LM394 on my Minelab PCB

[Mechanic]

Hi Ghaj,

The N fet is to block the flyback voltage, the P fet is to block the voltage from the TX on to stop the preamp going into saturation and to lower power consumption.

Cheers Mick

[ghaj]

Hi Ghaj,

The N fet is to block the flyback voltage, the P fet is to block the voltage from the TX on to stop the preamp going into saturation and to lower power consumption.

Cheers Mick

Hi Mick,
Can I replace the P fet with a N fet with the same characteristics (I mean the same max. voltage, Ron, C, ...)?

[Mechanic]

I'm not too sure that it would work. For a start the control signal would be wrong. You could fiddle and see, but don't come and see me if the smoke comes out!

Cheers Mick

[mario]

1G11



Hi guys,

I'm working on my frequency mod.

It seems like the circled part is the oscillator (VCO?).
Pin 1 is GND
Pin 2 is VCC
Pin 3 is sine wave output of ~50,000-60,000 Hz. (I could be be wrong about this. I had a hard time measuring the frequency with my analog scope. However, this frequency is very small.)

The signal is amplified to 1.2Mhz by the two caps (1nF and 6.8nF).

Can some one with experience identify this part? It reads "1G11" or "1GH".

I've interfaced with my detector by exerting a ~53khz to pin 3. The detector becomes fully operational.

However, I've tried removing those two caps, shorting them with a 1.2Mhz signal and I lose audio. The only way to rev up the frequency would be to rev up from 50Khz. I would like to have the DSS signal directly rev up from 1.2Mhz-2Mhz instead of the 50Khz+. Is it possible?

[detectormods]

Hi, That part is a BC847

[mario]

Thank you!

[ghaj]

Another question.

I'm not too sure that it would work. For a start the control signal would be wrong. You could fiddle and see, but don't come and see me if the smoke comes out!

Cheers Mick

Hi there Mick. I have a question. Do you know how shielding of SD2000 coils are connected to the PCB? There is no information about it on ZED schematics.
Another question. I have made a complete SD2000 PCB on my own. When I turn it on audio CCT starts to oscillate (scream) at 2~3 HZ frequency. It has confused me. Usually noises from surrounding are from 50Hz mains or Radio Frequency sources. I don't have any idea what the source of this oscillation could be. Interesting is that when I turn BUK455 FETS off, but rest of CCTs still on, the oscillation persits. But when I disengage the coil and short coil connector on the unit oscillation nearly, but not completely, diminishes. I have used two different coils, with different number of turns and diameter, but frequency of oscillations is the nearly the same and in 1~3Hz range. I have not used any kind of shielding on my coils. could it be because of that?

[Mechanic]

Hi Ghaj,

When you are in town it is almost impossible to run the machine with a coil on. If you get away from interference it will most likely settle down to the level of what the noise is with the receive wire connected to ground. The shielding is hooked up to the coil ground wire at the connection down in the actual coil. Make sure the shielding is well spaced from the windings and is not too conductive. Something like aqua-dag or a good mix of graphite in some paint(test it to make sure it is conductive) paint the drain wire in. Make sure it is less than .2mm wire. Make sure your coil wire is made from strands that are .2mm as well. Tin coated works well as does litz, but good luck finding some of that with the right specs. The coil needs to be about 300uh and .4ohm. Make sure that you have used decent caps in the integrators(polypropylene) and the rest of the filter caps seem to work well with polyester. Make sure everything is bypassed properly and you have a good ground plane. What is the amplitude of the signal on the output of the stage after the dc blocking cap with the front end connected to ground and the coil connected? use a very slow sweep speed, .5seconds/div. It should be around 15mv peak to peak.

Cheers Mick

[ghaj]

Another question

Hi there Mick,
I have implemented every part of the machine on separate PCBs; That is one PCB for Power, one for front end, and so on. I have done it to be able to modify each PCB separately without need to modify whole the job. My PCBs are double side boards. Do you know how much layers have been used on original Minelab boards? (I mean just simple double side PCBs or three or four or more layer PCBs). I have decoupling caps and inductors for each PCB board and bypassed front end opamps. One thing more, I have used SD2200 front end CCT (AD797) instead of original SD2000 front end CCT.
At the moment, my big problem is oscillation of audio output of the machine. As you have mentioned environmental noises are very critical. when I ground input to front end receive CCT, I have an oscillating voltage about 200mv peak to peak on output of dc blocking stage, that is input to full wave rectifier stage.
Do you know how much effective is shielding the loop for reducing external noises? Could you explain your method in more detail? How and where the shielding should be connected to, on the transmit CCT? Do you know how Minelab shields its coils?
Another question, Why are 10nf caps so important? Unfortunately I could not find polypropylene caps and I have used simple MKT caps instead. In addition, Nearlly all my components are through hole parts. Could these be really the source of oscillations on my job?
Thanks a lot.

