Thru hole vs surface mount R's at ~100MHz.

Nov 09, 2010 108 Replies

The holes on that array are too small for those leads. If they were slightly bigger, the added solder mass would allow the solder joint to reflow properly. It also looks like the person that soldered it did not preheat the entire board to assist the solder joint formation and reflow. I can say this because the thermal damage to the PCB is apparent at each of those locations, and I have solved this problem myself by making the hole bigger before, and have even done so on an 0.090" multi-layer PCB.

So, instead of sizing the hole to hold the element being soldered, forgo that for achieving a proper solder joint that does not stress the hole during formation.

Multi-layer boards soak up so much heat that thru-hole parts can be a pain in the ass. The last thing you want to do is undersize the hole and attempt to rely on the device lead to feed the heat through the solder joint or use the hole as a fixturing or location aid for the part. It is supposed to be a specific size *bigger* than the lead, not a tight fit AT ALL, and with multi-layer issues, it should be slightly bigger still. This provides more thru-hole plating on the larger hole, and a bigger annular ring, as well as a larger solder mass.

It will be easier to build, even if it requires a new method for exact placement of the device(s) if that was the goal in making them so tight.

Plenty of space there from what I saw. Bigger holes = easier build, no performance hit.

I figure it is either an array of discreet sensors of some type or a photo-multiplier tube exciter array that one finds under the socket, splitting up 1500 volt feeders. At least, it looks similar to one I used to make. Can't tell though since you are not showing that side of the board.

Hey, on your board here:

...do you know if those little rotary switches are washable? I used something that looks nearly identical

formatting link
-- Mouser stocks them) a couple months ago and the manufacturing guys were complaining that they aren't washable...

Thanks,

---Joel

You're babbling. Stop that.

And apparently zero knowledge about fast stuff.

John

this

All possible SOT-23 pinout mistakes can be fixed by rotation and flipping... if the leads don't break. Trust me, I've tried them all.

John

this

Idiot. This board sells for about $10K, and it works fine.

Nothing like that.

John

Those are trimpots, Bourns 3314G, and they are washable. They make a SPDT switch in the same package, essentially the same pot with no resistance element.

Actually, a 1M trimpot makes a pretty good SPDT switch in most applications.

John

Looks like 1/3 the series inductance! The next step would be to snake a small ground trace under the body, between the end caps. I'd guess another 2:1 improvement could be had.

This resistor, heat sunk as it is, would probably be happy dissipating a watt.

I should try Vlad's 100 ohm idea, namely two in parallel as a 50 ohm terminator, maybe each 90 degrees to the trace, at the end. Two watts!

Beats me. Just guessing.

Yum. Make fresh whipped cream.

John

this

frequency

been

;-)

...unless it has more than three pins. :-(

In our case, only two of three pins were wired (LM385).

You'd be surprised what a little residue can do.

Look what the world left us with in your case.

Sad world.

Great, thanks!

Ha... good point!

1/8 watt metal film, radial lead. Snip, curl, and form the leads to create a surface mount component That stands on its own. It can be hand soldered or actually reflowed like a normal SMD.. +-[===]-+ .+- -+

To even get a slightly smaller form factor, use an English (UK) brand device.

to 100

.

rbon

se

s

John, your resistor measurements look way too high.

I pushed them down across the gap of the surface-mount fixture of my AADE c-meter. The 1 pF ceramic cap checked fairly well.

On a PC board, with pads, capacitance might well be different, probably more.

Got better numbers?

John

They are not his, and he noted the anomalous numbers to the author.

I made those measurements.

John

Then, the response I saw was correct. That the numbers are too close to each other to be real, so you messed up in your testing.

Again, I refer back to simple things like cleaning AFTER the assembly has been made.

Also, the chopped of right at the body resistor leads I saw are not real world in any way either.

I thought it was you mentioning this to the guy that measured, but I guess it was him, mentioning it to you.

He was right. Your numbers are not real world. You did not perform the test correctly, AND you should have spotted that fact right away.

There's a reason the numbers are close. Guess what it is.

What does cleaning have to do with the capacitance of resistors? I didn't solder them.

English, please? What resistors are you referring to?

???

You're babbling again.

John

It should vary with form factor. If it doesn't... or at least doesn't enough, it is a sign that you are not getting valid readings.

F*ck off, you retarded twit.

You posted a pic of a resistor soldered down on a piece of clad FR-4

I picture your work-up of this "test" series to be just as bad.

Maybe you soldered them in another language.

And they aren't credible.

Think about it. Winfield Hill doesn't trust your results, I don't trust your results and the Gibberman agrees with me and Win.

It might be worth your while to get yourself a test fixture which looks a bit more like a real situation.

-- Bill Sloman, Nijmegen

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required