PI CHIP

[Mechanic]

Hey Guys,

Just another thought. It would be worthwhile setting up a dummy program that just turns a pin on and off as fast as it can and then look at this pin on the scope. Try for a sweep speed that allows a lot of oscillations, so you can almost not identify the waveform, then use the X10 function on the sweep speed and look at the waveform at the far left of the trace and see if there is jitter there. If there is, it is oscillator jitter, if not, then something in the program. Thats why assembly is good. The damn chip does what you want exactly when you want, providing your code is correct!

Also I would be aiming for much higher resolution than 1us on your sample pulses. How are you effectively going to adjust for earth field cancellation? If 33ns is what the chip can do, then thats what the sample pulse resolution needs to be.

Cheers Mick

[Mechanic]

G'day Guys,

That is jitter you are seeing. Scrap the internal oscillator, go crystal. It is important for a good low noise platform not to suffer from jitter. Once on my sd2000 I tried to make a vco so I had infinite adjustability on the operating frequency to be able to accurately tune out interference. The problem with the vco was it jittered. Although you could not notice this jitter in the threshold, the detector was not as sensitive as it should have been when compared to using a crystal of the exact same frequency.

Cheers Mick

Ps Digital crows are not the be all and end all! They will not pick up jitter easily, where as with an old clunker analog scope it will be very obvious, same goes for noise.

[mickstv]

Paul, i'll try to explain what I'm seeing. Ok i've set the TX to 100us then on the CRO set to .1ms/Div I have 3 TX pulses on the screen the first 1 looks still the second pulse in the middle is moving about 8us then the last pulse is moving approx double.

It's hard to explain it, wish I had a digital CRO.




Mick

[moodz]

I'll post up a schematic as soon as I can.



First things first, I have noticed jitter on the TX and sample pins, I'm only using an old 50mhz CRO but looking at say the TX output from the micro it seems to be jittering on the RHS of the screen. So I made up just a standard 555 timer circuit to check the cro and it displayed correctly. So I'm having some sort of issue with the micro section.



Mick

Mick how much jitter are you seeing .. it should be 33 ns maximum .... I will have to rebuild the source with the crystal option for comparison .... the TX signal is not generated by code but chip hardware .. if there is still a jitter problem with crystal drive then that would mean there is a design flaw in the chip itself
All the other timing is taken from the TX timing so this will jitter to some extent also.

Paul.

[mickstv]

I'll post up a schematic as soon as I can.



First things first, I have noticed jitter on the TX and sample pins, I'm only using an old 50mhz CRO but looking at say the TX output from the micro it seems to be jittering on the RHS of the screen. So I made up just a standard 555 timer circuit to check the cro and it displayed correctly. So I'm having some sort of issue with the micro section.



Mick

[6666]

Mick
would you be interested in posting your schematic here ?

[moodz]

Hi Paul, I'm actually back working on a diff front end at the moment. I'm running the FET at 5volts and an OP37 running as the diff amp then into the 4066 and LM358. I'm now getting the about 15cm on a piece of tin plate about 5cm X 5cm, still poor.

What program are you using to draw the schematic's your posting because I might do my schematic and PM it to you so you can have a look.

Mick

Eagle.

moodz

[mickstv]

Sorry to hear that Mick ...are you sampling at 15 us ? ...... I am working on the frontend blocking FETs at the moment. Ferrous / non Ferrous descrim is working really well .... I am trying to work out how to encode the descrim AND the target ( yes / no ) in the audio channel.

moodz.

Hi Paul, I'm actually back working on a diff front end at the moment. I'm running the FET at 5volts and an OP37 running as the diff amp then into the 4066 and LM358. I'm now getting the about 15cm on a piece of tin plate about 5cm X 5cm, still poor.

What program are you using to draw the schematic's your posting because I might do my schematic and PM it to you so you can have a look.




Mick

[Vortxrex]

HI Moodz, Mick, everybody. I spent most of the day working on a Prototype board for the dspic30F4011 Schematic in post #85. I am using the RadioShack 276-170 (I like these because they are set up like a breadboard) Looking forward to experimenting

Dave Frank

[moodz]

Hi Paul, I've had no luck getting good range yet so I might wait and see what your schematic is like, then have another go at it.


Mick

Sorry to hear that Mick ...are you sampling at 15 us ? ...... I am working on the frontend blocking FETs at the moment. Ferrous / non Ferrous descrim is working really well .... I am trying to work out how to encode the descrim AND the target ( yes / no ) in the audio channel.

moodz.

[mickstv]

Hi Paul, I've had no luck getting good range yet so I might wait and see what your schematic is like, then have another go at it.




Mick

[moodz]

Hi Paul, what I'm actually using is a OP37 into a 4066 and LM358 as the diff integrator. I actually suspect my problem might be timing, after going back through previous posts I missed some info. Which if I read correctly suggests for instance that if I have a 100us TX and I want to sample at 15us then the S1 start sample would have to be at 115us. I hope that is correct.

The coil damping option sounds interesting another good option to test after I get this frontend finished.



Mick

Mick .. you got it ..everything is referenced from Tx ON not OFF .... that would mean you must have ben sampling during TX on.

Moodz

[mickstv]

Hi Paul, what I'm actually using is a OP37 into a 4066 and LM358 as the diff integrator. I actually suspect my problem might be timing, after going back through previous posts I missed some info. Which if I read correctly suggests for instance that if I have a 100us TX and I want to sample at 15us then the S1 start sample would have to be at 115us. I hope that is correct.

The coil damping option sounds interesting another good option to test after I get this frontend finished.



Mick

[moodz]

Hi Paul, found problem number one I made a mistake with two of the 10 k resistors had them going to gnd instead of 5 volts. Then whilst working on the front end I somehow killed the processor chip it only comes up with rubbish on the screen, i erased the chip and reprogrammed but still the same. I had a second IC which I put in and it functioned ok.


Next problem to work out what the problem is with the front end as I only have shallow depth on big objects.



Mick

Hi Mick, the ADC expects to see an increasing voltage for a target .. check polarity of target signals from the diff integrator if that's what you are using. I'll put up the damping code and circuit so you can tweak the coil damping for optimal setting.
Moodz.

[mickstv]

Hi Paul, found problem number one I made a mistake with two of the 10 k resistors had them going to gnd instead of 5 volts. Then whilst working on the front end I somehow killed the processor chip it only comes up with rubbish on the screen, i erased the chip and reprogrammed but still the same. I had a second IC which I put in and it functioned ok.


Next problem to work out what the problem is with the front end as I only have shallow depth on big objects.



Mick