VDI-RODERICO

[roderico]

it might work if it is compatible with Microchip XC8 compiler, give it a try or else you might need mplabx, I also tried with SDCC but it gave many problems, also C18 might work...

[praktik_1974]

OK. I'll try.

[waltr]

Sorry, I have now re-read the whole topic. And I realized that the readings of the VDI number also depend on the position of the DISK regulator. Did I understand correctly? I asked the same about the GB regulator. Many thanks to you for the work done and the open source. Now I will look for a free C compiler. Can you tell me on this issue?
It is interesting to experiment and understand the mathematics of calculating VDI ....

I use the Microchip MPLab X IDE, IPE and compliers. This work fine for the VID code published in this thread by Rederico.

for each value of channel X and Y it just computes atan2(y,x) and fills a histogram, vdi number is extracted as the maximum value of the histogram, so it depends on what you plug in channels X and Y, however, you have more information with the whole histogram, X vs Y and/or X,Y vs time than with vdi numbers, I remember that waltr found some patterns in XY view to discriminate some coins that were mixed with junk but I do not really know the procedure, so if you are doing some systematic tests it would be interesting to see your results. There are already precompiled firmwares so you do need a c compiler unless you want to change the code...

Yes I have.
I use the X-Y graph to help ID target.
A single 'clean' target, like a coin horizontal, will show a straight diagonal X-Y plot. The angle is the ID.
The Disc setting will change this angle so I simply set Dis to a hair below Foil.

For complex shaped targets the X-Y plot will not be a straight line. It can also be not symmetrical around the 0-0 point. This then requires sweeping while moving the coil into and away from the target and sweeping from different angles. Watch the X-Y plot while doing this.
A good target, non-ferrous will give an X-Y plot line that is positive (upper right to lower left quadrants.
Ferrous gives a negative angle, upper left to lower right.
A mix of good and bad targets or a target that is odd shaped will give a curved or kinked X-Y plot line.
If I see any part of the X-Y plot line that has a positive angle I will dig. This has resulted in some very nice finds that have a rusty nail, ox shoe or other junk in the hole. Best was a War of 1812 1st Regiment of Artillerists button that had an Ox shoe 6 inch above and about 2 inch to the side. The X-Y plot had a lot of the curve negative but at some sweep positions and angles there was a definite positive in the X-Y plot.

The shape of the line in the X-Y plot gives much more info (almost like a 2D scan) then the simple VDI number. A almost never look at the number, only the angle and shape of the X-Y ploy.

[roderico]

Yes I have.
I use the X-Y graph to help ID target.
A single 'clean' target, like a coin horizontal, will show a straight diagonal X-Y plot. The angle is the ID.
The Disc setting will change this angle so I simply set Dis to a hair below Foil.

For complex shaped targets the X-Y plot will not be a straight line. It can also be not symmetrical around the 0-0 point. This then requires sweeping while moving the coil into and away from the target and sweeping from different angles. Watch the X-Y plot while doing this.
A good target, non-ferrous will give an X-Y plot line that is positive (upper right to lower left quadrants.
Ferrous gives a negative angle, upper left to lower right.
A mix of good and bad targets or a target that is odd shaped will give a curved or kinked X-Y plot line.
If I see any part of the X-Y plot line that has a positive angle I will dig. This has resulted in some very nice finds that have a rusty nail, ox shoe or other junk in the hole. Best was a War of 1812 1st Regiment of Artillerists button that had an Ox shoe 6 inch above and about 2 inch to the side. The X-Y plot had a lot of the curve negative but at some sweep positions and angles there was a definite positive in the X-Y plot.

The shape of the line in the X-Y plot gives much more info (almost like a 2D scan) then the simple VDI number. A almost never look at the number, only the angle and shape of the X-Y ploy.

I see now how you did on a mix of good and bad targets, very clear and very good explanation, thank you! I wonder if this can also be seen in the histogram, e.g. a small peak on the positive angles while you have many entries on the negative side with possibly a negative vdi number and thus missing the target and/or if the histogram loses important information to ID a target, i.e. that can only be seen in the X-Y plot...

[waltr]

I don't think the histogram would give as much info.
Seeing the shape of the curve I think gives you more.
And the shape an change as you move around while doing short sweeps back and forth.

The Eye/Brain does better with the curve in the plot.

I would need to try switching between X-Y & Histogram to see if the histogram can give enough info.

Of course I have dug many targets that looked promising only to find a piece of junk.

[roderico]

yes probably, but from your findings I was thinking that if an irregular shape in X-Y plot might also indicate the presence of multiple targets by sweeping in all directions you found a certain direction and/or an orientation of the coil that maximizes the signal of good targets over bad targets and a positive ID might appear in the histogram, off course, you might also find the direction that maximizes the bad target signal and hides the good signal, this maybe checked easily e.g. by making a small test site

[praktik_1974]

I don't think the histogram would give as much info.
Seeing the shape of the curve I think gives you more.
And the shape an change as you move around while doing short sweeps back and forth.

The Eye/Brain does better with the curve in the plot.

I would need to try switching between X-Y & Histogram to see if the histogram can give enough info.

Of course I have dug many targets that looked promising only to find a piece of junk.

I totally agree with this. On trash spots it will be correct to use a small coil.

[praktik_1974]

Who has? Please give me the latest up-to-date circuit, source code and firmware for the display with the 7565 controller in one go.Already got confused in stitches and diagrams ... (((

[praktik_1974]

I have such displays. I want to change the source for this display. https://www.buydisplay.com/2-9-inch-...black-on-white

[waltr]

I have such displays. I want to change the source for this display. https://www.buydisplay.com/2-9-inch-...black-on-white

Don't know that display.

When I build this VDI circuit I downloaded and mod'ed Roderico's source code to add support for a display I bought. I then uploaded the new display code to be included in the sources files.

This is what we expect you to do if you are going to use a display currently not supported.
The current source code supports a number of different display as is.

[ArchibaldSTM]

Hello! I also want to repeat this project, but I haven't found the processor PIC18F25K22. Will PIC18F25K80 work? Thank you!

[roderico]

Who has? Please give me the latest up-to-date circuit, source code and firmware for the display with the 7565 controller in one go.Already got confused in stitches and diagrams ... (((

please see post https://www.geotech1.com/forums/showthread.php?23823-VDI-RODERICO&p=287366#post287366

there is already a driver for 7565/7567 displays in the code, but you would need to build an additional pcb to configure the 7565 chip as it is in the firmware for that display, if you have not done this before, I would suggest you buy a 7565/7567 display with a pcb already attached, then it would only be a matter of wiring the display to vdi...

[roderico]

Hello! I also want to repeat this project, but I haven't found the processor PIC18F25K22. Will PIC18F25K80 work? Thank you!

Hello, then maybe you can find PIC18F26K22 which is also currently supported, PIC18F25K80 probably will work but as waltr explained with the display you would need to change the code for currently not supported pics...

[ArchibaldSTM]

I understood. Thanks for the quick response!

[praktik_1974]

Thank you waltr Thank you roderico!!
Yesterday I made a prototype board of the VDI indicator for 3 LEDs. I'll try to connect to Golden Mask 4