Barracuda Legend Kit Problems

[raygdunn]

Hi Qiaozhi,

My new 40106 is working well with all the mods in place. ☺ Audio send is more stable. Many thanks for the capacitor mods. Haven't tried it for movement sensitivity yet. First stage of 412 shows an offset of something like 1.25V. I guess this is likely to show some earth field sensitivity when off balanced from zero. This varies somewhat with the Delay setting. I take it that the ceramic magnet test is better than plain zeroing. Both by pulse width adjustment.

ICL7660S is a different story. Still overheats at 11.8V without 40106 in circuit. But runs cool on 9V (so useable). Other chips the NE5534A and LF412ACN seem to be running normally. I've checked all pins connect through the DIL socket OK. I still have the option to change the 79L05.

I finally figured out the basic workings of the 412 stages. Not bad, once I focused on the op. amps. I now see why the control is called Threshold. I had wondered.

Ray

[Old cart]

I've added your request to the REV-3 change notes.

Would there be any value in moving the battery input terminals directly to the main storage cap? Maybe even putting a 1 ohm resistor in series with the plus battery feed. This, along with referencing all the power supply caps to ground should help supply noise a lot.

[Old cart]

Hi Qiaozhi,

My new 40106 is working well with all the mods in place. ☺ Audio send is more stable. Many thanks for the capacitor mods. Haven't tried it for movement sensitivity yet. First stage of 412 shows an offset of something like 1.25V. I guess this is likely to show some earth field sensitivity when off balanced from zero. This varies somewhat with the Delay setting. I take it that the ceramic magnet test is better than plain zeroing. Both by pulse width adjustment.

ICL7660S is a different story. Still overheats at 11.8V without 40106 in circuit. But runs cool on 9V (so useable). Other chips the NE5534A and LF412ACN seem to be running normally. I've checked all pins connect through the DIL socket OK. I still have the option to change the 79L05.

I finally figured out the basic workings of the 412 stages. Not bad, once I focused on the op. amps. I now see why the control is called Threshold. I had wondered.

Ray

Did you adjust both the main and delay sample pulses to the same width before tweaking with a magnet? There is another way to adjust the 412 sample integrator for better balance. It involves adding a 100ohm trimmer with the two outer pins going to the two fet sources and the wiper going to the junction or r20 and c11 ( KT315 schematic). This involves cutting the trace between the two sources. By the way the schematic mentioned above does not agree with this. The layout has the sources and drains exchanged. This should allow very fine balance of the FETS wither the magnet test and the sample pulse widths equal.

[Michaelo]

One more inconsistency between bara PCB (rev2) and schematic...

On the schematic, both the 78L05 & 79L05 have 220uF caps on inputs (even though both are connected incorrectly)... however, the 79L05 actually has one 220uF connected to its input and another 220uF connected to its output... The 78L05 has no 220uF cap on it input as indicated by the schematic... And neither regulator have the 22uF caps on their outputs as indicated on the schematic...

The input to the 78L05 comes from a battery so chances are it really doesn't require a 220uF cap on its input (good practice is to include one) but more importantly voltage regulators almost always require an electrolytic or tantalum cap on their outputs...

So... in a nut shell, the board arrangement is close with the following exceptions... 1) the fact that the 79L05 has both cap's positive terminal are connected to Bat+ when they should be connected to Ground (Ground is positive with respect to -5) and 2) the 78L05 has no output cap (~220uF)...

Just thought I'd add this for those making changes to the caps and as a reference for next build...

Mike

[raygdunn]

Did you adjust both the main and delay sample pulses to the same width before tweaking with a magnet? There is another way to adjust the 412 sample integrator for better balance. It involves adding a 100ohm trimmer with the two outer pins going to the two fet sources and the wiper going to the junction or r20 and c11 ( KT315 schematic). This involves cutting the trace between the two sources. By the way the schematic mentioned above does not agree with this. The layout has the sources and drains exchanged. This should allow very fine balance of the FETS wither the magnet test and the sample pulse widths equal.

Hi Old cart,

I added a 1nF across the 4.7nF (C9) associated with the Delay preset and pot. This seems to have improved the first sample pulse shape some, but across the adjustable Delay range the pulse width varied from about 44us to something over 50us. This did allow me the ajust the pulse width through the matching point, so I tried the ceramic magnet. I found a minimum sensitivity point around where the pulse widths matched. Interestingly the first stage of the LF412ACN was offset at around +1.3V from earth. One would expeRay zero offset.

