Arduino Nano PI Main Discussion

[Qiaozhi]

Thanks a lot for your replies, well appreciated.

Yes, at some point, i will put a damping resistor that makes it closer to the critical value. Was just wondering if a little bit of underdamping was ok.

I am trying to get to the finish line and was wondering if you could tell me if those scope traces look ok, in particular, the signal at TP10. I am at step 13 in the book. I don't understand why the sample pulse is so large, time wise.

See here the traces i am getting at TP6. TP5, and TP10: https://imgur.com/a/1jamqw8

The book says I should get 700mV DC at TP10, but i am nowhere near that with this sample delay as it's actually negative.

Not too sure what I am doing wrong here. Again, thanks for your patience.

TP6 is the preamp output. This looks OK.

TP5 is the main sample pulse. This should be 50us wide, as defined in the Arduino sketch. If your timebase is set to 8us (which it looks like it is) then it seems to be OK. Measure it to make sure.

TP10 is the output of the 1st integrator. The 700mV is a guide in case someone attempts to build the detector without access to an oscilloscope. In the first image it looks to me like the start of the sample pulse is higher than halfway up the decay curve. Try reducing the sample delay with the pot connected to PL4. The 1st integrator stage inverts the signal from the main sample gate (TP7), which I suspect is currently slightly positive.

[ugocapeto]

Thanks for your suggestions. Really appreciated.

I have just realized that there is video in youtube land that shows what the traces are supposed to look like:

https://www.youtube.com/watch?v=G1gUYHVw20k

I do get about the same thing, so that's good.

[ugocapeto]

Hi all:

About the coil, does the type of epoxy resin matter much?

Is it ok to use a cheap epoxy resin commonly used in craft projects?

Or do you really need an electronics grade epoxy?

[Carl-NC]

It can matter. Usually you want low shrinkage and low dielectric constant. Low shrinkage is more critical in a DD coil and the dielectric constant is important for high performance. For some coils you can get away the cheap stuff, you'll have to try & see.

[ugocapeto]

I ended up using a epoxy resin for doing molds and the result is very good, at least, esthetically.

About the audio tone, it's a bit chirpy at rest.

When an object gets close to the coil, the tone become continuous (and louder).

Is the tone at rest supposed to be a bit chirpy, that is not continuous (I am testing indoors)???

[Qiaozhi]

I ended up using a epoxy resin for doing molds and the result is very good, at least, esthetically.

About the audio tone, it's a bit chirpy at rest.

When an object gets close to the coil, the tone become continuous (and louder).

Is the tone at rest supposed to be a bit chirpy, that is not continuous (I am testing indoors)???

PI detectors tend to be sensitive to external noise, especially when tested indoors. This can cause the input signal to jump around so that it keeps crossing the audio threshold.

[ugocapeto]

Thanks a lot for your reply.

Is the cabling from the coil socket to the pcb supposed to be shielded?

Also, in the book, you say that the delay can be set by waving a coin in front of the coil but isn't it better to set the delay with a scope so that it's at the middle of rising signal from the preamp and leave it there. Even if you would want to exclude pull caps and put more delay, wouldn't it make sens to have the pot directly on the pcb and not make it an outside pot that can be easily bumped into. Just wondering.

[Qiaozhi]

Thanks a lot for your reply.

Is the cabling from the coil socket to the pcb supposed to be shielded?

Also, in the book, you say that the delay can be set by waving a coin in front of the coil but isn't it better to set the delay with a scope so that it's at the middle of rising signal from the preamp and leave it there. Even if you would want to exclude pull caps and put more delay, wouldn't it make sens to have the pot directly on the pcb and not make it an outside pot that can be easily bumped into. Just wondering.

You could shield it if you find it makes a difference. Personally I didn't bother.
Ideally you could put the electronics in a metal enclosure, but this makes it heavier.

Waving a coin in front of the coil is just an alternative method of setting the delay. As you said, it could also be a fixed setting if that meets your requirements.

[ugocapeto]

Thanks for answering my newbish questions

I am a bit confused by the 2nd integrator to go from tp10 to tp11.

At tp10, I have a nice dc signal when at rest (no metal around) but at tp11, the signal is quite wavy (at rest) and the signal at the headphone is not continuous. Is it supposed to be like that? Do you get the same thing indoors?

I kinda understand the first integrator to get tp10 (not fully tbh) but i don't really understand why a 2nd integrator is used to get to tp11. If the input signal is dc, shouldn't the output signal of the 2nd integrator be a ramp?

[Qiaozhi]

Thanks for answering my newbish questions

I am a bit confused by the 2nd integrator to go from tp10 to tp11.

At tp10, I have a nice dc signal when at rest (no metal around) but at tp11, the signal is quite wavy (at rest) and the signal at the headphone is not continuous. Is it supposed to be like that? Do you get the same thing indoors?

I kinda understand the first integrator to get tp10 (not fully tbh) but i don't really understand why a 2nd integrator is used to get to tp11. If the input signal is dc, shouldn't the output signal of the 2nd integrator be a ramp?

The first stage is a sampling integrator, and the second stage boosts the gain and provides threshold adjustment.
The second integrator will not produce a ramp at the output because there is a resistor in the feedback loop. If there was no resistor present, then you would be correct.
Try putting the circuit in LTSpice to understand how it behaves.
https://www.electronics-tutorials.ws/opamp/opamp_6.html

[ugocapeto]

This is what I see on my scope for TP6 (2v per subdivision on y-axis) and TP11 (1v per sub): https://imgur.com/a/L0wEAVc

This jumping around doesn't seem right to me. I would expect a little of bit jumping but not that much.

What do you think?

[Qiaozhi]

This is what I see on my scope for TP6 (2v per subdivision on y-axis) and TP11 (1v per sub): https://imgur.com/a/L0wEAVc

This jumping around doesn't seem right to me. I would expect a little of bit jumping but not that much.

What do you think?

The output at TP11 will jump around a bit if you test it indoors. Running from a bench power supply can also make it noisier, so it's better to test with a battery pack.
However, your TP11 output does seem particularly noisy. As a comparison, have a look at this video.
https://www.youtube.com/watch?v=G1gUYHVw20k

[ugocapeto]

If I use 2v subdivision for TP11 instead of 1v (to match what's in the youtube video you linked to), the jumping is not as bad. I am indoors and i am using a battery pack. I think i am gonna let it be.

Now, I am still trying to understand that double integrator stuff.

What i don't understand is that high pass filter at TP10 going into the 2nd integrator. Isn't the output at TP10 pretty much a dc voltage? Shouldn't a high pass filter get rid of that dc voltage?

Thanks for your help in all this. Trying to learn as much as possible from that build.

[Qiaozhi]

What i don't understand is that high pass filter at TP10 going into the 2nd integrator. Isn't the output at TP10 pretty much a dc voltage? Shouldn't a high pass filter get rid of that dc voltage?

As it says on page 13 of the book ->

"C14 (470nF) and R25 (220k) form a passive high-pass filter that provides the detector
with a self-adjusting threshold (SAT). Therefore, if you hold a coin (for example) near the
coil, the audio tone will gradually reduce in volume. The idea of the SAT is to prevent the
dc level of the received signal from drifting over time, mainly due to temperature changes
and self-heating effects that would require the operator to constantly adjust a manual
threshold control to compensate."

[ugocapeto]

Ok, thanks, I kinda get it, you are interested in the DC variation, not the actual DC value.

As you said previously, simulating all that good stuff in LTspice would be useful in understanding how the whole thing works.

I just don't know enough about LTspice to be able to simulate going over metal and make the decay curve vary at the sampling points.

Any pointers on how to do that?