I've enjoyed the combination of TinyCAD (free schematic editor) + VeeCAD (now free stripboard / veroboard layout editor).
Pleased with "VeeCAD" software for Stripboard / VeroBoard
Pleased with "VeeCAD" software for Stripboard / VeroBoard
I think a pen & paper layout would be more than sufficient for most of us, no need to go into any program's learning curve
I have managed to enter the circuit into DYLC (I gave up on the TinyCAD/VeeCAD combo, it was an intractable mess):
I intend to make a new one, on ordinary perfboard, to validate the design.
I also include the DYLC file, in case someone wants to customize the circuit, add holes, etc.
Note that you will still need to fabricate the hardware holding the CUT, and the flexible sensing electrode.
The layout could easily be translated into a regular, one-layer PCB too
Edit:
There are two ICs on the board, because DIYlc only supports DIP8 and 16, and to make a DIP14 I had to make two DIP8 overlap
I intend to make a new one, on ordinary perfboard, to validate the design.
I also include the DYLC file, in case someone wants to customize the circuit, add holes, etc.
Note that you will still need to fabricate the hardware holding the CUT, and the flexible sensing electrode.
The layout could easily be translated into a regular, one-layer PCB too
Edit:
There are two ICs on the board, because DIYlc only supports DIP8 and 16, and to make a DIP14 I had to make two DIP8 overlap
I just realized that the only strap of the board is not required. By default, all the components in DIYlc are completely opaque, which doesn't ease the layout procedure. I didn't find a global transparency control, and I had to tweak each component individually, which is not particularly convenient. Switching between layers is possible, but it is completely binary and is not helpful either.
Only when I had painstakingly made all the components transparent did I realize that the strap had already been duplicated by a track.
Here is the updated layout:
Only when I had painstakingly made all the components transparent did I realize that the strap had already been duplicated by a track.
Here is the updated layout:
I have tested the above layout, and it works perfectly:
Some differences with the project are visible, but initially it followed the layout exactly.
In fact, I took the opportunity of a new prototype to make some tests and tweaks.
I wanted to see if improvements in sensitivity were possible.
During the test, I realized that I had made a mistake: the sense amplifier is a bandpass filter, and I had (wrongly) assumed that at the center frequency, the phaseshift would be zero; it is in fact 90°, which is the worst possible outcome, since the coincidence detectors react for 0° and 180°, 90° being indeterminate, thus useless.
Although the circuit worked, I wanted to correct the phase issue, and added phase-shifters, but it didn't work well: near the center-frequency, the phase has a very steep slope, meaning the slightest frequency error results in a large phase error.
When the filter is slightly offset from the measuring frequency, the phase becomes very close to 0° or 180° quickly, and the loss of amplitude caused by off-peak operation has a minimal effect.
Thus, operation with a small frequency offset is much more stable and reliable than being spot-on.
In this implementation, the capacitor C2 is 1pF instead of 1p5, for two reasons: it is mounted on a small socket, having ~0.3pF, and the perfboard is naked, without a top shieldind screen.
Without the shield, the capacitance between the two nodes is not shunted to GND and remains active.
The phase accuracy is sufficient to let the circuit operate without the shunt capacitors across the LEDs, but I have included them anyway, meaning the inactive LED remains perfectly dark.
It also works with white and blue LEDs, but with the 1K5 resistors, you need sunglasses to operate the circuit comfortably.
Ideally, the board should be enclosed in a metallic case, to reduce unwanted influences.
Since I have no need for two Folio's, I'll give away this one.
If you are interested, send me a PM and you'll get it completely for free. Since only a single piece is available, don't waste too much time! First come, first served.
One exception to the rule is if someone commits to the creation of a PCB for the community, and publishes freely gerbers etc. on this site
Some differences with the project are visible, but initially it followed the layout exactly.
In fact, I took the opportunity of a new prototype to make some tests and tweaks.
I wanted to see if improvements in sensitivity were possible.
During the test, I realized that I had made a mistake: the sense amplifier is a bandpass filter, and I had (wrongly) assumed that at the center frequency, the phaseshift would be zero; it is in fact 90°, which is the worst possible outcome, since the coincidence detectors react for 0° and 180°, 90° being indeterminate, thus useless.
Although the circuit worked, I wanted to correct the phase issue, and added phase-shifters, but it didn't work well: near the center-frequency, the phase has a very steep slope, meaning the slightest frequency error results in a large phase error.
When the filter is slightly offset from the measuring frequency, the phase becomes very close to 0° or 180° quickly, and the loss of amplitude caused by off-peak operation has a minimal effect.
Thus, operation with a small frequency offset is much more stable and reliable than being spot-on.
In this implementation, the capacitor C2 is 1pF instead of 1p5, for two reasons: it is mounted on a small socket, having ~0.3pF, and the perfboard is naked, without a top shieldind screen.
Without the shield, the capacitance between the two nodes is not shunted to GND and remains active.
The phase accuracy is sufficient to let the circuit operate without the shunt capacitors across the LEDs, but I have included them anyway, meaning the inactive LED remains perfectly dark.
It also works with white and blue LEDs, but with the 1K5 resistors, you need sunglasses to operate the circuit comfortably.
Ideally, the board should be enclosed in a metallic case, to reduce unwanted influences.
Since I have no need for two Folio's, I'll give away this one.
If you are interested, send me a PM and you'll get it completely for free. Since only a single piece is available, don't waste too much time! First come, first served.
One exception to the rule is if someone commits to the creation of a PCB for the community, and publishes freely gerbers etc. on this site