XL500 sample timing

[porkluvr]

Micheal, you should look at the VCO schematic I drew in LTspice and compare against the perfboard you constructed. Maybe I made a mistake transcribing into Eagle, or maybe you made a mistake in wiring. In any case I'm glad a whole lot of time and energy wasn't wasted. But we haven't learned anything because the problem remains unsolved.

There are plenty of things I don't know but every auto threshold adjust I remember seeing is between two capacitively coupled amplifier stages. Unfortunately, the XL500 is DC coupled throughout. Look at the Hammerhead and see its auto circuit.

http://www.geotech1.com/cgi-bin/pages/common/index.pl?page=metdet&file=projects/hammerhead/index.dat

It is only two added parts, but again it is between two capacitively coupled stages.

It occurs to me that the auto threshold circuits I have encountered are intended to prevent the output from saturating and causing false target - the opposite of what you previously described (signal fading).

Perhaps what you need is a switching regulator to maintain your power supply at 8V throughout your battery's discharge. For my detector, I would not have it any other way. I used LTspice to devise my power supplies, but neither of my projects is finished so I can't say "for sure" that my switching power supplies will do all that I want. But all my eggs are in the switcher basket.

Have you measured the battery voltage to see if it is actually sagging as you suspect?

Do-it-yourself switchers aren't expensive but they require a carefully crafted PCB for good results.

edit: I just now looked at your previous post. You should clean the flux off the circuit board where you soldered. Alcohol works pretty good.

[Micheal]

the 2 parts to remove to achieve a mono tone front view

[porkluvr]

Here is a variation of a power supply I plan to use. The main difference is that I changed R4 to produce 8V instead of 10.5V as in
my original circuit. Although I show efficiency report numbers, they may not be accurate. They are what LTspice calculated, but then
LTspice reported 97% efficiency for a 14V input which I highly doubt. So I "tested" at 13.5V input and got a more believable 88%.
LTspice's power efficiency calculator will sometimes make measurements at the wrong time and produce incredible results.

If you don't like wasting 15%~20% of your battery power for a regulator, then maybe you'll prefer your tx voltage dropping steadily
throughout the battery's discharge cycle? Take your pick.

I added a a filter section to make a pi filter at the output. Although the output is not rock steady I prefer a pi filter to a linear regulator
because there is less power loss and I figure the output is steady ENOUGH. I think so, anyway. This is not the crown jewel of
electronics technology we're working with. (No offense?)

Since I have already designed a PCB for my PI detector it is not much of a problem to remove some extra components and adapt my previous
work to the XL500. I make no claim of suitability, so use at your own risk. My detector is still in work, so this circuit is UNTESTED.
I'll post a PCB layout in a day or two.

[Micheal]

I have attached a picture of which ones to remove and add jumpers to...please let me know your opinion and the other question ..is it safe for the ciruit to run on direct 9volts from the battery without the voltage being dropped down by the regulators .

[porkluvr]

I have attached a picture of which ones to remove and add jumpers to...please let me know your opinion and the other question ..is it safe for the ciruit to run on direct 9volts from the battery without the voltage being dropped down by the regulators .

Huh?

NO.

I do not think you can safely run from 9V without having some way to adjust the TX pulse width intelligently.

You are going too fast for me. Do not remove the regulators, that would be very bad. (OW!!!)

Step back and take at look at what you're doing (you are scaring me). Without an oscilloscope and some more technical knowledge you are in very dangerous territory messing with the power supply.

Without having a schematic and PCB layout for YOUR model there is high likelyhood for error unless proceeding VERY carefully. Good communication is important.

I'll have some information tomorrow. It's too late today. I need to crash before getting too technical and making a mistake .

SLOW DOWN. THINK things through. Don't get over-confident.

I've been working on this all day and I need a break. (I'll be back. Tomorrow.)

edit: For what purpose were you asking about removing components? You gave no explanation. And remember, we have no documentation for your model so that means GO SLOW. Isn't your board different from Homefire's? How can I know what you are pointing at - a bunch of wires in the way... I'm lost!

