Future: 0603 versus 0402 parts

With 0402 on a hand placed board, a noticable amount of parts were only soldered on one side. 0603 behaved much better.

--
Uwe Bonnes                bon@elektron.ikp.physik.tu-darmstadt.de

Institut fuer Kernphysik  Schlossgartenstrasse 9  64289 Darmstadt
--------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------

Oh. I thought at those prices you were getting them individually bagged or something.

Best regards, Spehro Pefhany

--
"it\'s the network..."                          "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com

On 3/13/2007 8:01 PM, The digits of Joerg's hands composed the following: > Interesting. 0402 caps in my case would only be 0.01uF and smaller

If I remember correctly it was more an issue with .1uF than .01. One conclusion of a report I read said to over-voltage the caps more than normal. For example for 3.3V I would normally be o.k with a 6.3V cap, but they suggested using a 10V part or more for 3.3 if using 0402 (which greatly limits what is available).

Another comment. I rarely use 0805s. I try to use 0603 and 1206 only, unless space pushes me to 0402 or 0805 or requirements push me up to

1210 or something. The reason for this is to avoid duplication. Why should I maintain PCB libraries and inventory for three sets of capacitors? (1206 - 0805 and 0603). So I use 0603 mostly then go up to 1206 if I have to. If the customer does not want 0603 or the board does not require them then why not go 1206. The biggest exception to this is larger MLCCs where I can't get want but the PCB is small. For example I I want a 10uF 10V MLCC I may go 0805.

Hawker

Perhaps that's the placed cost?

--

John Devereux

That would not be a good sign for the assembler. I have hand-soldered

0402 and didn't have a problem. Under time pressure BTW because one of their engineers was working up a serious flu and had to go home (and drop me off at the airport). Except for a little eye strain from wearing 3x glasses for too long it wasn't a problem and the solder joints were ok.

--
Regards, Joerg

http://www.analogconsultants.com

For lab stock I can understand this but for the CAD library? My CAD allows to generate a part and then designate as many package styles as you want. So, for example, I can place 100K pull-up resistors in 0603 and when it ends up looking like L.A. at rush hour I can tell the CAD software "Let's make all these 0402".

--
Regards, Joerg

http://www.analogconsultants.com

My headbone is having somewhat of a time keeping up here - "0402" - isn't that, like, 0.040" x 0.020"? Isn't that like 1 mm x 1/2 mm?

That's smaller than what comes out of my pepper shaker. =:-O

I guess you need a microscope and a non-magnetic tweezers or something? (or hire some kid who can actually see the things!)

Thanks, Rich

Yep, that's the one. The real fun starts with 0201 and smaller. Yes, there is even smaller SMT. Recently I did a design where I could not place a SOT-23 reference. Because that was too big...

Tweezers yes, microscope no, at least not for me. 3x glasses suffice, so far. Most engineers and technicians at my clients prefer to solder such sizes under the microscope. It's hard to come by good tweezers in a lab and TSA rules don't allow me to take mine along anymore. So at times I have ventured outside my client's buildings, picked up a few twigs, filed them into "micro arrows" and used these for hours to solder really small SMT. Tooth picks are ok but a bit short, gives me cramps. But before that my wife educated me to distinguish good plants from poison oak. Of course you'll also have to watch out for rattlers before reaching into a flower bed.

--
Regards, Joerg

http://www.analogconsultants.com

Yup. There are two smaller sizes: 0201 and 01005. The 01005 are

0.016 by 0.008 - they're shorter than the width of my solder!

Here's a microphoto of a 01005 soldered on to ~8 mil tracks (the tvsop lands on the right are 0.4mm pitch):

For comparative photos, this project of mine has everything from 0805 down to 01005 parts on a 0.5"x0.5" board:

(if there's enough interest, I could do another run of those)

The 01005s are more like dust. It's even hard to see them when they're still in the paper tape.

Here's what 0201's look like:

And by comparison, 0805 0603 0402 0201 01005:

I use a 2x magnifying visor for everything down to 0401's, and a 3.5x for 0201 and 01005s, and tvsop (0.4mm pitch). I have an IntelPlay QX3 microscope for verification.

I used a combination of steel tweezers and sharpened toothpicks to manipulate the 01005s.

I came to like the fine titanium tweezers that I found kicking around in my lab. Most "non-magnetic" stainless steel tweezers I have are too magnetic for SMD stuff.

