Bandido 2 Modified

[Qiaozhi]

I have posted a video of my FireballCNC milling my new PCB for my Bandido. I'll try to post a picture of the finished product. The best part of using CNC is that it drills all those holes for you and it does not miss any pads.

Merc

http://www.youtube.com/watch?v=kqA4ohzdEus

That's cool.

[Merc]

Here is the finished board after drilling. Not perfect but getting better. I learn a little each time.

Merc

[Merc]

I have posted a video of my FireballCNC milling my new PCB for my Bandido. I'll try to post a picture of the finished product. The best part of using CNC is that it drills all those holes for you and it does not miss any pads.

Merc

http://www.youtube.com/watch?v=kqA4ohzdEus

[Merc]

Well, I can confirm it is a ground problem. I powered all electronics except the audio output from the internal battery. I then used an external bench supply set for 9V to feed power to the audio output circuit through R14. Then I tested while moving where my ground lead from the bench supply was clipped to the PCB ground. Attaching the ground to one end of the board caused obvious oscillations, even worse than before. Attaching the ground lead to the area closest to the audio output circuit produced normal operation. Now, weak targets simply sound faint with no hint of breakup. So it looks like I need to pay closer attention to grounding, especially in the area of greatest current drain such as the audio output.

Merc

[Merc]

Well I tried using JFETs on a separate breadboard. It worked, but was no better or worst than before. I had the same break-up on weak targets. I have gone back to using the 4066 on the pcb. The problem that I'm seeing is occurring at the output of U9. This should be a shifting DC level as the target moves past the coil. With a scope, while moving a coin 5-6 inches away, I get a slight oscillation or bounce which is what is giving me a broken response. It seems as though it is related to the audio output circuit. If I lift R14, the 20ohm resistor feeding power to the audio output, the oscillation stops. So either the supply is sagging or the ground is not low enough impedance when the extra current demand of the audio output occurs. Next, I will try using a separate battery to power just the audio output circuit. I think the original Bandido used a single battery, so I should be able to as well.

Merc

[Merc]

I have fixed the mistakes that have been pointed out. Here is the updated schematic. I think I need to test before building a board for it. I should be able to breadboard the JFET sampling circuit and run signals between my current Bandido PCB board and the breadboard. I'll keep the leads as short as I can to keep the 60Hz noise down, but since I'm dealing with basically audio frequencies, I think it will work. I'm hoping I can at least figure out if it will definitely not work (using JFETs) or if I just need to change a capacitor here and there to shift phases a little.

Merc

[cr15]

Working circuit

Hi Guys. Time for me to weigh in. Just built the PCB as per Bandido_Paul-Panzer_03-2006. I own an original Bandido, and a Rapier and find that the circuit values and designations are pretty much all the same! It looks like Tesoro developed one circuit and just tweaked some bits to make a new/different model.
Anyway, this gave me something to refer back to when I found problems. One of these was that I could not get a threshold. I finally pinned this down to not using the exact same Op-Amps as Tesoro used either because I could not find any or I thought they were too expensive. In this case I was not using TLC2262 for IC3, 6 and 10. As pointed out in above postings, when in All Metal and manual mode, the input of IC6a pin 2 is left floating. Not a problem if using TLC2262 as it goes to the rail voltage giving a solid 0V at IC3b pin 3. But in my case using LM358 it floats high giving 0.06V at IC3b pin 3 instead of 0V. Enough to screw things up. Now I'm going to be using CA3240 Op-Amps that have similar characteristics to TLC2262.
The other problem I had was that the negative volts generator didn't work properly, giving about -0.5V instead of the -3V I was seeing in my Rapier. The Bandido uses a 7660 circuit giving -6V from a regulated +6V. So as my box is big enough and roomy enough to take the PCBs and more, I removed that section completely and added a second battery and negative voltage regulator. Soooo much simpler.
And it does work. I also modified my Rapier to the Bandido II configuration so that I could manually ground balance and I'm getting more depth. But I haven't quantified that yet.
Next item on the agenda is to experiment with the output Tx power. Connecting the master oscillator to the battery rails will increase depth. Or maybe put an inductor in place of the Tx coil so that the oscillator still works, then put an audio amp between oscillator and the Tx coil. That should shove some current through it! And more current equals more magnetic field equals greater depth.
Good luck all!

[nick_f]

It Is CORRECT on the Circuit Board. Sorry, It is a 2N3904, NPN Transistor and Backwards on the Schematic.

And YES IT WORK. Also Tested Outside.

However I don't find the Descrimination very good.

Would be great if Someone could Varify Cap Values.
I have Recieved Numerous Opinions on these!

Hi Gary,

Does that NPN transistor have any role in your circuit? To me a jumper between C-E of that transistor could do as good a job as the transistor.

I believed the original Tesoro Bandido uMax diagram from Carl contains two errors:
1. the supply rail for the negative voltage generator should be connected to +B, not to +V!
2. R5, R6 should be 1kohm, while R65, R7 should be 22kohm.
3. The PNP transistor at the output has the role to reduce the voltage slightly, from approx -7V to approx -5V. The capacitor C42 in the base of the transistor filters the voltage on the resistive divider. This part is very similar with a voltage stabilizer with Zener diode and one transistor, except for it does not use the Zener diode (no need for stabilization of -5V). For reference, see fig 4 from http://www.bristolwatch.com/ele/zener_power_supply.htm

PS - It is RECEIVED, not recieved!

