Sucessful Hammerhead project (finally)

[Kevco]

I had been having several problems, and then it looked like one of the CMOS 74HC221s was acting badly. But I had no extras to swap with it (all chips are socketed for now). I replaced those two 74HC221s with newly purchased ones, and now I am getting proper timing signals.

The detector is working in a "noisy" house. It's sensitivity seems unimpressive now, but I have not optimized the coil's damping resistor yet, and the coil is not a "fast" one. I used a 1k ohm for damping now. The coil was just thrown together to get it running - and it's self resonant, undamped frequency is 277 KHz. I'll make better coil(s) after the main circuit is working.

Although I said it's sensitivity is "unimpressive" now - it seems promising. I can detect large metallic objects (i.e. 10 cm x 20 cm Al foil) at about a meter distance. Coins and rings for now only can be detected at about 10 cm (but I do not have the timing variables optimized either). I expect mainly improvements from here.

I plan on posting pictures and how well it performs, after fine tuning the coil damping, and then mounting everything for field use, in a shielded enclosure.

Kevco

[dfbowers]

I had been having several problems, and then it looked like one of the CMOS 74HC221s was acting badly. But I had no extras to swap with it (all chips are socketed for now). I replaced those two 74HC221s with newly purchased ones, and now I am getting proper timing signals.

The detector is working in a "noisy" house. It's sensitivity seems unimpressive now, but I have not optimized the coil's damping resistor yet, and the coil is not a "fast" one. I used a 1k ohm for damping now. The coil was just thrown together to get it running - and it's self resonant, undamped frequency is 277 KHz. I'll make better coil(s) after the main circuit is working.

Although I said it's sensitivity is "unimpressive" now - it seems promising. I can detect large metallic objects (i.e. 10 cm x 20 cm Al foil) at about a meter distance. Coins and rings for now only can be detected at about 10 cm (but I do not have the timing variables optimized either). I expect mainly improvements from here.

I plan on posting pictures and how well it performs, after fine tuning the coil damping, and then mounting everything for field use, in a shielded enclosure.

Kevco

Kevin,

I built a HH1 a few years ago. You should get be getting at least 30cm for a gold wedding band. In my basement, I get a clear signal at 36cm on a gold band if everything is adjusted right. (I will post a video for reference if you would like). I'm not using a "fast" coil either. That should give you something to compare your results to. Component selection is important as well as shielding material.

Let me know if I can help.
Don

[Kevco]

Thanks Don,

I wouldn't mind seeing the video, if it's already out there and no trouble.

You must have a huge wedding band Just kidding!

My band is 3mm wide, and after the prior posting, I was able to get close to 20cm for my ring.

It might be nice to know what variables you are using, such as Xmit PW, Rcv PW, Pri delay, Sec delay, and freq.

Here's more details of what I've got, in case you see any problems...

I'm using HH1, ver C, option 1 (using VCO audio).

I used 1% metal film resistors throughout, and poly for most caps, except tantalum or electrolytics for the larger caps. I panel mounted six pots, which have hookup wires going to them, instead of PCB mounted:

Volume / On /Off
Threshold
Xmit pulse width
Rcv pulse width
Primary delay
Secondary delay

I used CMOS for the TL072 (TLC2272?), for longer sample & hold for static searching. And I used an LT1054 instead of the 7660 for voltage doubling.

I used LessEmf.com's "canopy Mesh Fabric" for shielding, which is a woven 4 mil copper wire screen, which is tinned. It's mesh is finer than normal window screen. I would have used Scotch 24, but I don't use ebay (where most of scotch 24 is from), and ended up using this instead. They claim 40 dB attenuation as shielding, from 1 MHz - 2.5 GHz.

I'm using an 11" mono coil, wound with PVC insulated solid strand copper wire, with a plastic wire loom between it and the shield. It was just to get this detector to run, before I use teflon for higher quality coils. And I'm using Mogami #2947 coax (18pf/ft) as feedline for the coil.

TNX!

Kevin

[dfbowers]

HH1 Video

Hey Kevin,

Attached is my attempt at a video. It doesn't look that good but you can get the idea.
It was really tough to get the sound to pick up from the headset, but the HH1 will indicate a target the size of a gold band at around 35cm that I would definitely consider digging. The audio is modulated since I did not do the VCO option
http://www.youtube.com/watch?v=Lx47Fmt8E_A

One thing I did notice.. A noisy environment had a very detrimental effect on testing, just like we found out with the TGSL project!! So, maybe nothing wrong with your setup at all. Your setup seems pretty reasonable.