[porkluvr]

Micheal. Your detector PCB looks like it is about 80% the same circuitry as the first circuit and PCB that Homefire posted in this thread. So it would be a good idea, and probably not to difficult for you to trace out the remaining 20%. That way you would know where the original circuit leaves off and your model begins.

I realize that what I just said may sound like a daunting task, but without a good idea of exactly what is on your PCB, instead of just guessing, I don't think it is wise to go hacking a power supply modification.

The VCO modification that I came up with before was simple and only involved a small portion of the entire circuit, and so things could fairly easily be put back the way they were if it did not work - as was the case. That was a bad omen.

Adding a switching regulator power supply is something of a different scale which would affect the entire detector. It would be irresponsible of me to try and walk you through such a task without my having more knowledge of your exact model, AND, your having a good deal of skill, knowledge and also, test equipment.

By your guessing at what components do on your PCB suggests to me that you need to gain more skill and knowledge or else have somebody do the work for you, or else forget about a power supply modification.

Between us, I don't think we have what is required to do the job succssfully, and I don't need a failure of such magnitude as this may turn out to be. I do not have time to teach you basic electronics and try and walk you through a tricky modification on a detector that I don't have a layout and schematic for. I'm sorry to have to tell you that.

Maybe if you got a new battery you wouldn't have trouble with the threshold fading so much

[porkluvr]

Possible mistake in parts placement diagram

Micheal, I have compared the picture of your board against Homefire's parts placement diagram. I think that the FET in the picture was installed backwards. Here's why:

I made an incorrect statement very early in this thread; that if an IRFxxx were to replace MJE3055 then it must be reversed 180 degrees. That was incorrect. I appologize.

Another issue is that the parts placement diagram shows the MJE3055 power tab towards the edge of the board. After looking at a MJE3055 datasheet it appears to me that the parts placement diagram is wrong in this respect. The power tab should be shown towards the inside of the board. This may have been partly the cause of my getting the pinout relationship wrong... but I should have know better. The power tab is shown backwards, RIGHT?*

I would like to point out that I still believe that any coil driver replacement should not be attempted without having a scope handy. Installing an FET in place of the MJE3055 will result in a different pulse width driving the coil (hence, coil current will be different) and will also require the damping resistor value changed. Also, drive requirements for BJTs and FETs are somewhat different so more parts besides just the MJE would need replacing.

*Feedback would be appreciated. I don't have a XL500 to look at.

[Micheal]

Hi porkluvr...Thank you for your patience and thank you overall....I totally agree with the way you feel about the Auto threshold mod...I have not got enough knowledge to proceed with this and I dont know anybody around here that does know anything about these..
I have had a look at the mje3055 and its not the same as what is on my machine....the one on my machine is RFP8N1NL and it is in the same way as the first drawing you show ,that is the heatsink is on the boards edge side . I will try this mod again becuase nothing bad will happen to my machine ...But what I am going to try this time is the IRF 540 and I will see what happens and I will try it both ways and trust me xl 500 can handle my abuse .
The other question is which one is the damping resistor (I hear a lot about this and cannot locate it )?is that the 100k resistor directly on the right of the mje 3055
I know you dont really want to educate me on the topic but please hang in there ....who else is there that can help us newbies and just think of all the xl 500 owners waiting to see the results.
And I am also going to retry the VCO and go right through it again I am just wondering if I do this a different way. I really think that if I take the source signals from the signals that go to the speakers (headphones)I am going to have better luck at it ,because the source signal voltage at the speaker(headphone)increases as the target is closer to coil ...hence isnt this what drives the vco and make it do what it is meant to do,the higher voltage signal the higher the pitch in the tone...So I will try to incorporate the existing vco unit that you showed me to somehow work in this way... Like the guitar effects pedal I have shown in previously....and it all then goes into a separate amplifier box rather than using the headphones ....

I,ll give all of this a go and let you know how it goes ...cheers

[porkluvr]

...the one on my machine is RFP8N1NL and it is in the same way as the first drawing you show ,that is the heatsink is on the boards edge side .

