pick-and-place

If you're talking about bypass for digital then SRF is what you to know as it translates directly into supported rise times and damping.

Three.

It's 1930's poured concrete side walls with wood floors, ideal to pancake in a quake. We added a bunch of steel and stuff.

The old concrete interior columns would likely crumble in a decent quake.

I do digital and mixed-signal stuff, from KHz PWM to picosecond timing, and I never worry about SRF.

With nanosecond edges, it makes no difference whether SRF is 5 MHz or

The real bypass is parallel PCB copper planes, sprinkled with almost any ole ceramic caps here and there.

I took a high speed digital design class with Lee Ritchey once and that is pretty much what I took away. He even proved the the location of the bypass caps is not so critical as the power planes form a transmission line which supplies the needed current to the chip while the wave front is traveling to and from the cap. He even built a test board to demonstrate this.

The only issue which may require some consideration is the issue of mixing cap sizes/values. There can be interaction resulting in peaking of impedance. However, the ESR of the caps usually limits these peaks. Even when just one value/size of cap is used it resonates with the power planes with a similar result. So oddly enough some ESR in the decoupling caps is your friend. :)

I'm one step removed from manufacturing as I use a contract assembly house, but if they caused any real problem I would have heard I'm sure. I use the ones with the mouse bites. Back when I started using these there were two kinds, mouse bites with the leads in the cavities and notched with the leads on the protruding edges. I think the industry ended up going for the mouse bites and the protruding edge parts are not so common now.

I use the 1206 sized ones with 8 resistors tied to a common pin. That is a big improvement in part size compared to 8-0603 resistors. But it still requires a lot of routing with a savings in the common pin connections. They get used as digital pullups mostly where the value is very non-critical.

Yes it usually is, lowering the Q helps remove crap like that.

Jamie

The last big soldering problem that we had was on the Linear Tech LTM-series of switcher bricks.

The part has flat lands on the bottom, and solder tends to squish between the lands and short things. Linear said that

1. It's our fault, bad soldering process

and

1. They are changing the package to BGA, so return the LGAs for exchange.

They had another problem internally. They used conductive epoxy to attach one diode, and it unglued itself when reflowed to ROHS sorts of temperatures. See 1, 2 above.

LTI didn't do them selves any favors with the LGA modules. I was told by the CM without prompting to avoid them because of soldering problems. I think the change to BGA says it all.

• Who is the pregnant guy?

for you, sheesh. Isn't the SRF of the 330nF 0603 rather low for bypass?

as it translates directly into supported rise times and damping.

Okay, you have a fundamental misunderstanding of the function of the caps i n the case of power plane layouts. Looking into the parallel combination of capacitor and transmission line power plane, the capacitor diverts the hig h frequency current loading off the line thereby decoupling the bypassed de vice from other devices on the same power plane feed. This is only good wit hin the working frequency range of the cap, so a 100KHz or a 1GHz SRF almos t certainly does make a difference in that regard. Unless you're working at a few tens of GHz, there's no wavefront hitting capacitor, it is a lumped element. That 330n with max 10MHz SRF is only slightly better than worthles s. A 3.3n would do much better.

The point is that the "working range" of a cap is not defined by the SRF. What is important is the impedance vs. frequency. The lowest point is at the SRF, but it is still low well above the SRF just as it is low below the SRF where it is capacitive, but increasing impedance as the frequency drops.

The working range is defined by the inductance curve which is largely a function of the size and shape of the cap package. The capacitance only defines the impedance below the SRF where the problem is less important since there are other, larger caps to provide decoupling.

btw, I'll let one of the Google groups users clean up the double and quad and even octal spacing mess.

I don't need you or your kind explaining the frequency characteristics of a capacitor. Thanks for the thought though.

If you can make the boards with parts only on one side, you can dip solder the through-hole parts. I've just gotten into this for one product, and it worked amazingly well. I bought a solder pot from Honk Kong, some SAC305 bar solder on eBay, and used some liquid flux I had around for rework. I used Kapton tape to mask the places that shouldn't be soldered. It worked amazingly well, and saves many minutes over hand soldering.

Jon

Most decent CAD packages can provide a basic XYR file with a component description. I wrote a C program to read the placement file from my Protel 99 CAD package, and a component file that tells what feeder and rotation the part comes on the tape, and produces the file that our particular machine needs. I manually add a header that contains board origin and fiducial locations. This is usually quite easy to do.

Jon

n for you, sheesh. Isn't the SRF of the 330nF 0603 rather low for bypass?

e does. And the impedance in the frequency range that mainly matters is dom inated by ESL, which is about the same for all 0603 ceramic caps.

w as it translates directly into supported rise times and damping.

and I never worry about SRF.

500 kHz. The high-frequency impedance will be about the same.

y ole ceramic caps here and there.

in the case of power plane layouts. Looking into the parallel combination of capacitor and transmission line power plane, the capacitor diverts the h igh frequency current loading off the line thereby decoupling the bypassed device from other devices on the same power plane feed. This is only good w ithin the working frequency range of the cap, so a 100KHz or a 1GHz SRF alm ost certainly does make a difference in that regard. Unless you're working at a few tens of GHz, there's no wavefront hitting capacitor, it is a lumpe d element. That 330n with max 10MHz SRF is only slightly better than worthl ess. A 3.3n would do much better.

And a 1nF porcelain microwave capacitor would be even better.

Not cheap, but good.

I was only responding to what you had posted. I'm glad you understand better now. :)

You don't actually design or build electronics, as I recall.

I knew a guy, in the aerospace biz, who didn't use bypass caps at all on his multilayer digital boards. They worked.

Insufficient info for that to be of any value.