New Bipolar Boost TX and Front End

[KingJL]

... Are the forward and backward pulses sent close in time and then summed together or how does bipolar drive work?

Yes, Carl is right. The method of obtaining bi-polar operation in this instance is via a modified H-Bridge. The pulsing and timing of the pulses can be observed and obtained from the LtSpice schematic contained in the zip file of post #4 of this thread.

[moorejl57]

Yes, Carl is right. The method of obtaining bi-polar operation in this instance is via a modified H-Bridge. The pulsing and timing of the pulses can be observed and obtained from the LtSpice schematic contained in the zip file of post #4 of this thread.

OK, I looked at the sim and saw that the pulses at L1 alternate in current direction about every 500us. Thanks!

[waltr]

Thanks for the quick response. I guess my instincts were good. Are the forward and backward pulses sent close in time and then summed together or how does bipolar drive work?

Bipolar TX pulsing has been discussed in a number of threads. Carl has even posting basic timing diagrams.

Basically, Pulse TX then do the samples. Next pulse is at the normal pulse rate but in opposite polarity. Then samples are polarity reversed to add to previous sampling cycle.

The advantages is no need for EFE sample since reversing polarity of TX and RX cancels EF.
The disadvantage is more complex Coil drive circuit and the need to reverse the RX sampling. RX is not bad if using a microprocessor for the RX sampling, discrete timer chips would make this much harder.

[Mdtoday]

pushing... I did not take it that way. It is good that you have VHDL/Verilog experience... makes it much easier to understand what is going on. You will readily understand what we are building with the VHDL... an autonomous TX/RX concurrent processing unit that will be treated as a hardware peripheral by the embedded cpu.
..
I am not posting these changes yet... I will incorporate some other minor changes and remove some redundant VHDL first. I am also starting the RX module... need to wrap my head around the XADC to acquire the samples. All will be posted as they progress.

No problems JL, I appreciate the amount of work that you have put into this project, so thanks again for sharing.
For some reason my Vivado download has stalled so I'm currently trying to figure out what's going on.
The XADC looks interesting

cheers

Mdtoday

[Carl-NC]

Are the forward and backward pulses sent close in time and then summed together or how does bipolar drive work?

In a normal detector you space them just like a unipolar design. Positive -- process -- negative -- process -- etc. The negative pulse signal is inverted and summed with the positive.

[moorejl57]

In a normal detector you space them just like a unipolar design. Positive -- process -- negative -- process -- etc. The negative pulse signal is inverted and summed with the positive.

Thank you for the explanation, metal detectors are providing a lot of new stuff for me to absorb.

[moorejl57]

Bipolar TX pulsing has been discussed in a number of threads. Carl has even posting basic timing diagrams.

Basically, Pulse TX then do the samples. Next pulse is at the normal pulse rate but in opposite polarity. Then samples are polarity reversed to add to previous sampling cycle.

The advantages is no need for EFE sample since reversing polarity of TX and RX cancels EF.
The disadvantage is more complex Coil drive circuit and the need to reverse the RX sampling. RX is not bad if using a microprocessor for the RX sampling, discrete timer chips would make this much harder.

Thanks waltr, I had searched for posts on this and read until my eyes were watering. I guess the concept is taking a while to absorb (why does the signal remain and the noise is canceled?). After I complete my MPP Beach project (just 3D printed my coil form and housing to go with the Whites rod kit) I will start on designing a microcontroller based unit and was looking into various TX pulsing schemes. I currently design paintball electronics (hobby business, I am an software engineer at HP in Corvallis Oregon) for a major manufacturer so I have some pcb design experience in Eagle and a good amount of firmware experience with PIC's. My intent is to use an arduino compatible Teensy 3.2 board so that people can modify the firmware without having to read a 100 page data sheet. Most of what I have done has been surface mount since it is so much faster using a toaster oven than hand soldering.

[Carl-NC]

The advantages is no need for EFE sample since reversing polarity of TX and RX cancels EF.

And this also allows for higher pulse rates.

The disadvantage is more complex Coil drive circuit...

An H-bridge driver allows the use of a simple mono coil, but has higher parasitic C so it's tough to get super-fast speeds. I use H-bridge drivers where speed isn't critical.

[waltr]

Thanks waltr, I had searched for posts on this and read until my eyes were watering. I guess the concept is taking a while to absorb (why does the signal remain and the noise is canceled?). After I complete my MPP Beach project (just 3D printed my coil form and housing to go with the Whites rod kit) I will start on designing a microcontroller based unit and was looking into various TX pulsing schemes. I currently design paintball electronics (hobby business, I am an software engineer at HP in Corvallis Oregon) for a major manufacturer so I have some pcb design experience in Eagle and a good amount of firmware experience with PIC's. My intent is to use an arduino compatible Teensy 3.2 board so that people can modify the firmware without having to read a 100 page data sheet. Most of what I have done has been surface mount since it is so much faster using a toaster oven than hand soldering.

