Improved Simple PI - all BJT front end for greater sensistivity

[Beenthereuk]

Improvements over Russian simple PI circuit

1. Replace positive edge clipping diode for a low VF fast schottky type to further limit the input signal and target the smallest part of the fly-back voltage's trailing edge.

2. Replace first op-amp with a fast BJT linear amplifier.

3. Replace 20 k variable resistor (R7 in my schematic) with 47 k.

4. Replace 1 nF (C2 in my schematic) capacitor with 10 nF.

5. Use just about any op-amp for the final part, slew rate not particularly important. I prefer the MCP6023 by Microchip.

The resulting circuit has oodles of gain (you can use higher gain devices for even more, e.g. BC547C, BC557C) that should translate into greater sensitivity and hence greater detection range.

[Waikiki_Sweep]

Interesting ideas. Can you make it working in real life?

Finally best way to improve that schematic is to use sampling switches and integrator as Carl told already. Without that functionality it is useless toy or in the best case pinpointer.
Task was to create something with minimum components but working somehow. With more components it goes away from the purpose of its creation.

Probably you can add transistor and remove second opamp too to at least make PCB small.

[Beenthereuk]

Why would it not work in real life when it does what the other circuit did, only better?. Yes, the next thing is to replace the final opamp for a BJT based circuit too for cheapness and size. Or incorporate a micrcocontroller to switch the precise point on the trailing edge into the amplifier to try to discriminate metals better. I'll have another go once I've built this version.

[Carl-NC]

Comments...

Generally I've found that fast Schottkys are no better than plain ol' 914/4148 diodes in this app.

You probably should put C1 after the diode clamp, not before.

Darlingtons have notoriously slow transient speeds.

I'm failing to see an improvement over a NE5534, just a lot more parts.

[Davor]

That is actually a Sziklai pair, and it can be convinced to become very fast in case the first transistor is allowed, say, half the current through it. It would require a resistor between a collector and a ground.

[Beenthereuk]

it is not a darlington pair and the circuit simulates a lot faster than the TL082 opamp and certainly faster than what is needed to follow that signal without adding a collector load.. that is good enough for me.

The diode clamp works well where it is because the 1uF capacitor couples more than enough signal into its anode - simulations showed no difference when placed before the capacitor.

And the diode used is not only fast, it is also low VF, meaning it clamps more of the input signal, limiting the signal passed to the amplifier further than a normal Si diode can.

This circuit is more sensitive than the op-amp based circuit - you will see the improvement if you build it.

[Carl-NC]

Sziklai pair, yes, I couldn't remember the name... we used to call it a complementary Darlington. It has the same bad traits as a normal Darlington, which are vastly improved with DC bias currents. Tony, if you haven't built this yet I urge you to give it a try, as it does look interesting (and, if it does work, I see a possible valuable improvement). However, I've designed a lot of circuits which worked fabulously in Spice.

[Beenthereuk]

I know.. I used to doubt it too. Then I had to learn CST Microwave Studio and Microwave Office for a couple of projects I became involved in and they forced me to look into modelling in far greater detail.
I always build what I simulate - breadboard then PCB and I've learnt to trust the simulators now. But I now have a new version - see new post.

[6666]

Hi Beenthereuk, good to see new ideas, when you say - see new post - do you mean the first post in this thread ?

[Beenthereuk]

No, new post is up there now. Just posted it. All transistor PI.

[Waikiki_Sweep]

Great!!

Ooops....

Last eye sore remains - NE555.

Lets kick it out with transistor multivibrator.
And sure lets see it on Youtube kicking a penny from, let say, 8 inches then celebrating a creation first solid state ever created Simple PI

[Qiaozhi]

Great!!

Ooops....

Last eye sore remains - NE555.

Lets kick it out with transistor multivibrator.
And sure lets see it on Youtube kicking a penny from, let say, 8 inches then celebrating a creation first solid state ever created Simple PI

If the circuit contains ICs, then it's also solid state. I think you meant to say the first discrete PI design created.
Although I seem to remember that another member has already done this.

[Davor]

Last eye sore remains - NE555.

Lets kick it out with transistor multivibrator.

Here it comes. Dbogus and Vbogus are not real components, but only a few bits to check what happens at Vbe breakdown. Actually nothing happens, but that's the way to simulate it.
PRR is about 700/s

[Beenthereuk]

Sorry, I didnt bother with the 555 because I plan to use a PIC with mine to enable GPS pinpointing and a visual display. A 12Fxxx PIC has the same footprint of a 555 but is capable of soooo much more..

But yes, a good challenge would be to produce the smallest component count, 100 to 500 Hz (better if its variable), 50 uS pulse duty oscillator to drive the Simple PI.

[sinclairuser]

If the circuit contains ICs, then it's also solid state. I think you meant to say the first discrete PI design created.
Although I seem to remember that another member has already done this.

deemon was the first his had energy recouperation .
thats not to diminish what the op has done, well done and fair play
come to think of it that might have been vlf, did'nt it have GB and DISC?.