Is that a typo? It looks like RFP8N18L in the foto. I think they have modified your board layout compared to Homefire's. They turned your part around! Check it out. The tab IS wrong in the parts placement diagram but on your model's board they have reversed the pads and that component's orientation! If we were looking up from the bottom of the board then the tab would be correct, but we're not.

http://www.datasheetcatalog.com/datasheets_pdf/M/J/E/3/MJE3055T.shtml

No big deal, except that I was looking at that PPD picture and comparing with IRF640 pinout when I made the incorrect assertion that IRF and MJE pinouts were mirror images. (Stupid me.)

Moving right along. Yeah, please try to get the VCO working. I really believe it should be compatible with the XL500 but I may have "wound" the component values too tight because I tried to eliminate unused potentiometer travel. You should get something out of it if it's wired correctly, i think.

I'll share the switching regulator -uhh soon. But after our success with the ultra simple VCO I have a troubled conscience. I roughed out the switcher PCB layout but it needs checked for errors. (Now, that's going to be a trick, checking my own work AGAIN.) I may etch that PCB after I get my Looboo PCB layout completed so I can do them both at same time. I haven't seen Looboo for a couple of days, so I'd better get busy. It's almost finished but there a subcircuit I still haven't quite de-bugged.

edit: Ohhh, about your damper. Look for a 1K resistor connected between the coil output and ground. Maybe your coil is different so that they may have changed the value, but it should be near the coil (unless it's IN the coil!!?) or maybe mounted on the connector.

[porkluvr]

Micheal, since Garrett has replaced the MJE3055 with an FET they would have almost certainly changed the value of damping resistor unless they have changed the coil specs, which seems ludicrous. (Changing coil specs to match a damper? Naaah. I don't think so, anyway.)

You should be able to trace the drain circuit and find the damping resistor.

But wait. You already have an FET. What possible benefit can be gained from replacing it with a different FET unless you have the equipment and know-how retune your system? Take the time you would spend doing that counter-productive action and get to know what you already have.
If you hack your PI's coil and driver (the FET) without having the basic tools AND WITHOUT BEING FAMILIAR WITH THE CIRCUIT you are wasting time. It's foolish, so leave the FET alone.

Search through the forum on topics of interest and do some browsing.

And remember: you should buy an LCR meter if you don't already have one, and a dual trace oscilloscope. It doesn't need to be a 500Mhz quad trace Tektronics. 35Mhz or so should be plenty good. Those are two basic and instruments. They are indespensible for metal detector hacking.

edit: Please wash the solder flux residue off your board if you haven't done so already.

[Micheal]

I tried the VCO again ...and I just cant get it to work...Can be kind enough to mark where the 5 wires have to go on the photo of the board I have provided please ...I have gone through the VCO board and it seems to be as your diagram.Without faults.But just can get anything to happen.I feel that I am not connecting the wires to the correct locations on the board.If you think you know locations please mark the locations on the photo I have provided earlier this will be much easier for me to understand.

Also the photo that I have provided previously above that has the remove indications on them are the parts I have removed so that the xl 500 has mono tone and it is working as such perfectly(thanks to you porkluvr).


By the way I am learning and making a little progress.(i can now mesure resistors with my multimeter).

[porkluvr]

.If you think you know locations please mark the locations on the photo I have provided earlier this will be much easier for me to understand.

Simply put, I can't easily follow your fotos and trace out your board when I don't have an accurate schematic. I would prefer the board in front of me; those pictures wont do. I would need higher resolution pics.

I pointed out on the schematic where to make connections. I will try to explain.

Start with ground. Ground would probably be a fat trace running around the edge of the board. The (+) side of the large electrolytic should connect to it. Look on Homefire's schematic and find it on your board. Ground is where you would connect the black probe of a voltmeter to make voltage measurements.

(+5V) Look at the schematic and see where +5V is. Find the LM741 pin 7 and CA3140-7 (also LM358-8 and LT1037-7) and use a voltmeter to confirm +5V on those pins. Then look at the bottom of the board and trace back to the voltage regulator LM78L05. You should be able to read the part number on the regulator. LM78L05 Pin-1 is +5V regulator output.

(-5V) Find LM741 pin-4 and CA3140 pin-4 (also LM358-4 and LT1037-4) and use a voltmeter to confirm -5V on those pins. Then look at the bottom of the board and trace back to the voltage regulator LM79L05. LM79L05 pin-3 is -5V regulator output.

