U5 quality issues

[elutionsdesign]

I seem to be having a real hard time with voltage regulators not regulating, has anyone else seen this? I've got a fresh batch of 5 from Digikey and none of them regulate, all show 12V in and 12V out. LM2937ET - 5. I've seen this happening randomly for a year now but all 5 bad is crazy.

[racingmini_mtl]

The only time I've seen that is when the ground connection was bad. So check if the regulator is correctly grounded. I would suspect either the solder or the board rather than the regulator itself. Especially since you've seen the issue with numerous units.

Jean

[scott87star]

They're not yet assembled, just using gripping test leads to connect everything.

[racingmini_mtl]

They need capacitors to regulate: look at the data sheet.

[elutionsdesign]

OK, I see that now, but I just started a V3.0 board build using a new U5 from DIYautotune which when tested with just leads like I've done all the time showed around 10 volts out with 12.9 volts in. I went ahead and soldered it in place and tested the power supply per the V3.0 assembly guide step 23 and the output voltage is 5.6. That is the zener diode voltage D19 AND is out of spec high per the data sheet. I am not comfortable with the long term reliability of such a device, should I be concerned? And now this is six U5's in succession? My testing of every U5 came about because I was seeing high out of spec voltages during the step 23 assembly guide test and I've always used ones that test good (regulate voltage in spec) with my bare leads test (not on the board, no capacitors). I don't think I'm so special that all the bad ones get shipped to me, I'd be curious how many faulty U5's are out there unbeknownst to the builders because of D19.

[SymTech Laboratories]

There's nothing to limit current through D19, so if it were subjected to high voltages it wouldn't act as a regulator, it would create a short and trip resettable fuse F1. Can you rule out all other possible culprits (faulty test equipment, a problem elsewhere on the board, etc.)?

[racingmini_mtl]

As I said, I've never seen that or heard of that happening other than with a ground problem. And if that were common, we would have seen a lot of them and it would be obvious due to the fact that there is no current limiting through D19 as mentioned by Symtech.

So you are either incredibly unlucky for having had all those defective parts or there is something no quite right with the way you're checking things out.

Jean

[jamies]

any [linear] voltage regulator needs a load for it to regulate correctly.
if you are testing it open ended with only a multimeter, it can show any random voltage, put a load on it and it will stabilise.

if you look at the schematic http://www.megamanual.com/ms2/v3pcb_6.gif without a cpu in the socket, the regulator is operating open ended.
generally you wont get a stable voltage in this state.

ie try a back to back test any 1-10k resistor across the S5 and SG holes near the DB9 and it will show you a stable voltage.

[elutionsdesign]

Thanks for the replies Jean, I'll go over the test equipment again.

[elutionsdesign]

An update, it was the test equipment. In the back of my mind I knew it had to be something I was doing/causing, in this case it was a low battery in my DVM. Odd how it read the supply voltage fairly accurately and the regulated voltage not well at all.