IB Metal Detector Project, Part 9

[Robert Hoolko]

Now that the big interrupt routine is done (except for fine tuning) perhaps I should review what it does. This routine performs two functions. It provides a 6600 Hz transmit signal, and it digitizes and demodulates the receive signal.

The transmit signal is a 6600 Hz rectangular pulse. This is a logic signal which can be used to control a current source which drives the coil. This method is suitable for White's BM coils which use a parallel resonant circuit (tank circuit). It is not suitable for other coils that have an un-tuned transmit winding (like Tesoro). The transmit pulse has a variable duty cycle. I plan to use this to try to get one or two more bits of dynamic range from the system by cutting the transmit signal to 1/2 or 1/4 of the normal strength. There are some practical problems to overcome to achieve this, but I wanted it designed in to the front end so it is available later.

The received signal is digitized by the on board 10 bit A/D converter, and demodulated in a dual synchronous demodulator which produces two output channels. The demodulation is achieved by sampling the signal at 4 phase angles (P, P+90, P+180, P+270). The P and P+180 samples are subtracted from each other to produce the signal for one channel. The P+90 and P+270 samples are subtracted from each other for the other channel. Both signals are run through a median filter then added into 24 bit accumulators. Sixty times a second the contents of the accumulators are dumped into memory and the accumulators are reset. This produces two signal channels at 60 Hz. With full scale inputs, these output signals can be +/- 37851, which is a little over 16 bits.

Robert Hoolko

[Carl]

Re: IB Metal Detector Project, Part 9

Robert,

Sorry for not giving feedback on your project... this is something I've been wanting to get into, but I'm up to my eyeballs in a GSM basestation design.

I'm not qualified to discuss the uC aspect as I've never dealt with one of these. But what you're doing sounds pretty reasonable. With the ADC, I see where you're coming from. Since you're using the same ADC for both the I and the Q sample, you're effectively getting 9x9 bits in each phase quadrant. That's 512x512 discrete amplitude/phase positions. And, like you said, that's plenty of phase resolution for large signals, but as the amplitude drops you rapidly lose resolution. Also, by using an accumulator you get better SNR, which effectively looks like more bits. I'm with you on that, I do something similar in radio receivers.

Seems like a better approach would be a true polar converter. Ever run across one of these? I have an idea how to do one, but never tried it. Another idea is to use a gain-compression ADC, which uses a fast-attack AGC to give progressively better resolution as the signal gets smaller. I designed one of these a few years ago.

You mentioned a delay in the ADC outputs. Most ADCs have some amount of latency delay as it takes at least one complete cycle to digitize the input, then the bits are latched out during the next cycle. I design multistage converters which can have delays of several clock cycles. They're not meant to be used in a burst mode. Which reminds me, are you trying to use the ADC in a burst mode, or letting it run continuously?

You made a comment about coils, that Tesoro uses an untuned coil whereas White's is tuned to 6600Hz (roughly). I think that Tesoro, like most or all makes, use a Colpitts oscillator for the TX. Thus they have to use a tune TX coil. Maybe they don't put the tuning cap in the coil whereas White's does. If so, you can always add your own tuning cap which makes the Tesoro coils more flexible. Didn't understand your comment about adjusting the TX duty cycle to vary the signal strength.

It maight be too late, but did you ever look at the Analog Devices ADuC812? It has a 16MHz 8051, a 5us 12-bit ADC (8 channel, plus 0.5us T/H), and (2) 12-bit DACs. Also 8K program/640b data eeprom. Let me know if you're interested.

- Carl

[Robert Hoolko]

Re: IB Metal Detector Project, Part 9

Carl

I am not familiar with a true polar converter, but I think the I,Q approach is going to work out ok.

The XLT implements the gain compression approach by reading the signal with an 8 bit A/D at four gain levels (1,8,64,512). Then they use the one that had the highest gain that did not overflow. That gives them good resolution on the weak signals, but still lets them handle large signals (at poor resolution). They immediately convert the input into an 8 bit logarithm. But I want to do the filtering in software so I need to have good resolution on the strong signals so I can separate the small target signal from the large ground signal.

I am using the A/D in burst mode. It is a successive approximation device that take 13 clocks to complete. (4 usec clock). The latency time I was talking about is the time from when I issue the command until the sample and hold switch opens. That time can be from 1.5 to 2.5 clocks. As long as I get the same latency every time I am ok. I am actually operating the A/D out of spec. The minimum clock is supposed to be 5 usec, but my only other choice was to go to 8 usec which is too slow for me.

My comment about the Tesoro coil not being tuned was a warning that you cannot just plug any search coil into this circuit without making some design changes.

My comment about changing the TX pulse width is because a 10 bit A/D by itself does not have anywhere near the dynamic range required. I need some way to extend the dynamic range. I will try a couple of things. One of which is to reduce transmit power when strong signals occur, such as having the coil over strong ground, or a manhole cover. I will also try variable gain before the A/D. Because of the TX coil design I can reduce TX power by just sending a narrower pulse.

I did look at the ADuC812. It looks like a good converter, but I did not think I could do the whole detector in it. I also wanted to buy a complete development board and development software package.

I am really impressed, in fact thrilled, with the performance I am getting so far. I can run an AC signal of +/- 1 LSB into it and measure a reasonably steady phase angle (within a couple of degrees) and then see the average angle shift a couple of degrees when I put a piece of ferrite on the coil. This is better than I expected for such a weak signal.

Robert