Using 0402 when plenty of room for 0603

[Faringdon]

Hi,
We are doing a LNK3209 based Buck converter. The designer has used all 0402 resistors for the auxiliary resistors around the control chip. There is bags of room for 0603's but he used 0402 instead.
Do you agree this is a very bad idea?
0402 doesnt allow the value to be checked by checkers in the PCB assembly house.
Also, 0402 needs the board temperature to be within a much tighter tolerance than 0603 resistors during surface mound soldering.
"Head on pillow" is far more common with 0402 than 0603...partly as the 0402 is lighter and can  "tomb up" much more readily.

[floobydust]

Hi,
We are doing a LNK3209 based Buck converter. The designer has used all 0402 resistors for the auxiliary resistors around the control chip. There is bags of room for 0603's but he used 0402 instead.
Do you agree this is a very bad idea?
0402 doesnt allow the value to be checked by checkers in the PCB assembly house.
Also, 0402 needs the board temperature to be within a much tighter tolerance than 0603 resistors during surface mound soldering.
"Head on pillow" is far more common with 0402 than 0603...partly as the 0402 is lighter and can  "tomb up" much more readily.

How is it that "we" are doing the "converter" but an idiot is spec'ing 0402's? That sounds like your preferred way, too small for your own good in a SMPS with large through-hole parts. 0402 saves nothing at all.

[SMTech]

Hi,
We are doing a LNK3209 based Buck converter. The designer has used all 0402 resistors for the auxiliary resistors around the control chip. There is bags of room for 0603's but he used 0402 instead.
Do you agree this is a very bad idea?
0402 doesnt allow the value to be checked by checkers in the PCB assembly house.
Also, 0402 needs the board temperature to be within a much tighter tolerance than 0603 resistors during surface mound soldering.
"Head on pillow" is far more common with 0402 than 0603...partly as the 0402 is lighter and can  "tomb up" much more readily.

No, I don't. 0402 is increasingly standard for many designers, also don't want to break your bubble but you may have noticed the values disappearing from 0603 and 0805 resistors too. However if you are doing very low volumes and not using a proper assembly line, sure you'd be better off with larger parts just because they are easier to fiddle with or handle with tweezers. Your assembly concerns are unfounded unless your process sucks.

[asmi]

Hi,
We are doing a LNK3209 based Buck converter. The designer has used all 0402 resistors for the auxiliary resistors around the control chip. There is bags of room for 0603's but he used 0402 instead.
Do you agree this is a very bad idea?

Absolutely disagree, and I use 0402 everywhere I can because 1) they are the cheapest 2) they take the least amount of space on a PCB.

0402 doesnt allow the value to be checked by checkers in the PCB assembly house.

PCB assembly house can't afford a multimeter?

Also, 0402 needs the board temperature to be within a much tighter tolerance than 0603 resistors during surface mound soldering.

"Head on pillow" is far more common with 0402 than 0603...partly as the 0402 is lighter and can  "tomb up" much more readily.

[Feynman]

Maybe it's your company's default size, because of price, availability and space constraints on other boards. A consequence might be that your company's in-house stock for prototyping and repair is mainly 0402. Or the EMS is storing a large quantity of 0402, because most boards your company is ordering are 0402.  Maybe your company or the designer has some well established fan-out patterns for 0402. For example, how vias are placed at a decoupling capacitor, how a resistor network is placed around an operational amplifier, etc, etc, ...

There could be many more reasons for using 0402 although there is enough room for 0603 and even bigger ones.
But having a standard size you don't deviate from without strong compelling reasons is definitely advantageous. And 0402 is becoming increasingly more popular as a standard size.

You have to be a little more careful with the parameters of 0402 parts compared to 0603, especially with capacitors. However, you cannot be careless with parameters of 0603 either...

For most EMS 0402 is as "jelly-bean" as 0603 and there shouldn't be any soldering issues if you stick to the EMS's recommended land pattern.

[Faringdon]

Absolutely disagree, and I use 0402 everywhere I can because 1) they are the cheapest 2) they take the least amount of space on a PCB.

Thanks...on that basis, then why not use 0201 instead of 0402?

[mikeselectricstuff]

Absolutely disagree, and I use 0402 everywhere I can because 1) they are the cheapest 2) they take the least amount of space on a PCB.

