PIC16F series for PI detector timing

[waltr]

This thread is to discuss the use of the mid-range PIC micro-controllers for timing of PI detector
circuits. There are many other small micro-processors that also could be used like the Atmega series from Atmel.
These should be discussed in another thread as there are some threads for Atmels & Audinuos in this forum.
The PIC replaces the 555, 221, etc timing chips typical is many Pi detector circuit like the
Hammer Head, Mini Pulse Plus, etc. The two major advantages to using a uProcessor are:
1- one chip with a few non-critical parts replace several chips with critical R's and C's.
2- Changing timing, or even the whole timing sequence can be changed without a soldering iron.

I have been using PICs since the early 1990's when the PIC16F84 with Flash program store become popular.
Great little 18 pin DIP that is easy to use and very capable. Over the years the PIC18F88 become available
with the Same Pin out as the F84 but with added features such as UART, SPI/I2C, ADC, Analog Comparator, DAC,
Voltage Reference and more timers.
This PIC16F88 is now my go to PIC for small projects.

Years ago I did a small universal PCB for the 18 pin PICs. It has all the IO pins out to headers and an ICSP header (for PICKit2 or 3) as well as a header for the UART. I will post the PCB files for this board.

I will also post and discuss the code I have been using in the PIC16F88 for my version of the Hammer Head PI detector.

Please keep posts here on subject pertaining to the PIC16F mid-family range of processors.

[waltr]

Lets get directly into what I have coded. Attached is the source and hex files to do the simple two sample timing
typical of simple PI detectors. This has been posted in my thread on the HH2. Please read that thread for hardware details.
https://www.geotech1.com/forums/show...ake-on-the-HH2

The code sets up TMR0 to interrupt every 665usec which is the TX pulse rate chosen.
In the ISR all the timing for the TX and sampling switch pulse are done. Note: this is just like what Carl does in
the PIC code listed in the book ITMD and available on this forum.

Once in the TMR0 ISR, TMR1 is used to time the all the pulses. The sequence is:
TX on
wait TX on time
TX off
wait for delay 1 time
Sample 1 on (inventing input to integrator)
wait sample time
sample 1 off
wait for delay 2 time
sample 2 on
wait sample time
sample 2 off

The times of pulses and delays are adjustable with external pots (5k) connected the analog inputs on the PIC.
The analog inputs are connected to the ADC in sequence to measure the Voltage on the pot wiper and used to
calculate timer values. This makes all the timing easily adjustable and adaptable to the detector circuit and coil.
The first delay adjust pot was mounted on the case for adjustment during use whereas the others are small trim pots
inside the case for adjustment on the bench.

Since the pot value can not change quickly and the code does not measure the pot values every pulse cycle (665us rate).
Instead there is a 'tick' flag in the TMR0 ISR that gets set. Then in the "main" loop this flag is checked and if set
a counter is incremented. This counter is then compared to a fixed value. If counter exceeds the fixed value the code
goes and measures the ADC values of each pot then calculates new timer values.

I apologise for the crude coding with incomplete and wrong comments. This was my first cut of code for the PI detector and
was started as code for something else. Better code is coming.
16F88_detector.zip

[waltr]

I little about Tool chain and code.

I do like assembler code for the 8 bit PICs. It is not too hard and one does have full control of what the PIC does.
I did try C once with the Free version of the HiTech compiler but this creates very bloated code that for four times
larger and run four times slower. I have not try the newer XC8 compiler (GCC Based) for the 8 bit PICs yet. Lets keep this
thread on PIC assembler and techniques for controlling PI detector circuits. Someone can start a thread on "C code for PIC16F chips". Maybe even different threads for the
various C compilers available.

I also use the last of the older MPLAB IDE, version 8.92. This works very well and still available from Microchip
(download in the archives). Haven't tried the newer MPLAB X for the 8 bit PICs yet.

I use the PICKit2 for the PIC16F88 since it works well. The PICKit3 also works.
I posted the source code but not the Hex file for a reason. I prefer everyone to create an MPLAB project with the
source and build it. Then any issues like typos can be fixed and one can be assured the hex matches the source. This also give
one the opportunity to run the Simulator and step through the code or run it with break points.
The simulator in MPLAB 8 is excellent and is what I use to eliminate 99% of bugs and coding errors.

[Mdtoday]

Good thread Waltr and thanks for sharing.
There are many good Pics to choose from so will be interesting to see how many variants and sub threads are started from this


cheers

mdtoday

[waltr]

Thanks mdtoday. Yea, lots of PICs but what I do like is unity within Families. That is why the 16F in the title. This includes many variants and the code I post will run with very minor changes in many of these PICs. It also allows easy upgrading to larger/faster PIC with very little code change.

