Felezjoo Pi- adjusting automatic delay, A,B and C values to perfection!!

[Nightryder]

Hello everyone!

I'm sharing this because I believe no one should go through the waste of time, money and energy which I had over this!! I also would like to thank everyone who's been making this forum the way it is. THANK YOU from the bottom of my heart. I've been learning a lot from you guys.

Before we begin, I should tell you that I'm not an expert in electronics. I'm more like a learner from mistakes kind of guy

Here we go:

A,B,C numbers have to be at their correct values which will eventually lead to a correct value of the Automated Delay.

A = Ideally 300, however any number between 200 to 400 is fine
B = Ideally 700, but any number around it is ok
C = Ideally 800, but any number around it is ok
Automatic Delay = Ideally 20, but any value from18 to 22 is perfect, and you will get it right once you adjust the A,B, C values.

To do so, follow the below steps in order:

1. Assuming the pcb components are with the designated values, check pin# 23 of the microcontroller with your oscilloscope (not a multimeter) and it should be 4.8V (between 4.75 and 4.85 volts) if it is not, you need to change R8 (1K8. To increase voltage you lower the resistance, and you do the opposite if you need to decrease the voltage. In my case,
   R8= 1.75 K and now voltage is 4.84v. Now you should notice an effect in the "A" value but keep in mind that is not its final value.
2. Replace R9 (1M) with 1 mega potentiometer then turn the device on while the coil is connected to it. And start lowering the potentiometer to increase all of the three numbers but keep an eye mainly on "B" until you get it to its sweet spot. In my case, my numbers were about (170, 580 and 690) and I had to lower R9 to 560K to get (299, 700 and 802) and my amplifier is (LF157N). So if your numbers are higher than their supposed values, Instead of replacing R9, I would assume you need to connect a 500k potentiometer in series with R9 and start increasing resistance to lower the ABC values. (Do this step away from any metal objects even if it is not detected by the coil it still can affect the values)
   
   * Now with the corrected ABC values the automatic delay should be between 18 to 22. And with this, iron discrimination should be functioning.
3. If your device is giving too much false signals, try to lower the sensitivity setting and see if that works out for you (this was my case). If it does not, replace R10 (100 ohms) with a higher value which I do not know by how much since it was enough for me to lower the sensitivity setting, but I would've increased the resistance staring from 150 ohms up 300 ohms or so. I mean until I gain a complete control to get rid of the false signals by using the device settings).
4. Since I did not have to change R10, I'm not sure whether it has a significant effect on ABC values or not. If it does, go ahead and repeat step# 2. Also, double check pin# 23 of the microcontroller before soldering your new resistances.

Now you are all set.

A few words on Felezjoo Pi:

• This project is worth making due to its uniqueness in features and simplicity in components. Thanks to its maker and if he ever come across here I would like to say : Thank you so much Mr. Hamid!!
• Throughout my search on the internet, I found some people suggest to replace R11 with a potentiometer to increase/decrease ABC values and sure enough it does change the values especially "A", but once the detector recalibrates itself every time you turn it on or when you exit the settings menu, the values go back to where they were.
• The pre-amp you are using is not necessarily fake just because you didn't get the right ABC values by using it for the first time. Take it from someone who bought so many amplifiers to the point where he is thinking now to open an LF357 store
• It's preferred to the you make your final coil first, so you can optimize your device to it with the steps above.
• I do not recommend the pcb version I used (see photo attached) because it lacks enough space for the mosfet and regulators to have heat sinks, and for some of the components to not be in touch with each others, which I think it affects the detector performance.

Happy hunting everyone!

[ivconic]

Well done! Very good and detailed set of tips!
Even on one of the first pages of the FelezJoo topic, I did not hide my enthusiasm for the project.
Because I know the 328P architecture and capabilities pretty well.
The author got the most out of that processor.
The work was so well done that even today, after so many years, this detector is still desirable to make, I have done dozen of them so far and I always periodically do another more.
I have always put more effort into making the coil so that it fits the detector perfectly. And that is not an easy job. Many attempts, many failures and some success.
I would always solve the problem of instability or ("no coil") errors by first removing the tantalum capacitor 2u2 and putting in a regular polypropylene 100nF, and replacing the 100 ohms with 180-220 ohms.
If frequent instability occurs with increased Sens; in the menu I would turn off GEB (set it to "Off") and that would alleviate the problem a lot.
Increasing the pulse width to 180-200uS also affects behavior improvement.
But these are all changes from case to case, from detector to detector, from coil to coil.
Frankly speaking; to this day I have not had a universal and permanent solution in advance. Whenever I build that detector; I did not know in advance what would await me at the end of the problem.
But after a little trouble I would always bring the detector to a state of very good usability.
I have to mention that I always make coils with a smaller diameter, no bigger than 40cm. Because I'm not interested in big coils.
And now as I read your advice; makes me want to make one again. This time to draw my pcb on which I will provide enough space for elco with larger capacities, even if I put several smaller ones in parallel.
I agree, that pcb design (which I used most often) has a few questionable details that have always bothered me about layout and space.
I'm busy at the moment, later in the summer I'll do it, this time following all your advice. I am very interested to see the end result.
Thank you so much for such a clear and detailed set of tips, I'm sure it will make many hobbyists here happy.

