had to format it and so i lost all my work and datas from this year

You are a very stubborn man. You always insist on not using Anti-Virus software. If you must use Windows then download AVG - it's free and works great! http://www.avg.com/product-avg-anti-virus-free-edition

Otherwise use Ubuntu Linux. Much faster, stable and doesn't need Anti-Virus software.

[ivconic]

Ehhhh! With the respect, but it is not like that! 2 devices made so far and both already showed some results on the test fields. It is useable.
Ok...i can have remarks only on resolution at EPE mag. Not good enough.
But i found out the way how to overcome resolution problem; simply to make more often samples - meaning to walk slowly and move sensor tube in steps for about 5-10cm. So each 10cm take one sample. I made detailed test on this and it showed as pretty benefitial. I have small drain channel in front of the my house. It is located at 3m depth and local road is crossing over.
First time i tried device there, not knowing about resolution issue, i walked fast and took sample per meter almost (one step). Later i got pretty obscure graph on pc. Nothing clear and nothing recognizeable! I was pretty disapointed at the time, probably same as you now.
But....than i made another test by taking samples at each 20cms. Wow! Improvement! Graph was more logical now. Than i made another measurement with taking samples at each 10cm. Graph was pretty clear, showing almost the perfect position of channel. Ok...not clear as on some x-ray shot, but clear enough to beat any doubts about it.
I am very sorry cose i dont have those shots right now to post them here. Recenlty i got problems with my pc, i founded 92 trojan (userinit..or something) copies spread on my hard disk and i had to format it and so i lost all my work and datas from this year .
But i intend to make another device this winter and than i will record more shots from that place and post here if need.
So..chear up! You havent wasted your time.It is workable and useable. Just perform more tests in you backyard and try to take samples for short distances. Start form 20cm and shorter. Make several shots from same place (use some matrix on the field to know exact position for next measurements) and than compare gained results later on pc.
Regards!

[stephane.mg]

At least :
to me, as it is, this project cannot work efficiently.
the fgm-3 specs, on the speake's site talk about different resolutions / sensitivity for two different sensors, so one sensor moves in frequency faster than the other, even if correctly aligned.
I have for example made the tests, one of my sensors moves from 4000hz for e/w to n/s, and the other 1500hz for the same movement.
Test made in a magnetic perturbance free environment, sensor first put on a e/w alignement, rotated to have the max frequency value.

You must reprogram the pic to put some constants to take this sensitivity difference, so to me this project is unusable, and simply doesn't work.

Dont build it, dont buy the components, it's a pure waste of time, instead learning how it works, but you will not find anything with the epe (or maybe be very lucky and have exactly two fgm-3 exactly the same, but I wasn't)

[stephane.mg]

I finally managed to get rid of this fgm-3 problem.
There are two spikes on the signal, and a decay between them, this causes the flip flop to trigger when the fgm-3 frequency is above 42-43 khz.
Under those values, the second spike gets lower, and the flip flop doesnt trigger, this is really problematic as the flip flop goes directly to a frequency twice as less as it was before.

a single sine to square converter resolves the problem, I've tested it and it works, the fgm-3 is well squared, the spikes and decay are eliminated by this converter, so the fgm-3 signal is stabilisated. On the schema I've provided, the hex inverter is a 74hc14.

I've mailed Bill Speake 2 times, three days and two days ago. I dont have any answer from him and his company, so if you buy the fgm-3, you know what can happen to you, but you can make the little converter and the problem will be gone.

[stephane.mg]

Yes I did. I made a final test yesterday : remove completely one sensor, use another +5v power supply, and simply make a measurement.
Exactly the same results !

I'll try this evening to solder a 100n capacitor between fgm-3 output and ground to reduce the decay, if it 's too long, I'll try a 2,2n.

I'll report here the results and picture of the signal !

[olsteffe]

ground loop / isolation

Hi
Just adding to the think tank, have you considered an isolation or ground loop problem. To eliminate sources of common ground loop problems between sensor and measurement circuitry, always power and measure sensors with the same
local supply.

[stephane.mg]

I see. Hell... never occured to me to examine that more detailed. Look..i made two devices so far, checked and later sold. In process of testing i noticed something simillar - pretty obscure to me at the time. Number which presents difference between 2 sensor values was never right!? Maybe close but never exact?
At first device i got "default" values between 60-70000. At second 80-90000!??!?
I thought must be something changed in surrounding conditions.

->Normally you have a sensors that gives 40khz, the flip flop divides to 20khz, the pic compensates, so finally a 40000 value.

Each time i put effort to calibrate sensors properly. I think those were calibrated pretty good. In both cases i used Al pipe in which sensors matched exactly, maybe with 1mm free space. So when calibrated properly i later used super glue to fix those. So it was not because bad calibration.

->Aluminium tube, very good idea ! I'll buy one right away to align them, much more easy than the alimunium corner and blue tack, thanks for the info.

