Re: 5V @ 3.5A for automotive - best way?


colin wrote:
<ferrari.secret.santa@xxxxxxxxx> wrote in message
news:1167757609.733221.4530@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
I'm trying to make a little project work here... trying to use seven
5x7 dot-matrix displays to make a little "digital readout" that I can
display text on based on feedback from various things in a car (like
temperature, etc).

The 5x7 displays are of course multiplexed and have a 1/10th drive
current of 100ma. I am using shift registers to control the state of
each row, then sequentially updating the rows to produce the display.
That's all fine.

Problem is, given 7 rows, 35 dots per row, at 100mA per dot, that's a
potential max current of 3.5A! The duty cycle doesn't help me because
I drive the rows one after another, so if all elements were on, there's
a potential 3.5A constant current requirement, thats a lot.

Can anyone suggest solutions for getting such a drive current from a
12V automotive source? The problem is, I do not want the display to
flicker in brightness when the vehicle is started, so I really want to
use a regulated source, also if I just ran the segments off the 12V,
the power dissipation in the resistors becomes huge and I am trying to
keep this small. I have looked at the 1084 regulator which can do 5A
with up to 30V input, but this is a linear regulator.

I am a bit of a noob when it comes to circuit design, but am I correct
in believing the power dissipation of the regulator is the voltage drop
from Vi to Vo times the current? If so, and if you assume 13.8V in,
then that's around 30 watts which seems to be a ton. I have used
switchers in the past but the problem is the cost and number of
additional components (board space is tight).

Am I missing some obvious answer that would solve my problems here?
Would you guys go a different route in this design other than a
traditional linear or switching regulator, and if so, what would you
use?

you can make a buck type switching regulator quite small these days,
if you go SMD of course,
one small 8pin controll chip and a small mosfet and
with a high switch frequency a small inductor,
thats probably the way to go,
but dont forget the high voltages surges that can apear on car electrics.

however 245 LEDs at 15ma each is going to be a heck of a lot of light,
I doubt if you seriously need 3.5 amps if its inside the car,
or is it some sort of external message board ?

Colin =^.^=


Thanks for the info. It sounds like a switcher is the way to go. Is
there a preferred source for these? Pricing them out on Digikey (just
for reference) they are much more than linears. The least expensive
part they have is a 5A/5V switcher at $3.75 or so in 100's. That's a
BIG step up from something like a 7805. The part I used before (7856
IIRC) required a decently large inductor, some caps and diode, and even
going all SMT it was of reasonable size, certainly very hard to
shoehorn into the board size I have available.

On the wattage, the displays I would like to use are red and blue (not
together). The Vf of the red is around 2.6 and around 4 on the blue,
so of the ~17 watts going to the LED's, around 8 are being dissipated
in 35 resistors for the red variant, and 3.5 through the resistors in
the blue. I grant that even at 9 watts we're talking about a lot of
light, but driving at 100mA (the max current for 1/10th duty) is the
"worst case". I can go dimmer in software easily with PWM, but it will
only ever be as bright as the hardware allows. Bear in mind that in a
convertible in full sunlight in July, its a different animal than on a
moonless night in an enclosed car :). Also, the display will rarely
ever use all segments, but of course it's possible, so I want to design
for that. I would imagine generally around 1/3 of the elements would
ever be on simultaneously.

I was thinking of ways I could split up the segments so instead of a
3.5A rail, I could have four .85A segments or something, but those
solutions seem inelegant and however you slice it, there would be 30W
going *somewhere* which is going to get a bit toasty (then again, thats
a max design level, that would rarely ever be reached and if so would
be for a few seconds at most).

.

Relevant Pages

  • Re: 5V @ 3.5A for automotive - best way?
    ... then sequentially updating the rows to produce the display. ... but this is a linear regulator. ... traditional linear or switching regulator, and if so, what would you ... but dont forget the high voltages surges that can apear on car electrics. ...
    (sci.electronics.design)
  • 63 Impala charging problem
    ... the problems was the wiring harness was ancient on the car. ... GM with the external regulator, which would be correct for this car. ... you're supposed to use when you install an internally regulated alternator. ... The battery in the car was measuring around 11.8 volts, ...
    (rec.autos.tech)
  • Re: 5V @ 3.5A for automotive - best way?
    ... then sequentially updating the rows to produce the display. ... but this is a linear regulator. ... I am a bit of a noob when it comes to circuit design, ... traditional linear or switching regulator, and if so, what would you ...
    (sci.electronics.design)
  • Re: 5V @ 3.5A for automotive - best way?
    ... then sequentially updating the rows to produce the display. ... but this is a linear regulator. ... I am a bit of a noob when it comes to circuit design, ... traditional linear or switching regulator, and if so, what would you ...
    (sci.electronics.design)
  • Re: RO/ARM in car MP3 player.
    ... For control of the mp3's on the dash, I thought about a 2x16 module LCD ... (I've got a panel under my radio where it's a double DIN hole). ... Perhaps a PIC micro to control the display (I've already developed some ... In a car, it might be best to hide the gubbins in the boot. ...
    (comp.sys.acorn.hardware)