Let's made a PC-base metal detector with usb interface !!!

[simonbaker]

Hi SB,

this actually isn't critical. But I see, the analog ground of the IC isn't optimal connected to the bypass capacitor due to an obstacle. I have found a much better way of directing the auto router instead of putting obstacles now. I'm sure, I can have a much better result this time.
Cheers,
Aziz

I assume the pad and trace were on opposite sides of board so no real danger!

I hope later you can tell us some of these Eagle auto-routing tricks you use, because it can be difficult to control auto-routers.

Route-on!

-SB

[Aziz]

Hi SB,

this actually isn't critical. But I see, the analog ground of the IC isn't optimal connected to the bypass capacitor due to an obstacle. I have found a much better way of directing the auto router instead of putting obstacles now. I'm sure, I can have a much better result this time.
Cheers,
Aziz

[simonbaker]

Hi all,

this is what I get from the dumb auto router. This version has an improved high current ground and supply path above. The voltage regulator is in its better position too.

I think, I've played enough. It's time to solder the bread board. Then I know, what must in and what out of the board.

Cheers,
Aziz

Squeek!

[Aziz]

Hi all,

this is what I get from the dumb auto router. This version has an improved high current ground and supply path above. The voltage regulator is in its better position too.

I think, I've played enough. It's time to solder the bread board. Then I know, what must in and what out of the board.

Cheers,
Aziz

[Aziz]

Yeah, better get one built in case you find you want to change the design!

It looks snazzy!

-SB

I do not hurry up. The PCB is getting better and better. No via's at all at the moment. Just starting the auto route script (in the order of the signal classes) and the PCB is nearly perfect. Well, I have to "tell" the auto router: "Damn! Don't go this way! Go this way, where I have made you a path without obstacles!".

I hope, I won't be unhappy about the pre-amps. There is some space for a two single op- amp modification. And I hope, I won't have to change it into a bipolar power supply configuration (that would cost more PCB space and parts). I hope, the active power supply filter does a good job (a simple transistor shunt regulator as AC filter).

If everything goes well, we will have the possibility to make three different configurations with the circuit:
- high power VLF (series LC tank configuration)
- efficient PI pulser (parallel LC tank configuration)
- standard PI pulser (parallel LR tank configuration, damped coil configuration)
(and more configurations possible ... )

For the simplicity of the circuit (restricted PCB area and beeing a K.I.S.S principle), TX pulse width is the half of the pulse period time (50 % pulse width). I'll show you, that this will do a good job too (bye bye MPS, DVT, multiple TX width, .. ).

Cheers,
Aziz

[simonbaker]

The longer I see the PCB, the more optimization possibilities I find.
It's far from optimal solution however.

I'll start soldering the bread board soon.

Aziz

Yeah, better get one built in case you find you want to change the design!

It looks snazzy!

-SB

[Aziz]

Procrastination is the thief of time.

Just go and build it. We're all waiting to see the results.

The longer I see the PCB, the more optimization possibilities I find.
It's far from optimal solution however.

I'll start soldering the bread board soon.

Aziz

[Qiaozhi]

The PCB looks pretty well but I will play a bit more.

Procrastination is the thief of time.

Just go and build it. We're all waiting to see the results.

[Aziz]

Hi all,

latest modifications just built in and tested with LTspice. The high current section is further improved now. High current ground and coil power forced to the upper/left corner section now. This will form only a small "air coil loop" and hence less EMI radiation to the circuit.

The PCB looks pretty well but I will play a bit more.

Aziz

[Aziz]

Aaaarrrrrrrrrrrrrgggggggghhhhhhhh!

Forgot something very important: Pulse disable watch-dog timer.
This is important to parallel LC/LR tank configurations as the mosfet can drain a lot of power if the mosfet is left in its on state.

I was thinking for the series LC tanks and this configuration does not need it (no DC current flowing to the TX coil).
Well, this addition requires one IC (schmitt-trigger inverter or NAND) and some passive components. This will fit in the board, if I remove the lower two buffer capacitors. The mosfet driver (IR2184) has fortunately a disable pin, which should be driven by the watch-dog timer.

It seems, that the mosfet driver enable pin (shut down) is pulled up to +5V internally. The whole watch-dog timer part can be left unpopulated for the series LC tank configuration in this case.

Back to the black board..
Aziz

[Aziz]

PCB getting better

Hi,

well, this is what the auto router gives with some little helper (putting restriction regions). Only small cosmetics must be done finally.
I am quite happy now, that the high current path (left/top region) and mosfet driver stage is good enough. This little beast shouldn't be underestimated. It can really push high current to the coil (therefore thick signals).
I am also quite happy about seperating the standard ground and analog ground. The analog section shouldn't be disturbed much now. Well, I have only one via on the PCB and this is really well invested.

BTW, this PCB isn't finished yet. Still optimizing it.
Cheers,
Aziz

[simonbaker]

Hi all,

the auto router seem to be quite good. But one have to auto route supply wires first. And one needs several different net class types.

Ok, I have only two ground planes. So I have seperated analog ground plane from standard ground plane to get a better performance.

The PCB output seem to be quite good, if the parts are placed good enough (see below).
I'll play a bit more to get a better PCB.

Cheers,
Aziz

Very neat.

[Aziz]

Hi all,

the auto router seem to be quite good. But one have to auto route supply wires first. And one needs several different net class types.

Ok, I have only two ground planes. So I have seperated analog ground plane from standard ground plane to get a better performance.

The PCB output seem to be quite good, if the parts are placed good enough (see below).
I'll play a bit more to get a better PCB.

Cheers,
Aziz

[Aziz]

Hi,

there is still some space for more buffer capacitors.
Still not optimal but it's getting better. I'll add two different ground planes and two power planes as well. But the PCB won't have more than two layers (cheap).

Aziz

[Aziz]

Yep!
Active power supply filter makes really sense and it improves the SNR (according to spice simulation: 20 - 30 dB). It only needs 6 additional parts and they can be left unpopulated on the board, if not really necessary. I'll support this feature in case of having more power supply problems (high ripple voltage, noise, etc.).

Well some more improvements are also made to the circuit.

Aziz