Propeller CPU Board running off 12VOLT Battery.

DavidM

Hi,

I am currently designing my own CPU Controller board using the PROPELLER CHIP which will have a 40 pin header for I/O interfacing as well as 7 x 7 pin dedicated connectors for specific interfaces.

The CPU board is designed to be connected to several different other boards ( interfaces) such as..

1) LCD DISPLAY
2) USER PUSHBUTTONS ( up to 64 buttons)
3) THUMB-WHEEL CONTROL
4) HIGH POWER RELAY/TRANSISTOR OUTPUTS ( 16 -24 outputs )
5) LED'S for displaying on a control panel up to 64 LEDS)
6) SERIAL INTERFACE ( RF MODULE such as the XBEE PRO or equivalent)
7) RS485 Interface ( two wire )

The CPU BOARD will be powered by a sealed 12 VOLT rechargeable battery.

In regards to the HIGH POWER RELAY/TRANSISTOR OUTPUTS BOARD no 4 above, will require several AMPS of power at any one time.

The 12 volts will be supplied through the CPU board as it has the 5VOLT and 3.3 VOLT regulators.
All the other boards will require 5V or 3.3 volts except for the OUTPUT BOARD which will require 12 VOLTS for switching

I am concerned that when the TRANSISTORS/RELAYS are activated that they may interfere with the supply to the CPU Board,

So my question is..

What kind of design requirements are required to ensure that 12 VOLTS at several ( 1-3 amps?) can be supplied to the TRANSISTOR BOARD without interfering ( spiking) with the CPU BOARD and the other boards?

What kind of capacitor? Do I need a 12 volt regulator? etc

regards

Dave M

mahjongg

From the 12 Volt supply feed a 10.000µF 16V elco through a high current schottky diode diode such as a 30BQ040, after that use a low drop regulator for the 5V supply, such as the L4940V5.

The diode and elco ensure that a dip on the 12 Volt supply does not directly reach the power input of the voltage regulator, the low drop regulator ensures that the 5V will still be stable when the input voltage drops as low as six volt during a long "brownout" at the input.

The L4940V5 can deliver 1.5 A so it won't be a problem to feed the 3,3Volt regulator (such as a LM1117-3.3 ) from the 5Volt. This would normally not be a good idea, when the 5V regulator is of similar capacity as the 3.3V regulator, but in this case it's acceptable. Use a heat sink on the L4940V5 because with 12 Volt at it's input it has a lot of heath to dissipate.

Also do not forget the protection diodes across the coils of the relays. Use a relay driver such as the ULN2003 multiple darlington driver, it has the protection diodes built in.

Mahjongg

DavidM

Thanks Mahjongg,

I will investigate, I already know about the ULN2803 , and the use of the HEATSINK for the first REGULATOR ( 5V )

Would a 1N4007 Diode be the correct one to use?

I was planning to use the LM2937 for the 3.3 VOLT regulator and the 7805CT for the 5VOLT regulator as I have these in stock.

When I get closer to my design completion I will post an image.

regards

Dave M

Peter Jakacki

12V to 5V linear regulation is very inefficient and only suitable for light loads. A linear regulator will end up dissipating 7V*I which could get around 2 watts and then you definitely need a good heatsink. A better way to handle the +12V is to use a switching regulator (SMPS) which will not require any bulky heatsinks or capacitors. One such beast that I use is the trusty LM2594 www.national.com/pf/LM/LM2594.html.

As a guide here's a picture of what it could look like on your pcb.


You can of course use other switching regs perhaps even in DIP if you find that easier though mostly smd is a piece of cake to manually proto.

As regards the noise that may be generated by your load I would always make sure you suppress all you practically can at the source, that is right across the coils etc. The SMPS is very good at filtering out all those noisy little spikes due to the LC nature of the output but don't count on it, design with a margin. If your +12V battery drops down because of a load then that is some mighty load and you might have to think about beefing up the battery or staging the power delivered to the load. I had one design which used a nasty 12V motor for grinding and it would bottom out a 14V/14A PSU as it started up. When it was run off a battery then there was hardly any droop at all but the twin 110A MOSFETs could get hot at times.

By the time your 5V is then regulated to 3.3V you shouldn't have much worry about loads interfering with the CPU. Just observe good layout practices especially with respect to grounds and ground planes and the crystal oscillator. If you are "vero boarding" rather than doing a PCB then take extra care and run separate grounds for load and logic back to the battery (do this anyway).

If you use ULN2003/2803 style drivers don't forget to connect the diode common (pin9/10) to the positive supply for most applications. If you don't know what you'll be driving then could I suggest that you use discrete N-channel MOSFETs as these are pain-free and have an internal clamp diode from drain to source.

I have to keep this short (no, seriously) but if you have any questions please don't hesitate to ask.

*Peter*

Stan671

Thanks for the link to the LM2594 switching regulator.· I am going to give the LM2825 (http://www.national.com/pf/LM/LM2825.html) a try.· It has all of the additional components (resisters, capacitor, inductor) already right in the package.· So, all it needs is In, Out and Ground to regular 1 Amp at 80% efficiency.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Stan Dobrowski