Thanks...on that basis, then why not use 0201 instead of 0402?

Because of the tradeoff between size and useability - 0201 needs much more care for assembly, more expensive feeders , more precise placement and paste print, and is much harder to rework.
The sweet spot depends a lot on other factors - what sort of machine it's assembled on, volumes etc. For hand assembly, 0805 is the optimal compromise between size and assemblability, for low-end pick/place, 0603, for higher-end pick.place 0402.
The only reason to use 0201 or smaller is where you absolutely need the smaller size

[loki42]

The default size thing I think is very sensible.  I don't stock any 0603. 0402 fit more on a reel,  take up less space on the board and are easy to place / paste. Having a reel of the same value in a different size would be annoying.  I do use 0805 for larger caps like 22u.

[Reckless]

Absolutely disagree, and I use 0402 everywhere I can because 1) they are the cheapest 2) they take the least amount of space on a PCB.

Thanks...on that basis, then why not use 0201 instead of 0402?

Because of the tradeoff between size and useability - 0201 needs much more care for assembly, more expensive feeders , more precise placement and paste print, and is much harder to rework.
The sweet spot depends a lot on other factors - what sort of machine it's assembled on, volumes etc. For hand assembly, 0805 is the optimal compromise between size and assemblability, for low-end pick/place, 0603, for higher-end pick.place 0402.
The only reason to use 0201 or smaller is where you absolutely need the smaller size

Are they still more expensive than 0402?  I would like to move to 0201 but they have been more expensive than 0402. 

[tooki]

Absolutely disagree, and I use 0402 everywhere I can because 1) they are the cheapest 2) they take the least amount of space on a PCB.

Thanks...on that basis, then why not use 0201 instead of 0402?

0402 are the cheapest. Ergo, 0201 must cost more. (And they do.)

As you go smaller from 0402, cost increases, so you only use smaller components when board space or parasitic inductance need to be minimized.


Why does price increase as you go smaller or bigger than 0402?

You’ve got two size/cost curves: one for materials cost, and one for handling cost (both in component manufacturing and in board assembly).

Obviously, materials cost goes up with size, since you need more of it for larger components. So making them smaller reduces the materials cost. But at some point, the components get so small that they’re harder for the machinery to handle, so that curve goes up as the size shrinks. The cheapest components are where those two curves intersect, and today, that’s 0402. Anything smaller requires more advanced pick and place machines, which not all assembly houses may have, and those that do might not have those machines on every production line. And IIRC, they may not be as fast with the tiny components. Sticking with 0402 and larger means the older and/or faster PnP machines can be used, lowering assembly cost.

[schmitt trigger]

If there is something that the past year’s component shortage taught everyone, is that the design must follow a package’s “sweet spot” in terms of price AND availability.

Right now, the 0402 package seems to have consolidated that sweet spot.

A company I consulted for, had a strict policy of 0603 components, back when these were the sweet spot.
You may ask: what happens when the applied voltage causes the component to exceed its voltage or power rating? Then you use two or more devices in series.

There were exceptions to the rule, of course. Supply filter caps were 1210 in size. Likewise some mili ohm shunt resistors.

[asmi]

Thanks...on that basis, then why not use 0201 instead of 0402?

1) 0201 parts are more expensive than 0402.
2) the tech to handle anything <0402 is too expensive at this time.
3) I can handle 0402 manually (under a stereo microscope), working with 0201 is much harder.

[Zero999]

If you've got loads of space, then how about designing the board so it will take either 0603 or 0402? That way you can use whatever you can get hold of.

Incidentally, I'm working on an H-bridge with opto-coupler MOSFET drivers and will design the PCB, so it will fit different types of opto-couplers and possibly MOSFETs, so I don't face supply problems. The marginal increased cost of the PCB will be worth it.

[Faringdon]

Thanks, do you mean a single footprint that will take either 0603 or 0402...ive been told off for doing that in the past...told the 0402 could slip about during solder bath , and not be soldered properly?

