'KROT2-XM' (mole), new project Xmega32

[ivconic]

Continuation of this very interesting topic will be at Serbian-Metal-Detecting-Forum where Ivica is main technical administrator:
http://hobidetektori.serbianforum.info/
Good health and Happy Prospecting
all regards:
Don Corleone Caffe De La Skobalj
REPUBLIKA SRBIJA

No i am not. I am not member of that forum.
You confusing me with somebody else.
I don't visit that or any other local forum.
I don't follow those forums nor i have any interest in those forums.
If there is somebody on those forums using the nick "ivconic"; be sure it is not me.
I am the member of Geotech forum only.
...
Ne nisam to ja. Ja nisam član toga foruma.
Pomešao si me sa nekim.
Ja ne posećujem taj i takve forume.
Ne pratim te forume niti imam interesovanja za te forume.
Niti imam pojma šta se piše i dešava po tim lokalnim forumima.
Ako se neko tamo pojavljuje pod nick-om "ivconic"; znaj da to nisam ja.
Ja sam član samo ovog Geotech foruma.

[Koala]

How to boost KROT (MOLE) targets recovery speed ?

Lets see what recovery speed is on example of Makro Racer 2.

Ok what is wrong with MCP 601 ?

1st. Very slow Slew Rate 2.3V//µs or in prospecting term slow target response. This simple means that small signals from target in the soil will be very probably dispersed in the highly mineralised soil before front end receiver will even handle them.

2nd. Very miserable Gain Bandwidth Product of only 2.8MHz. In prospecting term this means that for very short in time and small signals from targets we will have very small gain of Front End receiver because as signal is shorter for given open loop gain amplifying is smaller. It further means poor detection depth. These are MCP 601 engineers design restrictions.

To solve problem we need Rail-to-Rail OP with extremelly high both Slew Rate and Gain Bandwidth Product. Proble is easy solvable and depend only from money. There are many excellent high speed rail-to-rail operational amplifiers on market.

For example Texas Instruments High-Speed Op Amps >=50MHz
http://www.ti.com/lsds/ti/amplifiers...-products.page

We need to find properly one "GOLDEN OP" !

Does anyone try these tests for themselves


Coin too big. Iron to small. All my older analogue detectors pass this test easily using in the field settings. However if you use a larger rusty forged nail and a cut half penny you will get a completely different result.

Also seen a funny video of someone digging a deeper and deeper hole in the wet sand placing a coin in it to show how good there detector works on the wet sand. In the end they are something like 18" down with the sea filling the hole. I would put money that if they left the coin out of the hole it would still register.

Its easy to exactly match the discrimination level to small iron target so any slight difference gives a good sound. Great for a YouTube video nail test (or nail board test) not so great if you are digging a lot of iron in the field.

Separation : how close to iron you can detect is pretty much down to the coil not the machine
Recovery speed : only affects how fast you can swing.

[Geo]

I changed MCP601→OPA1611

What about new results??

[ujosvec]

What about new results??

improved sensitivity

[scrungy_doolittle]

I think you are missing something.
Hex files are ascii encoded, so they are more than twice as big as the binary.
so a line like this:
:10567300300020362E30300020362E3530002037D3
takes 44 bytes to encode 16 bytes of flash.
D3 is a checksum
:10 is the length 5673 is the address.

[ujosvec]

Hex files are ascii encoded

All right!!!
HEX not ASCII
built ten KPOT2 and no problem

[kt315]

guys, i have not relation to the project,
all infos about his new 3-freqs http://www.radiodom.ucoz.ru/index/md..._khm_quot/0-20
and again - write DIRECTLY on his email if you get problems radiodom2010@mail.ru

PLEASE DO NOT ASK ME ABOUT THIS PROJECT

[ivconic]

I think you are missing something.
Hex files are ascii encoded, so they are more than twice as big as the binary.
so a line like this:
:10567300300020362E30300020362E3530002037D3
takes 44 bytes to encode 16 bytes of flash.
D3 is a checksum
:10 is the length 5673 is the address.

Shame on me! You are absolutely right!
What was i thinking!
Ok here is the case:
K3_48_Turbo.hex with size 88.466
when stripped to bin gives
K3_48_Turbo.bin with size 31.435
...
Right in a range!

[ivconic]

...PLEASE DO NOT ASK ME ABOUT THIS PROJECT

Nobody is asking you!

[kt315]

asking in PM. i do not want quote them here for privacy.

