AVR based laptop startup/shutdown controller

Shad0wca7 · 03-27-2008, 02:51 AM

I'm going to be using a laptop in my latest carpc, unfortunately its battery is totally dead so I'm designing a startup/shutdown controller which will act in very much the same way as the normal pc car power supplies.

When it senses ignition it will start the pc up by connecting some relays, when ignition is lost it will shut the pc down and after about 30 seconds will kill power totally to stop the pc draining the battery.

This is my schematic so far, anybody have any suggestions / comments?

Rob Withey · 03-27-2008, 04:33 AM

Instead of daisy chaining the relays for the main power like that, why not just use a transistor to switch the larger relay. I assume you did it that way because you didn't think the AVR output could drive the large relay directly. It's clunky the way it is.

I think you chose a device with a few too many IO pins, also. My first shutdown controller used a 12c508 (8 pin device), my next one (which includes a bunch of I/O and servo controllers) uses only a 16 pin device (16f88 IIRC). A 40 pin device is huge for such a simple task.

D2 is kinda redundant. I usually do reverse polarity protection by relying on the fuse popping when the zener forward conducts (or another diode in parallel). This just reduces the voltage drop before your circuit gets the juice (all kinda academic since you are using a linear regulator anyways).

You might want to put a pull down resistor on ignition sense too.

Relays are noisy for switching. A lot of power-switch-press circuits just ground out one of the power switch leads with an NPN transistor or FET. Much cleaner design that way, and less noisy, lower power, smaller, more reliable, etc... Similarly for the main power switch. Might I recommend a high side driver or FET.

Shad0wca7 · 03-27-2008, 06:56 AM

Thanks for the input Rob, some great suggestions which I'm going to use so standby for a redesigned version using transistors. Although won't there be the risk of a massive transient going through the transistor which will be activating the laptop PSU relay?

The Mega16 is WAY overspecced for this I agree, there are 2 main reasons I am using it at the moment: 1) I have a few of them lying around 2) I have a Tiny84 which I will probably be using (even that is overkill) but I don't have an Eagle library for it.

I was going to use the internal pull-down resistor for ignition_sense, I also figured that D2 won't harm (and won't cost me anything) so any extra protection is worth a squirt.

Rob Withey · 03-27-2008, 07:10 AM

Quote: Originally Posted by Shad0wca7

Although won't there be the risk of a massive transient going through the transistor which will be activating the laptop PSU relay?

Snubber diode? I just assumed you omitted them because they are always taken for granted with inductive loads (even your small relays will have some inductive kick on release).

Shad0wca7 · 03-27-2008, 07:13 AM

Actually I forgot to add them for the relays!

What I meant on my previous post was possible transients coming from the actual car power source which the transistor will be connected to?

Rob Withey · 03-27-2008, 07:25 AM

Quote: Originally Posted by Shad0wca7

Actually I forgot to add them for the relays!

What I meant on my previous post was possible transients coming from the actual car power source which the transistor will be connected to?

If you're not comfortable designing with transistors then stick with the relays.

I don't know what you are referring to with the transient thing. Just get a transistor that can handle sufficient Vce (assuming bipolar) involved. I'd switch that relay with an NPN bipolar with the emitter grounded, suitable base resistors and relay coil between the collector and power source. When the base saturates, the collector goes to about 1 diode drop above ground and the relay switches. Whatever you do, don't go for an emitter-follower when switching with logic levels - big waste of time.

Actually, I wouldn't (and didn't) swtch a relay. I use a high side driver.

Rob Withey · 03-27-2008, 07:30 AM

More info on the high side driver I use. It's a VN30N. This is the datasheet :http://www.ortodoxism.ro/datasheets/...cs/mXryrqt.pdf

Shad0wca7 · 03-27-2008, 07:33 AM

The high side driver is pretty impressive, I think for now though I'm going to stick with using transistors and relays, I have not really used transistors before so now is a goot time to learn and apply. New schematic coming soon.

Shad0wca7 · 03-27-2008, 08:03 AM

Ok, how does this look:

I decided I didn't want to rely on the internal pulldown so added one. The cascaded relay is no longer present, instead driven by a transistor that should have no problem driving the coil. Snub diodes have also been added.

I'm not sure how to go about removing the power switch relay alltogether and replacing it with a transistor to short the contact... do you mind giving me a few pointers with that?

Rob Withey · 03-27-2008, 03:47 PM

Quote: Originally Posted by Shad0wca7

Ok, how does this look:

Looks good.

Quote: Originally Posted by Shad0wca7

I'm not sure how to go about removing the power switch relay alltogether and replacing it with a transistor to short the contact... do you mind giving me a few pointers with that?

Most power switches are between a signal line and ground. You can replace them with an open collector NPN transistor just like the ones you are using to switch the relays.

The relay is more portable in that it will work when the power switch is not configured in this way - might be best to stick with the relay.

Shad0wca7 · 03-27-2008, 05:34 PM

I think I might be getting slightly confused about using the transistor to short the wires, would this work correctly or am I on totally the wrong path:?

Curiosity · 03-27-2008, 07:07 PM

I think you should add a few simple things that you could implement later in code, like a little resistor divider to constant on ADC0 to monitor the battery voltage since yours is dead. Then put IGN on ADC1. A button for manual startup/shutdown. An LED for some feedback. And use the serial I/O for talking to the PC. You'll have more ideas after that.

dupa2 · 03-27-2008, 10:42 PM

wow, I'm surprised you ended up using such a weird solution for power button driver. First of all, the power button will accept standard ACPI signal, and since you work with 5VDC already, all you need is just one tiny transistor with pull-up resistor that will make ~0.5ms 0V pulse.

Instead of regular relays, I'd use solid state relays that typically can handle at least 10Amps, you don't need any high-end SSR, just get the cheapest one with triac output.

Your avr may die if you use relays directly, either from overcurrent on IO line or voltage peak from coil transient.

Since you use avr, you can actually implement your battery voltage monitoring (avr has ADC inputs), I think it could be cool feature.

Look also for AVR Butterfly, it's a cheap development board with LDC display and atmega169, it would simplify whole project a lot.

Curiosity · 03-27-2008, 10:55 PM

Actually, all you need is a diode for the power button. If it works like a PIC, and I think it does, when off and set to output a digital line pulls to ground with the internal transistor. You just have to make sure it's on the whole time when the main relay is on. I used that for 8x8 LED matrix displays.

Rob Withey · 03-28-2008, 02:28 AM

Quote: Originally Posted by Shad0wca7

I think I might be getting slightly confused about using the transistor to short the wires, would this work correctly or am I on totally the wrong path:?

That's about what I use. The pullup is on the motherboard though, and the grounded switch line is not needed (only one line is needed).

Quote: Originally Posted by dupa2

wow, I'm surprised you ended up using such a weird solution for power button driver. First of all, the power button will accept standard ACPI signal, and since you work with 5VDC already, all you need is just one tiny transistor with pull-up resistor that will make ~0.5ms 0V pulse.

Like the circuit he posted above?

Quote: Originally Posted by dupa2

Instead of regular relays, I'd use solid state relays that typically can handle at least 10Amps, you don't need any high-end SSR, just get the cheapest one with triac output.

Like the high side driver I linked to (VN30N).

Quote: Originally Posted by Curiosity

Actually, all you need is a diode for the power button. If it works like a PIC, and I think it does, when off and set to output a digital line pulls to ground with the internal transistor. You just have to make sure it's on the whole time when the main relay is on. I used that for 8x8 LED matrix displays.

True, if you can get it to pull low enough.