Voodoo Main Discussion

[m.shamim]

Hi everyone
I want to make a coil for a voodoo metal detector
What should be the resistance and inductance of each transmitter and receiver coil?

[Qiaozhi]

Hi everyone
I want to make a coil for a voodoo metal detector
What should be the resistance and inductance of each transmitter and receiver coil?

Read Chapter 6 - Coil Testing & Construction, and in particular look at Table 6-1.

[m.shamim]

Thank you, but in the sixth chapter, nothing is mentioned about the resistance of the coil, inductance is mentioned, I am asking because it is not possible to buy loop Tesoro and Troy in our country, and I have to make the coil myself. I guess that the resistance The transmitter loop should be at least more than 4, but I don't know the resistance of the receiver to use the right wire

[Carl-NC]

Neither coil has a critical resistance requirement. I would use no smaller than 24AWG for the TX coil to keep the R reasonably low. The RX coil is in series with a 5.1k resistor so RX-R is even less important, use 30AWG if you like.

[aft_72005]

For those who still want to make pcb; this is by now the latest version:

Hi ivconic

Thank you for this circuit and PCB design .
Are you built this project?

[Sean_Goddard]

The PCBs are usually in stock at -> https://diydetector.co.uk/store/index.php
but I think the online shop is currently closed for holidays.
Otherwise, you can send the Gerber files to JLCPCB.

Nope, Closed for EVER.

[Qiaozhi]

Nope, Closed for EVER.

Yes, unfortunately that's true.
Andy told me that after 10 years of running the Silverdog shop, things have moved on and he no longer has the time.

[Gunghouk]

Hi all, I'm a recently retired automation engineer and have decided to return to the soldering iron as my retirement hobby/pastime and have started building the Voodoo (PCB version C and software 2.02) . Please let me say up front I have zero knowledge of detector technology but have some general observations regarding issues I have found so far. I apologise if some have been mentioned or resolved previously.

1. Regarding the Lt1054 - pin 7 and R48 should not be connected directly but be separated by a 100pf capacitor, as in the mini pulse plus detector, otherwise the voltage doubler doesn't work. An alternative solution I'm using, until I cut the pcb trace and fit 100pf, is to lift pin 7 from the ic socket. Of course this could just be this manufacturers device version (I've tried 2 of the same).

2. ADC averaging is far more effective at eliminating noise if the ADCs are sampled at a regular and precise timing interval such as in an interupt of sufficient high sampling rate. Sampling every program loop suffers from the vagaries of timing due to different pieces of code running each loop cycle. Think audio sampling at 41Khz etc.
I'm not sure what the ideal rate for Voodoo is but would like to ask if the 1mS TX rate interruot couldn't be used to sample all ADCs, all of the time, regardless of operating mode?

My 2 cents (so far).

Thanks George and Carl for a great project and for giving me a chance to revive my enthusiastism for electronics and my new found interest in metal detecting.

BTW I'm currently taking a long vacation in China and have been busy purchasing various detector PCBs from JLC while I've been here to feed my hobby upon my return to the UK. Obviously I won't need all of them (5 is minimum qty) so there could be some available in the New Year if anyone is interested. No guarantees on version no.s etc.

Looking forward to engaging in discussions on the Voodoo et Al.

Regards to all

Neil

[Gunghouk]

Hi all, just to add to my previous post regarding ADC sampling.

The average value of a single sinewave cycle is zero. If you take enough samples during a single cycle which is most likely to give you the true value of zero, randomly timed or regularly timed samples for the same number of samples for each method ? The more samples you take the closer to the real average you can get.

From experience I can tell you the regular sampled time gives smoother, noise free results.

Opinions please

Cheers

Neil

[Carl-NC]

Thanks George and Carl for a great project

Voodoo is 100% George.

The average value of a single sinewave cycle is zero. If you take enough samples during a single cycle which is most likely to give you the true value of zero, randomly timed or regularly timed samples for the same number of samples for each method ? The more samples you take the closer to the real average you can get.

From experience I can tell you the regular sampled time gives smoother, noise free results.

This is generally true. In the case of a Voodoo, the ADC is sampling relatively slow signals so I doubt the sampling jitter will make a big difference. I don't recall the processing loop rate of Voodoo so I can't say for sure.

My personal preference is to run the ADC at a specific sample rate (say, 5ms) and use the data ready to trigger an interrupt, whereby this sets the loop rate of the software. But then, I also run the timers for the TX & demods autonomously. Voodoo uses a low-end PIC micro so it is pretty limited for doing stuff like this.

[Gunghouk]

Hi Carl thanks for the speedy reply. The TX interrupt of 1 ms would appear to be a suitable point to sample all ADCs and run the averages too. I'm going to play with that. There still appears to be jitter in hybrid mode so there's something else to investigate.
I give you credit in this project due to your software conversion to XC8 which allows nerds like me to use Mplabs X-IDE to tinker with the software.

The appeal of this project is not to make the best detector in the world but to improve, if possible, by sitting on the shoulders of giants like George, yourself and many others on this forum who provide the happily obsessed with technically challenging and thought provoking projects and which, ultimately, provide many enjoyable and fullfilling hours of entertainment.

Kudos

Neil

[Qiaozhi]

1. Regarding the Lt1054 - pin 7 and R48 should not be connected directly but be separated by a 100pf capacitor, as in the mini pulse plus detector, otherwise the voltage doubler doesn't work. An alternative solution I'm using, until I cut the pcb trace and fit 100pf, is to lift pin 7 from the ic socket. Of course this could just be this manufacturers device version (I've tried 2 of the same).

The correct method (according to the datasheet) is the one used by Voodoo. See Fig. 12 -> https://www.ti.com/lit/ds/symlink/lt1054.pdf

The MiniPulse Plus design works slightly differently. The MPP TX pulse rate can be adjusted over quite a wide range, and the LT1054 does not function properly at the lower pulse rates, which makes it difficult to synchronize the voltage converter to the TX. Instead of attempting to use the TX to drive the LT1054 directly, it simply pulses the converter (via the OSC pin) to produce pseudo-synchronization.

2. ADC averaging is far more effective at eliminating noise if the ADCs are sampled at a regular and precise timing interval such as in an interupt of sufficient high sampling rate. Sampling every program loop suffers from the vagaries of timing due to different pieces of code running each loop cycle. Think audio sampling at 41Khz etc.
I'm not sure what the ideal rate for Voodoo is but would like to ask if the 1mS TX rate interruot couldn't be used to sample all ADCs, all of the time, regardless of operating mode?

The Voodoo design is completely open-source, which means you are free to experiment and make improvements.

[Gunghouk]

Thanks George. Fig 20 shows the +ve voltage doubler with no connection on 6(osc) or 7(Vout). So no sync used. That's why I lifted pin 7 to make it work but with sync still connected to pin 6.
Regards
Neil

[Qiaozhi]

Thanks George. Fig 20 shows the +ve voltage doubler with no connection on 6(osc) or 7(Vout). So no sync used. That's why I lifted pin 7 to make it work but with sync still connected to pin 6.
Regards
Neil

Although Fig. 20 (+ve voltage doubler) does not specifically show the sync circuit, you can still add it as indicated in Fig. 12.

The converter definitely works as connected in the Voodoo schematic in Fig, 3-1, page 16 of the book. The only proviso is that the PIC needs to have been programmed, otherwise it definitely won't work since it's needs the sync pulses from the processor.

[phred01]

Hi all
Just become aware of this interesting project. Without having a batch of PCBs to be made has anyone one got for sale?