impossible stud

I need 0-80's; don't have room for huge 2-56's.

Flatheads in big plated holes would be almost planar, but the contact force in the sharp corner would be high, and FR4 could cold flow and loosen up.

I once did a board with plated countersunk holes. The board house thought it was weird, but it worked. I might ask our current board houses how they feel about that.

I do need electrical and ideally thermal conductivity.

As you can see, we're still playing with ideas here.

Anything with small loose parts is an assembly irritation, and a did-it-get-loose maintenance issue.

If you go instead with some plastic snap-in standoffs, the baby board can have six pogo pins for its electrical connection, or if plating is pricey, just put landing spots on the (small) baby board and pogo pins in motherboard.

fredag den 20. april 2018 kl. 05.51.29 UTC+2 skrev John Larkin:

why male threads rather than female?

Press-in from _backside_ (with a shoulder). ...Jim Thompson

-- | James E.Thompson | mens | | Analog Innovations | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | STV, Queen Creek, AZ 85142 Skype: skypeanalog | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at

To those of us in my age bracket...

GREEN means inexperienced and/or incompetent.

The snap-ins would take too much room. We considered pins on the baby board and pin sockets on the main board, also a hassle.

Currents and speeds are high for pogos.

If I make the holes in the baby board a tad smaller than the screw size, the screws will thread into the plated holes a bit and become essentially captive. That makes assembly a little less messy.

I can't find a 0-80 fastener like that. There are all sorts of PEM-ish things in larger sizes.

There would be no pull-out stress on my knurled part. We might give up the knurl and solder the pins onto the baby board.

I'm getting quotes from screw-machine companies, but they are expensive. Over $2 each at first, over $1 in quantity.

I'm mostly confused about what you want to do. Is the thermal conductivity of the connections important? If there's not enough room for some header. Then some mechanical mount for stability, and hand solder in bus wire through plated holes. (?)

(with a loop on top so it doesn't fall through, how many are you making?)

George H.

Me too! Design usually starts that way.

The baby board will get hot, so I may as well dump some heat into the main board. But the main heat transfer path is baby board to gap pad to baseplate.

The baby board is 0.47" square, has parts, has six big connections to the main board. Tight. Hand soldering (and desoldering!) would be messy.

Just a few now, until we determine if the world wants it. We have one potential customer who might.

The shortest 0-80 flat head that I could find was 3/16". You wanted

The fiberglass won't cold flow but the solder plate most certainly will move under pressure. If installed with a torque limiting pneumatic or electric screwdriver, you could limit the damage. Getting a screw with a "flatter" head, using a 110 or 120 degree head instead of 82 degrees will also help. If you countersink the hole BEFORE solder plating, you can get a decent electrical connection. However, if you countersink after the board is plated, all the contact material will be cut away, leaving the threaded part to make the electrical connection. Also, if you want a decent electrical connection, forget about using glue to mount the screw. You probably will get an insulator instead.

Yeah, it would be weird. I tried to so something like I describe once and it was vetoed by manufacturing who claimed that it was too difficult to assemble properly. My "solution" was to mask the hole and insert a brass rivet. I then tapped the rivet. If you did the same with 0-80, 80 threads per inch with 0.063 thick PCB will yield at least 80 tpi = 0.0125 in/thread 0.063 in / 0.0125 in/thread = 5 threads which should give sufficient retention. To sorta lock the screw in place, squash the brass rivet threads in an arbor press.

However, finding a suitable tubular or semi-tubular miniature brass rivet is going to be a problem. 0-80 has a pitch diameter of 0.050" which would be the minimum center hole size for the rivet. The major diameter is 0.060", so something around 0.055" (1.4mm) should give sufficient thread depth for the screw.

If nothing can be found, maybe just make your own out of brass or copper tubing.

With a sandwich of dissimilar metals, I don't think it's going to be suitable for low level signal paths. Copper or brass hardware should offer tolerable thermal conductivity. However, if you need something better, perhaps a pin and socket type connector would be more appropriate? Something like one of these:

The trick to engineering is knowing when to stop engineering.

After seeing what info you've added in replies but abstained from in the OP:

How about clamping the board between two plates, with gap pad on either side? No screws needed at all.

