Tweet
Thell us somthing more abaut interesting and amazing findings...
Regards!
Regards!
-
-
See topic "New PI technology proposals."Originally posted by miki73 View Post
https://www.geotech1.com/thuntings/s...2260#post82260
AzizComment
-
Hello friends,
I am planning to experiment with PI technology using the laptop as the processing front-end platform (in the first stage for experiments). So a mcu based PI board is necessary to extract the signals for further processing by the laptop. Laptop is capable to digitize the analog signals with 24 bit resolution (big advantage). The fast decaying signals will be converted to a low frequency spectrum, which can be extracted with the lock-in amplifier in the laptop (another big advantage). The PI board will be synchronized with the laptop (lock-in amplifier reference clock - a must have feature).
This could take several month in development. I hope, I will not be prevented from this interesting task by the criminal government here.
Take care,
AzizComment
-
Good luck, keep us informed of progress and discoveries!Originally posted by Aziz View PostHello friends,
I am planning to experiment with PI technology using the laptop as the processing front-end platform (in the first stage for experiments). So a mcu based PI board is necessary to extract the signals for further processing by the laptop. Laptop is capable to digitize the analog signals with 24 bit resolution (big advantage). The fast decaying signals will be converted to a low frequency spectrum, which can be extracted with the lock-in amplifier in the laptop (another big advantage). The PI board will be synchronized with the laptop (lock-in amplifier reference clock - a must have feature).
This could take several month in development. I hope, I will not be prevented from this interesting task by the criminal government here.
Take care,
Aziz
-SBComment
-
Hello friends,
as the LC resonant tanks used in previous coil configuration produced heavy ground effects and mechanical instability of coil, which degraded the unique possibilities of the laptop MD to a standard VLF MD. I therefore have decided to follow the PI technology using the laptop as the processing front-end. The laptop has huge processing power and could do ground balance and discrimination. It gives unique possibility to digitize the small signals with 24 bit of resolution using the external USB sound card. Additionally, one could use the lock-in amplifier to get rid of noise and mains hum and recover the small signals from the PI front-end. Furthermore, boxcar averaging techniques will increase the signal-to-noise ratio (SNR). It will allow more pulses per second to get more samples which also will increase the SNR (up to 12000 pulses per second).
The 24 bit sound card would very likely have at least 18-20 number of effective bits. With some good sound cards, we can achieve even more accuracy and thus sensitivity.
The previous LC resonant tanks were used in range of milli Watt transmit power which will be heavily increased by the PI method to several Watt's producing more target signals. Also the ability to be almost immune to ground effects will improve the overall performance (compared to VLF technique).
As the ADC in the sound card is compared to the signal decay very very slow, the PI front-end will extract four channels (four samples) and encode them to a low frequency analog signal output (AC coupled, can be heared with headphones), which they can be extracted from the laptop using lock-in amp for each channel. The laptop then will do ground balance and discrimination on these four channels and will provide target signal beeping.
A small and common micro controller (probably ATtiny2313) will be used for the PI front-end. The PI front-end needs an AC coupled clock input (from the sound card output) for the triggering the pulse, encoding the signals and the DC/DC charge pump (single battery supply). I will try to prevent using additional PLL circuit. I think, the micro could do this job as well for the right synchronization.
I also will try to keep the whole circuit as simple as possible. The PI front-end will incorporate a synch clock generation, transmit pulse controlling, sample controlling, low noise pre-amp, sample and hold (S & H) or integrators for each channel, channel encoding, DC/DC charge pump, signal output amplifier and simple user interface for mode of operation.
I will also do some circuit simulatons using SPICE to speed up the whole project. So my solder iron will not be fired up until the model is running. I do not exactly know yet, how the PI front-end will work at the end. So it depends on many experiment results. The end of the project will bring an external PI front-end module with external power supply (very likely 12 V). The +5V/500 mA power supply from the USB port could also be worth to try out for reduced sensitivity but practical usage.
You are welcome to contribute to this new project.
I am sure, this will give some really surprizing results.
AzizComment
-
Let's make a PC base metal detector with usb interface !!!
Hello and Happy New Year Aziz , I hope you will be able finish your goals this New Year , and I am looking forward to your New P.I. Micro-controller Design that will be extremely sensitive to small Gold Nuggets !!!!!.....Maybe you will be the first , to be successfull at Extremely fast tx pulse shut-off and a less than 10 micro-second delay on sample rx return signal .....Regards............Eugene52Originally posted by Aziz View PostHello friends,
as the LC resonant tanks used in previous coil configuration produced heavy ground effects and mechanical instability of coil, which degraded the unique possibilities of the laptop MD to a standard VLF MD. I therefore have decided to follow the PI technology using the laptop as the processing front-end. The laptop has huge processing power and could do ground balance and discrimination. It gives unique possibility to digitize the small signals with 24 bit of resolution using the external USB sound card. Additionally, one could use the lock-in amplifier to get rid of noise and mains hum and recover the small signals from the PI front-end. Furthermore, boxcar averaging techniques will increase the signal-to-noise ratio (SNR). It will allow more pulses per second to get more samples which also will increase the SNR (up to 12000 pulses per second).
