How to access signals on BGA?

Oh, tough one. What are your limits in cost and time? If it were just one signal, I'd look to see if back drilling it were feasible. Which balls? If they're on the edge, it's not that hard. Do you have a microscope?

1. JTAG

1. Connect the BGA via the patchboard with all signals accessible.

Vladimir Vassilevsky DSP and Mixed Signal Design Consultant

All BGA packages have JTAG? News to me.

On an existing board with the IC installed? Non-destructively? Good luck.

You could remove the IC, install a test adapter and attach an IC to that.

You might be able to flip the board over and drill through from the opposite side. If there are no components in the way, or traces. I'd venture a guess that this is a multi-layer board, so the probability of traces passing under the pins of interest is pretty high.

Both of these methods I'd classify as destructive.

Next time you build a board, dog-bone and via every single ball? (I once requested this on a prototype board, but some enterprising layout guy deemed it unnecessary... we ended up spining the board again, to the tune of another $1k or so... and used a TQFP part instead.)

At this point, unless the part if something like an FPGA where you can just shuffle signals around internally, I believe only option is to drill a hole to the ball you want to access from the opposite side of the board. Be precise?

You didn't do a layout review?

Sounds like a cable box hack attempt to me where the JTAG pins are not taken out for anti piracy :) No Chance m8, been there done that.

Not at the time. There was a mistaken notion that layout guys do what you ask them to do, or ask you if they'd like to do something differently

We do now. Because they don't.

You should review _everything_. You should even respectfully insist that anything you do gets reviewed.

Trust, but verify.

Yeah, I've learned that lesson the hard way I guess.

JTAG is the usual method to see if the pins are at least toggling - great for inputs, but outputs need to be probed elsewhere in the circuit to confirm connection.

Also, it's common practice to fanout every pin (including the outer ones) to full vias so that they can be probed from the bottom of the board. For prototypes you can even leave off the solder mask on the bottom vias so you get easy test points. If it's a denser board with blind vias then you have to put test points elsewhere.

Probing options exist but they must be soldered in place with the chip.

Dave.

JTAG is the usual method to see if the pins are at least toggling - great for inputs, but outputs need to be probed elsewhere in the circuit to confirm connection.

Also, it's common practice to fanout every pin (including the outer ones) to full vias so that they can be probed from the bottom of the board. For prototypes you can even leave off the solder mask on the bottom vias so you get easy test points. If it's a denser board with blind vias then you have to put test points elsewhere.

Probing options exist but they must be soldered in place with the chip.

Dave.

Which box did you try to hack?

Stick to bit banging, where you know something about the subject. BGA is one of many style of IC packages. They can contain any type of IC circuit, including analog applications. JTAG is a useless concept on those parts.

One in the UK

Maybe someone already invented something like a electric field beamforming probe. If not now may be a good time to start with it.

What video chip p/n is used in it?

Did you get a datasheet or was it proprietary?

< What are the options for accessing a half dozen signals on a 26x26 (1.0 mm < pitch) BGA? Some of these signals are not connected to he PCB (ie, some of < the IC's features not utilized) so tapping into traces is not an option. < Thanks.

I think I'm speaking out of my leage here. I don't even know what 26x26

1 mil is (unless thickness), but I just saw the coolest idea from the phone man (I probably shouldn't be saying anything).

First, I do not know what a BGA is. Unconnected signals seems to imply radiation - RF/EMF/? Is it possible to ground them (non-leaded chips).

You might try tapping the traces with an induction type magnet. Speakinh again out of my leage, the self sealing gel phone taps are like totally out of style these days...

I have been using a simple lm272 opamp circuit, from a book, that acts as an electromagentic signal reciever and field strength indicator. It is able to play a tone based on the amount of electromagnetic radiation passing through the induction coil, also it displays magnetic strength via a 10-segment bar graph (276-081). These items are store bought at the radio hut. 10k pot and 100k pot. Circuits was not designed by me.

Some investigating should come you across more sensative instrumental amps. Perhaps piercing the skin of these traces without tap, possibly even activating them? Sounds like a fun project but very tedious and frustrating.