Update on Basic stamp and Hammerhead

[Ray-NM]

This is an update on using a Basic stamp microcontroller on the Hammerhead.
the purpose of this was to test different coils and cables and find out what the
best combinations would be. Also to learn how to interface it and control timing on the Hammerhead using a basic Stamp.
First off I breadboarded the basic circuitry for pulsing the mosfet and coil.
The first coil that I used would not create the flyback pulse. So I switched to
another coil and it worked as expected. Need to figure out why the one coil did not
work. Anyway I tried 3 different cable types to see how the delay was affected.
The first was a cable with shield/drain and 3 wires it showed a delay of about 7 usecs
after the begining of the high voltage pulse - 250v aprroximately.
I then tried some RG6 cable that I had left over from my Sat/dish receiver.
It read the same. I then tried 2 single strand wires left over another project.
I had the same signal and delay of 7 usec on the scope. I ran out of time to check
out some other cables. My conclusion so far is really unclear as I need to check everything
again and verify what I got.
but I guess the best thing is using the Basic stamp controller. It allows me to use precise timings and enables me to verify differnet settings. I have set up several small programs
to test various aspects of timing the pulses that way I know I am comparing apples to apples.
If any of you have suggestions let me know.
RayNM

[bbsailor]

Some Possible Answers

Ray-NM

If you used 3 different cables that had close to the same capacitance, then the MOSFET capacitance being so much larger swamped any slight difference in cable capacitance. I suspect that if you measured the shield to center conductor capacitance that is would be less than 100pf compared to 700pf to 1000pf for the MOSFET operating at about 12 volts. Measure the capacitance of each cable and post it, if you can.

When you measure the delay, try using an external scope sync connected to the MOSFET gate signal at the point of pulse turn-off. This, I believe, is the way Eric Foster measures delay. There are many ways to measure delay but they can vary by a few us. So, to obtain a delay value that is consistent with accepted measurement techniques, measure the delay from the pulse turn-off point on the MOSFET gate.


Coil type, mono or DD?
What is the value of the damping resistor across your coil (mono) or coils (DD)?
What is the inductance of your coil (mono) or coils (DD)?
What is/are the coil(s) resistance?
MOSFET you are using?
Any series TX resistor?
Is/are the coil(s) shielded?
Frequency of operation?
TX pulse width?
Cable capacitance?

With a little more information we may be able to provide more help?

bbsailor

[Ray-NM]

Have some of the answers tomorrow.

I will have to wait until tonight to get some of the answers.
Will Post them tomorrow.
RayNM

[Ray-NM]

some answers

BBsailor
It was too cold in my shop to do anything last night but I can post a few of the items from memory.
Coil is mono not shielded at this time.
the coil is from Carls specs on the hammerhead.
8" 30 turns of 26 awg enamelded wire.
No TX resistor
Pulse width when I tested it was 2 usec. 100 usec between pulses I think
Damping resistor is a 1 k pot.
do not know cable capacitance but using RG6 coax probably less than 100 pf.
This is all in a breadboard as I did not want to kluge up the pc board with the working circuit.
RayNM

[bbsailor]

Ray-NM

It appears that living in NM and making a PI operating at 10,000 PPS, you are making a sensitive gold machine?

Your 8" diameter, 30 turn AWG26 wire coil I calculated out to be:
602uH, 2.588 ohms, with about a 232us TC (time constant) (602 divided by 2.58 ).

This means that a pulse width of 2us would not let the current grow very much. The pulse width would need to be 232us to allow the current to grow to about 63 percent of max potential. This short pulse width would mean that the peak flyback voltage would not be very high.

With a PPS of 100us, main pulse frequency is 10,000 PPS.

Here is what I would suggest doing to improve your results or expand your experiments.

Make the coil with 21 turns, closer to 300uH.
Add a 39 ohm resistor in series with the coil (between the MOSFET and coil connection) to get a shorter TC.
Run the Pulse Width in the 10 to 20us range.
Try to make a shield to minimize noise pickup by using conductive mylar such as is found in decorative balloons. Aluminum foil is too thick and conductive, and will extend your delay.
Look on ebay for AWG30 Teflon wire to make a coil.
You may need to lower the PPS frequency to the 2K to 3K PPS range to keep the total current draw in the 100 to 150ma range or something resonable for your battery pack size.
Use a low capacitance MOSFET with a COSS of between 30pf to 75pf.

