DIY with commonly available Microcontrollers

[Aaron Smith]

Has anyone played with the idea of using microcontrollers to build your own detector?


This is what the big manufacturers seem to be doing. The choice of which Microcontroller is big, getting bigger, and getting cheaper. Even if a proprietry coil was used the other electronics are 'off the shelf' items and a fully programmable detector should only cost a few hundred top$ to build. A simple design should afford ful reprogrammability via a PC with several programs kept on board via flash ROM.


I have a few design ideas at the moment yet my knowledge of such things (Microcontrollers) is still in its' infancy.


Further discussion is welcomed.

[Carl]

Re: DIY with commonly available Microcontrollers

>Has anyone played with the idea of using microcontrollers to build your own detector?


The Stuart PI design uses a microcontroller. Complete article and PIC code are available here.


I have the Basic Stamp kit but it's still in the box. Also just ordered the eval board for Analog Devices' ADuC812 microconverter. It includes an on-board 8-channel ADC plus DACs and 8051 uC. Hope to learn how to use them soon.


- Carl

[Piotr]

Re: DIY with commonly available Microcontrollers

Hi All,


First of all let's ask ourselves what the microcontrollers would be for. What are we going to achieve using these chips? What kind of singal analysis would be usefull and informative?


The computerized detectors I know use controllers for:


- calculating and displaying target IDs (many types)
- analyzing target size (GTIs)
- ground cancelling (Minelab detectors)
- generating waveforms (PIs)


What feature you are going to include in yours prototypes?


Personally, I would welcome non-motion target identifying detector (regardles of object depth and size) that totally ignores ground.


Any ideas?


Piotr

[Carl]

Re: My thoughts

>Hi All,


>First of all let's ask ourselves what the microcontrollers would be for. What are we going to achieve using these chips? What kind of singal analysis would be usefull and informative?


>The computerized detectors I know use controllers for:


>- calculating and displaying target IDs (many types)


>- analyzing target size (GTIs)


>- ground cancelling (Minelab detectors)


>- generating waveforms (PIs)


>What feature you are going to include in yours prototypes?


>Personally, I would welcome non-motion target identifying detector (regardles of object depth and size) that totally ignores ground.


>Any ideas?


>Piotr


In my opinion, uC chips are more useful as a control method, not as a processing method. For example, in the PI design it controls waveform timing (both TX & RX) but does not actually generate the waveform or process it.


For truely analyzing targets I would use a high speed sampling ADC and DSP. For true waveform generation I would probably use a DDS chip (at least for IB). For example, with DDS in a chirp mode and DSP processing you could do spectrum analysis and have a hell of a target ID machine. For PI, you could sample and process the decay curve to a much higher degree and possibly get more target info.


My goal is to learn how to use the uC chip to control all of the other devices. Otherwise, you end up with A LOT of knobs. I expect that, until I figure out how to set up an LCD interface, I will program my prototypes through a laptop.


- Carl

[Tom Delaney]

DIY with microcontrollers

Do you have any idea how fast you have to sample in


order to be able to do DSP processing on the signal?


There is a nice chip by Analog Devices which has an


8051 CPU core, an on-board a-to-d convertor and flash


memory, but the fastest a-to-d rate is one sample every 5 microsec. I don't think this is fast enough but maybe I am wrong. If one sample per microsec or faster is required, have to use an external a-to-d chip and DMA it's output into memory.