Let's made a PC-base metal detector with usb interface !!!

[Aziz]

Hi all,

a new month and a new and a small project budget (max. 80 EUR). But I can not order all the parts, that I want to have.

I don't know, which project should be forced forward:
- PI Controller project?
- IB-PI project? (latest)
- Measurement apparatus for measuring magnetic viscositiy, relaxation and eddy currents (to understand mineral ground response and ground balancing)?

The IB-PI project has still to be built. I have an old modified board, which runs sufficiently for IB-PI experiments. But the board is awful populated and there is no space for other modifications. It even began to smoke due to short circuit failure yesterday (runs again now).

So I want to go forward with the IB-PI project with a clean and new board optimised for high coil current drains (not battery drain). I will add a synchronised step-up DC/DC converter to push more current into the coil. So the coil can be powered up to 50V. I hope, it won't vaporise totally.

Another issue is the efficiency: How much copper should I invest into the coil? 500g, 1kg, 2kg? Copper is really expensive these days. Less copper in the coil means more heat and less efficiency. Heat in the coil is critical in the induction balance configuration.

I will think of to make more progress with the small project budget.

Aziz

[simonbaker]

Hi all,

a new month and a new and a small project budget (max. 80 EUR). But I can not order all the parts, that I want to have.

I don't know, which project should be forced forward:
- PI Controller project?
- IB-PI project? (latest)
- Measurement apparatus for measuring magnetic viscositiy, relaxation and eddy currents (to understand mineral ground response and ground balancing)?

The IB-PI project has still to be built. I have an old modified board, which runs sufficiently for IB-PI experiments. But the board is awful populated and there is no space for other modifications. It even began to smoke due to short circuit failure yesterday (runs again now).

So I want to go forward with the IB-PI project with a clean and new board optimised for high coil current drains (not battery drain). I will add a synchronised step-up DC/DC converter to push more current into the coil. So the coil can be powered up to 50V. I hope, it won't vaporise totally.

Another issue is the efficiency: How much copper should I invest into the coil? 500g, 1kg, 2kg? Copper is really expensive these days. Less copper in the coil means more heat and less efficiency. Heat in the coil is critical in the induction balance configuration.

I will think of to make more progress with the small project budget.

Aziz

Hi Aziz:

All your projects are interesting. One thing that would really help everybody is this project, if you can do it well:

- Measurement apparatus for measuring magnetic viscositiy, relaxation and eddy currents (to understand mineral ground response and ground balancing)?

If you can use that apparatus to make LTSpice models of actual targets, such as coins, caches, iron objects, etc., it would be very interesting and valuable.

It would be great to be able to model the sweeping of a search coil over a target at different depths, and to model the signal coming into the receive coil, for pulse and IB detectors.

I know it is not easy, but would be very interesting and allow many experiments using LTSpice to try different designs.

-SB

[Aziz]

Hi friends,

I just ordered the new parts. I hope, I did not forget some important parts this time. Well, I have strained my small budget heavily: total 120 EUR.

Here are some items of the order:
0.5 kg 0.4 mm magnet wire (0.4 mm is the smallest wire diameter I could get at this 0.5 kg package)
ribbon cables (to make low capacitance shielding: coil and coil leads)
lots of different ferrite cores (quite expensive)
step-up converter (the expensive one from Linear Tech)
a lot of capacitors (high voltage)
some schottky diodes
a lot of different voltage regulators

It should be possible to push forward the IB-PI and PI Controller project a bit. Still missing lots of parts but I will get them the other time of course.

Aziz

[Aziz]

Hi Simonbaker,

If you can use that apparatus to make LTSpice models of actual targets, such as coins, caches, iron objects, etc., it would be very interesting and valuable.

It would be great to be able to model the sweeping of a search coil over a target at different depths, and to model the signal coming into the receive coil, for pulse and IB detectors.

I know it is not easy, but would be very interesting and allow many experiments using LTSpice to try different designs.

-SB

I saw some interesting videos on youtube, which inspired me:
http://www.youtube.com/watch?v=kI7FXReF-5g
http://www.youtube.com/watch?v=oRcK6jE42PA

Such an apparatus would be fine. But I think, it isn't really necessary. I should push the PI Controller project a bit, which would save effort and money. The PI Controller should be able to measure this with high accuracy.