[coromonadalix]

avoid 0402  if possible .......... thoses pesky ...

but for part number or values written on them

for sure they love to makes a mess   before on bigger parts you had say written "101"  for the value  who was easy to decode etc  ...

and now on some parts it "a21"   or something like this 

 

[asmi]

Thanks, do you mean a single footprint that will take either 0603 or 0402...ive been told off for doing that in the past...told the 0402 could slip about during solder bath , and not be soldered properly?

Don't do that, you will have problems during assembly with such footprints - unless you intend to assemble those boards manually.

[mikeselectricstuff]

You have to be making a colossal quantity of things if the relative cost of resistors matters. 

[Zero999]

Thanks, do you mean a single footprint that will take either 0603 or 0402...ive been told off for doing that in the past...told the 0402 could slip about during solder bath , and not be soldered properly?

No. I was talking about separate footprints, not one for both.

[nimish]

A fun corner case is that decoupling capacitors are principally limited by package inductance, so shrinking them is required for them to do much at higher frequencies. The capacitance itself doesn't matter that much. So going from 0402 back up to 0603 might cause more problems than you solve.

[Faringdon]

I think if you have your own in-house SMD machines, and do you own board stuffing, then 0402 is ok...but if you are using external board houses, and your volumes are nt massive, then they never do the temp profile right, and you end up with 0402 with "head on pillow"...this happens less with 0603........its just the truth.......board stuffers dont know how to perfectly set their machines, and in truth, every different board needs a different temp profile, and unless they are dedicated to that board, they wont get the temp profile exactly right, and  you get head on pillow with 0402

[SMTech]

I think if you have your own in-house SMD machines, and do you own board stuffing, then 0402 is ok...but if you are using external board houses, and your volumes are nt massive, then they never do the temp profile right, and you end up with 0402 with "head on pillow"...this happens less with 0603........its just the truth.......board stuffers dont know how to perfectly set their machines, and in truth, every different board needs a different temp profile, and unless they are dedicated to that board, they wont get the temp profile exactly right, and  you get head on pillow with 0402

Horseshit. I can run a pretty much the same profile for just about any board of similar mass, the measured profile will be broadly the same. The thermal headroom on modern reflow is massive, if you see issues outsourcing, your layout is bad.

[HighVoltage]

I think if you have your own in-house SMD machines, and do you own board stuffing, then 0402 is ok...but if you are using external board houses, and your volumes are nt massive, then they never do the temp profile right, and you end up with 0402 with "head on pillow"...this happens less with 0603........its just the truth.......board stuffers dont know how to perfectly set their machines, and in truth, every different board needs a different temp profile, and unless they are dedicated to that board, they wont get the temp profile exactly right, and  you get head on pillow with 0402

This was a problem 10 or 15 years ago but not today.
I use 0402 all the time, especially for buck converters and there is really no problem in manufacturing.

And 0402 can still be soldered by hand for prototyping.

[nali]

We are doing a LNK3209 based Buck converter. The designer has used all 0402 resistors for the auxiliary resistors around

I can't be the only one wondering what "doing" means in this context..?

I've worked with 0402 in smart meter products on boards with extensive copper pours and connections to fairly massive busbars, no problem although can be a pain to rework. IIRC the only 0603 & 0805 were ceramic caps.

[Faringdon]

Analysis of failure returns.......surely easier with 0603 as you can see what value they are.
Ditto Board stuffers QA dept doing 10%_of_batch checks....they can check the 0603 value, not so 0402.
When in cct, often they cant be reliably measured.

I worked in one place, where the guy soldering up the 0402's in the prototypes, used to use a weller iron, with the wrong tip in it.....so as to reduce thermal contact.....which stopped the pesky 0402's from getting sucked into his soldering iron when he soldered them..(when he removed the soldering iron, the solder would "follow" the iron, and dislodge the 0402 he had just soldered.)

All would agree, they are much worse for product prototype testing point of view.

Never yet seen a "head on pillow" 0603......seen many H.O.P 0402...seen more tombstoned 0402 than 0603.

Admittedly people get bullied in to 0402 by the board stuffer, who wants to keep stock room low.

[tooki]

You know they’re now getting rid of markings on 0603’s, right? (Which I hate, but so it is…)

As for the rest of those arguments: I’m sure that people who were used to THT resistors used to say the same things about 1206, then 0805, and then 0603, too.

And as for the soldering technician: I guess nobody told him about these magic things called “tweezers”.