Idea for for a voltage sensing relay

[billhig]

I'm putting this here because it applies to either C10 or C14.

I'm building a mircroprocessor controlled relay, mainly for myself and friends but I may decide to produce it.  The idea is that it will turn off accessories at low system voltage (idle) so you can power up stuff that only works at RPM's that will generate enough power.  (Heated gear comes to mind)

It will also turn it off if there's ever a an over-voltage condition to protect it from damage. Imagine your I'phone and GPS runnign of this! There's a dash mounted indicator to act as a voltage monitor.

It has ground, battery, and ignition connections and a pair of relay connections for the device being powered.  Here's how it works

Ignition off - the relay's off and the dash indicator blinks red once every few seconds to simulate an alarm (the please don't steal me mode used in various BMW cars for years)

Ignition on, battery voltage in normal range - Relay on, dash indicator solid on green.

Ignition on, battery voltage below normal but not critical - Relay on, dash indicator blinking green

Ignition on, battery voltage critical or below - Relay off, dash indicator blinking red

Ignition on, battery voltage dangerously high - Relay off, dash indicator alternating green/red.

So I've got a prototype built and software written and everything works. I need to settle on voltage points for Critical, Normal, and High.  I'd make them adjustable but then it becomes difficult to make the unit completely waterproof.  I've currently thinking

Critical - Below 11.5V
Warning - 11.5 to 12.5V
Normal - 12.5V to 16V
Overvoltage - above 16V
 
Any thoughts on the idea or the thresholds?

Thanks,  Bill

[Rick Hall]

Any thoughts on the idea or the thresholds?

For over voltage, sounds like a "crowbar" device for aero-planes. In an over voltage situation, a 'crowbar' is dropped across the feed wires, blowing the fuse(s) and protecting the avionics. But I wonder how many times this had happened on a Connie? Not many I'll venture.

If you're programming a PIC, encapsulate it but allow for ICSP. Just in case trip voltages need to be modified after production.

LV shutoff. Set it ever so slightly below idle state. But this may vary according to the battery chemistry. Flooded vs AGM vs Oddysey vs This way the user can shut off their bike and know the accessories on the relay will [eventually] shut off too. And hopefully not start back up? But see below.

If a person wired his GPS into this relay, would the relay shut down when said person has to wait at a stop light (at night) when he had his aux lights on too? Could be a bad scenario.

You code for Microchip/Atmel devices?

Rick

[billhig]

Rick,

I'm using a TI MSP430 for this project.  It's more user friendly (in my view) than PIC.  I use ATMEL arm 9 processors at work but I have a team of software people to do that for me.  I'm just a lowly hardware guy turned manager trying to keep my hand in.

Instead of a crowbar, it just opens the relay to disconnect the attached device, not protect the whole bus.
I would like to make the set points adjustable, or make the several different profiles selectable but it makes it hard to encapsulate with a dip switch or something hanging out.

Bill

[Pokey]

Sounds like another electronic gizmo that could go wrong IMHO.

[Rick Hall]

Crowbar is a carryover from when they probably used them Now the simple ones blow the field fuse, the more advanced open a relay (mechanical or electronic).

What about 2-3 exposed header pin(s) and a shunt?

Pokey brings us a valid point though, what are failure mode scenarios?

I'm doing a hack job on a 12F689. It works, but could be improved, as all software can

Rick

[billhig]

Pokey's right, all electronic gizmo's can go wrong.  A good solid electronic hardware design is the key.  I currently design control systems used to protect the high voltage power grid and if you don't trust software controlling your heated vest, you don't want to know how the power grid works.

I like the shunt idea.  I may use that.  Different resistor values setting up a divider which the Proc. will read to select different modes.

[Rick Hall]

Dump the resistor idea. You're programming in digital, ether the shunt is there, or it's not. Along this line, what about three pins? 1-2, 1-3, 2-3, 1-2-3, none.

And not knowing the chip capabilities, you'd be using up an additional A/D converter by using a resistor too. And what happens if the user installs a resistor that's out of range of your program? Never underestimate the capabilities of a user on a mission

Rick

[OCK913]

All I can think of is .....

220/221 Volts Whatever it takes?

[billhig]

I'm reminded of the phrase "when you think something's foolproof, you've underestimated the power of a fool and when you make something foolproof, somone will come up with a bigger fool"

I may include a terminal block instead of wires.  Then the shunt idea is pretty clean.  I've got more ADC channels than digital I/O.

If you ever looked at the V-strom fuel injection computer and tuner, that's really slick.  To program in new mixtures it uses three analog voltages to select ranges and mixtures and then it pulses a pin to program it.  That way you can make a programmer with a bunch of resistors and rotary switches.  That's all the famous Yosh box is.

[ZG]

All I can think of is .....

 
 
I love that scene, 221, what ever it takes...