Aziz

[Aziz]

The package is on the way and will very likely arrive tomorrow.

[Aziz]

Parts arrived today.

Hi all,

I am happy to get the parts today so I can start soldering the coming days. I will use the LT1170 CT step-up converter with the external synchronisation circuit. It should deliver enough current. I have choosen this step-up converter as Linear Tech provides good spice models and the operation and synchronisation can be modelled and tested well.

I also need to build a heavy duty coil lead for high current applications. To minimise the skin effect, it will be a self made Litz wire with double layered Litz shielding. I have bought some cable spacer, which will ensure the minimal gap of coil lead wire and shielding. The shielding is a thin ribbon cable, which is wound around the wire. The two shielding layer wires will have an angle of 90 degree to shield EMI better (vertical and horizontal polarisation EMI shielding).


Aziz

[Aziz]

Coil Leads

Hi all,

the coil lead is almost finished. With two layers of Litz shielding (ribbon cable), it is going to be a 16 mm thick cable with four internal wires. Each wire has approx. 4-5 mm² cross section area (Litz wire). For pure mono coils, two of them can be paralelled to achieve high coil currents. The four wire version will be used for differential coils.

The coil is quite stiff and has a big bending radius. But it should minimize the self detection and skin effect.

If I'm finished, I will show you the monster cable and measure it's capacitance.

I might focus to the coil section soon.

Aziz

[Aziz]

Alternate Litz Wire for Coils

Hi all,

it seems, I will need much much more ribbon cables. I have been thinking of using an appropriate cheap cable for the search coil.

The ribbon cables are also quite attractive to make the search coil. And it offers Litz-wire benefits. A rough calculation proofs it's suitability:

R ribbon cable (measured) 0,2 Ohm/(strand*m)
Number of strands 20
Total length 30 m
Parallel Resistance 0,3 Ohm
Cost 11,1 EUR

The cost calculations are made for my own supplier. It may differ to other suppliers.

The ribbon cable will be splitted into 5-10 strand stripes. A group of stripes will be connected together (all strands parallel connected). So making a 20-40 (or whatever) stranded Litz wire possible. Then the coil will be wound as a spiral coil. As all the strands are bound in the ribbon cable, the winding process is quite simple.

BTW, each strand of the ribbon cable is also stranded: 7 strands (seems to be silver or tin coated).

The new coil calculations show the benefits of spiral coils.

Just curious but could be well worth to try it out.

Aziz

[Aziz]

Ribbon Cable Coil Patent

Hi all,

the absurdity has obviously started much earlier. Patenting obviousy matter is not only made by the big ripping companies. Can somebody remember the Litz wire patent for the coils?


Have a look at this:
http://www.google.com/patents/about?...BAJ&dq=4866424

No worry about this. It is already expired.

What about shielding with ribbon cable?
Hurry up! Run to the patent office and lodge a patent for this idea!


Aziz

[Aziz]

Project update:

Missing parts:
- more copper for the coil (ribbon cables, magnet wire)
- VCA chip
- case (Euro format for 160x100 mm PCB's)
- some more ferrite cores
- heat sinks
- mosfets (IXYF high power products!)
- some conventional parts

Laptop IB-PI (simple version):
I am working on the next version of the Laptop IB-PI. The clock detectors will be implemented using the LM393 comparator (simplifies the design further, I hope it will run). Bloddy f*$$, I can not start soldering until I know how the case is constructed and bolted. A big heat sink is also necessary to be flexible enough (must sink up to 10 W power).

Laptop PI controller (advanced version):
Well, it is quite difficult to me to get the VCA chip. I propably will make a discrete version using the MAT02 transistor pair. I can get this but it cost's more. Just waiting to finish the simple IB-PI project. It also makes sense, to implement the TEM transmitter for this controller.

I assume, I will need 2-3 months more to afford all the missing parts. But I am looking forward to order some more parts soon.