Same with pin outs and why the prototype PCBs I did years ago are still useful. Which brings us to this post's subject, the PCB.
Attached is the schematic in pdf and the Gerbers in a zip file. This is actually two PCB's, one for most 18pin PICs and the other for most 28 pin PICs. I just mod'ed the 28 pin board with three cuts and three jumpers to use with a PIC32MX. That is how similar Microchip kept PIC pin outs. I am planning on starting a PIC32MX thread since I am currently upgrading to this one and it is a PIC Moodz is using is his latest project.

The PIC names in the schematic are PIC16F87 and PIC16F873. These are close to Generic 18 & 28 pin PICs.

Anyone is welcome to use these files or build there own PCBs.

Note: CAUTION-- the ICSP header pin out does NOT match the PCKit2 or 3 directly. I made a custom cable that matches the pins.

P-CAD EDA - [16F87].pdf

[Mdtoday]

Thanks for sharing waltr, looks good.

Yep, I have designed a lot of product around Microchip midrange & Dspic/Pic24/Pic32 gear too and have to agree, there is a fair bit of similarity in pin-outs which lessons the pain of upgrading and porting code.
Must admit, in the early years, it was all machine code for Z80 with all its registers, or Motorola for its ports and on board eeprom ..well as young engineers, we would brag about who could write the tightest bit of code and not waste Eprom space in the 2716 ,haha, showing my age...

It's been all C code for me for the last 25 years with a little bit of machine code as required for tight spots.. compilers have come a long way.

I will start posting a bit more in a few days, I'm busy duplicating my computer's back up drives, it hasn't been done for a while, upgrading PC soon so just double checking I haven't missed anything.
Then I have to re-install applications..its a drag...

cheers

Mdtoday

[waltr]

Oh yea, the Z80 or 8080 and a 2716. I remember those back in the early 80's.

Been thinking about what I posted about this thread and C code. Now thinking yes to discussing C code in this thread but restricted to the Mid-range PICs.
Dspic/Pic24/Pic32 processor should have their own thread.

[Mdtoday]

Oh yea, the Z80 or 8080 and a 2716. I remember those back in the early 80's.


Been thinking about what I posted about this thread and C code. Now thinking yes to discussing C code in this thread but restricted to the Mid-range PICs.
Dspic/Pic24/Pic32 processor should have their own thread.

How could we forget the 8080 lol

I for one, am happy to discuss either C or assembly, and agree the 16 and 32 bit devices should be seperate threads.

cheers

mdtoday

[waltr]

Here is a link to the PIC code form the ITMD book:
https://www.geotech1.com/forums/cont...57-source-code

These use the PIC12F1840 which is a Mid-range PIC. It does run from an internal oscillator at 32MHz so is very capable.
This code can be ported to another PIC quite easily by redefining pins and adjusting timing value for a different clock speed.
The code is written in C which hides the fine details and is easier to read and understand what the code does.

If this code is used with a different compiler, XC8 instead of HiTech, then there will be some other changes needed, like the syntax of the 'CONFIG' bit settings.

The assembler code I posted works much the same so study Carl's C code first.

The heart of Carl's and my code is TMR0 Interrupt. TMR0 is setup for an overflow at the desired Pulse Rate. The TMR0 overflow sets the interrupt flag and program execution goes to the ISR (Interrupt Service Routine).
In the TMR0 ISR the sequence of coil pulse, delays and Integrator sampling is done.
This simple code replaces the timing chips of the typical PI detector circuits.

TMR1 is then used to timing the pulses and delays.

Look through both Carl's C code and my assembler code to see how both do this basic timing the same.

[waltr]

Now to jump ahead to my latest code that I have been running since last spring.
The biggest difference is the addition of GEB (Ground Elimination Balance) using a Three Sample method. See the thread "my take on the HH2, link posted above for the details of the method.

The great thing about using a processor for the timing is there was no PCB circuit modifications, it is all done is the code.
I had a spare pot on the case which is used to adjust the GEB sample pulse width. Since this has a third sampling pulse the TMR0 ISR sequence starts with the third pulse. Then the coil, delay 1, sample 1, delay 2 and then GEB sample.


Attached zip has source and hex file.

[6666]

Hi Waltr, how well does the GB system work that you are using ?

[waltr]

Hi Waltr, how well does the GB system work that you are using ?

Very well.I don't run into much mineralized soil but have run into an old brick walk that set off the detector. The GEB adjustment eliminated the response from the bricks and then did respond to metal between the bricks.
The other good side effect is a Low/High conductor discrimination. This and the PIC VCO give different frequency tones for the different conductivity targets.
Check my thread for details.

The best I can tell is this works similar to the TDI.