chipKIT uno32 Arduino MD experiments

[hdphilip]

greylourie,

i just got the sound working out of my experimential QED,

their's a bit more to the QED, than it appears at first. both on the hardware and software level, right now, without any gain from the "frontend opamp" with just some crude timing, i'm getting about 2 feet of detection of a metal table lamp, and that's on my workbench RFI/EMI all over the place. So, so far I'm pleased with the project.

Philip

[greylourie]

Hi Philip

I still learning the basics of the circuitry that goes into various P.I designs. Having looked at schematics and patents until I am cross-eyed, I get a vague sense of what certain parts are supposed to do - although I don't really The "damping" and "snubber" are common enough terms. I looked up the IXDF604 switch and the ADG401 switch. Have no clue what the "FEF Drive" is for. Is the SMPS pin for controlling the switching frequency ? Then there is the PWM_PWR ?

I used to follow the originator's threads and posts closely, both here and on another forum. Sadly I cannot recall all that he posted on the now defunct original GPOZ forum. Its a pity that no one has an archive of those posts. There was so much discussion and openess. But I do remember terms like V2C and "box car" averaging. I wonder what sort of processing went into the final iteration of the QED. If you read posts at the other place, its stated that the drawings are to scale.

I joined Geotech when he first started his thread on the UPIM. It was exciting. I even went on to buy the ADUC 845 Quick Start board. Now look at what you can buy for half the price. UNO32. And Moodz has done a fantastic job, picking up from where Bugs left off, providing "pie and sauce" to run a pretty fancy P.I building block so that anyone, keen enough, can get started. I will keep reading through what you have posted. Maybe some of it will sink in by osmosis....

You seem to have gone on in leaps and bounds. Unfortunately, I don't know enough to help. One thing is for sure, if I did, I would be building alongside you.

Jon

[hdphilip]

hello jon,

the "FEF Drive" would be used to turn off the receive signal path turing the tx pulse, but you will also notice the adg 1419 is able to do it also with the enable pin, they seem to work very well.

the PWM_PWR is used to turn the smps (- voltage) power supply off during the target sampleing, by doing this it reduces the power supply noise from getting into the target sample. The frequency is set by either the microprosser as in Bug's version, or preset in the smps driver chip in my version.

in the attached picture you can see were i turn the power supply off. you can also see where i was taking my time before turning it back on.........


also here's the software code for doing it:

#define TX_PULSE 120*CLOCKus
#define FLY_DELAY 10*CLOCKus 
#define TX_DRIVE BIT_0

#ifdef __cplusplus
extern "C"
{
#endif  
  void __ISR(_TIMER_4_VECTOR,ipl6) Tx_timing(void)
     {
                             //TX pulse Pin timer
         // mPORTDClearBits(PWM_PWR);     // Turn off power supply   
                
          TMR3 = 0;                                   // reset the timer, we need it to time the TX drive
 
               while (TMR3 <= TX_PULSE) 
             {
             mPORTDClearBits(TX_DRIVE);       // Turn on TX mosfet
             }                                               // for a time described in 
                                                              // in TX_DRIVE varible      
               mPORTDSetBits(TX_DRIVE);        // turn off tx drive
               
               mPORTDClearBits(PWM_PWR);     // Turn off smps power supply  
           
               
                             //TX Flyback Delay timer

               TMR3 = 0;                                // reset the timer, we need it to time the flyback delay
                        

               while (TMR3 <= (8*CLOCKus))
                     {
                       mPORTDSetBits(FEF_DRIVE);     // Turn off sampling
                     }
        
            
             swindow();   // code for taking samples I'll turn on the FEF drive inside "swindow" subroutine
             
             mPORTDSetBits(PWM_PWR);       // Turn on smps were done sampling  
             
       mT4ClearIntFlag();  // Clear Flag
 
              }
#ifdef __cplusplus
}
#endif

when you first look at it it may seem complex, but it's really not, it just take thinking about it in logicial steps


Philip

[greylourie]

Hello Philip

Thanks for explaining the FEF drive (its an acronym for Front End Fet drive ? lol I am slow ). I thought the adg1419 was to invert the input/configuration of U2.. Are there advantages to using it to isolate the rest of the amplifier chain from the pre amp ? Is something being done further down, when the adg1419 is isolating ?