[Nightryder]

Well done! Very good and detailed set of tips!
Even on one of the first pages of the FelezJoo topic, I did not hide my enthusiasm for the project.
Because I know the 328P architecture and capabilities pretty well.
The author got the most out of that processor.
The work was so well done that even today, after so many years, this detector is still desirable to make, I have done dozen of them so far and I always periodically do another more.
I have always put more effort into making the coil so that it fits the detector perfectly. And that is not an easy job. Many attempts, many failures and some success.
I would always solve the problem of instability or ("no coil") errors by first removing the tantalum capacitor 2u2 and putting in a regular polypropylene 100nF, and replacing the 100 ohms with 180-220 ohms.
If frequent instability occurs with increased Sens; in the menu I would turn off GEB (set it to "Off") and that would alleviate the problem a lot.
Increasing the pulse width to 180-200uS also affects behavior improvement.
But these are all changes from case to case, from detector to detector, from coil to coil.
Frankly speaking; to this day I have not had a universal and permanent solution in advance. Whenever I build that detector; I did not know in advance what would await me at the end of the problem.
But after a little trouble I would always bring the detector to a state of very good usability.
I have to mention that I always make coils with a smaller diameter, no bigger than 40cm. Because I'm not interested in big coils.
And now as I read your advice; makes me want to make one again. This time to draw my pcb on which I will provide enough space for elco with larger capacities, even if I put several smaller ones in parallel.
I agree, that pcb design (which I used most often) has a few questionable details that have always bothered me about layout and space.
I'm busy at the moment, later in the summer I'll do it, this time following all your advice. I am very interested to see the end result.
Thank you so much for such a clear and detailed set of tips, I'm sure it will make many hobbyists here happy.

You are most certainly welcome!! I learned a lot from you.

I hope one of the attached PCBs (which I found on the internet) can save you the troubles of making a new one. I renamed them 1,2 and 3 so it's easier for me to tell you the following:

I only made # 1 twice which I destroyed due to soldering and desoldering too many times throughout my journey in shooting the troubles. It has decent space for components and also it has voltage check points which is crucial for such a device.(see the attached file named parts list regarding voltages) Just keep in mind that it has a little mistake which is R5, such that on the pcb it says 10K but the correct value is 2k2.

#2&3 which I did not use in any of my felezjoo projects; therefore, I'm not sure whether they have any mistakes or not. But they are very interesting to me because they contain a trim-pot which is not in the schematic and I'm curious about it's purpose.

I included a pdf file named parts list, which has notes from the maker himself worth reading at the bottom. Which suggests to use an 18 volts battery rather than a 24 volts one. I personally use a 12 volts charger (at my own risk) at the moment because I just ordered Li-ion batteries and waiting for them to be delivered. Therefore, I had to replace the 12v regulator with a jumber also I had to replace R19 (22k),which is responsible for telling the microcontroller how much voltage is left, with 10k resistor as you know the device is programmed to shut down at 14 volts. Now with the 12v charger, the screen shows voltage is about 21 Volts .

Good luck in making your next felezjoo!

[algan]

Hi, yes the Felezio ground balance does not work as advised by ivconic, it is best not to use it at all, but how do you deal with stones or roman ceramics, the detector reacts like iron and when the discrimination works correctly, the silver depth drops drastically, how did you manage this?

[Nightryder]

Hi, yes the Felezio ground balance does not work as advised by ivconic, it is best not to use it at all, but how do you deal with stones or roman ceramics, the detector reacts like iron and when the discrimination works correctly, the silver depth drops drastically, how did you manage this?

