chance pi build

[dantech]

@thx fo suggestion



Important thing!
I do not agree to distribute my firmware release in any commercial form (eg. Kits with pcb and atmega)

[6666]

Thanks Dantech for your project and its great to see you have improved the chance.

In reading the google translated text of changelog

For system optimisation the switching transistors Q1, Q2, Q5, Q6, Q7, Q8 need to be BC639 / BC640 which are rated at 1 amp.
and D4 and D6 need to be RL207 1000V 2A

R9 = 680ohm

Dantech how much testing have you done on the ground balance ? does it work in low medium or high mineralisation ?
I understand that you have done some testing at the beach, was it dry or wet sand , thanks.

[baum7154]

@baum7154
You say - you have
Tell me something more about 3DSS coil and your idea.


Try this one. The compiler did not report any error so in theory firmware should work well because change was not too important
Keep in mind that v163 was not tested by me at all!

Main pulse delay time range is 1-99us

_________________________________________

The 3DSS coil (3 dimensional winding pattern, Self Shielding) is completely detailed in the CHANCE PI COIL thread of the COILS forum of this site. I do not use the original 400uh coil proposed by FANDY as it is not suitable for small gold and would not operate below about 18us of sample delay. A 3DSS coil of 328uh will operate down to at least 7us which is the sample delay range I use for the smallest gold detection. Take a look at the thread for details on winding the 3DSS coil. Of course the damping resistor will have to be optimized for the 3DSS coil as well, typically in the range of 1000 to 1200 ohms.

So if I understand correctly the main pulse sample delay allowed in your new firmware would be about .9us at a setting of '1'?

Regards,

Dan

[dantech]

Hi I have used Chance on rather wet sand (Baltic sea) and moderate soil in garden. Worked perfectly. Plus thousand tests in my basin.
1 means exactly 1us.

Wait! Conversion takes at least 4us so I have to set lower limit on 5us! For an hour I publish corrected software

[dantech]

I have thought a moment and delay from menu should be the delay before start of conversion.

Test firmware below. Directly from compiler, not tested at all
Look into changelog.txt

[6666]

I have thought a moment and delay from menu should be the delay before start of conversion.

Test firmware below. Directly from compiler, not tested at all

Hi Dantech, the firmware loads and verify's , but I have no way to test it

[dantech]

Remember,
Every release I made above v1.4.0 has new fuse bit settings as below:
hfuse=C9, lfuse=3F

Watchdog function is permanently disabled by fusebit but older firmware should work well also with new fusebits

The same with ICL7660 power control. In my firmware pin 28 of Atmega is driven at start but ICL is not controlled in dynamic way
above v1.4.0.
Original dynamic control of ICL had no chance to work because of capacitors and RC time ratio on opamp power lines.
So that circuit can be disassembled permanently.

[6666]

Hi Dantec, I calculated the correct bits thats good news but

In my firmware pin 28 of Atmega is driven at start but ICL is not controlled in dynamic way
above v1.4.0.

What does this mean , there is no connection to pin 28.

[dantech]

It simply means that pin 28 originally was controlling switching on or off the ICL7660. ICL makes "-12V" for opamp
Like we see there are serial resistors and parallel capacitors in opamp power circuits. So there is no common sense to
switch ICL a few hundred times a second because of energy saved in capacitors. Since v140 controller drives pin 28 at program start
and ICL is enabled whole time without hundreds a second switching cycles.
Entire circuit for ICL control can be removed even if someone uses original firmware. Look onto schematics in my archive witch v162.
In theory dynamic control of ICL was thought to reduce sensitivity for hazards. At the moment Chance is stable as rock
BTW. Chance has original mechanism to measure own internal noise and correct measured signal

[6666]

It simply means that pin 28 originally was controlling switching on or off the ICL7660.

Thanks I can see that now on old schematic.

[baum7154]

Hi I have used Chance on rather wet sand (Baltic sea) and moderate soil in garden. Worked perfectly. Plus thousand tests in my basin.
1 means exactly 1us.

Wait! Conversion takes at least 4us so I have to set lower limit on 5us! For an hour I publish corrected software

__________________________________________________ _

So are you saying that the A/D conversion adds 4us of delay but does not lose any of its input to output signal due to this delay in which case it is no problem ...or you saying the A/D loses the first 4us of its input signal thereby adding additional minimum delay to the sampling? Even so, a 5us actual sample delay will be more than quick enough for very small gold. The challenge at even 5us is to build a coil that damps at less than 5us.

Regards,

Dan

[dantech]

Conversion takes at least 4us but it seems that fact should not interest us
v163 has delay correction to the end of conversion (-4 units) and it is possible to get negative delay value (1us-4us=-3us)
So v163 has bug in this issue.

In my opinion when we use delay to the conversion start the time of conversion is not important at all


Here is verified v164 containing range 1-99us and delay to the start of conversion like in the original v121

[baum7154]

Here is verified v164 containing range 1-99us and delay to the start of conversion[/QUOTE]

__________________________________________________ __________

I did not get the V164 attachment in your last post. Was it provided in an earlier post?

Thanks

Dan

[dantech]

That is it. v164 was attached with 5 minutes delay

[dantech]

@baum7154
I was thinking about shortening of main delay only whilst using faster coils than dedicated ones.
You have to take into account one potential problem.
When you overclock Atmega you will get proportional delay in the whole process.
But we have to remember that
there is 2nd series of 16 pulses and 2nd constant delay. That 2nd pulse series and 2nd delay is used to distinguish metal and balance ground.
So using fast coil and shortening main delay only we can
impair and even lose metal distinguishing and ground balancing features.
Entire process in Chance PI is still optimized for dedicated coils.

So, you have to verify the issue with v164.
I do not have faster coil than dedicated one and I do not have 2 stage preamp also.