Thanks for posting the screen capture. It truly is graphic ! I guess seeing that, would make any builder want to implement it into their P.I project. It cuts out a whole lot of noise rubbish.
Its really helpful to see the commenting on your code. It helps to envisage and run through the steps.

A huge number of people are following your progress and learning along the way. Have you seen how many views this thread has had ?

You are doing, as they say, " a smashing job".

Jon

[hdphilip]

hello jon,

I thought the adg1419 was to invert the input/configuration of U2

yes it does, if you look at the bottom green trace, you can see it with respect to the sampling.

Are there advantages to using it to isolate the rest of the amplifier chain from the pre amp ? Is something being done further down, when the adg1419 is isolating ?

yes their is, it's important to only have the desired target signal charge or discharge the integration cap, I don't know the exact reason, I didn't design it, but it's my belief that the adg1419 enable pin and the FEF pin gives 2 levels of isolation on such a small signal.

Thanks for posting the screen capture. It truly is graphic ! I guess seeing that, would make any builder want to implement it into their P.I project. It cuts out a whole lot of noise rubbish.
Its really helpful to see the commenting on your code. It helps to envisage and run through the steps

that's why i posted it, i use the chipkit mpide to program the board, it compiles the code, and downloads it with just 2 clicks of the button, that way i can make changes very easy. many times while the program is downloading the software, I'm already changing it...

A huge number of people are following your progress and learning along the way. Have you seen how many views this thread has had ?

I really hadn't paid much attention to how many views, the main idea is to show their are many ways to build a metal detector, and i think Bugwhisker's QED is the most insteresting.

and it has very high reviews from top leading metal detectors companies...........

Philip

[IBGold]

0

[IBGold]

and it has very high reviews from top leading metal detectors companies...........

Philip

Thats one way of putting it Philip by the way I am back.

Regards, Ian.

[greylourie]

and it has very high reviews from top leading metal detectors companies........... This is very true.

About the thread views. Its all about learning.

[greylourie]

Hi Philip

How is your audio going on ? What hardware have you currently got running ? Think that somewhere I saw a special 2 channel audio implementation, with interesting filtering. I think it would be perfect for your project.

[hdphilip]

it's going well, after looking over the source code, it seems i needed to open the sample window after i took the target sample, so i did that and wow!!!!

i forgot to save the code, oops, now it's time to mow the lawn..................

[greylourie]

Open the sample window ? Go ahead and slap me.. but after doing that, could you explain please ? I'm going to duck now and do some research.

Edit. Just looked at some page on window functions. Way over my head. But I did see mention of boxcar, which I have heard mentioned before. Also know as "Rectangular" window and Dirichlet window. Is this what you are using ? How are you filtering ?

[kt315]

is this thread really QED project or not?

QED on Geotech:it lives!!!
Don't believe all the lies and garbage you read by pro ML anti anything else fanatical jihadists goons on other forums. The QED is alive and well not only on Geotech but elsewhere as well.
http://www.geotech1.com/forums/showt...MD-experiments
doug

[greylourie]

Hi kt315

Also this thread : http://www.geotech1.com/forums/showt...riment-Project

I think HD showed a lot of scope screenshots in this experiment thread. I think he was quite happy with his results. Maybe he will tell us more about it if we ask him ?

[hdphilip]

Hi kt315

Also this thread : http://www.geotech1.com/forums/showt...riment-Project

I think HD showed a lot of scope screenshots in this experiment thread. I think he was quite happy with his results. Maybe he will tell us more about it if we ask him ?

I got pretty far with the project, far enough to realize the potential of bugwhisker's QED technology. I was having some noise issues with the front end. It'll just take more time than I have right now, I also posted about it on the aegp forum.

Philip

[sinclairuser]

i gave up on qed a while ago, i'm afraid the hammer fell for that md long ago, the detector itself was a brilliant piece of engineering, the problems arose due to those connected with the project.
it should have been developed quietly with encrypted software and novel processors, when the patent(uk parent would have been better) was granted, it should have been sold to a massive manufacturer with no silly bragging and pseudo promotion.
from what i understand it was the software that caused problems(minelab wise), if this thing had been developed properly in the first place and those involved had not got the attention of the competition the story would have played out differently.
if the qed has taught us anything its commercial projects have no place being discussed on the internet.
p.s. the source code is useless, what is "out there" has been sabotaged(atleast the one posted on a particular invite only torrent site is, all other qed rar's are missing code completely).