Can you tell me what the ABC numbers are in your machine? (with no metal objects at all are around)
also, what level of discrimination are we talking about here? I mean when you notice the significant depth drop for silver (1 to 30)

you see, discrimination in Pi detectors is a tricky thing. I remember when I was trying different parts and changing things here and there in felezjoo and also in my delta pulse 2 (which has a different approach for discrimination) of course disc did not work with every change I tried, but there were a few times when discrimination worked although it was not my aim because I had bigger issues. I noticed it did not work at same settings every time it functioned. For example, I use one particular drill bit to test disc in all of my detectors, and my felezjoo was discriminating it at 25 in the settings (at this time I noticed a big loss in depth) and after a change or two in the circuit disc for the same bit was at 16 and then at 13. Now it discriminate the same wood drill bit at 4 and I have not tried setting at 3 or 2 just yet. I remember once my felezjoo was discriminating an aluminum lid (about 8 cm diameter) at a certain distance but when I brought it closer to the coil the detector went off.

bottom line is that discrimination can affect on depth especially in the high settings (like 25 in felezjoo). And just because discrimination works at a certain point, does not necessarily mean it is set up right.

I hope this makes some sense to you

[algan]

And for Grund balance, how can I adjust it so that it can work on heavy soils mixed with ceramics and there are no signals of mineralized stones?

[Nightryder]

And for Grund balance, how can I adjust it so that it can work on heavy soils mixed with ceramics and there are no signals of mineralized stones?

again, provide me with some info like abc numbers. automatic delay, measurement of pin # 23.... etc there could be one, two or ten reasons why it is wrong.

also check out pcb named 1 in the attachment above. it will tell you locations of voltage checking points (blue layer)

[Detectorist#1]

Hi algan,
Don't have big expectations from this project. This project is only attempt to make digital version of simple analog PI projects. An other project with victory of the hope over the reason.

[algan]

A315 B686 C775

[ivconic]

For some the glass is half full, for some the glass is half empty. And for those who have never tasted water in a glass; the glass is always empty.
As for me, I said this so many years ago, repeated it yesterday and will say it again today; FelezJoo Pi is worth the effort.
Because it provides much more performance than the vast majority of Pi projects on this forum.
And even when it comes to the most banal reason; amount of money invested in making: FelezJoo is an absolute winner.
Ok, all that aside, let's get back to the meaningful stuff.
It's not that GEB doesn't work, it does, but when the coil is made really very precise according to the author's description.
DISC works even when the coil deviates slightly from the strict requirements.
When I advised to turn GEB off (select Off); I didn't mean always and in all conditions. But only in conditions when it is clearly noticed that there is no benefit from it.
In most cases GEB will work solidly. It's the same with DISC.
DISC works much better, more accurately than say GS4, 5 ....
All this should not surprise anyone, even those whose glass is always empty.
Such people usually never do anything, but that's why they have excess energy to spread negative comments and bad energy on forums.
It should not be surprising because the author himself said that in writing the code for this detector he followed the methods described in the Minelab patent.
The three debug numbers are, according to my understanding, three vectors representing three time constants. Based on which everything else is adjusted in the detector.
After all, I don't have the slightest desire to rewrite here something that has already been said and repeated several times on the original topic.
I just chimed in here to thank Nightryder for the great job.

[algan]

There is still much to be desired from FelezJoo and the experiments will continue there is a glass and there is water, but who will drink water how and whether with a glass or without a glass is his business

[ivconic]

[Nightryder]

Just turned off my detector after a 24 hrs test run. It was stable all the way through. No heat in components or instability in performance what so ever.

settings were:

Freq: 200 ( I changed it to 220 and then 260)
Pulse width: 150 then raised it to 175
Intg. width: 68
GEB: 4 ( after 10 hrs I changed it to 6
Iron reject: 4
speed 7
delay: 0
sensitivity: 17

A= 302
B= 701
C= 799
Automatic Delay= 21

Coil= 390uH , 30 CM Diameter, not shielded or wrapped with an electrical tape, just bare enameled copper (AWG 22)

I will perform other different tests. not for 24 hrs though.

[gurdal]

hello dear masters,
When I made the device for the first time,
I tested it in the field.
excellent at discrimination,
perfect in depth.
The hat I use is 45 cm,
The problem is that it gives an iron signal in red soil.
I probably couldn't tune the coil.
I would appreciate it if you could share your fine calibration information with coil adjustment.

Best regards
Gurdal

[Nightryder]

hello dear masters,
When I made the device for the first time,
I tested it in the field.
excellent at discrimination,
perfect in depth.
The hat I use is 45 cm,
The problem is that it gives an iron signal in red soil.
I probably couldn't tune the coil.
I would appreciate it if you could share your fine calibration information with coil adjustment.

Best regards
Gurdal

Nothing really is special about my coil. If I were you, I would've taken an empty bucket, fill it up with that soil, bring it home with me and start experimenting different settings and maybe a smaller coil and compare results until I find out what's best for me. I hope you overcome this soon.