chance pi build

[baum7154]

@baum7154
technically it is not possible make VDI samples gap below 10us (even on 2,7MHz ADC clock). Real VDI samples gap in the v121 was 32,5us. (20us took conversion with 1,7MHz ADC clock + fixed in the code 12,5us)
More details in the changelog.txt[/QUOTE]
--------------------------------------------------------

dantech,

I am a bit confused by the phrase 'VDI samples gap'. Please explain what a VDI samples gap is. Are you talking about the main sample window width or possibly the window duration?

When I look at the scope trace of my V1.2.1 Chance the main sample delay at a minimum Guard Interval setting of '10', the sample delay with the original clock of 11.0592mHz, starts about 9us after the coil begins its flyback. The duration or width of the main sample window is 12.5us, after this the window closes. I do not see any other real delay of 20us on the scope.

Also using Google Translator I am unable to open your link and there is a message that there is no file.

Thanks

Dan

[baum7154]

OK now I am able to read the latest changelog and it appears that the VDI Sample Gap you refer to is the minimum time between repeated Sample Windows but the system will allow Sample window widths of 10us and no less at this time. Is this correct? What is the reason a shorter Sample Window width cannot be defined regardless of the VDI Sample gap?

Dan

[dantech]

3 issues:
-main sample delay. This is the distance between main pulse and sample in the odd cycle. It is clear...
-vdi sample additional shift. This is additional delay between main sample and 1st vdi sample. In the v121 it is fixed at 20us. So 1st vdi sample occurs 30+20=50us after main pulse
-vdi samples gap - like we know we have 3 vdi samples. The gap is simply equal distance between three vdi samples.
Sample conversion in the v178 takes 10us so it is technically not possible get distance between samples shorter than 10us.
Samples made one by one without any delay will have natural delay 10us. In the even cycle we make 3 vdi samples to get something like "angle of vdi slope".
In the odd cycle we make only one sample so time of conversion is not important for us.
In the even cycle it is important. Look onto rising edges of ADC clock signal in the v121. You will see 28us distance.
This is the sum of conversion time and 12us fixed delay in the code. It means that the ADC conversion in the v121 takes 28-12=16us

Is it clear now?

If phrase "vdi" sample gap" is not proper tell me how should I name it to be more clear.

[baum7154]

If phrase "vdi" sample gap" is not proper tell me how should I name it to be more clear.[/QUOTE]
-----------------------------------------------------------------
Since 'VDI Sample Gap' is a new phrase and I had not encountered it any where else in these forums I did not get your meaning. Now I understand what you mean and the phrase is no problem.

From the perspective of small gold detection capability, it would be good to have selection of the sample window widths of 5us, 7us, 10us, 12.5us not including sample conversion time.

Thanks

Dan

[dantech]

Like I said the gap below 10us is technically not possible. Not with 11MHz crystal. In the menu you can set the vdi gap in the real range 10-40us.
Remeber that in the v121 the real gal was 28us. So this is great improvement. Excluding conversion time this the range 0-30us

[baum7154]

Ahhhhh now that's what I like to hear! Perfect !

[dantech]

http://chomikuj.pl/dantech/Chance+PI

v179
There is only one change of one parameter in the digital filter. I need to know how v178 and v179 differs in working and which one is better.

At the moment I have removed from my PCB entire Q5/Q9/D6 circuit and replaced D4 with thick wire (simple shortcut).

Replacing D4 by wire strenghten signal a little bit on oscilloscope so I have decided to replace one.

The Q4 circuit I have removed much earlier.

Now on my PCB is so empty


@Baum7154
I have tested on v179 and dedicated coil thing you asked me.

Square aluminium can (from beer) 6x6 mm in size is detected from 1 cm and 10x10 mm from 5cm in the air. Signal mask was default (it means 3 segments)

[baum7154]

http://chomikuj.pl/dantech/Chance+PI

v179
There is only one change of one parameter in the digital filter. I need to know how v178 and v179 differs in working and which one is better.

At the moment I have removed from my PCB entire Q5/Q9/D6 circuit and replaced D4 with thick wire (simple shortcut).

