Misbehaving L7805

[krich]

When I first started on my $50 robot, the first thing I built was the 5V power supply.  Worked like a charm.  9V battery, hooked up to a L7805, and viola! 5.0V.  Happy camper.

I've moved on to other projects since then.  When troubleshooting a board I had just etched and populated, I found that the voltage coming from one of my trusty L7805 voltage regulators was 5.8V.  5.8!?!  That's way over what I'm looking for.  The datasheet says a maximum of 5.35V under any conditions.  If I hook up 12V bench supply, I'm getting 6.5V out of the L7805.  Not good.  That's over my poor atMega168P's maximum operating voltage.

So, I put an L7805 on my breadboard with nothing else.  Hooked up the 9V, measured the output voltage, and, yeah, 5.8V.     What's worse, the 9V isn't really a 9V anymore, it's more like a 7.4V.  (I need to dig up a new one)

So, what gives?  I feel like I'm missing something obvious.

[alessio136]

isn't it just a 7806? should produce 6V, but if you supply 7.4V it drops something...

[benji]

So, I put an L7805 on my breadboard with nothing else.  Hooked up the 9V, measured the output voltage, and, yeah, 5.8V.     What's worse, the 9V isn't really a 9V anymore, it's more like a 7.4V. 

first of all ,,try another voltmeter

next , is the 9volt source u are using a battery? if it really dropped into 7.4 then it may has been shortened ,,

and why dont u try another 7805 ? they are as cheap as hell

[krich]

Thanks for the replies guys.

I've also tested with my bench 12V power supply and get similar results. 

With a generic L7805, the 12V supply reads at 12.23V and the 5V out on the VReg reads 6.04V
Swapping the L7805 out with an ST version (likely better quality, right?), I get almost identical readings.  12.32V and 5.98V

So, the only thing in the equation that that I haven't swapped out, as suggested by benji, is the Voltmeter.  I'll swap out the batteries and then the whole voltmeter.

Thanks guys.

EDIT:  Swapped the batteries and now I'm testing at 4.91V.  Ahhh....much better.  I figured a low battery on a voltmeter would result in a lower voltage reading.  I guess that shows how little I know about voltmeters.  Alright, back to my man cave.

[Steve Joblin]

I don't think this should have any effect, but are you using filter caps on both the input and output as per the recommended design spec?

[benji]

actually it happened with me oneday , i had a cheap voltmeter with an old battery , i did test the 7805 out and it was 5.8 v
did test my 9 v battery and it was 10.2 v so i knew its the voltmeter

[krich]

Well, on the breadboard when I was trying to figure this out I didn't bother with the caps.  On my MCU board, I most certainly do.  Although, I've seen many opinions on the values of said filter caps.  The datasheet says 0.33uF on input, 0.1uF on output.  Many project circuits out there use values much higher than that.  Either way, that shouldn't affect the ouput Voltage that much since they're there to filter the ripple and I was dealing with a gross disparity between what I was measuring and what should be.

What tricked me, benji, is that my 9V was testing below 9V while my 7805 was testing above 5V.  Doesn't make sense until you put 2 and 2 together and realize it was testing with an error of +1V or so and my 9V battery was much weaker than I thought (~6V).  In fact, the 9V gave out shortly after I posted.

Thanks guys.