Problem with 2 different supply voltages on PCB

[pucca]

Hi everyone,
I have a 2 layer board with one GSM and one GPS module. The operation voltage is 3.7V and 3.3V for the GSM and GPS. Is there a problem for the two voltages to be on the same board?
The tracks are not very close to each other(4 cm at least) and the board thickness is 0.5mm. If the tracks are on different sides is it still a problem? Is the difference of the two voltages significant enough to cause crosstalk?
Thanks in advance!

 

[MacGyverS2000]

Difference in voltage is not what causes crosstalk, it's change in voltages.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[Noway2]

Based on the fact that you are using a two layer board, will increase your susceptibility to cross talk. Without getting into the specific details of the mechanism, cross talk is caused by field coupling in the signals. The tighter you can get the coupling between the signal current and the return current, the less cross talk you will have. Since two layer boards to not have a ground plane with a constant geometry, you need to be extra careful to avoid cross talk.

Of course, whether or not the cross talk causes problems, depends on other design factors.

 

[pucca]

You mean when changing from one voltage level to another level, for example when using a regulator or a level translator( that causes crosstalk)? So when you have two different supply voltages is not a problem?

 

[MacGyverS2000]

Placing two different DC voltages next to each other have zero effect on one another (within all reason). If one of those voltages is not actually DC (such as ripple from a DC/DC converter), crosstalk can occur.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[pucca]

The first voltage comes from a Li-Ion battery(3.7Volts) and the second(3.3V) comes through a connector of another board. On this other board the 3.3V comes from a dc/dc converter. So I think that this should be a problem according to you....(macgyvers2000)

 

[MacGyverS2000]

It is definitely a potential problem... what's the converter frequency? How well is it smoothed (what value caps)? Pay attention to the length of the wires, as well... short stubs can easily turn into antennas at your frequencies of interest.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[VEBill]

If the PS noise is so high that it can couple from one trace to another, then you have bigger problems than the coupling (i.e the noise itself).

 

[IRstuff]

Seems like a lot of worrying, but little in the way of analysis. Take the worst case ripple. Figure out the worst case coupling capacitance. Figure out the total amount of bypass capacitance. Figure out the coupling ratio. What do you get? microvolts?

A 2-in square FR-4 board with 10 mil layer thickness presents about 400 pF of capacitance, assuming solid sheets of metallization, which would be shared across both adjacent layers. Nonetheless, the bypass capacitance on the victim layer must be in the hundreds of microfarads. The coupling ratio would be at worst 1:100?

TTFN

FAQ731-376

 

[GonzaloEE]

A few cents:

- Pay attention to clock track layout
- Add ferrite beads and ceramic caps to your PS input.
- While possible, set up both modules for minimal 'chatter', i.e. half duplex, etc.

 

[pucca]

thanks for the tips.....