EEVblog Electronics Community Forum

Electronics => Projects, Designs, and Technical Stuff => Topic started by: Simon on July 11, 2010, 07:25:10 pm

Title: MTF of a PIC / 78L05
Post by: Simon on July 11, 2010, 07:25:10 pm
Hi guys, I've been reviewing a circuit design which was going to run on dual power supply (power side and MCU control side) and now intend to put the control circuitry which is essentially a pic powered on a 78L05 (or equivalent) on the main supply that is on 24/7.

I'm now going to have the 78L05 reg and associate caps powered off 12 V and the pic off 5 V 24/7, what sort of life expectancy can I be looking at ?. I suppose I'm concerned as I still treat the pic as a computer versus a simple digital chip
Title: Re: MTF of a PIC / 78L05
Post by: EEVblog on July 11, 2010, 10:21:06 pm
How are you getting your 12V?, that might be the weakest link.

Dave.
Title: Re: MTF of a PIC / 78L05
Post by: DJPhil on July 12, 2010, 04:59:20 am
I'm now going to have the 78L05 reg and associate caps powered off 12 V and the pic off 5 V 24/7, what sort of life expectancy can I be looking at ?

I'm just guessing from before, but is this an automotive application?

It's almost impossible to be sure without testing the whole design itself in quantity. You might be able to get part specific data from each manufacturer, but any given collection of components would have to be tested. Assuming no major electrical trauma and good build quality all you're left with is environment. Heat and moisture are the big enemies. If there's not a lot of heat generated by the circuit itself you might be able to pot the board (dunk it in black goop) to fight against moisture problems. Shelling out for parts with a military temp range could help as well.

In the end, consider how many microprocessors (from a few in the 80s to several dozen nowadays) are in cars and how often they die. Computers (desktop) are a bit different in that they have to dissipate relatively huge amounts of heat just to function.

Hope that helps some. :)
Title: Re: MTF of a PIC / 78L05
Post by: Simon on July 12, 2010, 06:05:29 am
How are you getting your 12V?, that might be the weakest link.

Dave.

Car battery which is charged by a dynamo in a vintage car, yea it could be a rough supply I'll be trying to filter it out as much as possible, at least when left standing it will just be smooth battery power. Yes it's the same old project, me moving house and my friend having his test car still in bits gives me lots of time to rethink and refine the design
Title: Re: MTF of a PIC / 78L05
Post by: Zero999 on July 12, 2010, 03:05:03 pm
It might be a good idea to add some spike suppression before the the LM7805, a 18V zener from 0V to Vin in series with a 22R resistor (to limit the current spikes) should do.
Title: Re: MTF of a PIC / 78L05
Post by: Simon on July 12, 2010, 06:05:49 pm
oh I'm planning better than that a 100+ ohm resistor in series with a 100uF cap on the 78L05 input, I think because I'm doing a smother regulation I'm getting smoother power out of the dynamo, apparently the dynamo alone and control box used to create some of it's own due to the crudity of it
Title: Re: MTF of a PIC / 78L05
Post by: TheDirty on July 12, 2010, 06:54:54 pm
If this is an automotive application, use a proper automotive spec'ed regulator.  They aren't that expensive and they protect against reverse voltage and high voltage spikes.  If it's a non-critical systems and you really don't care, then just use the 7805 and whatever protection you add on.  I've had gadgets running with non-automotive spec'ed regulators for a long time now, but any new design would have one for me.
Title: Re: MTF of a PIC / 78L05
Post by: jahonen on July 12, 2010, 06:55:23 pm
Put a (schottky) diode before the input cap (make it large enough) so that if the input supply momentarily drops below 7-8 volts your pic won't suffer a brownout situation (possible reset) so easily. Large input cap also helps to absorb transients.

For transient protection, TVS is more robust than zener, if you want to play safe.

Regards,
Janne
Title: Re: MTF of a PIC / 78L05
Post by: Zero999 on July 12, 2010, 08:18:06 pm
Is the leakage current important?

What about using a Vbe multiplier? It's not very accurate but simple and it should be effective, if the correct component values are chosen.
Title: Re: MTF of a PIC / 78L05
Post by: Simon on July 12, 2010, 08:41:14 pm
not sure what your referring to, some sort of pre-regulation ? I'm aiming for low consumption so that I don't have to worry about draining the battery, 1mA really, the reg will not take much and neither will the PIC until it detects an engine startup and the whole thing kicks off.
Title: Re: MTF of a PIC / 78L05
Post by: Zero999 on July 12, 2010, 09:31:35 pm
No just transient suppression.

Look at the transistor and the resistors, Google for Vbe multiplier and you'll see what I mean.

The leakage current is about 1.4mA at 12V but can be reduced by increasing the resistor values slightly.
Title: Re: MTF of a PIC / 78L05
Post by: Simon on July 13, 2010, 06:45:45 am
so basically a high voltage (spike) will turn on the transistor which will divert the spike to ground, sort of zenner diode style.
Title: Re: MTF of a PIC / 78L05
Post by: Simon on July 13, 2010, 06:49:14 am
If this is an automotive application, use a proper automotive spec'ed regulator.  They aren't that expensive and they protect against reverse voltage and high voltage spikes.  If it's a non-critical systems and you really don't care, then just use the 7805 and whatever protection you add on.  I've had gadgets running with non-automotive spec'ed regulators for a long time now, but any new design would have one for me.

any suggestions of parts ? yes this is highly critical, it is controlling the car dynamo and so is supplying power to the whole car and needless to say to itself ultimately, if it fails the only way to get the car going again (assuming it don't blow 50V through the system and wreck everything) is to replace it fairly quickly (ie before the battery does so flat you can't get it to self start again).
Title: Re: MTF of a PIC / 78L05
Post by: RayJones on July 13, 2010, 10:05:04 am
It has to be asked, WTF is MTF?

Or was it meant to be MTBF?  ???
Title: Re: MTF of a PIC / 78L05
Post by: joelby on July 13, 2010, 11:02:04 am
Or was it meant to be MTBF?  ???

It's MTF if it only fails once :(
Title: Re: MTF of a PIC / 78L05
Post by: Simon on July 13, 2010, 11:47:21 am
basically "how long can i expect it to last ?"
Title: Re: MTF of a PIC / 78L05
Post by: Zero999 on July 13, 2010, 02:11:48 pm
so basically a high voltage (spike) will turn on the transistor which will divert the spike to ground, sort of zenner diode style.
Yes, that's the idea.

Bear in mind that it's a very crude solution with a temperature dependant and current voltage characteristic but it should be fine for fairly low currents.
Title: Re: MTF of a PIC / 78L05
Post by: Simon on July 13, 2010, 05:29:36 pm
well it has the resistor in front of it so current is limited
Title: Re: MTF of a PIC / 78L05
Post by: jahonen on July 13, 2010, 06:10:41 pm
Why not simply the TVS, it is just robust zener-like component made to absorb surges, and leaks only microamps (if chosen correctly)? Probably it is faster too, although it probably doesn't matter. For example 1V5KE22A TVS would have standoff voltage of 18.8 volts, leakage 5 µA, and clamping voltage 30.6 volts @ 49 A.

BTW, if idle current matters, then 78L05 is wrong choice, LP2951 is much better (albeit a bit more expensive). With LP2951 you also get reset to pic (it has open-drain error-output which is low when output voltage is out of tolerance), if the output voltage drops too low. Also LP2951 has lower drop-out (~450 mV @ 100 mA), so your PIC won't get reset so easily, for example when starting the car.

Regards,
Janne