GPZ7000 and GPX6000 Bypolar square Wave

[dbanner]

There's also:
SoC FPGA Smart Fusion 2
STM32F 32 bit microcontroller
A Bluetooth audio SoC CSR8670

[dbanner]

The folks at minelab(website) say the GPX series uses MPS and MP sampling.

The GPZ7000 is given the initialism ZVT (zero voltage transmission). I guess that translates to bipolar CC square wave.

[dbanner]

Yes, I also saw a bunch of HCT4053 gull wings in a row, plus another row of what I'm guessing might be comparators?

Actually, there appears to be a row of two different chips paired together. (6 pairs)

Interesting to see teardown pictures of those detectors. Nice.

[Mr.Jaick]

Hi I'm also in

so 14 MUXs followed by 6 pair of 8-pin Analog goodies then all go to 2 x 24bit 8 channel Delta-Sigma ADCs.

I can't see the traces from those ADCs going directly to the SmartFusion 2 FPGA but they're probably doing that, I wonder if there is any thing on the other side of the board, isolation?!

now the thing is the last two 4053 MUXs on the left must don't have the analog pair sooo...not sure!

I don't find any info about those two SSOP8 packages but one of them seems like TI 640CP0 which is non sense.


now the question is:
1) are they using sample and hold technique and then reading the output of those SHAs with ADCs?
2) or just Multiplexing the Pre-Amp out puts to some multi order band-pass filters then to the ADCs?

the first analogy is the right answer I believe because the ADCs only have 64 ksps which is clearly not enough for direct sampling the Pre-Amps and if they wanted to do that, they would do it with the ADCs directly

so with that being said, how many samples are they taking ?! too many MUXs and each has 3 x SPDT switches, doesn't fit!

[Mr.Jaick]

this type of close together dual board construction has been common in the GPX6000 and GPZ7000 series recently
I wonder how they connect together, can't see any headers, maybe flex PCB ribbon? or expensive high frequency board to board connections?(don't think so)
here is a shot of GPZ7000 board, look at the point to point coax connection on the board with nilon stands "FULL SPEED"

can you see the regenerative module in plastic mold?
some more GPZzz....

GUI:

and the HEAVY coil:

[ZED]

I would say the 6000 is a six channel detector,and most of the 4053 chips are demod's for sampling.

[Mr.Jaick]

I would say the 6000 is a six channel detector,and most of the 4053 chips are demod's for sampling.

hi ZED
two things are very interesting to me
1) too many switches are there with those many MUXs , only for 6 channels !
and also we have overall 16 channels of ADCs, that doesn't make sense.

2) why minelab pretty much never goes with the direct sampling method ?? (what are the disadvantages?)
i'm sure there is a benefit with discrete SHAs , maybe implementing analog filters are easer, cheaper and much more effective than let's say digital filters?
because Lorenz uses this direct sampling technique almost all the time, but Lorenz is not a nugget detector...
I like to discuss Lorenz as well should i do it here or open up a new thread?
because these Germans are very mysterious , they never talk about the technology they're using if they're using Anything at all!
maybe just a simple PI with high sensitivity + high quality construction = 8000$

by the way I'm not trying to reverse engineer these things or anything like that by any means, i'm just excited in properly engineered electronics, I like to explore them and possibly learn something new from them.

[ZED]

To many mux chips you say,...not really i say....most P.I. tech on the web is just simple two samples after a single TX period and a later earth-field sample,its understandable to view all the mux chips as to many for just six channels when the mind set is two samples after a TX period.
Im like you, sticking my nose in top tier detectors in an effort to better understand P.I. ,having a better understanding of P.I. makes one more creative when designing P.I. circuits,which i like to do.
I have never seen a 6000 before so my comment is only a guess,i exist in a circle of modders and repairers and so its that experience that leads me to the six channel statement.

[Carl-NC]

Minelab does direct sampling with VLF, the Go-Find, X-Terra, Vanquish, and Equinox models all use direct sampling. They probably stick with demods in PI to maintain the highest sensitivity.
Also, using 2 E08s doesn't necessarily mean they are processing 16 channels. You can feed the same signal into multiple channels to achieve an SNR improvement.

[Mr.Jaick]

sticking my nose in top tier detectors in an effort to better understand P.I. ,having a better understanding of P.I. makes one more creative when designing P.I. circuits,which i like to do.

yup love the high performance stuff.

[Mr.Jaick]

Minelab does direct sampling with VLF, the Go-Find, X-Terra, Vanquish, and Equinox models all use direct sampling. They probably stick with demods in PI to maintain the highest sensitivity.
Also, using 2 E08s doesn't necessarily mean they are processing 16 channels. You can feed the same signal into multiple channels to achieve an SNR improvement.

thanks for the insight.
when it comes to VLF it's a no brainer to go with the direct sampling, and yes i meant the pulse machines with that.
the thing is if you have two different channels taking the same signal and you add them later on you get 2x sensitivity, but no because the noise would double too.
so how do you think you could remove the noise from that?

[Orbit]

Minelab does direct sampling with VLF, the Go-Find, X-Terra, Vanquish, and Equinox models all use direct sampling. They probably stick with demods in PI to maintain the highest sensitivity.
Also, using 2 E08s doesn't necessarily mean they are processing 16 channels. You can feed the same signal into multiple channels to achieve an SNR improvement.

Hi Karl ! I didn't quite understand you think rx goes directly to ADC ?

[Carl-NC]

thanks for the insight.
when it comes to VLF it's a no brainer to go with the direct sampling, and yes i meant the pulse machines with that.
the thing is if you have two different channels taking the same signal and you add them later on you get 2x sensitivity, but no because the noise would double too.
so how do you think you could remove the noise from that?

Direct sampling is more difficult than using demods. You need a good ADC and good DSP to achieve similar SNR.

The signal doubles, but the noise only increases by sqrt(2).

[Carl-NC]

Hi Karl ! I didn't quite understand you think rx goes directly to ADC ?

Yes, from what I've seen they put a preamp in the loop and that feeds a 24b ADC in the pod.

[Orbit]

Yes, from what I've seen they put a preamp in the loop and that feeds a 24b ADC in the pod.

Yes, I agree !