The 24 bit sound card would very likely have at least 18-20 number of effective bits. With some good sound cards, we can achieve even more accuracy and thus sensitivity.
The previous LC resonant tanks were used in range of milli Watt transmit power which will be heavily increased by the PI method to several Watt's producing more target signals. Also the ability to be almost immune to ground effects will improve the overall performance (compared to VLF technique).
As the ADC in the sound card is compared to the signal decay very very slow, the PI front-end will extract four channels (four samples) and encode them to a low frequency analog signal output (AC coupled, can be heared with headphones), which they can be extracted from the laptop using lock-in amp for each channel. The laptop then will do ground balance and discrimination on these four channels and will provide target signal beeping.
A small and common micro controller (probably ATtiny2313) will be used for the PI front-end. The PI front-end needs an AC coupled clock input (from the sound card output) for the triggering the pulse, encoding the signals and the DC/DC charge pump (single battery supply). I will try to prevent using additional PLL circuit. I think, the micro could do this job as well for the right synchronization.
I also will try to keep the whole circuit as simple as possible. The PI front-end will incorporate a synch clock generation, transmit pulse controlling, sample controlling, low noise pre-amp, sample and hold (S & H) or integrators for each channel, channel encoding, DC/DC charge pump, signal output amplifier and simple user interface for mode of operation.
I will also do some circuit simulatons using SPICE to speed up the whole project. So my solder iron will not be fired up until the model is running. I do not exactly know yet, how the PI front-end will work at the end. So it depends on many experiment results. The end of the project will bring an external PI front-end module with external power supply (very likely 12 V). The +5V/500 mA power supply from the USB port could also be worth to try out for reduced sensitivity but practical usage.
You are welcome to contribute to this new project.
I am sure, this will give some really surprizing results.
AzizComment
-
Hello friends,
after thinking a while, following details arise:
Max. pulses per seconds (PPS) will be 6000 instead of 12000.
The reason for this is, I need equal duty cycle clocking for the channel encoder. The synchronization clock will be divided by 2 to achieve this.
I will need both stereo channels (both input and output lines). The sound card output will provide the TX pulse clocking and channel encoder clocking. The latter one will also trigger the DC/DC charge pump. The PI module will have some user interface and the laptop needs to know about any changes. The sound card input therefore have two lines: channel encoder line (target signals to the laptop) and PI board control interface (frequency encoded status and commands).
The maximum channel encoder frequency from the sound card will be 12 kHz. The TX pulse clocking frequency have to be power of two (fpulse = fsynch/2^n, n = 1, 2, 3, .. ).
The channel encoder is the most critical part of the system due to pure analog interfacing. It will be realized with simple analog switches/MUXes. For each channel, it will generate a different frequency on the channel output line. The first sample will be coded with the lowest frequency (fsynch/16).
Due to the nature of the signal decay curve, sample1 has more signal level than the other following samples. Small signal levels will be encoded more times to achieve a better decoding. The channel encoder will be implemented pure in hardware (no micro).
Encoding frequency:
ch1 (sample1): fsynch/16 (high signal level)
ch2 (sample2): fsynch/8
ch3 (sample3): fsynch/4
ch4 (sample4): fsynch/2 (low signal level)
Sample timings:
t(sample1) < t(sample2) < t(sample3) < t(sample4).
Why analog interface?
No need for expensive ADC's. Just using the high fidelity sound-card in the laptop or external USB sound card. It speeds up the development. Simple to code in laptop software. Flexible in use. High resolution. Cheap.
I am not sure, whether I can achieve a high signal quality and good performance of the detector. I will check this in the coming weeks and month. It may be sure worthwhile to try this out.
AzizComment
-
Hello friends,
I am making good progress on this project. The soldier iron isn't still fired up.
Clock generation (synchronisation with laptop), synchronized DC/DC converter and coil driving stage were finished. Power supply is trivial. Micro controller is trivial. Analog front-end is quite trivial.
I am working now on S&H (sample and hold) and channel encoder stages. I will take JFET transistors instead of integrated analog switches/MUX's. Some JFET's are providing a very good isolation (10 pA leakage current). As the channel encoder will run for the whole cycle time, the sampled hold signal must not droop too much. So a JFET op-amp will also be used for the S&A. As a consequence of the channel encoder (four channels), I will need many op-amps. I won't use a PLL for the synchronisation. The micro will be triggered for a cycle start and the clock jitter will be low.
I will try to do also some improvements on analog front-end stages. I hope, I can get all the necessary parts. If not, I will do some discrete solution. The laptop software will do some impressive noise filtering and signal recovering.
Now, the run for a true 24-bit PI is started.
AzizComment
-
It seems I am making further steps forward... S&H and channel encoder (modulator) are running.
