Simplest PI detector possible?

[Carl-NC]

Why do people insist on using opamps as pulse preamps when single transistor stages are typically >100x faster ?!

Probably for the "good enough" reason.

[Tinkerer]

My circuit did not work because the integrator period is too short - I realise now having analysed the simplepi circuit that has a response to the pulse that is over 300 uS long.
Also, my main problem with this whole thing is that I cannot discern any movement in time in the signal on the scope when probing the first (coil) stage; I have a high speed scope and would expect to see something change as metal is brought close to the coil. Has anyone here with access to a high speed scope seen any movement on the trace when metal is brought close or is the movement so small that it is lost in the scope's trigger jitter?

Single transistor stages are faster, you may get away with opamps at low frequencies or for low response rate circuits like integrators, but you cannot beat single transistors for high speed signal amplification / pulse following. Power, noise and repeatability have absolutely nothing to do with this as any circuit designed with either type of component can be optimised for these. Why would you want to match transistors when designing separate single stages?, this isn't monolithic CMOS, one cancels out beta differences with robust topologies.. if you are a 'solid state guy', you should know that op-amps are solid state too.

And no, spice is accurate when you know what you are doing. Any inconsistencies with fabricated circuits are likely to be due to parasitics that are not modelled, like the length of component legs introducing inductances and capacitive effects due to the PCB and layout used. These only become a nuisance at higher frequencies.

I just took a few pictures today, of the response of different targets. The opamp is SSM2019 and the gain is 10. Since I have the circuit on the breadboard I can also take pictures of the signal before the amplification, if it is any help.

[Waikiki_Sweep]

Capacitive coupling. How bad it is for the noise?

Current schematic has a capacitive coupling and virtual ground for preamp.

How bad it is for the noise? Probably it is too bad. Who can comment it from the noise reduction point of view?

I have tried similar setups with capacitive coupling and found that polyester capacitor gives less noise than electrolytic in practice. What capacitor to choose?

I saw virtual ground IC - TLE2426 that you can buy and use in your projects. Is there reason to use it to reduce noise instead of usual two resistors and capacitor?

Also I see that everybody (including myself ) forgot to add BYPASS CAPS near chips power pins in all that Pirate schematics mods. Too bad even for beginners.

[Beenthereuk]

Hi Tinkerer, thanks for that. That is what I expected but for the life of me I cant get that response on my scope. I probed just after the limiting diodes like you did but could notice no change with any metal. I must have missed something so I'll try again on Monday.

[Tinkerer]

Hi Tinkerer, thanks for that. That is what I expected but for the life of me I cant get that response on my scope. I probed just after the limiting diodes like you did but could notice no change with any metal. I must have missed something so I'll try again on Monday.

Actually I am quite interested in a different option for the SSM2019. As I am using a balanced system it is quite difficult to find a fully differential low noise opamp or INA that is fast enough.

I probed just after the limiting diodes like you did but could notice no change with any metal.
The above pictures were taken after the opamp and, just looking at the breadboard, a LP filter. The extremely noisy signal is due to a variety of external sources, including the cheap switching computer PSU, fluorescent lightbulbs etc. nearby. This simple temporary rig helps me test and balance the coil and locate the noise sources on my new bench that I have to eliminate.

[Beenthereuk]

OK, I realise that now. I guess it is not possible to see anything before the amplifier in real time.

But yes, speed and gain are of the essence here, so.. see my new post of an improved PI that uses an all BJT front end.

[bernte_one]

hello waikiki

R2 is not 15K or 16K
in shematics there is missing a komma
R is 1K5 or 1K6 what you have laying around

with r2 1K5 i get around 100µs pulsewidth
and a frequency of 140Hz

please lokk a inserted all corrections you say



many thanks

[bernte_one]

http://www.ne555.at/timer-ic-ne555/rechner-ne555.html

[Waikiki_Sweep]

hello waikiki

R2 is not 15K or 16K
in shematics there is missing a komma
R is 1K5 or 1K6 what you have laying around

with r2 1K5 i get around 100µs pulsewidth
and a frequency of 140Hz

please lokk a inserted all corrections you say

[ATTACH]28118[/ATTACH]

many thanks

Yep! 1.5K is proper value.

Frequency: 140.097 Hz
Duty Cycle: 98.54 %
Cycle Time: 7.138 ms
Time High: 7.034 ms
Time Low: 0.104 ms

"0R" value near R4 looks like ZERO Ohm or just a wire jumper. If you like to remove R4 probably it is better to remove it completely or you can make it 10K too.

Everything else looks perfect!

Good luck!

[bernte_one]

Yep! 1.5K is proper value.

"0R" value near R4 looks like ZERO Ohm or just a wire jumper. If you like to remove R4 probably it is better to remove it completely or you can make it 10K too.

right mistake from me thx

[Waikiki_Sweep]

SSM2019

!!!! Audio preamps !!!! Great idea !!!! Sure, why not to search for special ICs instead waiting for good opamp.

off-topic: Just in case someone need low noise preamp for center tapped coil: http://www.electroniccircuitsdesign....it-design.html

[bernte_one]

members of german mikrocontroller forum say i should add some blocking caps at the ic 555 and tl072 of 100nf kerko ond vcc pins

[Waikiki_Sweep]

members of german mikrocontroller forum say i should add some blocking caps at the ic 555 and tl072 of 100nf kerko ond vcc pins

Yep. I already mentioned bypass capacitors just ten posts earlier (on page #8 ). Without capacitors on power lines it will play music instead of detecting probably.

Also nobody has ideas to answer on my question: Capacitive coupling. How bad it is for the noise?

Because most of new detectors will use microcontrollers next years using one power +5V opamps will be extremely convenient.

But most of such schematics will use capacitive coupling as you can see it on Pirate, PICKINI and many others.

So how to build detector with capacitive coupling between coil and preamp without damaging noise characteristics?

[Beenthereuk]

What kind of noise do you think will be the issue here?, noise due to the DC bias of the amplifier can be greatly reduced by using a very low noise op-amp or a discrete transistor stage as I have shown. The gain of the first stage, specially when using an op-amp, is not likely to be so high (most probably < 20 given the BW of most op-amps) so as to amplify such noise anyway.

Also, keeping the input impedance as low as possible will reduce interference from stray signals as the incomming signal from the coil is quite strong, so the amplifier's bias network resisitors should be of the lowest practical values.

[Waikiki_Sweep]

What kind of noise do you think will be the issue here?, noise due to the DC bias of the amplifier can be greatly reduced by using a very low noise op-amp or a discrete transistor stage as I have shown. The gain of the first stage, specially when using an op-amp, is not likely to be so high (most probably < 20 given the BW of most op-amps) so as to amplify such noise anyway.

Also, keeping the input impedance as low as possible will reduce interference from stray signals as the incomming signal from the coil is quite strong, so the amplifier's bias network resisitors should be of the lowest practical values.

Sounds promising! As I understand to keep noise low it is required to keep input resistance as low as possible.
Without coupling capacitor everything looks simple. Input connected to ground by 390 Ohm dumping resistor.
I just worried that capacitor rising impedance and as result noise rising too. But how much?

"incoming signal from the coil is quite strong"

May be not. As I know Surf PI metal detector amplifying signal in 1000(preamp) X 474(integrator) X 220(last stage) = 104,280,000 times
So we amplifying signal in nano Volts range and noise which is quite strong in comparison with signal is a biggest problem that reducing detecting distance severely.
That is why QED detector (most sensitive that I ever saw) using LME499990 opamp with 0.9nV/√Hz (1.3nV/√Hz for 10Hz.)