I make mistakes so you should download LM78L05 and LM79L05 datasheets and double check what I said. Actually, saying pin-1 or pin-3 is meaningless unless you know where to start counting from - so please get the datasheets and take a look.

So we have three of the five wires, right? Try to find connection points for those three wires as close to the regulators and their output filter caps as possible. After you locate a good tap-in point you should drill a small hole, (make sure you don't drill into a trace) insert wire from the top of board and then solder underneath. This affords some strain relief. Or, maybe you can find jumper wires on top of the board to solder to.

For the VCO input locate your CA3140-6 (the output) follow that trace and the resistors until you come to the LM741-3. You should probably solder in the mid-point, to the junction of the two resistors and capacitor between CA3140-6 and LM741-3. They may have changed your circuit so keep your eyes open. I doubt it, but maybe.

VCO output should be at the end of the 47k resistor that connects to the LM358-7. You would need to disconnect the LM358-7 because that WAS your sound oscillator output but is no longer.

Sorry, but I had trouble following your fotos. If you need more help, I need a clearer pic. Send to me directly by email.

[Micheal]

Ok porkluvr thank you ,this is great,I am starting to see things more clearer because I was way off i was looking in other places the way you are describing things I am making progress I am starting to understand ... the description is much better now porkluvr .I have downloaded the datasheets like you said and I am going through your directions at the moment ...and I feel I can decipher things easier now...(If I cant work it out I will email you hi res pictures)...its looking good I,ll let you know how I go .

[Micheal]

"VCO output should be at the end of the 47k resistor that connects to the LM358-7. You would need to disconnect the LM358-7 because that WAS your sound oscillator output but is no longer."

I just want to confirm that I understand this correct please correct me if I am wrong...ON the LM358 pin number 7 is not to be connected so I am assuming i will bend or remove that pin (leg) so that it doesnt connect into the saddle therefore it doesnt make contact...

[Micheal]

I have connected the vco today ...results are that there was sound but no vco ...the sound is a zzzzzzzzzzzzzzzzzzzzzzz type of tone ,the detector is working and detecting but like I said with a zzzzzzzzz tone and the output volume is weak or maybe I should describe it as low...I have gone through everything a few times and checking .W2 W3 W4 were connected to the right spots which was not difficult but w1 and w5 were a bit difficult because the is circuit around the LM741 and CA3140 are different on my machine,it is not the same as the older models (I cant find the schematics anywhere on the net for this thing)...the 5 volt power is actually +&- 5.69volts power and it is consistant at all the points you mentioned. Nonetheless I have gained a little more knowledge this time around.Things are starting to become more clear...
Thanks for your help porkluvr I appreciate your help very much .
While I was doing all this and the machine was open I added a switch to the 5k resistor mod so I can switch between 5k and 12k and it works nice.
I also added a switch for the tone so now I can switch between mono tone and the original pulse tone ,also works nicely...and the switches are inside the machine so no wholes are drilled on the waterproof housing.
Back to the VCO I am leaning towards the Idea of...where the headphones are connected there is a + & - terminal ,from this terminal feeds the speaker (headphones) and also a red LED ...this red LED gets brighter as you get closer to the target(same as the volume at the same time) which is a nice little feature ...when I connect my multimeter to this +& - terminal I have discovered that the xl 500 is actually a lot more sensitive than it is able to put out audibly.On a simple 20 cent ozzy coin I can see that the voltage increases on the multimeter noticably a good 5cm before it is audibly hearable and on bigger objects like a coke can It is much more.So the audio amplification is a weak part of this machine really hiding its true sensitivity...
I am thinking that,going back to my idea earlier a vco power from this source is from 0volts when there is no target to a bit over 3volts when a target is right near the coil.Do you think it would be easier to incorporate a vco from this +&- terminal...I am looking around on the web and I have found a few simple looking shematics but nothing in a pcb form ,and at this stage I dont have enough knowledge to turn schematics into pcb design so I am hitting into brick walls...any ideas or should I end this VCO segment.