Help to design a PCB for my PI design!

[1843]

hi 1843 schematic According to pcb Correct? why Q9 collector pin pcb connect to batt - schematic connect Q9 emiter pin batt- My head interfered Which one, correct, schematic our pcb Q9-Q2-Q3 I checked strong. I connected to According to schematic..

See the BC547 pinout:


It's ok in both the schematic and the pcb.

[night]

hi 1843 schematic According to pcb Correct? why Q9 collector pin pcb connect to batt - schematic connect Q9 emiter pin batt- My head interfered Which one, correct, schematic our pcb Q9-Q2-Q3 I checked strong. I connected to According to schematic..

[1843]

Night,
The layout of gate driver circuit (including Q9, Q2, Q3) is ok. Don't change R1 to 330 ohms or any other values.
The drain and source of the MOSFET need to be swapped on the PCB.
You may have burned your Q9, Q2, Q3. So change them.
On the components side, Q2 and Q3 are labeled wrongly! they need to be swapped too.

[1843]

hi 1843 schematic problem Q9 emiter connect to batt - top sik pdf Q9 collector batt - ? and Q2 emiter and collector ? schematic Different top silk pdf Different and R3 4k7 schematik 547 collector base connect topsilk pdf not connect r3 557 emiter and base connect Which one, correct, schematic our top silk pdf?

Hi Night, The MOSFET is laid out wrongly! And that caused your Q9 frying! as I said the schematic is right...

[1843]

Hi, The schematic is OK. Correct your pcb regarding that.

Thanks,
1843

[night]

hi 1843 schematic problem Q9 emiter connect to batt - top sik pdf Q9 collector batt - ? and Q2 emiter and collector ? schematic Different top silk pdf Different and R3 4k7 schematik 547 collector base connect topsilk pdf not connect r3 557 emiter and base connect Which one, correct, schematic our top silk pdf?

[moodz]

initial results.

Wellcome Moodz,

what about your designs depth results. air test results for coins, cola coke..etc.
regards

Well the design is not finished yet.

I am working on the descrim and earth field cancellation.

However this "single channel" model can detect a dime sized coin at 30 cm.
A fist sized nugget at 55 cm and a frying pan at 1 meter.

Thats why I need descrim as I dont intend to dig a 1 meter hole for old frying pans

[okantex]

Wellcome Moodz,

what about your designs depth results. air test results for coins, cola coke..etc.
regards

[moodz]

re balun mod

I was having the same problems with noise getting into the front end amp and though the switching 16 volt supply would be generating way more hash than a switched cap supply 7660 I could see from the circuit that essentially the amp has an antenna hanging off it because the coil is not differential.
ie one side is grounded and the other is fed to a high gain amp .

On that note why hasn't someone come up with a differential coil sensor for PI ? A differential coil sensor would be inherently self shielding from EM interference but still sensitive to magnetic fields.

OK the balun idea was tried on a whim ... I thought at the time that it would not work ... but then why not ... the balun is used in many situations where you need to couple an unbalanced Z ( pi circuit ) to a balanced Z (coil ) and the balun is ideal for broadband applications ...
ie keep broadband common mode interferance out ... but pass differential pulses ( also broadband ) which is what we want in this application.

I have not experimented with a voltage mutiplication scheme ie 1:4 or 1:16 balun however it may raise the possibility of driving coils with extremely low impedance ( ie inductance ) as the balun would appear to 'magnify' the inductance. Providing the benefit of an extremely 'fast' coil.

If anyone wants to use this idea you are welcome ....

To make a simple balun .... I find the best supply of the ferrite core comes from an old PC power supply where they are used as chokes . Locate a yellow / white toroid 1 " diameter ( doesn't really matter however make sure the toroid is not too small or the current pulse will saturate it ). Remove the original single winding. Wind on approx 10 turns of twin lead ( bifilar ) evenly spaced around the toroid. Secure with tape or glue and your done. connect the coil to one pair and you PI to the other. The balun should be as close as possible to the electronics to gain the full effect.

[Aziz]

Hi moodz,

Note the careful layout

Oh and I made some changes ... to remove alot of noise issues the coil feed line passes through a bifilar "balun" ... it does not affect the pulse.
ie reduces the unbalanced antenna effect of the feedline. ( I used to be an RF guy )

nice layout!
It seems, you have a 1:1 current balun. Isn't it? (it is difficult to see the details). What about using a 1:4 or 1:9 voltage balun? (almost inherent noiseless amplification).

Aziz

[simonbaker]

What an interesting read this thread is ....
I have just registered with geotech.

Here are some photos of my version attached ...

Note the careful layout

Oh and I made some changes ... to remove alot of noise issues the coil feed line passes through a bifilar "balun" ... it does not affect the pulse.
ie reduces the unbalanced antenna effect of the feedline. ( I used to be an RF guy )

I use a CA3130 as first amplifier ( FET inputs ... much higher input Z )

The uP is a DSPIC 30F4012.

A switch mode supply provides 16 volts for the analogue no sync.

A 12 volt batt pack from a $15 dollar drill powers the circuit for 4+ hours.

I love that construction layout! No PCB! It reminds me of the old time radio wiring where everything is attached to the chassis. Really nice! Very creative design. How well does it work?

I'm interested in the balun concept. Would that apply to VLF metal detectors also?

Cheers,

-SB

[Sean_Goddard]

Darn it.. Why didn't I think of that? That balun is a top idea. I'll incorporate that into my Surfmaster design if you don't mind

[moodz]

noise and careful design.

What an interesting read this thread is ....
I have just registered with geotech.

Here are some photos of my version attached ...

Note the careful layout

Oh and I made some changes ... to remove alot of noise issues the coil feed line passes through a bifilar "balun" ... it does not affect the pulse.
ie reduces the unbalanced antenna effect of the feedline. ( I used to be an RF guy )

I use a CA3130 as first amplifier ( FET inputs ... much higher input Z )

The uP is a DSPIC 30F4012.

A switch mode supply provides 16 volts for the analogue no sync.

A 12 volt batt pack from a $15 dollar drill powers the circuit for 4+ hours.

[Sean_Goddard]

The PCB I posted is a Whites Surfmaster (unfinished) clone. Just to show you what a PROFESSIONALLY designed PCB looks like and shows you what CAN be done to make this Atmel based design very quiet too (note the GROUND traces around ALL inputs AND outputs)!!

I have not divided up the analogue stage to add screening cans on this version, but MAY do that on a Mk.2.

The original Whites board could be so much quieter if it was designed in the way I am doing it. I have spent three months calculating placings and copper dimensions for the best possible board. When it is finished, I will post the PDF and the Protel format of the PCB here.

Laying out a board in this manner means that the analogue path MUST be done by hand with accurate measurements taken as you go along to ensure correct matching of stages.

This unit has the non-motion pinpoint mod added.

[Rov]

Surfmaster??? PI-111??Whats this?