Uni-PI - Universal Pulse Induction Metal Detector board for microcontrollers like Ardiuno, PIC, AVR.

[Aziz]

You can use a biased diode to **** it close to the rail. Using a low value resistor will probably give you less noise than a pair of resistors there anyway.

I cannot find a good model for the ISL28291 or I would simulate the noise for it but I have used an LT1028 instead in this simulation and compared it with a single transistor stage with dynamic load (pnp current mirror).

The LT2018 is a £12 chip (Farnell, UK) that has 10 Hz noise = 1nV / RootHz.
The transistor amplifier is the cheap and common BC337 (with the low base resistance).
I could only squeeze 800x gain out of the transistor (can do more but need more time to play) so I set the same gain and BW on the op-amp and compared noise plots.
Considering the cost difference, the results are amazing. See below:

[ATTACH]28173[/ATTACH]
[ATTACH]28174[/ATTACH]
[ATTACH]28175[/ATTACH]
[ATTACH]28176[/ATTACH]

Hi Beenthereuk,

nice circuit but the LT1028 will do better. That's the reason, why they exist.
The transistor stage isn't stable (highly temperature dependent, have a look at the hFE).
There is no feed-back-loop control for the gain (unstable gain).
The current mirror stage requires a matched transistor pair (preferably on the same die).
Power supply rejection ratio suffers!

Solve the issues and you are on the game again.
Cheers,
Aziz

[Beenthereuk]

Of course it will never be like the opamp - it has many transistors; this circuit only uses 3 ! and it is easy to stabilise it for thermal drift and gain. The main problem with it is I cannot seem to get gain above 800 without an extra stage. But will try again soon. Right now have realised the new coil I made is crap.. too low inductance, so am off to wind another.

[green]

Of course it will never be like the opamp - it has many transistors; this circuit only uses 3 ! and it is easy to stabilise it for thermal drift and gain. The main problem with it is I cannot seem to get gain above 800 without an extra stage. But will try again soon. Right now have realised the new coil I made is crap.. too low inductance, so am off to wind another.

How much preamp gain do we need? I've seen a gain of 100 being enough in some of the threads.

[Beenthereuk]

That is a very good question - it would be nice if that was clear as going above 1000x is not easy with few devices.

I've ordered some LM318s (noisy but fast) and opa37's (fast and quiet) to try with my new coil and see what if any difference the extra BW and low noise response make to the amplified signal.

[Beenthereuk]

OK, in case it is of interest, here is a 500x, large bandwidth, low noise, fully thermally stabilised 4 transistor amplifier:

[Carl-NC]

The main problem with it is I cannot seem to get gain above 800 without an extra stage.

Gain is fundamentally limited by Early voltage, namely VAeq/VT, where VAeq is the combined Early effects of Q1 & Q2, and VT is thermal voltage. It sounds like your VAeq is about 20V, which seems low for these transistors. What do the models say?

How much preamp gain do we need? I've seen a gain of 100 being enough in some of the threads.

This is a system design question. Generally, 50-1000.

OK, in case it is of interest, here is a 500x, large bandwidth, low noise, fully thermally stabilised 4 transistor amplifier:

Still not a stable-biased amp. It will spend most of its life railed out.

[Beenthereuk]

>> It sounds like your VAeq is about 20V, which seems low for these transistors. What do the models say?

Yes, it is the current mirror pnps with low VAF leading to low collector impedances. I stuck with them because I have a draw full at work.


>> Still not a stable-biased amp. It will spend most of its life railed out.

With 4.4 V of negative feedback on the emitter and a diode clamp fixing the current through it?.. I think its more stable than the pound. Still, have to build it to see.

[Carl-NC]

You're still trying to match 2 uncorrelated currents at a high-impedance node, with no overall feedback. You can tweak it in, but it's just not going to stay there.

[Beenthereuk]

You dont consider the DC voltage across RE feedback?, the circuit should need no tweaking.

http://www.zen22142.zen.co.uk/Design/bjtbias.htm

[Carl-NC]

RE has no useful effect on Q2's current. Build it and see what happens.

[Beenthereuk]

Very true, but it does regulate the current through Q1 thereby regulating its Q point (which is what matters most here), to a considerable extent, by auto-biasing to cancel out Vbe variations due to thermal drift in the mirror transistors. It simulates well for a range of input temperatures (-10 to +75 deg. C.) and works with every small signal BJT I've thrown at it, which qualifies the topology.

[Carl-NC]

Build it and see what happens!

[Beenthereuk]

Waikiki: just simulated this (took forever for some reason) and I understand the sampling, etc, now.

What range do you get from this metal detector?, for a coin buried in sand for example?

[Waikiki_Sweep]

Waikiki: just simulated this (took forever for some reason) and I understand the sampling, etc, now.

What range do you get from this metal detector?, for a coin buried in sand for example?

With coil from Surf PI it acting like Surf PI.
In air for overage coin or ring 20-25 cm or 8-9 inches. Big coins and heavy rings may give you 30 or more cm. or 12 inches+.
Range depends on coil diameter and quality also EMI noise and ground mineralization may reduce depth.
If frequency increased from 600 Hz to 1-2 kHz range increasing on couple inches as well as power consumption doubling.

[Tepco]

This project desperately need some, at least basic form of GB\disc control. Very little change is needed, two sample subtraction using same integrator. I have working hardware on table, analog, not uC controlled, will post more details, actually tempted to add this to minipulse project.