TD3X ( pi )

[DOOLEY]

ah ha,

thank you George !!



I shall have a ponder over it.

[Qiaozhi]

garrett XL500,
don't think ive ever seen the schematic for that,
if you have it available could you post so I can look ???

See attachment.

[DOOLEY]

thank you IVCONIC for your kind words and praise.


and there was me just thinking it's performance was "average"

the only other Pi ive built was the blind squirrel !!

garrett XL500,
don't think ive ever seen the schematic for that,
if you have it available could you post so I can look ???

[Orbit]

Is there somewhere pcb for this detector ?

[eclipse]

I will keep it simple, went back to the 4V, will probably try both options (with additional 20V, this way I get to keep the original design). Added 300V snubber.
Increasing C39 (snubber cap) decreases the voltage limit. Example for 400V cap should be 100nF (I don't fully trust the simo, correct me if I'm wrong).

Cap should be a "snubber" cap. Example: http://www.mouser.com/ds/2/88/940C-25653.pdf

TD3XM-1.0b.pdf

[ivconic]

It may be quite simple to add a snubber to the design and observe flyback top voltage there. It may do good in motion mode.

I could throw few ideas... but i will not.
There is one simple and nice idea, not so obvious on first glance... some of you will discover it soon.
Just make it first and play a bit with it.

[Davor]

...Now I propose a BRAINSTORMING to add discrimination to the design...

It may be quite simple to add a snubber to the design and observe flyback top voltage there. It may do good in motion mode.

[ivconic]

I don't think my detector design anything too radical.
...

Actually it is... speaking from the standing point of dedicated enthusiast, amateur and hobbyist.
Let me explain.
We have here on forum numerous similar "small" and "simple" (i completely disagree with this idiotic attitude from some "gurus" here on forum) PI projects.
I've been here for 9 years or so, i remember Garret XL500, Delta Pulse, dozen derivatives of those two, Surmaster PI and lately Barracuda.
I've made them all. I made 2 XL500, 72 Delta Pulses (yes you see it right, not a typo), 2 or 3 Hammerheads, 3 MPP's, 11 Surfmaster's and some 20 to 25 Barracudas.
I also made 20-30 other PI models that i can't recall right now.
Quite a list? Also quite experience with those "simple" and "small" projects? Right?
So that's giving me enough credit to come here and say that you did splendid job.
Because your PI detector is SUPERIOR over all mentioned above on small targets. Especially on very small targets. Not to mention that it performs SUPERB on medium sized and large targets.
Before this; Barracuda was my favorite. Another splendid design. Now this TD3X is my most favorite.
It beats Barracuda and others on small targets just like that. And i must point; properly made Barracuda is a hell of machine!
TD3X is better now.
And that's something i've been awaiting for long time; nice diy PI machine, simple enough to understand and to make. Ease of use, ease of adjustments, splendid and informative audio.
And most important; everything is so straightforward; no "space psychology" involved.
Most important; another proof of my claims from above is huge interest and attention by some rare and valuable members that we are having here on these forums. Right people gathered on the right place.
Vactrol question here is indeed very interesting! It was nice touch! It itches me to experiment more in future around it.
But... enough from me by now.
Cheers!

[Davor]

Perhaps a better idea is to connect something like 100 ohm to 4V and, say, 100uF to Batt+ (AC short), and use that as a reference. KISS principle.

[eclipse]

If R6 goes to 4V, R57 should do the same, but with coil referenced to Batt+, both should be referenced to batt+ as well, so they should both use the same reference that is AC shorted to Batt+. Of all voltage referencing options, red LED is the least noisy option, and zeners are the worst. Also C14 value may as well be 10 times lower and still work the same, but the settling time will improve.

You're right about C14 - in the doc I mentioned AN-581 is pointed exact value as you as your suggestion, 0.1uF-0.3uF for this setup
I will then connect R58, R6, R8 to Batt+

[Teleno]

Davor, I did as Teleno suggested, replacing Q4 with different types in LTSpice and noticed only vary minor changes, but I run all my sims at 11u time step.

You will notice a radical deviation from the target of 50% duty cycle, which was the aim of my design.

Now I propose a BRAINSTORMING to add discrimination to the design.

Let's what happens to the voltage at the coil when the target is magnetic and chages the inductance of L1 by 10% from 400uH to 440uH:





We see that if we sample the ramp-up voltage at about 70us the difference between ferrous/non-ferrous is 400mV. A change of 1% in the inductance would give about 40mV.

We can sample this voltage and pass it through a differential integrator with two time constants (slow and a fast) to detect this change in L1, indicative of a magnetic target.

[Davor]

If R6 goes to 4V, R57 should do the same, but with coil referenced to Batt+, both should be referenced to batt+ as well, so they should both use the same reference that is AC shorted to Batt+. Of all voltage referencing options, red LED is the least noisy option, and zeners are the worst. Also C14 value may as well be 10 times lower and still work the same, but the settling time will improve.

[eclipse]

Well I've taken completely working schematic and turned into... something that probably won't work

If I understood correctly... decoupling must be something like this:

TD3XM-0.9.pdf

The 6.2V zener, resistor and cap - ZENER DIODE BIASING (AN-581), suppose to add stability.

No more changes will be done, I only need to get the preamp in working state, to be honest I rely on you guys to tell if it looks okay, I appreciate it.
I plan to make some simos on the changes but the online simulator I use circuitlab.com is making alot of mistakes, so it's not so reliable

[DOOLEY]

I don't think my detector design anything too radical.

very similar to most Pi's,
555,mosfet,coil,diodes etc etc......

notable differences :
fast (ish) dual stage pre-amp,
variable pitch audio,
different earth field minus system ( actually utilized from another detector )
suppose the difference in sensitivity due to the op-amp/comparator for final detection.

just had a idea of doing a few things a different way to the norm .

looking back, suppose my key criteria were :

economical on power,
good detection distance,
pleasant informative audio ( I wasn't going to spend a day listening to scratching crackle and ear piercing clicks n screams !! )

and much admiration to ivconic for producing HIS TD3X

[Thomas]

WOW, ivconic that was impressive, I agree that Dooley has done a splendid job with his detector and helping with the mods thrown in, what a guy. I for one am glad he decided to share it with us, its all good.
Davor, I did as Teleno suggested, replacing Q4 with different types in LTSpice and noticed only vary minor changes, but I run all my sims at 11u time step.