TESORO GOLDEN SABRE

[Max]

Okay I've done more measurements:
The oscillogram below is measured transmit signal with settings 10uS/Div and 1V/Div. As you can see I have huge difference in the frequency and voltage.
I don't have F.counter so Im using the scope:
3.4 Div x 10 uS = 34 uS, so T = 34 uS = 0.034 mS
F=1/T; F = 1/0.034 = 29.4 Khz
Amplitude is less than 5 Volts Vpp (1V per Div)..
I measured supply voltage:
- Between: GND and Output of 7808: 7.98V (Good);
- Between: GND and 4 Pin of LF353: 0.8V (Not Good??);
---
Obviously the frequency is way too high. And considering my wire is 0.28 with varnish (I bought from the store, I haven't even measure it, it should be accurate), and the windings number which is 100, and the coil former is 28DD (255x137) I just have no idea why the freq went so high. Scope is calibrated,
2V/2Khz sine wave shows accurate on the screen.

I can make new Tx coil with lets say 135 windings, which I will, but this is not important.
Important is what I am missing here. Up to here:
Negative supply voltage suck (or Im not measuring it right)
Tx Freq is too high, too low Tx voltage signal (could be because of -voltage)

As for a start from now, I need to make new Tx coil with more windings (or attach the Rx to see the freq change since it's 105 windings).
Negative voltage - don't know..

Help needed , lets help the new guy ... Your ideas and criticism are most wanted now.

Hi,
frequency is wrong... really high. I don't know why... but I will check the capacitors at TX if the turns are 100 on TX coil. It's really strange you get so hi frequency there...

0.28mm with varnish is ok.

I made many coils... even different wire and get easy always around 14Khz with 100turns. Remember you need to tight the wires very strong on coils to get required inductances.

PS: check voltages too , you must get +8V at regulator, +12 or more from battery (otherwise works but not so good) and at least -6V at converter.

Kind regards,
Max

[Max]

OK I light the 'thing' up, no smoke (I'm suprised?)
But I am suprised, why there's no sound, no even chatters or beeps
from the not ajusted coil. Is it normal to do that ( or actually not doing anything when it's still not nulled)?
I did quick bench test on the scope - oops 5V Tx amplitude, argh.. not good, we need at least 10V...
Yep the batteries are not charged. I did plug them yesterday evening for charging, and what's wrong with them, wtf!!. So i found the problem, it was a disconnected wiring inside the charger..
I didn't do any further tests, but strange thing I noticed is that the Tx form seems just like those from simulations with LTSpice (pic taken from Max, page 57) than the other posted pictures. What I mean - the signal has a small noticable "bump", it's far not perfect sinus wave signal.
I guess the fun part begins.

Hi,
have you connected the 1K resistor on top that it's missed from pcb ? I mean R27 at Ivconic's schematic... you must put it in component side... make 2 extra holes for it if required.

You hear no sound at start cause there isn't anymore battery checker there... so sound could be produced just by signal path.

Also... which darlington you use ? If use BC517 you must reverse 180° cause pinout is different from MPSA13.

The small bump there is non linearity at oscillator... depends also on coil... good coils show less of it... but expecially transistor is involved.

Often is not much important unless it's very large bump.

Kind regards,
Max

[eclipse]

Okay I've done more measurements:
The oscillogram below is measured transmit signal with settings 10uS/Div and 1V/Div. As you can see I have huge difference in the frequency and voltage.
I don't have F.counter so Im using the scope:
3.4 Div x 10 uS = 34 uS, so T = 34 uS = 0.034 mS
F=1/T; F = 1/0.034 = 29.4 Khz
Amplitude is less than 5 Volts Vpp (1V per Div)..
I measured supply voltage:
- Between: GND and Output of 7808: 7.98V (Good);
- Between: GND and 4 Pin of LF353: 0.8V (Not Good??);
---
Obviously the frequency is way too high. And considering my wire is 0.28 with varnish (I bought from the store, I haven't even measure it, it should be accurate), and the windings number which is 100, and the coil former is 28DD (255x137) I just have no idea why the freq went so high. Scope is calibrated,
2V/2Khz sine wave shows accurate on the screen.

