Look at my voltage regulator circuit design

[HotCat]

I use TLE4275 as voltage regulator chip, it can sustain 45v, build in short circuit proof, reverse polarity proof. So that it simplifies the voltage regulator circuit design

cpu reg.PNG

(8.51 KiB) Not downloaded yet

[jharvey]

I seem to recall that .7 to .6 amps isn't quite enough for our needs when run fully loaded. I seem to recall our design spec is 1 amp min. However I don't think we've actually measured that, so less than 1 amp might work. 5 V ± 2% is a fair bit of wonder. Appears that's mostly temperature wonder. This is to much for us, as the A/D's reference that voltage, and it would induce 2% error in our calculations, for each sensor. The compounding tolerances would make for a buggy system. Also it appears to be a linear regulator, so it will get hot, and you'll have to deal with it thermally via heat sink. Remember the power drop will be around (13.4 - 5)*1 = 8.4 watts, while delivering 5 watts to the PCB. So a total power draw of 12.4 watt's to be dissipated by the board. The efficiency of a linear for this application is (5/12.4)*100= 40% efficient. The switching is upwards of 80% efficient, so total power dissipated by the board is around 6.25 watts to deliver 5 watts of power. Dropping that additional 6.15 watts is handy, as we have several other heat sources like the injector drivers and ignition drivers.

[Fred]

Less than an amp is fine! I think the power requirements for any given board depends on its intended usages. Reality for most setups is probably around 200mA max? Under 500mA supply potential would worry me, though.

The 2% tolerance is a bit more of a worry :-/ As Jared said, those errors will compound, and although it won't make it buggy, it will make it inaccurate. If they are only temperature related, then expect the tune to drift with ECU temperature, not ideal, to say the least.

As for heat output, it's not an issue in reality either. Putting that heat into a sink of some sort means it's not an issue. Putting it into the PCB can be an issue, if not managed well.

I didn't actually look at the design, however I like the idea of a simple and protected setup. I don't like the idea of specific niche parts being used, though. Or is it pin out compatible with a normal regulator? Perhaps others have such similar features and tighter tolerances?

What about always-on functionality, Hotcat?

Fred.

[nitrousnrg]

I see the schematic saying CPU regulator. If this one only supplies the MCU, you are fine. If it supplies the whole board, you might be a bit under the expected maximum. Anyway, it depends on how much stuff you plug in the board.

For example, the MCU alone can draw around 250mA, and analog sensors add some mA too.

The 2% can be improved using a 5.00v zener in the ADC reference pin, although you won't dodge the fact that your sensors will be supplied from the 5v +-2%. About the temperature, just keep it in mind, and plan to have a robust heatsink mount.

Are we going to see some schematics? I recall Puma license was Open Hardware :-)

[Fred]

Are we going to see some schematics? I recall Puma license was Open Hardware :-)

You may distribute products you make to third parties, if you either include the documentation on which the product is based, or make it available without charge for at least three years to anyone who requests it.

Just like with GPL there is no requirement to share what you're doing in a closed environment. Only if he sends out a board, does he have to send out the design files for it. Likewise, only if he sends out design files, must he maintain the license. If he keeps this stuff to himself, changes it, uses it, but never sends it out, that is his business. GPL is the same way. I'm on your side, it'd be nice to see what is happening with this.

If you need help setting up a git account and publishing your changes as you go, let me know, I can help.

Fred.