Minipulse Rev D first attempt

[Old cart]

ok so the scope is working well heres a pic of TP1 the pulse width is 46us how can i adjust this to be 58us?

[ATTACH]35821[/ATTACH]

I don't know Minipulse well. If that is the transmit pulse width I don't think that would make much difference, if any at all. The narrow TX pulse width will reduce sensitivity only to long time constant targets like iron or silver coins. Narrow pulses are better for gold jewelry.
Do you see how the voltage curve is starting to flatten near the end of the pulse? That indicates the coil current is no longer rising so the coil has achieved a steady state mode limited by the coil resistance. This is good.
The general shape of the pulse also looks good.

[Impulse]

ok so the scope is working well heres a pic of TP1 the pulse width is 46us how can i adjust this to be 58us?

[Impulse]

ok so i have made a simple coil out of my 0.5mm enameled copper wire 24 turns and it reads as 390uh i get a £2 coin at 20cm, this seems a little short? i have borrowed a laptop so i will get the scope installed soon!

[Qiaozhi]

So the plan them is to fix the existing model calling it perhaps Ver2.1 and then maybe do an enhanced version calling it GEOTECH Bara V1?

That's about it. I think we should patch up the existing Apberg Rev-2 board, and then create a new Geotech version.

[Old cart]

So the plan them is to fix the existing model calling it perhaps Ver2.1 and then maybe do an enhanced version calling it GEOTECH Bara V1?

[Old cart]

Right! Q it is then. It fits! I can barely speak American ( Texas) English, let alone the the Kings English.

[Qiaozhi]

I did not make it clear that I agree with Q. I sort of like that name, has a bit of a Bond ring!

Qiaozhi is a transliteration of "George" from English to Chinese Mandarin romanized form, known as pinyin.
Unless you know some Mandarin, Qiaozhi is difficult to pronounce. I think it was Carl that first shortened it to Q.
Of course, it's also a reminder of some other Qs, such as the Quartermaster from the Bond films, or the omnipotent being Q from Star Trek.
Take your pick.

[Old cart]

I did not make it clear that I agree with Q. I sort of like that name, has a bit of a Bond ring!

[Old cart]

[ATTACH]35776[/ATTACH]

So the chip supplies are -5 and 0 volts. May be best to stick with that.

[kt315]

[kt315]

[Old cart]

I'm somewhat inclined to take the Apberp REV-2 version and fix it so that everything functions correctly, without making so many changes that it's not the same design. Then create a "Geotech Baracuda REV-A" version, as I think simply changing it to REV-3 will confuse everyone. IMHO it would best just to draw a line under the existing stuff and start again.

As long as it is clear what REVISION anything is we have made a huge step in the right direction. I know how tough it is to take random data and make sense of it. That is exactly what I am trying do right now in creating a Surf PI build guide. It takes many hours of study...

[Old cart]

Exactly. I did point that out earlier when I first took a look at the Baracuda design. The confusing thing is that the so-called "original" schematic doesn't show that. Perhaps it's never been correct, and that's why it was changed to 0V and -5V for the supplies, but whoever did that forgot to correct the pull-up resistors.

Maybe we should just stay with the 0 and -5V design. Ray has proven that this works if the pull up values are adjusted. But as mentioned in my last post we should not conclude that this is engineering that will produce a repeatable design. Unless someone has and commits the time to verify it will work under all conditions, this is more of a variant.

As for the original design being good, I would suspect that this is more of a garage design and that the designer also hand built each one and MADE it work. I say this only because of the marginal engineering practices.
I do not mean to insult anyone here but we must recognize that this design has merit but in the current state needs some work.

[Qiaozhi]

Precisely the point I tried to make earlier. This is a common trend of many of these discussions. We often try to reverse engineer many detectors, this is not a perfect process. Then folks start to mod the reversed engineered design. That is fine but the degree of engineering skill varies widely. Multiple schematics are produced some have mods of mods. Components are changed to get the moded detector to work. Ultimately someone gets the detector to work, but only that specific detector. This is not a very effective way to produce a reproducible working detector.
While I myself like to tweak and adjust things I have no expectations that my modified designs could be easily reproduced. Indeed without a good scope, multimeter and LCR meter it is unlikely that someone could get one to work UNLESS they had a great deal of knowledge.
I don't know how to fix this process but it is my OPINION that something between the process we currently have and disciplined engineering could work. Any ideas or comments?

Maybe we should start a new thread.

I'm somewhat inclined to take the Apberp REV-2 version and fix it so that everything functions correctly, without making so many changes that it's not the same design. Then create a "Geotech Baracuda REV-A" version, as I think simply changing it to REV-3 will confuse everyone. IMHO it would best just to draw a line under the existing stuff and start again.

[Qiaozhi]

N-JFET sample switches are further proof of timing logic being supplied at 0/-5V. Their gates are directly connected to cmos logic outputs, they would forward bias the JFET gates if run between +5V/-5V. This would couple logic output into integrator input through the gate diode.

Exactly. I did point that out earlier when I first took a look at the Baracuda design. The confusing thing is that the so-called "original" schematic doesn't show that. Perhaps it's never been correct, and that's why it was changed to 0V and -5V for the supplies, but whoever did that forgot to correct the pull-up resistors.