I can make new Tx coil with lets say 135 windings, which I will, but this is not important.
Important is what I am missing here. Up to here:
Negative supply voltage suck (or Im not measuring it right)
Tx Freq is too high, too low Tx voltage signal (could be because of -voltage)

As for a start from now, I need to make new Tx coil with more windings (or attach the Rx to see the freq change since it's 105 windings).
Negative voltage - don't know..

Help needed , lets help the new guy ... Your ideas and criticism are most wanted now.

[simonbaker]

OK I light the 'thing' up, no smoke (I'm suprised?)
But I am suprised, why there's no sound, no even chatters or beeps
from the not ajusted coil. Is it normal to do that ( or actually not doing anything when it's still not nulled)?
I did quick bench test on the scope - oops 5V Tx amplitude, argh.. not good, we need at least 10V...
Yep the batteries are not charged. I did plug them yesterday evening for charging, and what's wrong with them, wtf!!. So i found the problem, it was a disconnected wiring inside the charger..
I didn't do any further tests, but strange thing I noticed is that the Tx form seems just like those from simulations with LTSpice (pic taken from Max, page 57) than the other posted pictures. What I mean - the signal has a small noticable "bump", it's far not perfect sinus wave signal.
I guess the fun part begins.

Yes, fun starts! For us too watching.

FWIW: Is that 5V peak-to-peak, or single peak? I get about 5V single peak Tx signal - if I want more I add a LED diode to the 4.3V zener to increase the voltage drop which increases the TX voltage - but maybe more distortion also.

TX voltage is sensitive to zener and also JFet Vgs-cuttoff. JFets have wide range of spec, Vgs-cutoff from -.5 to -5 ! If that is true, everybody have different TX voltage! But I think they are fairly consistently around -2 to -2.5 V. You could check yours to be sure.

I guess transistor non-linearity makes distortion, JFet tries to regulate it to make more sinus. If you disconnect JFet, nice big oscillation but big distortion too. So if too distorted, JFet not acting soon enough, need smaller Vgs-cutoff or different zener I think. I experimented with it a fair amount.

No chatters sounds good to me! My breadboard version is a henhouse! But good for certain experiments.

Keep in mind - it's a motion detector, so even if coil way off, no chatter with really tight circuit like you have. Probably super clean, you'll be finding coins from hang glider soon!

Looking forward to more reports.

Cheers,

-SB

[eclipse]

OK I light the 'thing' up, no smoke (I'm suprised?)
But I am suprised, why there's no sound, no even chatters or beeps
from the not ajusted coil. Is it normal to do that ( or actually not doing anything when it's still not nulled)?
I did quick bench test on the scope - oops 5V Tx amplitude, argh.. not good, we need at least 10V...
Yep the batteries are not charged. I did plug them yesterday evening for charging, and what's wrong with them, wtf!!. So i found the problem, it was a disconnected wiring inside the charger..
I didn't do any further tests, but strange thing I noticed is that the Tx form seems just like those from simulations with LTSpice (pic taken from Max, page 57) than the other posted pictures. What I mean - the signal has a small noticable "bump", it's far not perfect sinus wave signal.
I guess the fun part begins.

[eclipse]

Thank you guys Im glad you like it.
Still there is the most important part ahead - coil & electronics adjustment.
Without proper adjustment the device will perfom poorly.
There is enough information for the coil adjustments, I just need to follow it.
Great project!! Thanks! Hope I won't be dissapointed.
Cheers!

[Max]

If phase of RX at null critical, then it seems we should not null with voltmeter, instead connect our circuit and make sure on right side of null with scope -- or is it really that important?

