Goldscan Pulse IV

With regards to the battery meter, I believe it must be a 500uA (microamp), not mA. It looks like it's measuring the voltage drop across the -12V regulator, so with the battery full it will have 2.4V across the meter and resistor.

With the 8.2k (8200 ohms) resistor, even with negligible meter coil resistance, this will give only 293uA through the meter. Less if the meter's resistance is included. So, a 500mA meter wouldn't even notice the current.

Conclusion:- it must be a 500uA meter. A 250uA would probably be better, maybe with a resistance tweak.

All that's trivial though; it hardly stops it detecting metal.

I'm working on a parts list. Got all the parts listed but need to check which are SMD from the pcb diagram now.

Linkback:

You are not allowed to view links.
Please Register or Login

http://www.thunting.com/smf/index.php/topic,6515.msg45635.html#msg45635

Hi Chudster,

First, wrt the parts list, if you check the .pdf on page 4, there is a diagram showing placement of the SMD parts on the copper side of the board. I only realised this when looking on the board diagram (page 2) for the guard rings. I noticed there were a lot of little square pads hanging around... then I noticed the lack of passive components. Mostly it's resistors but a think a few capacitors as well. Just need time to sit down and focus for 30 minutes...

You might well be right about a higher split rail being a good thing. I've got a box full of different power regulator type chips: charge pump, switching etc. I'll have to take a look and see if any can be rigged to do the job better.

The 555 looks to be oscillating at 107KHz. There could well be some noise at this frequency being coupled into the rest of the circuit. Given that we are looking at events in the 10's of microseconds range, any noise at that frequency could be quite disruptive. Plus there is potential noise from the monostable switching. I would certainly agree with some extra decoupling at the front end opamp and switching circuit. I would like the switching circuit in a little tin box. And the front end in another.

I hadn't even noticed the 47 ohm resistors in its supply lines... never seen that done before!

The 555's will definitely be better swapped for something more modern anyway.

Trouble with me is I get completely obsessed with the accuracy, precision and purity of stuff, to the point that I often spend so much time 'improving' something that I never actually get a finished article. I used to design ultra high precision instrumentation circuits, but it's so long ago that I've forgotten most of the stuff. I also tend spend weeks solving some particularly interesting or difficult design issue in something I want to make, then get bored with the easy stuff and just stop.

I actually want a pulse induction detector for beach work, but can't afford a good one. Hence me thinking of making this. If I'm not careful though I'll end up doing my usual - I need to stop tweaking at some point and just make the thing and get out on the beach with it.

I'm also playing with my own design ideas at the same time... fast mosfet turn off, differential search coil, precision differential front end, fast ADC sampling of the waveform and a fast micro to drive it and produce a useful interpretation of the results. I've got some really nice 40Msps 12-bit pipeline ADC's and a box full of assorted MCU's...

Edited to add: I last made a metal detector (based on the PE Magnum design) in the early 80's. In that one, there were two separate, regulated supplies run from their own batteries. That seemed a better course of action than a single battery and lots of decoupling. Of course, that was an IB design without the big power drain of a PI. In my own, if I ever get far enough to need to work on the power supply, I think I would go with some small lithium-ion cells (4x RCR123's for instance), and linear regulators, for the main circuitry and a separate big battery for the coil drive. Just eliminate a lot of the noise and decoupling in one go!

Linkback:

You are not allowed to view links.
Please Register or Login

http://www.thunting.com/smf/index.php/topic,6515.msg45650.html#msg45650

Foxyrick,

You are right about SMD. I completely missed this!

I agree with you about finessing too much.  I think its the 80/20 rule.  20% of effort to give 80% improvement so that we don't get stuck in the weeds and lose interest (I am just like you).  The trick is to build one as standard and then make changes from the benchmark original because we don't have the rationale of the design decisions of the original designers.

I am looking at building this one because I think its a good base for the kind of DSP project you are talking about.  It seems to have the fundamentals.

Thanks for the parts list.  I will go through this in detail shortly when I have some more time.  The more eyes the better!

I think the only agreed change that is a given is to replace the old bipolar 555 timers with newer CMOS versions.  I would strongly suggest that we add the recommended decoupling for the 5534 op amp (or replacement) given the instability problem mentioned.  Also the TL064 and TL074 do not have the 47R like the 5534 but has 22uF decoupling and is a candidate for high frequency decouplers too.  It would be good to find a reference designs for these chips to confirm.  I will look.

I think that we are agreed that the 5534 amp circuit is high gain and is probably the most critical part of this whole circuit.

The battery solution you suggested is a good one WRT reducing noise but having multiple kinds of batteries to charge/ replace just offers the chance for more frustration in the field.  It is a strong disincentive for this solution to me.

I would suggest raising the rails by some means (batteries or otherwise) though.  It seems that the inverter circuit here is pretty similar to the one

You are not allowed to view links.
Please Register or Login

which specs a voltage drop magnitude of 1V from the input voltage.  If that is the case, increasing the rail voltage by replacing the regulators seems straightforward.  Can anyone who has one of these confirm the voltage at the input to the 78L05 regulator?

I note that the aaroncake.net inverter circuit oscillates at about 3KHz as opposed to the goldscan inverter that is around the 100KHz mark (as you mentioned) and it does not have the transistor output stage of the goldscan.

It would seem easy to omit the transistor output stage to implement the aaroncake.net version if it would be better- obviously also realising the inverted design of the GS for positive inversion from a negative supply.  I am not sure if the high oscillation frequency was a plus (putting the noise fundamentals high) or a negative and whether a 3KHz oscillation is going to give more noise problems.  100kHz seems kind of high.  Its all speculation without a circuit to test.

The RC filtering on the 5534 supply is not normal but I have seen it done as a bandaid to solve noise problems that were solvable by other means (like good layout or proper high frequency decoupling).  I note that the same was done with the 555 that drives the coil transistor and this might also be made redundant by replacing the 555 bipolar with CMOS.  The CMOS 4538 have a similar arrangement.  With great respect to the original designer (who I take my hat off to for the design and with my ignorance in not experiencing his problems and design decisions) it may have been to try and resolve the oscillation caused by the omission of the high frequency decoupling capacitor although it was also implemented for the previous op amp used in this position- the ua709.  Hard to say without having a working circuit to probe but I would favour replacing the 47 ohm resistors with zero ohm links on the 5534 rails after adding proper decoupling- as a test.

Your comment to place the driver circuit externally (in with the coil?) is a good one but there is probably some art to where you split the circuit.  It may be best to put the 5534 circuit there so that the amplification is done before the hike down the coil wire to the main electronics if you were to move out the driver circuit.

Hopefully this does not throw up more issues than resolutions!

I hope a few people check your parts list!  More eyes are better.

Thanks for your work!

Chudster

Linkback:

You are not allowed to view links.
Please Register or Login

http://www.thunting.com/smf/index.php/topic,6515.msg45833.html#msg45833

GD,

I am not sure that the board is designed for a faraday shield but you are right that the switchers are noisy.  I am not sure if the consensus will be to layout a new board.

Maybe we can make a "tin foil hats" and fit them carefully  :




Linkback:

You are not allowed to view links.
Please Register or Login

http://www.thunting.com/smf/index.php/topic,6515.msg46065.html#msg46065

سلام آقا حسام گلد اسکن رو ساختی یا داری میسازی؟

Linkback:

You are not allowed to view links.
Please Register or Login

http://www.thunting.com/smf/index.php/topic,6515.msg46072.html#msg46072