Help with new gold detector design

[dom44]

I am working on a new design for a gold detector based on the tesoro gold demon, I was wondering if any one could help to work out the components to change on the front end amp circuit to make it work with a 32khz transmit signal, that will be crystal locked,
and if there is any way to make the phasing veriable with a trimmer for final adjusting
thanks.

[porkluvr]

It would really help if you mentioned the L and R values of the coil you intend to use, or is that up-in-the air?

Also WHERE did you get your schematic?

1) You show 925Ω across pins 4~7. My copy shows 825

2) You show 84.5kΩ on pin-7. My copy shows 8.45k.

#2 looks accidentally wrong. #3 looks very wrong. What other surprises are in store?

I started to work on this but after a closer inspection it looks like trouble. Me? I already have enough trouble.

[dom44]

This circuit was copyed from a circuit I download from this site, the demon circuit front end is the same as the tesoro lobo, I have built and tested it on the lobo I am working on and does work with a centre frequency of about 13.5 khz I just want to move it up to about 32khz. I have bread boarded a lot of the circuit for this design and looks good so far.
The coil is the same one as the LOBO R=1.33mh and T=10.9 mh just need to adjust the resonant caps for 32khz, and I have a good sign wave from the recieve coil and the transmit coil At 32.768khz I am using 2 gates of a 4093 and a watch crystal driving the output transistor circuit as the original design and I will use the other 2 gates to make a VCO for target signal to vary amplitude as well as frequency on a target.

[dom44]

Sorry have the coils the wrong way round easy to do as with the lobo thread there is a lot of confussion about which one is right, I am using 10.9mh for the recive coil
and 1.33mh for the transmitt coil.

[porkluvr]

Sorry have the coils the wrong way round easy to do as with the lobo thread there is a lot of confussion about which one is right, I am using 10.9mh for the recive coil
and 1.33mh for the transmitt coil.

There you go... I'm glad we agree on that. The Lobo entry in the forum's "Coil Parameters" page is WRONG, WRONG, WRONG, it being backwards and all. It's bad WRONG.

I do not have a LOBO coil to check and that would be very scientific and the correct thing to do before I start saying it's all wrong and having no way to prove it, but here's my reasoning: it would make no sense to have a 1.33uH 11Ω RX coil and 10uH 51Ω TX coil the way it is listed in "Coil Parameters". Why? Because the 51 ohms suggests a lot of energy being wasted in the transmitter. Also, notice the Cx on the schematic... .047uF will resonate very nicely with 1.33uH at just about 20kHz. (What a convenient coincidence.) Also, (and here's the kicker): it would be a crappy system where you have a driven tx tank tuned to resonate with 10.0uH, and then change coils... with the new coil being 10.5uH. (doh?) That is what the data in the table suggests would happen by swapping the small factory coil for the larger one. That would be substantial de-tuning, and therefore I believe the "Coil Parameters" data to be incorrect.

I already pointed out two incorrect resistances in the picture you posted. Changing the 825Ω to 925Ω is not unreasonable, but that 84.5kΩ on pin-7 is not a good idea. It should be 8.45kΩ, or 8k45 depending on your preference, but not 84.5k.

Here's another thing: I've got to tell you that the Lobo and Diablo are designed to operate at 20kHz, not 13.5kHz like you hinted in your second post.

I'm not saying that you can't bump up the frequency to 32kHz - of course you can - but I would like to make sure that you and I are 'on the same page' so to speak. It makes it a lot easier arrive where you want to go, if you know where you started from before you start moving. oy

It's late here, on the East coast so I'd better knock off for the day. I'll work on this tomorrow (err, that's later today).

Note: While looking through my files I found a pdf copy of the Diablo schematic with parts values just like what you posted. However, it does not agree with any Lobo schematic I have on file, and it also does not agree with my other copy of the Gold Demon. I am sticking with what I have posted here (825Ω, 8.45kΩ) but some of those parts values around the rx preamp are likely to change at 32kHz.