[ivconic]

I was thinking to try to make it and see what is all about.
Looks interesting enough.
Yet i can't easy obtain damn MC34119 in local shops.
So i was thinking to try it in "phases".
First phase would be to transfer it to Arduino Micro and see the behavior, measure some signals etc.
Next phase is to omit damn MC34119 and put there something else.
MC34119 is actually good choice because of the 180 degree out of phase outputs which can later introduce more delicate coil, i really like that part!
But i was also thinking to try more simpler approach ... at least for the beginning. Until i obtain several pieces of MC...
It would be easy to mod it anyway i want, actually.
But... just skim looking on original schematic is horror to me! Nasty, sloppy, ugly and pretty stupid way of presenting such schematic.
Especially if you want to manipulate with it in different software than author did.
So... first phase would be to "transfer" that piece-of-crap-schematic to something more convenient. More easier to manipulate with.
Next phase would be to remove Atmega from schematic and re-wire Arduino Micro.
Etc... etc...
So far i done only first step.
Here is my trivial contribution to this story:

[scrungy_doolittle]

I would caution you about trying to convert this from a xmega to a mega. The X mega has a lot of peripherals that the mega does not, and it has a cross bar selection mechanism to map pins to peripherals.
It also has, differint ADC/DAC features.
That being said, I've been wondering for a long time, what the heck this guy is doing with DAC0 going through R14 to the op amp. what does that accomplish?
Obviously, he is applying a variable voltage to that circuit, which will cause it to operate differently, but what is the functionality he is trying to achieve here?

[scrungy_doolittle]

Personally, I would like to take this circuit and replace the XMEGA with a PIC24FJ256GCxx part. It has a 16 bit ADC at 60K samples, as well as a 12 bit ADC at 10 mega samples, with auto oversampling.
In the case of converting to a PIC or arudino, one would have to write the code from scratch, unless the author would be willing and able to port it, or release the source.
The XMEGA and the arduino are not binary compatible.
Perhaps you could walk through his circuit and try to explain what each block is doing.
https://www.maximintegrated.com/en/d...ex.mvp/id/3284 these are alternatives.

You can get the MC34119 from here. http://www.ebay.com/sch/i.html?_from...34119&_sacat=0
http://www.ebay.com/itm/MC34119D-Ori...UAAOSwPe1UIJio
Here is a modification to USBASP to program the XMEGA. These are cheap off ebay, but don't allow single stepping/debugging.
http://szulat.blogspot.com/2012/08/a...r-for-050.html
http://www.ebay.com/itm/AVR-10-Pin-USB-Programmer-3-3V-5V-51-ATMEGA8-w-Cable-adapter-USBasp-USBISP-/400763460232?hash=item5d4f5d1a88:g:CZwAAOSwstxU-RRp

https://ketturi.kapsi.fi/2013/05/pro...sbasp-avrdude/


PA4 and PA5 seem to be assigned as analog inputs, but I don't know what he would be trying to do with PA4..... or is he using PA4 to play with the input signal?
Without knowing how he has configured the xmega internally using the cross bar, you are pretty much guessing...

[ivconic]

Those are good points.
I was wandering if it would be possible to use Atxmega code on Atmega.
(http://www.avrfreaks.net/forum/usb-d...vrs-atmega32u4)
Why would i wish to do something like that... you may ask.
Simple reason; Arduino Micro is based on ATmega32U4.
Having all the things onboard, even power regulator.
DIL compatible.
Saves me from lot of torture from manually soldering such small package.
Shorter; much easier and comfortable to play with.
I already ordered few Arduino Micros.
Heck... if code will not work on them; i'll do something else with them.
But than i will have problem to solder Atxmega on diy pcb...
...
Good links! Thanks!

[ivconic]

PA4 and PA5 seem to be assigned as analog inputs, but I don't know what he would be trying to do with PA4..... or is he using PA4 to play with the input signal?
Without knowing how he has configured the xmega internally using the cross bar, you are pretty much guessing...

Yes those are interesting questions too.
But mostly i was wandering about PA6?!
Only thing comes to my mind is some sort of TX power sensing&regulating... through the ground pin?
...
I guess PA4 is taking analogue signal, buffered and amplified (low gain obviously) in ordodox manner.
While PA5 is using voltage taken raw directly from coil and used for "digital" coil balancing, displaying it as residual voltage on coil, additional RX gain control etc...
Really not sure, not payed much of attention on various video demonstrations, not sure what exactly is going on on LCD during initial setup.
Once i make it and perform detailed checks; i guess many things will be clear.
...
It would be easy to distinguish about PA5 role there. Once device is made; check the LCD data (coil) with and without PA5 signal and see the differences.