The reason why you can sample at 7us is that there appears to be a lower amplitude flyback voltage with a 2us pulse width, and consequently a lower amplitude of ringing voltage. When you make a new 21 turn coil with the series resistor, using a longer pulse width, the flyback voltage might be a little higher and the damping resistor adjustment would be more critical to sample at the earliest time.

Keep good notes about the value of the damping resistor as you make changes. Lower capacitance as seen by the coil (in the circuit, MOSFET COSS, coil capacitance, cable capacitance and the effective value of the input resistor in parallel with the damping resistor) will allow a higher value of damping resistor, indicating less capacitance that needs to be damped.

You will find that making a sensitive PI design for particular targets, such as gold, is a real balancing act between:

Coil inductance
Coil resistance, including series rsistor
Sample delay
Coil time constant
Integrator ircuit integration time
Main TX frequency, PPS
Pulse width
Current draw

With your new PIC design, you can easily change various parameters and quickly see the results.

I would enjoy hearing about your progress.

bbsailor

[Ray-NM]

Thanks for the comments and suggestions.

I will take better notes when I get to work on it this weekend and try to follow thru on some of your suggestions. I will let you know what I find out. And yes there is gold in New Mexico.
Thanks
RayNM

[Reg]

Hi Ray,

Are you saying you can control the timing while using a basic stamp type micro using basic? Things have sure changed from a few years ago, if you can. Some time back I built the Stuart PI and tried using basic to set the timing. Unfortunately, the inherent delay in the basic made it impossible for me to even come close to sampling properly. I had to do that in machine language to get it to work properly. Now, I was using a 4 MHz micro at the time, so some of the newer and faster ones might work better. Also, I was using a different basic so it might have been slower also.

Reg

[Max]

I agree

Hi,
I think that using a basic stamp is not a good idea in the microsecond scenario. I think one can use a simple microchip PIC16F628 that is really close to the PIC16F84 (used also in the Stuart's project) but with an internal osc and xtal (4MHz) , that minimize external components needed, and a USART port that simplifies in circuit programming using a simple RS-232 interface (a pc serial port) and more flash memory than the old 16x84.
The device can be programmed using C language , very simple using e.g a compile like the CC5X from B Kundsen Data (www.bknd.com/) that is free for non commercial use (students, hobbists, etc) that can be integrated in the MPLAB by Microchip.
Just write the prog in C, then compile and generate the .HEX file and then upload using RS-232 and a free program (like IcProg)

Best regards,
Max

[Ray-NM]

Reg
I am using a Basic stamp2. Not the fastest kid on the block but it allows me to learn how and what I need to control and sample. If I recall it will output a 2 usec pulse that is the minimum width.
Later when I have a much better understanding of what it will do and what it wont do, I can get a much better PIC Here are the specs.
24-pin DIP1.2"x0.6"x0.4"-40oC - +85oC (-40oF - +185oF) **Microchip PIC16C57c20 MHz~4,000 instructions/sec.32 Bytes (6 I/O, 26 Variable)N/A2K Bytes, ~500 instructions16 +2 Dedicated Serial5 - 15 vdc3 mA Run / 50 µA Sleep20 mA / 25 mA40 mA / 50 mA per 8 I/O pins42Serial Port (9600 baud)
Like I said it is a learning tool at this point.
the link is http://www.parallax.com/html_pages/t...tamp_specs.asp

RayNM

[bugwhiskers]

Universal micro for PI experiments.

For a while now I have been thinking about making a PCB with an ATMEL micro that is set up to be used in a variety of PI experiments. The ATMEGA48/88/168 would be the optimum choice for the following reasons.

Very inexpensive.
6 * 10 bit AtoD
Easy to use/replace 28 pin PDIP package available.
Enough spare I/O to make the micro configurable with jumpers and pots rather than the user having to buy the programming tools and learning a new language.
16 bit PWM for generating sound.
EEPROM for storing favoured settings.