Replacing D4 by wire strenghten signal a little bit on oscilloscope so I have decided to replace one.

The Q4 circuit I have removed much earlier.

Now on my PCB is so empty


@Baum7154
I have tested on v179 and dedicated coil thing you asked me.

Square aluminium can (from beer) 6x6 mm in size is detected from 1 cm and 10x10 mm from 5cm in the air. Signal mask was default (it means 3 segments)

---------------------------------------------------------
Congratulations, I am surprised that the 6X6mm target is seen at all with the original coil and single stage preamp and at anything over a 10us Sample Delay. If you could get that original coil to sample at 7 or 8us I'm sure the distance would improve.

Did the VDI indicate Aluminum/Gold?

The loss of D4 series capacitance by removal will allow the significant capacitance of the Mosfet to come into play in the coil circuit...it will slow down the in-circuit coil performance.
\
If you put D4 back in and test the 6MM X 6MM Al target does the distance improve?

Can you put the Signal Mask ('Barrier') down to '0' or '1'? Where is the sample delay set?

For reference my current Chance PI with 2 stage preamp V1.2.1 and a 328UH 3DSS coil overclocked and Sample Delay at 6.98us sees a .25" x .25" AL can target at 2.75" or about 68mm.

One more thing this test is with the coil resting on clean soil, not in air. To see complete data on these tests see post 334 in this thread.

Thanks,

Dan

[dantech]

I do not think that the D4 has significant capacitance issue because I can define my coil speed by means of main pulse delay setting in the menu.

There is no significant delay shift after the D4 removing. Maximum 0,5 us - less than available settings resolution.

BTW. Your 6,98us delay is in fact the distance between conversions and does not inculdes conversion time. In your case (overclocked ATmega) estimated conversion time is 10us.

So, real vdi sample gap is ... 6,98+10=17us!

My relase allows you to get the smallest gap 10us

[baum7154]

Hi dantech

If you put D4 back in and test the 6MM X 6MM Al target does the detection distance improve?

This picture was taken at the original single stage preamp out and A-D pin 5 with the firmware version 1.2.1. It shows a properly damped coil response and the 12.5us A-D sample window to the upper right. Sample Delay in this picture is about 8 to 9us. This picture also shows that the single stage preamp saturates and has a long recovery time of about 16us from start of flyback to the flattening of the response curve. This has been resolved by replacing the single stage amp with a 2 stage amp.

P1060274.jpg

The 10us conversion time you include is a constant that is always applied to both the flyback signal and to the target response. The Sample Delay is a relative measurement from the BEGINNING of the coil flyback pulse to the leading edge of the A/D sample window. If we set the Sample Delay at under 10us we have a much better chance to see target responses from small gold and aluminum test targets. If we set the Sample Delay (Guard Interval in the V1.2.1) to 20us we have far less chance to detect the rapid target response.

I hope this helps.

Regards,

Dan

[dantech]

There are no noticeable differences in working (with D4 or not)

[baum7154]

What is the shortest Sample Delay (Guard Interval) that your Chance will operate at with the dedicated coil and the latest firmware?

Are you able to run the Signal Mask down to a setting of 0-1? Is your Signal mask the same as the 'Barrier' selection or the same as the signal threshold before audio is allowed?

In using my Chance PI I generally do not set any threshold but run it at full sensitivity and it is generally pretty quiet in that mode with only the occasional chirp.

[green]

A chart I posted in another thread awhile back. The diode allows sampling sooner. The diode drop caused less peak current so the signal was less. Tx time could be increased for same peak current with the diode. The diode may cause more or less peak current drop with the Chance circuit. I wasn't thinking about the diode drop causing a problem before I ran the test. With the same peak current sampling sooner makes a difference with the 3usec TC target not the longer TC target.

[dantech]

My dedicated coil has 28us.

OK. I have restored D4 . I use UF4007. One has 25pF instead of 75pF (MUR460)

[baum7154]

Green, do you recall the TC of the 6mm square of AL can sidewall? Was it the 3us or even shorter?