Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

  • Congratulations waross on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!

Advice on filtering / regenerating digital signals

Status
Not open for further replies.

Morcego

Industrial
Apr 11, 2005
39
0
0
PT
Hi,

I'm designing a LED matrix for Outdoor VMS of a considerable size (about 5meters).
There are about 16 matrix. From matrix to matrix are traveling signals, like clock, data out, latch enable etc.

What is the best circuit to amplify / regenerate signals? And about noise filtering (I did see (in a commercial product) some inductors in series with the signals. Is this the best way?

Thanks in advance for the help.

Joaquim
 
Replies continue below

Recommended for you

What is "best" is dependent on the environment. If this sign is standing alone in the middle of a field, there's not going to be much interference, if any.

TTFN

FAQ731-376
 
The inductors you see are probably common-mode filters. They are required to keep noise from propagating from the switchmode power supply to the mains.

Filtering data and clock signals does not sound like a good idea. But, if the array is up to 5 m, you may need to run - say - a 12 V signal and use galvanic isolation. Either optocouplers or transformers. A differential mode system (think RS485) will probably also work.

The reason is that a large LED matrix consumes lots of power and if you use the same common for power and control signals, you will have severe ground shift problems.

Do not try to use 3.3 or 5 V signals with same ground as the power. You will be very sorry if you do that.

Gunnar Englund
--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...
 
Separate grounds are not possible. I use constant current drivers that use same common ground for logic and power. The only difference is Vcc (5v for logic) and VLED (12V for LEDS).

In my prototype sometimes I have one missing pulse causing the matrix to show the chars incorrectly. I had check (with scope) the signals. The level is correct, but the signals have some noise. The VMS will be in open field. I'm looking for a way to improve the design. Maybe I have to redesign the PCB using a star topology to connect the ground.


Thanks for the feedback

Joaquim
 
Joaquim, what troubles are you having with your current system? Gunnar's suggestion of RS485 for signaling is a good one. It sounds like you are using one control board for a large (both in quantity and distance) system, trying to use a single clock for all modules.

If that's the case, I would highly recommend changing your design to something a little more distributed. A small processor per module connected by RS-485, Ethernet, etc. to a larger control system is a much more tractable design. This removes the need for synchronous clocks and long distance comm issues.


Dan - Owner
Footwell%20Animation%20Tiny.gif
 
Sounds like you have a design problem, and not a noise problem. Your "prototype" is presumably running in a lab, which is relatively clean from a noise perspective. If you're having missing pulses, then there's more likely to be something wrong in the design or fabrication.

TTFN

FAQ731-376
 
Well, the prototype is running at a factory floor because is to large fr my small lab :)
In the factory floor we have some large metal working machines... Maybe they are the source of noise...

The RS485 approach is a good one. Small processor for each board (addressable) and a master controller for the entire VMS? I have 3 rows of 16 boards, so 48 nodes are no problem
for modern RS485 drivers right?

Tks

Joaquim
 
Not for today's RS-485 drivers, though if you imagine any setup that will increase the number of modules you may want to consider something a little more expandable, such as Ethernet. These days, even sub-$10 micros have built-in 100BaseT Ethernet stacks, and implementing them is not much harder than setting a few registers and waiting for interrupts to tell you messages have arrived.

Dan - Owner
Footwell%20Animation%20Tiny.gif
 
We have used the distributed addressable 485 method on our large displays. Break your design into some logical size elements and consider a processor for each.

If you aren't running video image rates then even a slow data rate is fast enough.

Keith Cress
kcress -
 
You may want to use cheap fiber and LEDs in a star-like arrangement for data/clock signals. RS485 + STP shielded cable for data/clock makes sense.
 
Status
Not open for further replies.
Back
Top