Yes, there is much to read. Check my HH2 build thread since I use a PIC16F88 for the timing and posted schematics and code as well as links to other threads.
https://www.geotech1.com/forums/show...ake-on-the-HH2

EF is not noise- it is a current caused by moving a conductor through a magnetic field which results in an offset Voltage. By reversing the TX and RX polarity this 'Voltage offset' is cancelled.
AC Mains 'noise' is also an offset during the much shorter TX cycle period so also cancels. Random noise such as from ignition systems can not be canceled.
.

[DonCaffeDelaSkobalj]

Big troubles in the mainstream physic !
Flux Cutting versus Flux linking (old queston) - Faraday versus Maxwell and Herring paradox as the experimental proof of 1st !

Well something that does not fit in many brains.
On pages (174-175) in the Berkeley Physics course (named Electricity and Magnetism, Volume 2, 1962), Edward Purcell (one of the greatest American authority in
physics and not only), has considered a sharp acceleration and deceleration of the electron (mass-charge particle).

What about the sphere EXIST around an electron in his OWN em. field ?

There exist a break in the electric field lines. Exactly at this place electron's E-component turn out to be 0(zero), same as on the walls of the waveguide. How unusual it is! And all this has happen in the open space, while you may hurry up to put a copper wall there - in detector coil.

Can you refute E. Parcella? I could not get.

E.Parcell preferred only X axis motion (sharp acceleration and deceleration or simply said pulsed X-motion) on the XoY plane. It is pretty narrow. oY axis is almost as good as oX. How to be here? This a place there Landau's 22 paragraph comes to help and rescue situation. Electrons need to be moved on cycloid. Only in this case, the electrons (mass-charge particles) are sharply accelerated and decelerated along the 0X and 0Y.

On this way we will get NON-UNIFORM, NON-LINEAR, PERIODIC CYCLOIDAL movement. And this is what big MD factories need to take into account if they want to make new MD breakthrough !

ps: Read physics courses.

Don.

[Mdtoday]

Big troubles in the mainstream physic !
Flux Cutting versus Flux linking (old queston) - Faraday versus Maxwell and Herring paradox as the experimental proof of 1st !

Well something that does not fit in many brains.
.........
........
On this way we will get NON-UNIFORM, NON-LINEAR, PERIODIC CYCLOIDAL movement. And this is what big MD factories need to take into account if they want to make new MD breakthrough !

ps: Read physics courses.

Don.

Um, yeah, OK, I think this is slightly off topic....
Please start another thread.

cheers

Mdtoday

[KingJL]

... I hope to verify the pulse transformer primary connections this week...

Well, I finally got around to verifying that the pulse transformer connections are correct... they are! The pulse transformer windings are in same direction on the toroid for primary and secondary. so there is a 180 degree phase shift from pin 1 to pin 3.
toroid.pdf

Mdtoday, did you find anything amiss with the boards?

[Mdtoday]

Well, I finally got around to verifying that the pulse transformer connections are correct... they are! The pulse transformer windings are in same direction for primary and secondary. so there is a 180 degree phase shift from pin 1 to pin 3.
[ATTACH=CONFIG]46547[/ATTACH]

Mdtoday, did you find anything amiss with the boards?

Excellent... thanks for the info JL

No, I did make some notes the other day , then got busy with work. The query was around U6 Pin 1 (NC) but I have just looked at the file again this morning and it all looks good JL.

The only other note I made was regarding the Artix-7 analog input range so I ran the Sim again just now checking the final output voltage swing and bias to see if it would exceed the Artix board analog input, it seems to be ok.

cheers

Mdtoday

[KingJL]

... The only other note I made was regarding the Artix-7 analog input range so I ran the Sim again just now checking the final output voltage swing and bias to see if it would exceed the Artix board analog input, it seems to be ok...

That is the purpose for R1 and R2 on the interface board... to bring the +20V down to 3.3V at pin 15 of the artix-7 cmod board. Then the resistive network on the cmod board brings it down to 1V.

[Mdtoday]

That is the purpose for R1 and R2 on the interface board... to bring the +20V down to 3.3V at pin 15 of the artix-7 cmod board. Then the resistive network on the cmod board brings it down to 1V.

Yes, spotted those after running the sim, thanks for clarifying this.