Other differences are that I used a TL072CP and I did not use the VCO option. I also built an 11" mono coil and used metalized mylar for shielding.

The only trouble I had at all was that during my first build I did not use a CMOS version for the TLC551CP. The HH1 worked, but not that impressive.

Settings that seem to matter the most:
* Minimum delay
* Maximum pulse width
* Maximum sample width

I have used it a couple of times a the beach and for relic hunting and it holds it's own.
I just got tired of digging staples, nails, sparkler wires and nuts and bolts!

Don

[Kevco]

Hi Don,

Thanks for the pictures and video.

Looks like you moved many of the pots to panel mount. I didn't move the sensitivity and frequency pots to panel, but wondered if I should have.

I have seen a TDI that was adjusted properly, to reject most junk, and zero in on silver & copper coins. It wasn't perfect, but far better than I had first thought a PI could do. We would compare results against my own White's analog detector, while checking the park.

I have made two observations on mine.

1) When I was getting ring detection at 10 cm, was when the coil was placed about a foot above where a ventilation duct was located under the floor

2) I noticed the noise increases in the evening, when the fluorescent lamps are lit up in the house. But during full sunlight, without fluorescents, and without TV, it's operation is more stable.

And, I did do some static distance (SW2 open) tests with my wedding band, and I can get it to detect that just over 1 foot, which is not far from your 36 cm tests.

So I agree, it may be OK. I am in the process of rounding up all of the misc hardware to use it outside. Nylon hardware is not as plentiful locally, as it once was, sadly. Looks like PVC pipe fittings will be the frame

Thanks for your input Don. It is appreciated!

Kevin

[Kevco]

Photos & results of HH1, Ver C, option 1

Below are various photos, showing the control panel, main enclosure, coil, frame support, etc.

Next post I will list some air test results.

Kevco

[Kevco]

I did do some air tests, indoors, with specific US coins. All were done in static mode, with SW2 open. Here's my results:

Gold wedding band: 33 cm for sure, but at 36 cm I can tell it is beginning to detect it.
Silver Dime: 26-28 cm (Clad dime about the same distance)
Copper Penny: 27 cm
New Copper Zinc Penny: 28 cm
Quarter (silver & clad): 28-29 cm
Silver dollar (1921): 33 cm (had a jewelry clasp around it - may have affected it)

Although I was doing bench testing, I doubt that the threshold I was hearing for these coins and rings, might be too low to notice in real world conditions, as the coil moves across the ground... This circuit seems to like gold better than silver

For the above targets, I used the following parameters for using it:
Xmit Pulse Width: 60 uSec
Sample Pulse Width: 5 uSec
Main Delay: 30 uSec
Sec Delay: 100-150 uSec
Freq (half way), about 1300 Hz (did not experiment with changing this - yet)

I provide those results for others who are considering the HH project, or are building one.

[Qiaozhi]

For the above targets, I used the following parameters for using it:
Xmit Pulse Width: 60 uSec
Sample Pulse Width: 5 uSec
Main Delay: 30 uSec
Sec Delay: 100-150 uSec
Freq (half way), about 1300 Hz (did not experiment with changing this - yet)

Did you really mean to write 5us, or should that be 50us?

[Kevco]

Yes, I was using 5 uS sample pulse. I did not vary that sample duration much, and spent more time trying the other variables, hoping to find optimum settings for different targets.

Why, does the HH work better using longer samples?

Kevco

[Qiaozhi]

Yes, I was using 5 uS sample pulse. I did not vary that sample duration much, and spent more time trying the other variables, hoping to find optimum settings for different targets.

Why, does the HH work better using longer samples?

Kevco

The HH sample pulse width can be adjusted from 3us to 50us. Try using a value near the middle of the range and see if you get any improvement.

[Carl-NC]

Why, does the HH work better using longer samples?

The "effective" gain of the sampling integrator is proportional to the sample time, so a longer sample time will usually give better results. I tend to use 20-25us.

[Kevco]

Thanks Carl, and all!

Increasing the sample period added an inch or two, or more (depending upon target size), on it's depth or distance sensitivity!

Kevco