I've yet to make it portable enough to test outside again.

Ray

[Michaelo]

I tried the instruction too... set pulses width to matched but I'm still not getting 12" for a UK pound... more like 6" ;(

@Ray, did you notice my last post (2 up) re caps on regulators?
Turns out the -5 volts supply has both the 220uF caps on it for rev 2 board... one on input and one on output... the +5 has no smoothing caps...
I think I may have disconnected the output cap (-5) which would be an issue... tomorrow I'll correct that and try again...

[Old cart]

Hi Old cart,

I added a 1nF across the 4.7nF (C9) associated with the Delay preset and pot. This seems to have improved the first sample pulse shape some, but across the adjustable Delay range the pulse width varied from about 44us to something over 50us. This did allow me the ajust the pulse width through the matching point, so I tried the ceramic magnet. I found a minimum sensitivity point around where the pulse widths matched. Interestingly the first stage of the LF412ACN was offset at around +1.3V from earth. One would expeRay zero offset.

I've yet to make it portable enough to test outside again.

Ray

Ray, I think all this would do is to increase the delay. Caps in parallel just add so 1nF plus 4.7 nF = 5.7nF. So did you match the main and EF pulse widths? That mod I mentioned really just adjusts the signal going to the samplers (main and EF) by adding
Resistance to make the channels match.

[Koala]

What's a good target for testing/setting the delay setting.

If I set for maximum distance on a small gold ring I get poor distance on a small fibula. Where as if I set for maximum distance on the fibula a only loose less percentage wise on the gold ring.

[Qiaozhi]

What's a good target for testing/setting the delay setting.

If I set for maximum distance on a small gold ring I get poor distance on a small fibula. Where as if I set for maximum distance on the fibula a only loose less percentage wise on the gold ring.

That's normal. It depends what you're searching for.

[6666]

What's a good target for testing/setting the delay setting.

If I set for maximum distance on a small gold ring I get poor distance on a small fibula. Where as if I set for maximum distance on the fibula a only loose less percentage wise on the gold ring.

This table was made by somebody else and tells a couple of stories.

[Koala]

This table was made by somebody else and tells a couple of stories.

Cheers thought I had a good compromise. Small hammered coin (14mm) at its maxim 6-7" (solid-faint) and Victorian Penney only lost an inch @ 9.5-10.5" and still had some discrimination. Fibula was still close to its max.

However when I went back to a gold or silver ring I was down to 8" down 3 inches.

Bit more testing needed but ran out of power in the batteries.

[godigit1]

Here are some scope shots from my Barracuda Kt 315 version. Im on the road again and all of my testing has been indoors in bad EMI and also lots of rebar in the floor.
Performance depth is pretty much on par with most other posters. Nickle at 10 inches in bad Emi.
At first I thought the motion non motion did not work but turned out I needed to move the coil 4 feet off the floor due to the rebar. It picks up large metal very well but strangely only has a depth of 4 inches on all my other test targets from gold rings to coins all hit the same.


Tx Pulse



Tx and main sample delay


Sample delay and main sample


Earth field delay and main sample


Earth field sample


Pre amp out and voltage across coil. (maxim 410)


And here is a picture of the peramp out with my dual front end plug in made for the HH (maxim 412). I know its not optimized for the two circuits but boy does it change the threshold to a nice clean hum similar to a Vlf and not hardly effected by the noise in my test environment. Depth response is about the same but for the deeper targets its a threshold worble or drop. Target sample range also changes, Good only from 7.2 us to about 30us no response after that.



This post may be out a little out of place but it seemed relevant to the discussion on the schematics working or not.
I'm still learning and may be missing things but this board seems to work Just build it and power it up and calibrate it.
Went as smooth as my HH Surf pi or Mpp. All of these kits worked from the get go.
And thanks to you all for that!
Best
godigit

[Qiaozhi]

What size battery are you using?

[godigit1]

What size battery are you using?

12 volt, never a blown chip and nothing is getting hot or even warm that I can tell.

[Koala]

Best compromise with delay setting I could find. 12V battery 8.5 coil

~
~

All maxim threshold change while sweeping. Loud beep is approximately 1" less

Can get nearly 14" on the gold ring but then the fibula is down to 3"


If I was only going to use it at the beach I think I would tune just to the rings both would be a solid beep over 10" which is about the max you can easily dig out.


Not sure how low the battery can go before the delay settings alter but at 8.75V there is about 1/2 a turn of the trim POT difference with no warning.