Outdoor, far away from any noise source, moment after reset pressed, both sensors are giving pretty close values and than normally changes due objective influences. So...i thought it is the way it should be.
Measured frequency at all 4 sensors i used so far never dropped bellow 35kHz and usually goes from 35 to 50kHz.
I did some field tests and am pretty satisfied with what i saw.
EPE do the job just fine for most of amateur needs.
But now reading your posts (thank you for that) i am starting to doubt. Next sensors i get i will put under more serious analyze.

=>I'm going today to my local electronic shop, and will add something to reduce the decay of signal that triggers the flip flop. This decay is reduced when the sensor goes to n-s where you've got the least frequency, and so the flip flop is not twice triggered anymore.

I can also suggest you to try to make your own sensor just for testing purpose. Basically it is sort of relaxation oscillator (reminds on Armstrong or Meissner). It has fb and tickle coils both wounded on mumetal core. Core is thin and flat, low mass so it doesnt need much power to saturate.
The way you explaining your problem looks like cores at your sensors are oversaturated at some point, shortly but enough to make problem!?

=>Yes, I think you're right. but they are satured for a frequency above 40khz, the decay is enough to trigger the flip flo. And normally this device can go from 35khz up to 105khz, so there's a problem with those sensors. Maybe one day, Bill Speake will explain, but now I've got to find a solution by myself, to make the epe work precisely as it shoud work. After this device works, I've planned to build a proton gradiometer, but that's another story.

Maybe trivial, but check power again

=>alreay removed the sensor, and putted a +5,2 v on it, got a measurement directly on the sensors output, same decay problem...

Tie cap and resistor as closest as possible to sensors pins.

=> it's the case, they are at 10millimeters...:-(

I couldnt help you much with this post, but hopefully i give you more ideas.
Regards!

I'll post here the solution that will be given to me, to correct the decay problem. Again, thanks for your help.

steph

[ivconic]

I see. Hell... never occured to me to examine that more detailed. Look..i made two devices so far, checked and later sold. In process of testing i noticed something simillar - pretty obscure to me at the time. Number which presents difference between 2 sensor values was never right!? Maybe close but never exact?
At first device i got "default" values between 60-70000. At second 80-90000!??!?
I thought must be something changed in surrounding conditions.
Each time i put effort to calibrate sensors properly. I think those were calibrated pretty good. In both cases i used Al pipe in which sensors matched exactly, maybe with 1mm free space. So when calibrated properly i later used super glue to fix those. So it was not because bad calibration.
Outdoor, far away from any noise source, moment after reset pressed, both sensors are giving pretty close values and than normally changes due objective influences. So...i thought it is the way it should be.
Measured frequency at all 4 sensors i used so far never dropped bellow 35kHz and usually goes from 35 to 50kHz.
I did some field tests and am pretty satisfied with what i saw.
EPE do the job just fine for most of amateur needs.
But now reading your posts (thank you for that) i am starting to doubt. Next sensors i get i will put under more serious analyze.
I can also suggest you to try to make your own sensor just for testing purpose. Basically it is sort of relaxation oscillator (reminds on Armstrong or Meissner). It has fb and tickle coils both wounded on mumetal core. Core is thin and flat, low mass so it doesnt need much power to saturate.
The way you explaining your problem looks like cores at your sensors are oversaturated at some point, shortly but enough to make problem!?
Maybe trivial, but check power again
Tie cap and resistor as closest as possible to sensors pins.
I couldnt help you much with this post, but hopefully i give you more ideas.
Regards!

[stephane.mg]

Hi ivconic, first thanks for your help.
I've already put my oscilloscope directly on the fgm-3 output, while deconnecting it from the flip flop.
I've played with n-s e-w orientation, saw the frequency changes of the signal, and saw that the signal is normally between 40khz and 50khz at my location.
But the main problem I have, is that the flip flop also have a 40khz-50khz output, on Q, fed to the pic16f877 on pins 2, and 3 for the bottom and top sensors, so I included this picture to show you the problem.
you'll see a first 0-4v, then a decrease, then a 1v-4v, then a decrease up to -1v, then stabilisation at 0v.
The second decrease seems to be enough to trigger again the flip flop, so the frequency is not divided by two as it should be.
So my problems comes from the two sensors itself, they muse have some defective components inside. I've mailed Speake two times since yesterday, still no answer.

Regards,

steph

[ivconic]

Look... to understand better FGM-3 behavior; i suggest you to wire up only power +5V (with cap an resistor) and output connect to oscilloscope.
Put it on table (with no metal arround) oriented N-S. Than observe waveform. Than turn it slowly in W-E orientation and see changes. Than use some iron piece to experiment. Close up iron to side of sensor and observe. Than close iron to front and observe. Play a bit with various orientations and various metals. All the time observe signal behavior on scope (phase, frequency and amplitude).
I think that will clear all doubts about FGM-3 sensor and its behavior.
Regards!