Regards,
Nicolae

[nick_f]

I believe the schematic I've been working from has an error. The modified negtive 5V supply does not work as shown. The PNP passing transistor does not pass any voltage when biased as shown. I believe another resistor is needed from collector to base. In the meantime, I have bypassed that transistor on my board and have good +5V and -4.77V.

Merc

Hi Merc,
I am having a laugh at how many people had problems with the -5V and yet apparently nobody picked up the errors in the circuit diagram. Maybe Carl intentionally created some faults with that circuit. I can tell you that by changing one letter and moving one decimal point between two sets of resistors, you will get the circuit to generate approx -5.6V on the -V rail Gary changed the transistor from a PNP type to a NPN type, but I don't think that transistor has any role in the circuit anymore, as compared to the one from Carl's schematic which has (at least) one clear function.

Cheers,
Nicolae

[Merc]

You are very correct about those JFETs and the Disc pot. I see a few other problems at some of the op-amps. They are flipped making signals go in the wrong inputs. I sure hope I didn't do that on my first attempt at the Bandido.


Merc

[porkluvr]

Hello, Merc.

I think you're right on with the battery test but I see a couple of other possible problems.

1) You should move the DISC pot connection to the other side of C9!

2) Remember that JFETS should only be pulled up to ground, not +5V as with the CD4066.

3) For the oscillator transistor, it won't make a ton of difference... but you should probably use 2N2907A instead of 2N3906. That might push a couple more mA through the coil.

Also, beware that the phase "sense" of your demodulators may be considerably different depending on whether JFET or CD4066 is used.

If you compare JFET an CD4066 demodulators side by side and rotate the phase of the input signals, you might find that the two outputs change from positive to negative at substantially different input phase angles. Unless I have it backwards (which is possible) the JFET side would be lagging the CD4066 side.

For example, I had a plan to use JFETS for GEB and DISC channels and use CD4007 as a full wave demodulator in an all-metal channel. Wierd though that may sound, that was my plan until I ran a side-by-side simulation and discovered that for a given receiver input to the demodulators, to "zero" the outputs the reference phase clock to the JFET needed to be almost +30 degrees relative to the clock on the CD4007. That ruined my day but at least I had not fully completed a PCB layout.

This was at 15kHz frequency but I think that (understanding that I was comparing apples and oranges for that test) that principle would probably still apply for your 10kHz Bandito II. It's no big deal so long as you don't want to use different styles of demodulators in the same detector but I thought I would mention it anyway.

[Merc]

Rev 1

I'm going to try again on my Bandido II. I have a nice case with a nicely printed label for the controls and the parts, I'm just not happy with the stability of the detector in the motion discriminate mode. During air tests, it gives a good clear sound on close objects...say 4-5 inches, however, as the object gets further out, there is break-up in the signal. The sound gets erratic and broken. The same if you place a coin 3-4 inches away from the coil and move it very slowly. Even my old Golden Sabre (factory original) gives a very clean and smooth signal on objects 7-8 inches out, it just gets weaker, fainter but not broken. After lots of measurement with a scope, the noise I'm seeing at the comparator stage looks like 60 cycle. Well, I am in-doors. That must be it. It seems to do the same standing in a field far away from electrical noise. Also, when I switch to all-metal mode, the detector is quiet. The all-metal mode uses the same receive front end and samples as the motion mode, at different phases of course, but the same ones so I should here it there. I'm suspecting that this is some type of phase jitter going on. I'm a little suspicious of my TC4066 switch I'm using. So I've decided to try using JFETs as in the earlier machines. I think it is easier to layout the pcb when you have separate JFETS instead of all the sampling done in one IC.
I've also made changes to the battery test circuit. I've decided that C7 (10uF tantalum) belongs where I show it on this new diagram. This is the only way to have a varying length tone based on battery voltage. This is also where it is on previous Tesoro designs. I decided to use a JFET here as well because they have an active region to give you an output that varies more in level. The 4066 is just going to switch on and off without varying. This is exactly what the Silver Sabre II uses.
Another change I have made is at the CD4024 divider. I have added an optional diode to give me a lower tone. I'm not really fond of the high tone that some of the new Tesoros used. By using D11 or D14, I can use a high or low tone.
I have also decided to sync the 7660 voltage mirror to the transmit oscillator. This should eliminate any problems with a "beat" interference happening.

[Merc]

The LM339 is just a 4 part version of the LM393. Yes, I believe it will work. It may make your pcb layout a little harder. Good luck.

Merc

[Rumcajs]

LM339

Hi all,

i planning building one Bandido panel.
Currently many parts are missing, and others are mixed.
I ordered three LM 393 dual comparator, but got three LM 339 quad comparator (pinout is other). The reclamation is long term and posting cost is bigger than 3 IC.
Therefore i'm replacing LM 393 with LM339.

What do you think from this replacement?

Brg

Rumcajs

[chemelec]

Corrected Schematic and Overlay Picture

http://www3.telus.net/chemelec/Projects/Bandido-Split/