But I didn't find spice models for some interesting analog MUX for the channel encoder (modulator). Anyway, the channel encoder is done with discrete solution using lots of transistors. It has only a big disadvantage: simulation time takes longer... I will break the whole circuit into functional parts and analyze them individually.
AzizComment
-
Hi all,
it seems, I will have following voltages:
° 12 V battery power supply (single power supply solution),
so seen as -12 V.
° -5V digital (using 79L05 or similar)
° +5V analog from DC/DC charge pump (using 78L05 or similar)
° -9V analog (using 79L09 or similar)
The analog signal level will be transferred into the negative power supply rail, having more dynamic range and supply current. -9V/-12V will also be used for switching some N-channel JFET's if necessary. The micro is providing -5V logic level and makes the controlling of JFET's easier.
I will implement an gated integrator in the front-end. The PI module will only be an experimental front-end, which will make lots of experiments possible and is a basis for finding improvements on PI technology.
AzizComment
-
Hello friends,
to make a flexible experimental PI hardware design, I will need 15 I/O lines at the moment. They all can be realized with the ports B and D of the ATtiny2313 (20 Mhz). If not, I will take a 16/20 Mhz ATmega8, which has more I/O lines and some additional functionality.
I will think of the two micros. But I tend to use the ATtiny2313 due to simplicity and cheapness (I have already 5 samples at home).
I will support 4 different modes of operation. One push button is reserved for ground balance. The micro does not do any real ground balancing but the micro is controlling the user-interface which informs the laptop in real-time. So any offset voltage errors in the circuit/ground balance will be tracked in the laptop automatically. The laptop software will be configured for each mode of operation.
The PI module will have a module interface (analog/digital) so this can be expanded with additional modules to a full PI machine later (without laptop). Laptop simplifies just the experimenting and development in the first stage.
AzizComment
-
Hello friends,
I have decided to split the experimental PI board into two parts:
- laptop/sound-card interface module (+power supply, +synchronisation)
- analog/digital PI front-end module (analog/digital interface, micro controller built-in, user interface, but no ADC built-in).
The laptop interface module will be thrown away later until many experiments are done. It will help to optimize the detector for the later versions. The laptop interface module has much more complexity than the PI front-end module itself. But there is no other way of this, when I want to use the 24-bit ADC of the sound-card and the huge processing power of the laptop. Also the possibility of collecting many field data, field testing and "on the field developing".
The channel modulator will mix 4 triangular wave forms from each channel having different frequency. The 4-channel signal demodulation will be done in the laptop software using lock-in amplifier of course.
The analog/digital PI front-end module will have pure analog and digital interfacing. This can be extended by using an add-on processing module into a full PI detector later.
I didn't find the spice model for CD4053 yet (CMOS analog MUX). So I took 8 JFET's for the modulator (two for each channel) until I get the spice model for the 4053 some time.
AzizComment
-
Way to go Aziz, you don't let the grass grow under your feet. We'll be interested in your progress!Originally posted by Aziz View PostHello friends,
I have decided to split the experimental PI board into two parts:
- laptop/sound-card interface module (+power supply, +synchronisation)
- analog/digital PI front-end module (analog/digital interface, micro controller built-in, user interface, but no ADC built-in).
The laptop interface module will be thrown away later until many experiments are done. It will help to optimize the detector for the later versions. The laptop interface module has much more complexity than the PI front-end module itself. But there is no other way of this, when I want to use the 24-bit ADC of the sound-card and the huge processing power of the laptop. Also the possibility of collecting many field data, field testing and "on the field developing".
The channel modulator will mix 4 triangular wave forms from each channel having different frequency. The 4-channel signal demodulation will be done in the laptop software using lock-in amplifier of course.
The analog/digital PI front-end module will have pure analog and digital interfacing. This can be extended by using an add-on processing module into a full PI detector later.
I didn't find the spice model for CD4053 yet (CMOS analog MUX). So I took 8 JFET's for the modulator (two for each channel) until I get the spice model for the 4053 some time.
Aziz
-SBComment
-
It isn't an easy job to pass 4 channel DC analog signal data through the one AC coupled line. I am not sure, whether all the efforts were worth to make. But the 24-bit ADC resoultion is really very very attractive. I will loose certainly some bits of the 24-bit ADC resolution or the number of measurements per second (max. PPS rate will be lower).
I will think of further modulation technieques in regards to easy to realize. The one I have developed so far is one possible way.
AzizComment
-
Hello friends,
Next excursion into topic: analog front-end, pre-amp, front-end blocking/clamping
I tried to improve the signal quality of the analog front-end. And I have found a method, which doesn't need the front-end blocking via switching FET's nor the normal diode clamping. Usually, a 1k resistor followed by anti-parallel diodes will be used to clamp the overvoltage signal from the coil.
A new clamping circuit will be used, which prevents the saturation of the pre-amplifier and allows a sampling at 2 µs. The pre-amp need no any clamping diodes on the feed-back path.
This method will also work with "ringing" coils (underdamped PI coils), which doubles the signal quality by rectifying and filtering the ringing signal.
AzizComment
Comment