The board will be configurable for either N or P channel MOSFET's via jumpers and a dedicated MOSFET driver chip.
Pot inputs will be available for setting TX pulse width, pre-sample delay, sample width 1, inter-sample delay, sample width 2 and battery voltage or threshold.

If there is enough interest I will go ahead and charge just the material costs and postage.

If you are interested or have any suggestions on features and functionality let me know.

It won't take long to set this up as I have done a lot of the work already for my own PI and have nearly 20 years experience programming a variety of microcontrollers.

regards
bugwhiskers

[bbsailor]

Great Idea!!!

Bugwhiskers,

A PIC experimenters board, a great idea. Is it possible to start out making this board be an add-on to the Hammerhead, replacing the main timer and having timing outputs that could interface with the various Hammerhead digital control points?

Once this basic PIC function is made and interfaced with the Hammerhead, a more advanced set of features could be added with forum input.

Can you explain some the chip limits so we can realistically set our initial expectations in the following areas:
1. Power Requirements
2. Potential PPS Frequency Range
3. Potential Delay Range
4. Potential Sample Width Range
5. Potential Secondary Sample Delay Range to fit within available pulse space,

and later more advanced capabilities,

6. A Second Channel
7. Possibility to add ground balancing with variable gain on a second channel
8. Memory required to store power-on defaults
9. Memory required to store user hunting session settings
10.Audio enhancements.


Thanks,

bbsailor

[bugwhiskers]

Hi bbsailor,
I had a quick peek at the Hammerhead design and it doesn't present any problems to the circuit I propose.

1. Power Requirements
4.5-5.5 volts, < 1ma + any current drive to outputs.

2. Potential PPS Frequency Range
At say 8MHz (the fastest speed of the internal RC osc) zero to 30KHz or more(depends on how busy the micro is performing other tasks).

3. Potential Delay Range
Unlimited.
4. Potential Sample Width Range
Unlimited
5. Potential Secondary Sample Delay Range to fit within available pulse space,
Unlimited

6. A Second Channel
No problem.
7. Possibility to add ground balancing with variable gain on a second channel
No problem.
8. Memory required to store power-on defaults
2 bytes(1 word) for each variable. the ATMEGA48 could store 128 sets and the ATMEGA88/168 double that.
9. Memory required to store user hunting session settings
As per 8. above.
10.Audio enhancements.
If the internal 16 bit PWM is used as the sound generator then their are all sorts of possibilities like rising pitch for increased signal coupled with vibrato and or tremelo indicating other aspects of the target.

Today I will CAD a broad overview showing integration in the Hammerhead.
It could replace all the timing chips and associated passive components around IC1, IC11, IC12, 1C5.

regards
bugwhiskers

[bugwhiskers]

UPIM overview

hi bbsailor,

Attached is a broad overview of the UPIM.

The number of I/O (ie keyboards, displays etc) is virtually unlimited if the SPI is used to drive shift registers.

The number of AtoD inputs can be expanded also by multiplexing.

The chip can be driven to 20MHz but that requires an external xtal. My experince has show that the std 8MHz is sufficient with the micro executing delay loops most of the time.

I agree with your idea of a Hammerhead type start with the option of ground balancing. With all key parameters variable by the user via adjustable pots & switches then virtually any coil can be accomodated.

I have added an option called "Pre-amp coming out of saturation". This optional input would allow auto adjustment of the first sampling delay.

The TL426 chip is a MOSFET driver (30 ns to switch 1000pf) The spare gate is used to generate the +5v (above battery plus) negating the 7660.

regards
bugwhiskers

[Tinkerer]

Bugwhiskers,
very interesting project. Would it be possible to make a PIC (Microchip) version?
Tinkerer

[bugwhiskers]

Hi tinkerer,
Long time no see !

I dont have any programming tools for PIC. Also, I have many routines written already for the ATMEL.

At the end of the day the idea is to pre-program a micro that has pots and switches to set all parameters so it doesn't really matter what the chip is.

My personal preference is to use one of the new Freescale chips that is a cross between a micro and a DSP. They have lots of goodies like 60MHz speed, 12 bit A to D converters, inbuilt Math functions etc etc.
The down side is they are very small packages and SMD.

The ATMEL chips I am proposing are standard PDIP packages with 0.1 inch pin spacing.

regards
bugwhiskers