Super TGSL With H Bridge TX Diff RX Quad Sampling

[ashefr10]

to gain depth

[Davor]

Increased what?

[ashefr10]

hello , can be increased slightly thee TGSL scope by acting on some component of the oscillator? thanks

[jladre]

For me tgsl is a very good project..why dont try to build one? "Caution not for beginner project"..

[Davor]

I promised a way to deal with phases so that the clocking gets right. I know exactly how to fix those, but I'm terribly busy lately. The vast majority of actual work here was done by Silver Dollar.

[Mike]

is this project dead? :\

I hope not. It looked very promising. We will have to wait and keep our fingers crossed, I guess.

[TreasureHunter]

is this project dead? :\

[TreasureHunter]

Helo everyone!

I'm new here, and very interested in this project? Are there any improvements? How is this project going?
Can we make the Super TGSL to be Super IGSL instead, to make two sounds instead of just one?
Something like this:
https://www.youtube.com/watch?v=3SPuarDiRds

[Davor]

scrungy_doolittle, the attachment is not correct and I don't know what it's about. Sorry.

[scrungy_doolittle]

It is no mystery. I found that I like proportional audio much better than the common binary beepy response. Because in an intuitive soundspace the volume/depth relationship is too steep for linear law amplification of target responses (the 6th power etc.) I found that equalising the response by means of a log-compression works the best. Hence a diode compressor. As a bonus, I can tolerate more gain for the small signal so my dynamic range is wider.
I also kept a similar arrangement for gating, as it keeps the opamps out of saturation. Because all the opamps in a same package draw bias from the common circuitry, and they all share common rails, any saturation would affect all the opamps in a package, at least by introducing a delay after a saturation event. The overall response is dead on punctual with diodes, and I can easily discern multiple tergets.

Thanks Davor, for me it is a mystery. I just need a guru to help me understand each block in some of these. The micros actually simplify a lot of stuff. The pic24 that I am looking at has 3 comparators, and two opamps built into the processor.
I am trying to understand part of the attached schematic. Specificaly, the receive amp. You will notice that this guy is running a DAC to the - input of this device. What would be the reason for that, and what would that accomplish? Additionally, why the small inductor to ground and the one to + voltage. I think he is abusing an audio amp for the transmit. Seems to me like it is the current you really need, and that could be done with an opamp, feeding t

[Davor]

It is no mystery. I found that I like proportional audio much better than the common binary beepy response. Because in an intuitive soundspace the volume/depth relationship is too steep for linear law amplification of target responses (the 6th power etc.) I found that equalising the response by means of a log-compression works the best. Hence a diode compressor. As a bonus, I can tolerate more gain for the small signal so my dynamic range is wider.
I also kept a similar arrangement for gating, as it keeps the opamps out of saturation. Because all the opamps in a same package draw bias from the common circuitry, and they all share common rails, any saturation would affect all the opamps in a package, at least by introducing a delay after a saturation event. The overall response is dead on punctual with diodes, and I can easily discern multiple tergets.

[6666]

Please enlighten me a bit. On U105, what is the purspose of the back to back diodes?

You will have to ask Davor, its his schematic.

[Silver Dollar]

I had some pcb's made to build an ARM M4 with Wolf chip to do something like the Xterra;


The processor can do the waveforms with DMA. And direct in and out for low mess and noise.
Maybe this winter I will start developing this board. I think it's easier to do analog but you
might be able to do some magic and see deeper with a good digital design. If nothing else
it's fun to play with!

http://www.geotech1.com/forums/showt...dec-Bare-PCB-s

[Davor]

Your problem with me is not my location, but my reluctance to mess with any micros. It is not that the micro can't be used even for a front-end, but the lengths you need to pass just to get there, and in my case a complete disgust with micros in a low noise environment of an amateur project. The very front end of a detector is best handled with pure analogue stuff, period.
Can it be done by a micro? Sure. To do it properly you should be able to sense the Tx coil's zero crossings and supply it with a corrected frequency - PLL-style. There is no need for extreme precision there, but so far such approach is not seen in any of the professional gears. You may supply it with a constant frequency, but you lose on power efficiency, and analogue solution proves better, again. Tx coil voltage is what matters.
Sampling Rx coil voltage by an ADC is also possible, but futile endeavour. What you need from a front-end are in-phase and quadrature synchronously demodulated components, as the rest of the magic happens on these. This magic can happen in analogue or digital world. Pick digital if you don't mind lugging lots of batteries. Or alternatively you may invest heavily to a digital part optimised for power consumption. Good luck with that.
And lastly, once you are finished with your design, so will be the lifecycle of your beloved micro. It happens more often than not.

If you look at what micros bring to the table, you'll see that it is mostly a way of obscuring your IP, human interface (bells and whistles) and precious little on top of that. Even multiple frequency gears can be realised in a manageable manner in analogue design, yet for some odd reason they are all designed with micros, and they all are battery hogs.

I'm considering making a dual frequency gear of my own with non-harmonic independent frequencies to avoid detection holes, active air signal cancellation, completely redone IF stage, and a completely different phase detector for direct VDI. All analogue. It can be done in a micro, but it would be very, very computation intensive. I'll take my time to do that, but I don't expect any of the commercial implementations in that direction by the time I'll finish it. Even by the template laid in this very paragraph. Why? Because commercial detectors operate under laws of marketing, and I don't.

[scrungy_doolittle]

All of this grief with the opamps seems strange. I have a project that I want to kick off. Davor, and Silver Dollar, I would like you to be involved in it. Basically it will be a hybrid detector.
A potent microprocessor (I'm considering the PIC24FJ128G010 (about 6.50 in singles). This sucker has two amps, comparators, 3 different resolution A/D's and enough I/O to hand the rest. The idea is to generate the frequency from the microprocessor. I would like to take that square wave, and feed it into the balanced H bridge for a coil driver.
Ground balance, and discrimination are easy then. On the rising edge of the oscillator pulse, inside of processor, you start two timers running. One is set for a count to produce the necessary delay for the ground balance, the other is set for the necessary delay for discrimination. When they time out, drive a line high, and reload them with the width of the sample time. When that times out, drive the line low, and wait for the next timeout on the oscillator. This eliminates the problems with opamps trying to provide a timing pulse for the sampling. From there, the initial phase of development would simply use the rest of the analog processing, and bring the X,Y and G lines into ADC's for VDI.
OF course there is a lot more that can be done. The problem is that I don't understand the *function* of the remaining analog stuff. I would bet however, that much of it can be done in software. I would want the project to be open hardware/open software. My original idea was to start with the IDX pro, but the TGSL would also work I think.

I realize that Davor is in croatia, and that might make things harder. If I an get this started, I'll have to get some microprocessor PC boards laid out. Not to hard to get done, infact, I might just use some of the quick prototyping boards for these i.e. they do make circuit boards that have footpads for a 48 pin quad pack, and bring all leads out to stake pins. I don't know if you guys have any programming expertise, but I have almost 30 years of embedded programming under my belt. I believe that the PIC can be debugged and programmed with either the pickitII or pickit III (cheap interfaces). The programming software is free.
I was thinking originally about using an 8051 because I already have a bunch of boards that would work well, but the feature set on the Silabs 8050F040 is not quite as good as the pic 24 series. The plus is that I have boards with displays, and already have tons of code running on these. (They would be scrapped boards from the product I work on, with most of the non applicable parts removed.

P