A closer look at the AF108 bi-polar drive circuit

[waltr]

Can somebody please have a look at the AF108 schematic I posted and tell me that there is a quad nand gate chip(4011) mislabeled as CD4069?
Or am I missing something?

Yes, a CD4069 is a hex inverter. The schematic for U8 is definitely a quad AND (not NAND since the output pin does not have a small circle to indicate inverted, Negation).
So probably a CD4081.

https://en.wikipedia.org/wiki/List_of_4000-series_integrated_circuits

[dbanner]

Can somebody please have a look at the AF108 schematic I posted and tell me that there is a quad nand gate chip(4011) mislabeled as CD4069?
Or am I missing something?

[sinclairuser]

The AF108 was specifically designed as a mine detector, but in the end it failed qualifications (as I recall) and so the limited number made were sold off. Many were bought for the beach.

I think the AF109 is strictly a utility model, an unfortunate naming similarity to the '108.

For those unaware, White's UK often design their own machines and sell models not available in the US. They still make old models like the 1970s Beachcomber.

yes carl, that link i posted has some of the beachcomber machines two from memory, they must still sell or they wouldn't still stock them.
one interesting thing i noticed when looking at the different machines was one which looked like a coil with a control box mounted on a 3" stem was said to operate at 0.3 khz.

[6666]

No, it is not my file. porkluvr posted a screenshot here:
https://www.geotech1.com/forums/show...ighlight=xl500

Thanks for the thread link, yes thats the one I was thinking of, gee 2009, but no .asc file, mr pork does not post here now, oh well maybe one day.

[dbanner]

I just opened up my ltspiceIV program and to my pleasant surprise, I found a vast library of third party models in a folder labeled as "EXTRA", including CD4538, CD4024.
So I have everything needed to simulate the drive and sample pulses. I have never done anything like this before, so its going to be a steep learning curve. Stay tuned.

[dbanner]

Hi Dbanner do you still have the .asc file for that xl500 ?, saw it years ago but lost the file, thanks.

No, it is not my file. porkluvr posted a screenshot here:

https://www.geotech1.com/forums/showthread.php?14823-XL500-sample-timing&highlight=xl500

[6666]

I found this model for a CD4538 here on the forum, does anyone have other models, or can offer any advice?
I also see somebody did this simulation of garrett XL500, you can see 14538 in it. Now if only I can find the thread where this came from.

Hi Dbanner do you still have the .asc file for that xl500 ?, saw it years ago but lost the file, thanks.

[Carl-NC]

i hadn't realised these were P.I., they were available on the whites uk site, though they are not any longer there is an AF109, that one is 124 quid but is vlf 12.5 khz.
from memory the AF108 was a similar price cheap for the beach, i just looked and all the industrial stuff is vlf.
i also noticed they still have 3900 pro plus for 235 quid a very underrated and overlooked machine, i might get a new one at that price, mine is battered and tired but still works great.
https://www.whiteselectronics.com/pr...ty-gb/?lang=gb

The AF108 was specifically designed as a mine detector, but in the end it failed qualifications (as I recall) and so the limited number made were sold off. Many were bought for the beach.

I think the AF109 is strictly a utility model, an unfortunate naming similarity to the '108.

For those unaware, White's UK often design their own machines and sell models not available in the US. They still make old models like the 1970s Beachcomber.

[Carl-NC]

I think the comparators are supposed to monitor the symmetry of the drive train pulses, after all, you want fairly equal energy dissipation about the centre tap of the coil.

Yes, for detection of magnetic mines (as used for anti-tank) you need a net-zero magnetic transmission so you don't trigger the mine.

The next perplexing question is is the coil centre taped by inductance or number of turns ? good work btw.

Normally both. The best way to accomplish this is to wind the coil with bifilar wire. I don't know how the '108 coil is made tho.

[sinclairuser]

i hadn't realised these were P.I., they were available on the whites uk site, though they are not any longer there is an AF109, that one is 124 quid but is vlf 12.5 khz.
from memory the AF108 was a similar price cheap for the beach, i just looked and all the industrial stuff is vlf.
i also noticed they still have 3900 pro plus for 235 quid a very underrated and overlooked machine, i might get a new one at that price, mine is battered and tired but still works great.

https://www.whiteselectronics.com/product-category/metal-detectors-gb/security-gb/?lang=gb

[dbanner]

I found this model for a CD4538 here on the forum, does anyone have other models, or can offer any advice?
I also see somebody did this simulation of garrett XL500, you can see 14538 in it. Now if only I can find the thread where this came from.

[6666]

The next perplexing question is is the coil centre taped by inductance or number of turns ? good work btw.

[dbanner]

You need to add a center-tap ground on the coil... the '108 uses bipolar pulsing by alternately pulsing each side of a center-tapped coil. The SK2 side therefore generates both a negative flyback and a mutually-coupled positive flyback that are used as the RX signal. Same thing happens on the SK1 side, except those drive comparators used for (it appears) fault detection.

I appreciate your feedback Carl, so the coil is definitely centre tapped, I have amended the schematic.
I think the comparators are supposed to monitor the symmetry of the drive train pulses, after all, you want fairly equal energy dissipation about the centre tap of the coil. I think I just confused myself.
On second thought, the unit fault L.E.D appears to be triggered by two conditions, 1) A low battery condition exists( in conjunction with replace battery L.E.D) and 2) when the TX power drops below a predetermined threshold set by VR6 and VR7. I maybe wrong.
Anyways, I now move on to ltspice simulation of MC14538's, and associated circuits. I have to look for a model for the 4538 and others.
Maybe I should take out the slide rule and some graph paper and get going. No need for spice rubbish, I'll do it the old fashioned way.
On second thought I need to preserve every hair follicle that I can, at my age is danger zone for rapid balding, I'd better stick to spice.

[Carl-NC]

Now here is a clearer view of the bipolar TX drive circuit, much easier to follow. Seems uncomplicated. Now to figure out just how it works.
Carl, anything to say?
I need to know what are the pulses driving this array. That's going to require modeling the timing circuitry.

You need to add a center-tap ground on the coil... the '108 uses bipolar pulsing by alternately pulsing each side of a center-tapped coil. The SK2 side therefore generates both a negative flyback and a mutually-coupled positive flyback that are used as the RX signal. Same thing happens on the SK1 side, except those drive comparators used for (it appears) fault detection.

[Carl-NC]

Here is the modified power supply circuit I intend to implement. I use a 555 timer as a negative voltage generator. Now somebody can tell me if this circuit can work in place of the original. I think I'll be needing a replacement for the BUY92 transistor.

Looks about right.
I use 2N3904 & 2N3906 for all my generic bipolar needs.