Aziz

[Aziz]

Ribbon Cable

Hi all,

just found the details of the ribbon cable now:

It's an AWG 28 ribbon cable and one wire consists of 7 tinned strands of 0.127 mm wire diameter (=0.087 mm² cross section) (silver would make it more expensive). The measured resistance matches to the calculated resistance.

The 7 tinned strands (=1 wire of the ribbon cable bunch) would have 0.213 Ohm/m (specified)
Measured resistance (not accurate enough): 0.2 Ohm/m
Calculated resitance: 0.201 Ohm/m at 20 °C

Ok, I am going to buy 30-40 m ribbon cable.

Another alternate would be using copper metal strips, which can also be made in such a way, that imitates Litz-wire principles. Parallelling more copper strips (0.1 mm thickness, 5-10 mm width) with some insulation between two strips. This would make very low resistance and cheap coils possible (spiral wound coil of course).

Aziz

[Aziz]

Stucking...

Hello friends,

It's sooooooo much frustrating again. My last parts order package is damaged by the parcel service DHL and is causing lot's of delay time. So the new parts still not arrived yet.

PI controller project:
It's also frustrating not to get the desired parts (VCA chip, super matched transistor pairs, MUX chips, OTA chip (operational transconductance amplifier), and some more other parts).

There are some alternates to the VCA chip:
- discrete OTA implementation (lots of current mirrors necessary, need matched transistor pairs)
- OTA-chip implementation (CA3280, f$§k it's beeing discontinued)
- discrete differential amplifier with voltage controlled current mirror (need super matched transistor pairs either)
- four quadrant multiplier (no, not really)

The OTA LM13700 is cheap but quite noisy for VCA applications. The CA3280 is not available (probably one of the best OTA chips ever).

Ok, the strategy for the coming month's is:
a ) obtaining some of the available super matched transistor pairs
or
b ) buying several hundreds of low noise transistors and hand matching them
or
c ) processing the temperature dependency in the software (requires temperature sensors and complex calibration procedures)
or
d ) f$$king up the VCA and replacing it with discrete PGA implementation

Solution a ) is a matter of money only. I have found another source for such transistors. Damn, the transistors are quite expensive for my low budget. These transistors could also be used in the pre-amp stage either.

Solution b ) is quite easy for me. Could be used until I have got matched transistor pairs.

Solution c ) would be the best (additional to a ) and b ) ):
Two independent temperature sensors are necessary: One for the high power parts (placed on the heat sink) and one for the ambient temperature (for the other electronic circuits). Temperature gradients cause usually drifts in the measured voltages/currents. They can also mismatch a calibration. Main cause of it is the offset error voltages and non-linearity gain changes. These temperature caused errors should be corrected in the software.

Regulating the temperature inside the case is not making much sense: need power for this.


It also seems to me very important to make discrete solutions as the chips have a short live time. Another issue is the pricing policy: a discontinued chip will be bought from a monopole and they are making profits with extremely high prices then.

I will focus to the integrators and sample & hold amplifier section soon to push forward the project a bit.

IB-PI project:
Waiting for the new parts.

Aziz

[Tinkerer]

With a high power detector, heat becomes a problem. Maybe a cast aluminium enclosure, of the type they use for automotive audio power amplifiers could be the solution.

These enclosures have external heat dissipation profiles.

It might also be useful to use 2 separate enclosures, one for the Power supply and TX circuit, the other one for the RX and controls.

Does anybody have some recommendations?

Tinkerer

[Aziz]

Case

Hi Tinkerer,

I have already ordered a modular ALU case for the IB-PI project (see below). One can combine different height. So the hole PCB will be a 160x100 mm bread board, which fits into the case.

Aziz

[Aziz]

Hand Matching Transistor Pairs

Hi all,

it seems, that hand matching transistor pair is more easy than I have expected. Even the thermal coupling is easy to do. I have found an easy solution for this. So a super matched transistor pair can be built in a DIP-8 pinout (compatible to the LM394, MAT02/03, SSM2210/2220).

Now I need hundreds of the following transistors:
BC550C/BC560C (cheap, low noise)
2SC1815/2SA1015 (cheap, low noise)
only 4 EUR-cent each!!!

I am sure, I can find several super matched pairs and I will have a live time stock for these transistors.


Aziz