[dom44]

I say the amp is tuned to about 13.5 khz as that is the result I got from sweeping a signal at its input from a signal generator while veiwing the signal on a scope peeks at about 13.5khz and drops away lower and higher could be the parts i used but i do remember some one else on the lobo thread saying the same thing.
there are so many versions of that circuit its hard to tell which is right have to experament with it a bit.

[dom44]

And i do agree with you the transmitt coil would be 1.33mh and recieve coil 10 mh doesn't make sens the other way round it would go against 90% of the metal detector designs out there that are VLF, I think a lot of people mean well but as you know in electronics its easy to get it wrong lost track of the transistors i have soldered in the wrong way .....

[Carl-NC]

My apologies on the coil parameters, the Lobo numbers were reversed, and I fixed them.

Dom, do you understand how to calculate RC and LC time constants? All you need to do is readjust these for the new frequency.

[dom44]

thats where fall down a bit I know it is a simple thing for some one who works with spice simulator or who has a lot of experiance in circuit theory but mine is mainly practical being in the hobby of electronics for more than 30 years.
Just thought it would save some time asking, or I could learn to use spice but I don't really have the time is the main problem.
but the coil mix up is no big deal like i said most of us would have guessed it was wrong from the start, I did but the biggest problem with the LOBO i can see is there is not one circuit that is 100% correct or some one has added a reisitor here or changed another there.

[turtlebowl]

Hope you don't lose hearrt mate. Lots of us here in your boat with decades of tinkering but just practical gradual learning. No degrees. Someone will come along with the right skills to make it quicker. Sounds like your pretty handy with an iron yourself.
PorkLover is on it and he certainly is one of the boys that can getter done if he has a little free time.
Do you have a drawing of the "Hot Rock" circuit found on the newer Diablo? That would be a hoot to have on that machine. Maybe someone will share it or PL will whip one up?
I have an interest in your project becuase I too have this machine on my list of near future projects. Toying with either this one or IGSL machine on this site next.
Please keep us posted?

[dom44]

Thanks turtlebowl I am not wired that way if some one says it can't be done just drives me harder to do it, I know its actually easy to do and i would be happy to present this project if I ever get it going right, but I wont put it up till i have proven the circuit or at least the parts i know work.
and no matter what stage it is at, it can allways be added to or improved I have gotten past the stage of just copying some one elses circuits thats what I love about electronics its as much an art form as a rock solid science and only limited by your imagination.

[dom44]

Update just worked out the VCO circuit for the audio and I am very happy sounds just like the minelab verying tone volume and pitch

[Carl-NC]

thats where fall down a bit I know it is a simple thing for some one who works with spice simulator or who has a lot of experiance in circuit theory but mine is mainly practical being in the hobby of electronics for more than 30 years.
Just thought it would save some time asking, or I could learn to use spice but I don't really have the time is the main problem.

You don't need Spice, you have a sig gen and an oscope; you can easily teach yourself some of the basics, including filters. Get one of those plug-in breadboards ─ Radio Shack used to sell them ─ and a bunch of components, and build up some simple circuits. Vary components and see what happens.

Your circuit consists of two band-pass filters in series. They can be considered independent. The 39.3ohm-0.22uF combo set the high-pass frequency at 18.4kHz. You get this by simply multiplying:



Same thing for the 825ohm-0.01uF combo:



except this is a low pass frequency. As you can see, the low-pass f is just slightly higher than the high-pass f, so it's a pretty tight band-pass filter. The band-pass region suggests to me that the transmit frequency is around 19kHz.

The second stage is:





This is a wider band-pass filter, but the high-f cutoff of 13kHz is too low if the transmit frequency is 19kHz. Something looks amiss.

- Carl

[dom44]

thanks carl-nc thats very helpfull I will play with the math when i get a chance have to work it out for 32.7 khz as it is crystal locked no drift apart from groun effects which should enhance the detector a bit.

[dom44]

Audio sample of the VCO

http://hotfile.com/dl/121494312/8255...32900.amr.html