next topic »

[on:]

I'm new to smd work, but I've tried to be as quick as possible about soldering my 0.1 uF 0805 smd caps on.  Using my Fluke meter, they measure 100 nF before soldering, but only 2 out 6 still measure that after soldering them on the board.  The others read 2 to 4 uF.  There's no other components on the board yet...  I got them off eBay from hong kong, like 30 different sizes of caps and resisters for $8.00.  I didn't think about them being poor quality. Is there a quality issue in these tiny smd's or is it that I got them too hot...  I cleaned the flux off with alcohol.
Maybe I should order some quality ones from digi key.  But they sell like a thousand different types, how do I know if I'm getting good ones.  These 0.1 uF caps are the ones you use with the crystal, on the Arduino and one at each IC, like my H bridge IC...

[Reply #1 on:]

Are you sure you read 2-4 uF? Or is that 0.2-0.4uF? It would make sense if you are soldering them to the board and they are in parallel (4x caps at 100nF --> 0.4uF).

I would be very surprised if you have bad capacitors. Either think about the above, or maybe let them cool down a bit? I don't think getting them too hot would cause a huge capacitance, just the opposite.

If you're going to go the Digi-Key route I'd say the 'good ones' are C0G/NP0/X5R/X7R rated (not the crappy Y5V or Z5U types). Beyond that...get the cheapest ones

--
The Ruggeduino: compatible with Arduino UNO, 24V operation, all I/O's fused and protected

[Reply #2 on:]

First:  Don't forget that once you measure in-circuit, you are measuring the entire circuit, not just the component.

One of the little known characteristics of ceramics (barium titanate like X5R, Z5U) is that their capacitance ages.  The clock starts when they've been heated up to about 150°C, which is generally well within soldering temperatures.  Every decade hour from that point, their capacitance decays by a few percent.  Decade hours are 1, 10, 100, 1000, 10000, etc.

It actually means a ceramic will see a large increase in capacitance immediately after soldering.  Give it 48-100 hours and it can be >5% different.

[Reply #3 on:]

Thanks.  I desoldered a couple of them and they read 106 nF again, so I guess things are okay.  I confess, I said no other components on the board yet, but I actually soldered two 0805 10k pull ups on, maybe them plus the traces themselves are giving the false reading...
Thanks.

[Reply #4 on:]

maybe them plus the traces themselves are giving the false reading...

No, not "maybe".

[Reply #5 on:]

I find on some of my bigger boards the 5v trace for example actually had a decent capacitence, I got something like .3uf, and most other traces had between 20-50nf
so it most likely was the entire circuit doing it

[Reply #6 on:]

Many of the capacitors on a modern microcontroller board are effectively in parallel, between the power and gnd rail.
(see "bypass capacitor."  I used to wonder why I couldn't replace all the (expensive!) 0.1uF caps on TTL boards with a single
2uF cap (cheaper.))
So once you solder them on and try to measure the capacitance, you'll actually be measuring the parallel capacitance of all the caps connected between those signals.  (for instance, the "Freeduino" has a .1cap at the output of the voltage regulator, a .1cap for the power rails of the FTDI chip, and two .1 caps on the power rails of the AVR.  Those are all essentially connected in parallel, unless you have a really fancy capacitance meter.)