I have just finished my *first* SMD project and I must say I love it. I have shied away from SMD thinking it'd be too much of a hassle, but the work is no slower than with through-hole, possibly even faster. I have to admit, though, that I didn't use anything smaller than 0805, SOIC, and SOT23.

I don't have proper rework tools though, which makes prototyping less fun.

--Daniel

SMT is faster because you don't have to turn the board around and snip off excess lead length.

What helps is to fire up two irons. I touch one side of a resistor with each, then flick it into a bucket. But in a safe way so the solder doesn't fly anywhere it could cause harm.

--
Regards, Joerg

http://www.analogconsultants.com

Try sharpened kebab skewers. Much longer, less cramping in the fingers.

--
Regards, Joerg

http://www.analogconsultants.com

I picked up a Metcal Talon on eBay. I'm quite pleased with it, especially for fiddling with 0603 parts. Just grab them and remove them from the board! It takes off tssops almost as easily. The wider jaws are useful for through-hole parts too; you can, for example, use just one jaw to heat a row of leads all at once.

Even the smallest tips are a bit big for 0201 and 01005 parts, though. However, those I can get with my regular chisel tip, since it spans both pads.

Yes, that's the hefty 0402. There's also a metric 0402 which is 0.4 x

0.2mm (01005), which is getting pretty dern small. The regular 0402s can easily disappear into a blob of solder or go flying into nothingness. With massive 0805s you generally should not need to buy any extras even for hand assembly, but I always allow for a few extra of the tiny ones to account for evaporation.

Or if you're myopic, just get close enough without searing your corneas with splashes of flux.

Best regards, Spehro Pefhany

--
"it\'s the network..."                          "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com

what you are alluding to here is actually the behaviour of the dielectric. so-called "high-K" dielectrics are appalling s**te - eg Z5U (worst) and Y5V (not much better). either of these need to run at about

10% of rated voltage to get close to rated capacitance. at 40% rated voltage its about 40% rated C.....

and for a given capacitance, reducing volume requires higher Er....

and the C-vs-T curve is just as bad (actually a bit worse), in either direction. So even if you massively over-rate the component voltage, unless your design sits at 25C all the time, you will still take a beating.

Several times I have done an exercise where I compare Z5U/Y5V with X7R, and for a real product with real temperature swings, X7R wins every time (most capacitance in a given package). I really dont know why anyone ever uses Z5U/Y5V.

Cheers Terry

I agree, X7R (at worst) every time in my designs where relatively high capacitance is required.

I suppose if you're building crap consumer products that don't really have to work much outside of room temperature and you can shave a teeny bit of cost off it, the 'semiconductor' dielectrics might start to look attractive-- kind of like the beer goggle effect.

Best regards, Spehro Pefhany

--
"it's the network..."                          "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com

Same here. Still difficult with multi-pin parts. But TPH can also be a botch to rework because you need to liquify (and get rid of) all the solder that is ob both sides of the PCB and inside the hole. Single-side is the easiest to rework, but I haven't done that since my hobby days anyway.

--Daniel

As Jim would say "Sn_o_ooort!"

They used _students_ in _Glasgow_?!! Where'd they find the sober ones?

Regards Ian

(My daughter graduated from one of the Glasgow Universities a couple of years ago. The social life is legendary ;-)

Well, Cadsoft Eagle is quite smart in that domain: You click on, say, a resistor, transistor, whatever. Something that comes in various packages. Provided you created the library correctly that'll give you a drop-down menu just like Windows Explorer does, with that little plus sign. Click that and you'll get the package versions. If you discover after placement that you could get away with a larger version you can use a routine "change" -> "package". Then it replaces the original part with a smaller part that has a different ID. This change will also be reflected when you run the BOM.

Unfortunately Eagle does not yet allow additional part fields. This will come with the next release, after some petitioning by us industry guys. So currently it is difficult to do the linking to the partmaster database. But it is possible via special BOM routines. In Eagle you can write your own user language program to create, say, a "Hawker-BOM". In it you can specify how it will merge with part numbers from the partmaster. The next release should make things easier, then all you need is the actual partmaster number in a third part field. Since I work for several clients at a time it would have to be a cross reference number in my case, or I just create one dedicated set of libraries for each client.

--
Regards, Joerg

http://www.analogconsultants.com