soldering mk20 mk22 - power test problem

[Rafael Ignacio Zurita]

I have soldered pin 6 (VREGIN) and 23 (VSS) :

Loading Image...
(red line is pin 6, black is 23)

I put 5v from USB pc port in pin 6 (VREGIN) and GND from USB pc port on pin 23.
Then, I tested pin 5 (VOUT33, white line in picture), guessing that I
would get 3.3v from internal regulator.
But no :( on pin 6 and GND I get 5.02V, and testing pin 5 and GND I
get 4.96V, but no 3.3V as expected.

Should I put capacitors first before test? Is it needed all the VSS
soldered before this kind of test?
Could somebody with knowledge on this kind of MCUs tell me if this is
the proper effect with this minimal test?

Thanks in advance and reading to realize if this is correct.

Rafa

[Werner Almesberger]

Post by Rafael Ignacio Zurita
Is it needed all the VSS
soldered before this kind of test?

It certainly can't hurt. I don't know how they connect internally,
but at least all the VSS on the left side (pins 2, 11, 12) should
be connected to ground. (And connecting pin 23 won't be much of an
effort either :)

Post by Rafael Ignacio Zurita
Should I put capacitors first before test?

It can't hurt either. Didn't you have that problem a while ago
already ? I.e., that you didn't have capacitors and the chip
performed erratically.

Also, is your MCU actually powered ? There doesn't seem to be a
direct connection between VOUT33 and VDD*, so I don't know if
you're even in a defined state.


Regarding soldering, the solder will flow towards hot surfaces. So
you don't need to heat up a single pin at a time but you can heat
up several pins at once. Plus, make sure you also heat the pad
under the pin. Only then will you get a proper solder joint.

I usually solder QFNs and such in two passes: the first pass is to
just establish enough solder joints, no matter how untidy, that
the chip stays in place. The second pass is to reflow the areas
where things didn't work out. In the second pass, I can add or
remove solder.

Also make sure your iron is hot enough. If it's just at the point
where solder will melt, it will get too cold as soon as you touch
something. A range of about 280 C to 300 C should be good, but
you'll have to find what works best with your iron and your
station.

Unless you've cranked up the heat far too high, you also don't
have to be in a rush to remove the iron from a solder joint. You
have to keep it there long enough to fully melt the solder and to
fully heat pin and pad. Only then will you get a good solder
joint.


Another thing to check is that VOUT33 doesn't accidently make
contact with VREGIN. Also, the regulator may need a minimum
external load of a few uA before it actually can regulate.

- Werner

[Rafael Ignacio Zurita]

On Mon, Aug 31, 2015 at 12:18 PM, Werner Almesberger

Post by Werner Almesberger

Post by Rafael Ignacio Zurita
Is it needed all the VSS
soldered before this kind of test?

It certainly can't hurt. I don't know how they connect internally,
but at least all the VSS on the left side (pins 2, 11, 12) should
be connected to ground. (And connecting pin 23 won't be much of an
effort either :)

Still nothing :( I soldered all the VSS on the left side and bottom side.
But, VOUT33 still gives me 5v :( I have checked that there is not
short between pins.

Post by Werner Almesberger

Post by Rafael Ignacio Zurita
Should I put capacitors first before test?

It can't hurt either. Didn't you have that problem a while ago
already ? I.e., that you didn't have capacitors and the chip
performed erratically.

Yes. I had. But it was powering the MCU with 3.3v from nanonote,
without using the internal regulator.

Post by Werner Almesberger
Also, is your MCU actually powered ? There doesn't seem to be a
direct connection between VOUT33 and VDD*, so I don't know if
you're even in a defined state.

No yet. There is not connection between vout33 and vdd.
I wanted to check the vout33 first, before to do those connections.

Post by Werner Almesberger
Regarding soldering, the solder will flow towards hot surfaces. So
you don't need to heat up a single pin at a time but you can heat
up several pins at once. Plus, make sure you also heat the pad
under the pin. Only then will you get a proper solder joint.
I usually solder QFNs and such in two passes: the first pass is to
just establish enough solder joints, no matter how untidy, that
the chip stays in place. The second pass is to reflow the areas
where things didn't work out. In the second pass, I can add or
remove solder.
Also make sure your iron is hot enough. If it's just at the point
where solder will melt, it will get too cold as soon as you touch
something. A range of about 280 C to 300 C should be good, but
you'll have to find what works best with your iron and your
station.
Unless you've cranked up the heat far too high, you also don't
have to be in a rush to remove the iron from a solder joint. You
have to keep it there long enough to fully melt the solder and to
fully heat pin and pad. Only then will you get a good solder
joint.

I will save all these tips. Thanks Werner.

Post by Werner Almesberger
Another thing to check is that VOUT33 doesn't accidently make
contact with VREGIN. Also, the regulator may need a minimum
external load of a few uA before it actually can regulate.

It does not. At least when checking with tester. ANd I have
put 5v in vregin for a while using two sources in tests:
usb port and using 5v from a pc power supply.

It will be continued ...
Rafa

[Werner Almesberger]

Post by Rafael Ignacio Zurita
No yet. There is not connection between vout33 and vdd.
I wanted to check the vout33 first, before to do those connections.

If you want to play it extra-safe, you could power VDD from a lab
supply and power 5 V from USB (or another lab supply, if you have
one). With VDD powered, the system will be in a defined state.
Then the regulator has no excuse for not working :)

But I wouldn't worry about VDD possibly getting a bit too high,
and just make sure the system can't burn too much power. See
below.

Post by Rafael Ignacio Zurita

Post by Werner Almesberger
Also, the regulator may need a minimum
external load of a few uA before it actually can regulate.

It does not. At least when checking with tester.

Hmm, the tester may not provide enough of a load. If things
still don't work after powering VDD, maybe add some 1-100 kOhm
resistor between VOUT33 and GND.

Post by Rafael Ignacio Zurita
usb port and using 5v from a pc power supply.

For such things, I'd strongly suggest to use a lab supply. If you
set a low current limit (e.g., 20-50 mA), there is little risk of
anything getting damaged, even if you made some bad mistake.

Up to a certain point, a low current limit also helps to prevent
damage if you run with an incorrect (or reversed) voltage.

If you're working with a supply that can provide several ampere,
you can easily burn up your circuit if something isn't quite
right.

- Werner