[Mechanic]

Hi Ghaj,

I believe there are some errors on the schematic that was put up on page 2 of this thread. Looks to me like the components around the 2nd and 3rd opamp are set up wrong? This could be causing oscillations. To keep things right, I would use that schematic for the front end up to the output of the ad797 and from there set the rest up the same as the sd2000 cct, so feed the ad797 into the 3.3k resistor and the front end will be fine.

What you can try with your setup is ground the preamp with the coil connected and then ground the output of the 2nd opamp in the front end. If your noise goes fron 200mv down to about 5~10mv then the problem is defiantly with the front end/preamp setup.

The reason I do the tests with the coil connected and the preamp connected to coil gnd is, everything fires up the way it should, and there is all of the correct currents flowing in the system. Then when the preamp is connected to coil ground we listen to the ground side of the coil as this is the same as listening to the detector with a coil on when it is not receiving emi or any signals at all. So when setup like this you will only hear internal noises., which will be the same noises heard away from all sources of emi.

The boards are 2 layer, make sure you use a ground plane on the bottom. Good luck!

For the coil shield, it is connected to the coil ground which is batt+, but it is connected to gnd in the actual coil at the end of the coax where you connect the coax to the coil wire. See here : http://s590.photobucket.com/albums/s...coil%20repair/ I don't actually have a photo of the connection in the coil but it is there!

Cheers Mick

[Mechanic]

G'day Ghaj,

You might want to check the 5+ power supply. If the inductor you have used is a low R one it could do what your machine is doing. The original inductor uses fairly fine wire so it would be a few ohms resistance, later I will check the exact size and resistance.

Cheers Mick

[ghaj]

Dear Mick,
You mean the 150uH inductor on power supply CCT, above 2N7000 FET? I have checked the +5V rail. It is nearly fixed on +4.92v and the inductor fires at about 200Hz (I mean nearly 2.5ms switching and 2.5ms off). But I can see switching noise on output of front end at about 100mv! But not on output of filtering stages. It seems to be filtered out in low pass filter stages. A question, Do you think dynamic response ( I mean slew rate, gain-bandwidth product, ...) of AD797 is similair to combination of LM394 and NE5534 amplifier? I was not able to get results from LM394 front end to compare the speed with AD797 front end. So I was forced to use AD797 CCT. By he way, I am going to do your suggested test. I will connect input of last NE5534 opamp in front end to +Bat and check output of DC blocking stage and tell you the results ASAP.
I have seen in photos of SD2000 board on this forum that three wires (black, red and yellow) goes towards coil connector. Is the yellow one going to shield?
Tnx.

[Mechanic]

Gday Ghaj,

No THE SHIELD IS CONNECTED TO THE COIL GROUND WIRE(outter coax) INSIDE THE COIL HOUSING!!!

The yellow wire is the front end receive wire. You will see noise from the power supply in the front end and about 100mv seems about right.

It is the second last opamp in the front end that you need to short to gnd, as the last one is just there to invert the signal so the integrator can subtract samples taken from that opamp from the samples taken from the opamp before it. so if you short the second last opamp, both the + and - samples will have the same potential.

As far as the ad797 giving problems because of slew rate, I doubt it! I am running a front end similar to what you are using and it works just fine!

What have you done with the trimmers at the integrator stage? If these are set wrong the detector can become noisy, but I doubt that they would cause the noise you are seeing?

Keep at it, you got this far....

Cheers Mick

[Mechanic]

Ooops here is some photos of the shield connection in the coil housing.

Ghaj, I would suggest you purchase a second hand coil from ebay for the purpose of getting your machine up and running. Coil building is an art in itself and when you are building the machine and the coil sometimes it could be very hard to find where the problems are. At least if you know the coil is good, then you are looking in your machine.

Also does your machine oscillate when there is no coil connected?

Cheers Mick

[ghaj]

m