Maybe I can try to test with "nulling machine" and see how much difference which side of null makes. Probably will need to wait until have proper PCB and shielded coil (with breadboard, unshielded coil, I can detect clad US quarter at about 15 cm -- maybe not impressive, but quite a victory although required lots of work to suppress noise, spikes as much as possible, including some extra caps. Many more experiments to come, I can put in separate thread).

-SB

Hi,
what I think of nulling ?

Truly... I think is possible doing it without scope in a good way... but with enormous patience and efforts and not sure results...

It's worth that way ? Uhm... if you're a good DIY man yes... you can do that way and it's worth, no dubt. If you're really motivated and have no scope (not your case)...too , it's worth... many here did coils without scope I think... but great motivation is a key then.

Truly... for me it's just by scope: I have no time to mess without see what I'm doing... requires too much time and efforts I simply can't spend at now. In past years I made similar crazy stuff too sometimes... but now I will never start a project without required instruments...

So... have scope !? Use it: save time, save work, save nervs.

Kind regards,
Max

[simonbaker]

more nulling

Hi,
this is the right side... if you're looking at output of preamp: there's something (I cannot read it well from picture) around 200°

That's where you're supposed to "null" the coil: be aware you'll probably cannot mantain that phase also at minimum voltage... you just need to keep voltage smallest as possible matching that phase... other phase will false disc and geb... and your device will be completely unreliable.

Kind regards,
Max

If phase of RX at null critical, then it seems we should not null with voltmeter, instead connect our circuit and make sure on right side of null with scope -- or is it really that important?

Maybe I can try to test with "nulling machine" and see how much difference which side of null makes. Probably will need to wait until have proper PCB and shielded coil (with breadboard, unshielded coil, I can detect clad US quarter at about 15 cm -- maybe not impressive, but quite a victory although required lots of work to suppress noise, spikes as much as possible, including some extra caps. Many more experiments to come, I can put in separate thread).

-SB

[Max]

Almost ready, tomorrow - big day

I'm trilled... seems really nice!

[simonbaker]

Almost ready, tomorrow - big day

Gawd that is gorgeous! I think you should make a separate thread with the whole story of building your detector, experiments, waveforms, pictures of components, why you chose each, field tests - don't forget schematic, pcb files -- so we can drool! Really nice!

-SB

[Qiaozhi]

Hi,
look at that: it's Qiaozhi's picture of signals at TX, RX (output of preamp) using an original coil (maybe it's an 8'' round ? I've tested an 8'' round an was exactly the same stuff too...I suppose they are about all the same... no matter of size or widescan...spider... etc).

I read about the same on my 9x8'', also... you'll get left shifts of RX signal nearing any metal, but no shift in pure mineral response (e.g. ferrite crystals)

Kind regards,
Max

This picture was taken using a Tesoro 9"x8" coil, but the result was the same for an 8" and a 4" coil. These were all concentrics.

[rdec4]

I think: sync on TX signal. Measure width of one cycle. Measure distance RX signal is shifted by comparing peaks. Then make fraction and multiply by 360 degrees.

Example: one cycle on scope is 4 divisions. RX signal is shifted 2.2 divisions. 360 * 2.2 / 4 = 198 degrees.

I hope I did that right...

Cheers,

-SB

There is a good graphical explaination of what simonbaker says, hope it helps.
Here:
http://www.hobbyprojects.com/oscillo...surements.html

Thank you for your help.

[eclipse]

Almost ready, tomorrow - big day

[eclipse]

There is a good graphical explaination of what simonbaker says, hope it helps.
Here:

http://www.hobbyprojects.com/oscilloscope_tutorial/waveform_measurements.html

[simonbaker]

Given the above published photos have the following question:

What exactly is the method of calculating phase difference between TX and RX with two channel analog oscilloscope, so the result is in degrees?

I think: sync on TX signal. Measure width of one cycle. Measure distance RX signal is shifted by comparing peaks. Then make fraction and multiply by 360 degrees.

Example: one cycle on scope is 4 divisions. RX signal is shifted 2.2 divisions. 360 * 2.2 / 4 = 198 degrees.

I hope I did that right...

Cheers,

-SB