Coils

[Cody]

Yea, you are sure right on this. I think that is part of the thinking on the X-terra, light, single frequency, easy to use, etc, etc, etc, and on past the PR it looks like a good machine. I recall hauling the big bulky electron tube Whites that had the old A and B battery supplies. Man talk about needing a mule to haul the batteries.

I think I am stuck on the changing Q of the coil as that is a major obstacle. I may be beating a dead horse but think I see something here with the smart coil. However, it may be just a cute way to tell the control box what signal to send down. I think I said before it seems like engineers would put that concept to better use but then it is a lot easier to say than to do.

I don't see anything in the near future but time domain and frequency domain, square wave and sine wave, so my bet is most of what we see is a refinement of that. Seems like a safe bet since that is where it has been for 40 years. If the changing Q of a coil due to soil minerals is overcome then we are going to have a better machine. If we can drive the coil with a better square wave or sine wave and don't induce trash we have to clean up on the receiver side we are better off. I don't know how to quantify these gains. We will find out over the next few months.

At 7.5khz we are looking at a very small signals changes in the target to ID the target. In essence at 7.5khz the soil and target is almost saturated all the time. The target also appears to stand still at this frequency. It is that little dip for time domain or phase shift for frequency domain that tells us what the target metal is.

 

[Cody]

It is hard to let analog go but I am afraid that is where it is. You are sure correct on this and what I think the point is with the X-terra. Somewhere someone has to consider chip count and cost.

 

[Varmint]

otherwise you would have seen me mention that the square wave is generated by an output compare function on the u-controller.

I see also you ignored the fact that the DAC generating the psuedo-sine would need filtering much like I have here to get rid of the fast glitches that always occur when a DAC goes from step to step.

You also ignored that having the u-controller have to generate the DAC pattern takes away time from it's primary job, to process the signals that are in response to the TX signal you are spending so much time generating.

Amazingly arrogant reply you posted for someone who obviously didn't even read the post!

DAS

 

[Rudy]

n/t

 

[Rudy]

otherwise you would have seen me mention that the square wave is generated by an output compare function on the u-controller.

And instead I gave you an alternative implementation that wouldn't need the external components.

I see also you ignored the fact that the DAC generating the psuedo-sine would need filtering much like I have here to get rid of the fast glitches that always occur when a DAC goes from step to step.

Yes, I glossed over that, because the ucontroller DAC cannot drive the coil directly. It would have to drive some kind of a power amplifier. Since this amplifier only has to drive a sinusoidal current signal of a specifc frequency, or two frequencies assuming ML offers only two, it is cheaper to bandwidth limit the power amplifier to something a little over the detector frequency. The DAC glitches that may exists are in a significantly higher portion of the frequency spectrum (by definition of a narrow glitch) and would therefore be beyond the frequency capability of the power amp. The Tx coil's inductance would also not allow any remaining voltage glitch that got through the power amp to affect the sinusoidal current needed.[b/]

You also ignored that having the u-controller have to generate the DAC pattern takes away time from it's primary job, to process the signals that are in response to the TX signal you are spending so much time generating.

This can be done quite efficiently in several ways. For example, all the values (e.g. 64 or 128 steps in a digitized sine wave table should be more than adequate for this purpose) can be precomputed and stored in a table in memory and a pointer used to index through the table to the next value. There are other ucontrollers available with on-board DACs specifically designed to efficiently produce sounds. We could also use this type to produce the single tone needed for a specific coil. These types require even less software overhead to produce the tone.

Your assumption that the ucontroller would be totally consumed in just processing the received signals and have no horsepower left to generate the sine wave is .... an unproven assumption on your part. Might have been the case years ago, but modern day ucontrollers have an amazing computational ability, without even having to go to a 32 ucontroller, which are still more expensive than what MD designers would want to use.

Amazingly arrogant reply you posted for someone who obviously didn't even read the post!

Sorry, maybe I should have posted all the above thought process instead of assuming that you'd be able to interpolate it from what I wrote.

DAS

[Reply To This Message]

 

[Rudy]

I am not saying that this is the way ML chose to implement the XTerra, but that it is largely a digital design. If you are still troubled by an implementation using a ucontroller with an on-chip DAC, one could also use a ucontroller that includes some user programmable PLLs whose output can be taken off chip. Then all they would have to do is program the PLL to the frequency needed for the particular coil and use the output to drive a pre-amp and power amp. No DAC glitches to filter either.

So, I just explained two of many possible digital centric, low cost implementations.

 

[Varmint]

which you original post indicated you had not. Instead you decided to focus solely on my comment about analog electronics, then go off on a tangent about (paraphrased) how stupid I must be since u-controllers are part and parcel of modern detectors.

DAS

 

[Varmint]

and save the cost of having to have a DAC equipped controller when all they really need is a single ADC to process the RX info. ADCs are a far more common function in garden vartiey u-controllers than DACs, and certainly more common (and cost effective) than DAC equiped controllers with access to ROM tables, auto-sequencing for the DAC, or PLLs.

DAS

 

[Monte]

Eureka Gold, XT-18000 and Golden Hawk had three choices (6.4 kHz, 20 kHz and 60 kHz)

XT-17000 had dual frequencies (6.4 kHz or 32 kHz)

Eureka Ace (the earliest dual frequency unit I think dates to about 1989 (8 kHz and 19.5 kHz)

The GT-16000 and FT-16000 that replaced it were single frequency gold nugget units and their first dual frequency selection model was the Eureka Ace (as I recall) and that was followed by the XT-17000 that replaced the 16000 series.

Monte