Tim

-- Seven Transistor Labs, LLC Electrical Engineering Consultation and Contract Design Website:

And then SMTs did it with every connector imaginable. They usually come loose, but a lot of them manage not to.

Understanding the reasoning behind such rules of thumb, is far more valuable than the thumb itself (which is often little better than useless).

Namely, drugstore resistors with corroded, flaking plating, combined with amateurs wielding irons well before the days of IPC certification, makes for very questionable bonding. Hence the advice to wrap the leads around a hole in a terminal strip, providing some strain relief in a much larger blob of solder. Hopefully, what little surface actually becomes soldered in place, is held by the bulk of solder and other leads wrapped up together. Still, you'd better not exacerbate it by applying a repetitive mechanical load.

Tek's terminal strips surround the joint just as well, or better than, a punched metal eyelet, without having to wrap leads around them. Good soldering is still needed. I would imagine Tek employed much better platings and QC than, say, Muntz did. They also used stronger solder (silver bearing).

Speaking of platings, I have some relatively contemporary 5702s (submini pentode, wire leads). Mfg date in the 80s I think. I expect the leads are solder dipped, copper plated invar. They are bright and clean, and hardly need flux at all. Also have some tubular ceramic capacitors, probably older, and they solder just as well.

Modern manufacturing relies on quality. Component platings decay over time (oxidation); reflow ovens are tuned to the degree; metallurgy in general is much better understood than back then (we don't even need lead to deal with whiskering today, though not so much for metallurgical reasons as also having a range of mitigations to choose from).

Tim

Isn't that just an off-label use of a socket head cap screw? Press the knurled SHCS into the PCB and secure in place with epoxy. There are also PEM style studs, but I don't know how small they get.

Is there some edge you have access too? Then you could solder some bus header there... and clip the wire to take it apart. I saw some 2-56 pem nuts upstream can you fit two of those for support? Maybe the thermal conductivity of pem nuts is not so good.

George H.

Paul Revere was a silver-smith. Silver-soldering, also known as hard soldering or brazing, produces much more robust joints.

Nobody is going to argue with Paul Revere, but you've invoked him in an area where you should have known that he wasn't the relevant authority.

When you need the impossible screw or stud or net or bracket it's time to step back and re-think your mechanical. Is the baby board size cast in stone due to some other limit? It may be cheaper to increase the size of that board and use more standard screws. As you said, this product may be a no-go and to get a decent price you may have a bunch of inventory you will never use.

What is the max component height on the baby board? Keep your screw head height at or below that. If it's getting that hot I suspect there is something larger than a 0402 on there. Put the PEM on the motherboard and find a 20 thou or close thick washer to use as a standoff.

How well do they work? Years ago one of the large connector/wiring compani es made a press in pin for IDC wiring (like wirewrap but quarter inch tall pins instead of inch and a half). It was rather tricky to get the thing in serted without pushing out the via barrel. One PCB fab house I visited sai d the only way to get a final hole diameter that would meet the tolerance r equirement was to drill after plating. Even then they said it would be tri cky to meet the spec. These parts were "interference" fit with a complaint pin meaning the pin was fatter than the hole, but had a degree of springin ess to the width which would allow it to fit very tightly. The knurled fit ting works by reshaping the metal it is being pressed into. Can this work very well with PCB material?

From the PEM product page for a broaching PCB fastener...

Broaching and broach/flare types are designed for unplated mounting hole ap plications. If used in plated mounting holes, the stresses involved can damage the plating, push out the plating entirely, or break any traces inside the board that might be connected to the plated hole. When installing into non-plated mounting holes there may even be issues with del amination, measeling or crazing in some instance

I think they are saying this mounting type doesn't always work well.

Rick C.

How about a MillMax press-fit turret?

Can you separate the electrical from the mechanical? What about using flex or flex-rigid to link the baby to mother board and then a spring or other compliant device to keep baby pressed to the heat sink?

piglet

der

If you used thick enough wire you could poke it through holes in the daught erboard, bend them over & solder. Now if you run those bent over ends along the board & solder them somewhere else again you get a more pull-through r esistant fixing. Or perhaps fold them over the daughterboard edges, not tri ed that.

NT