Does anyone have an opinion on using a soldered-together assembly? I'm visualizing something that has right-angle connectors on the "big" boards, that just go into holes on the end board.
It would mean that repair would have to happen as a unit, but it would save me some space.
Again, comments welcome. The more I think about it the more horrid it sounds, but -- comments welcome.
That looks cool if the sockets could absorb any alignment mismatches of the opposing pins -- I'm not sure how to insure that.
Well, it's done. It might be less icky if you used a 2mm pitch right-angle header.
The boards will be difficult, if not impossible, to separate without damage and you should budget for a jig to hold them during hand soldering, but maybe those are not deal killers.
Years ago I designed some assemblies that were made with through-hole ribbon jumpers and wave soldered, then the boards were depanelized and assembled into the final configuration (two parallel boards and another at right angles). It was very cheap and effective, and the
0.1" pitch ribbon cables were UL rated etc.--sp
If voltage is a problem you can Mill out the adjacent pin on the PCB.
Cheers
On Fri, 15 Jan 2016 13:15:08 -0600, Tim Wescott Gave us:
Remember the mil spec DIP sockets that had round PCB insertion pins and spring loaded internal IC chip pins? They stack really well, so we noticed this 'feature' and we used to cut them in half and put one on the circuit board and whatever wires into the connecting element and it would fit right in like a serviceable plug and jack pair socket thingy. Heating one up with a solder tip allowed it to be removed from the plastic DIP socket frame it was in, so you could remove every other one and get 200 mil spacing. It could be hard attached to a board via mag wire "stitches" bent at the 90 degree mode you like. I have attached these to the edges of a PCB too.
We also used to use little "christmas tree" PCBs with foil at the edges and slots to mount 90 degrees to the edge of a perpendicular mounted daughtercard type PCB for extra rigidity or as a guide for the daughtercard.
You can also buy those in SIP form at like 25 pins and cut them into whatever number you like and space them however you like.
Oh! I found it!
On Fri, 15 Jan 2016 16:48:51 -0600, Tim Wescott Gave us:
The pins are rounded and they nest and the sockets have larger diameter entry openings.
Mil-max is the perfect choice.. A bit pricier but worth it and serviceability remains intact with hard (soldered) integration between the elements.
We have never had alignment issues with the 7305 sockets but we did make sure our board mounting holes were big enough to allow for some play. I've never used the 8806. Pad design could be a little tricky to get them to reflow correctly.
If the boards are self supporting you might look at a FPC/FFC type ribon connector. I've used those before to run from a main board to a small display board. Not sure you can get your 200V though. How many pins are you thinking you will need? Vibration an issue? High volume cost issues?
I would just go with 2mm headers on one board and a RA socket on the other.
Other option might be a board edge connector and make your daughter card with edge pins. I've done that also.
That's an option and not as bad as it sounds but disassembly can be hard. Since it's rigid your case mounting can be trickier due to assembly variations. Just make sure and deal with it up front in the case design and mounting holes.
The price I got was 0.33 USD for a connector with 8 pins, but that was for quite low volume, so I guess it will be significant lower in high volumes
Cheers
Klaus
d
We use this a work for a high volume product. The angle PCB has copper pads , and the main board PCB has holes, so the angle PCB goes some millimeters into the main one. Biggest issue I know of is the support must be spring de coupled, so that aluminimum for example doesn't fight the expansioen coefic ient of FR4
Cheers
Klaus
On Sat, 16 Jan 2016 00:50:09 -0800 (PST), Klaus Kragelund Gave us:
All the more reason to NOT make it a hard coupled pair.
That looks good, how do you buy it. I don't see specific part nos for various choices, are purchases individually negotiated with the company in China? If so, what kind of minimums are involved?
100 mA is not much current. I can't remember the last time I saw a connector pin that was rated at less than 1 amp. As to the spacing, I think Tim can figure out how wide a spacing he needs to use. if the 200 volt rating sets the spacing of the pins, it won't save him space to use a connector with pins spacing smaller than the voltage rating he needs unless there are only a small number of 200 volt pins compared to the logic level pins. I forget how many he needs, but I think it was maybe 5 out of 20 ballpark.
This is one from Sumida:
This is the TeknoFlex:
Cheers
Klaus
Connector is not the problem. The copper traces need to be wide enough, perhaps 30 to 40 mils. and 50 to 60 mils spacings.
He said 8 200V 100mA power and 8 corresponding return pins. Just trying to be conservative and be prepared to hear from the layout guy, if not himself.
I am the layout guy in this case.
Interesting.
ISTM that's probably not something that your average assembly house is going to feel good about, unless the boards are in the same panel. Some
0.1 inch FFC connectors are rated to 500V.Cheers
Phil Hobbs
40 mil traces for 100 mA??? I don't think so. I've designed boards for current a bit higher than that and only needed 10 mil, IIRC. No, even that is very conservative. A calculator says 200 mA with half oz copper thickness and a 10 Deg C rise only needs 7 mil trace on an internal layer and less than 3 mil on a surface layer. So the current is virtually a noop.
On a surface layer a 50 mil spacing supports 250 volts, so traces shouldn't be a problem. Internal layers can provide this voltage with just 8 mil spacing. A 0.1" pin spacing for a through hole part, the solder side would need pads no more than 50 mil wide which is a bit small with 40 mil holes. Either an oblong pad could be used 60 mil long and 50 mil wide, or the pad can be omitted on every other pin.
Air is 1000V per mil and 500V doesn't even generate corona. So it is pretty easy math that any 100 mil spaced connector pin can handle 500V differential to any adjacent pin and plastic is usually more than 1000V per mil, so practically any plastic the industry uses to make a pin carrier/shroud from will handle it as well.
Until one gets up around 1000V, it doesn't require very big balls to connectorize less, and he isn't pushing power line level amperages either so it is a no brainer that damned near anything will work, but I would shy away from surface mount applications as they typically are less than
100 mil spacing and dirt and such can make for creepage pathways to be generated. Much less the tin whisker RoHS stupidity the european idiots foisted upon the world. Conformal coating could take care of that for an extended period though.RoHS was likely the most stupid, most costly, most unnecessary change over event to have ever happened to the entire electronics industry.
63/37 took decades to formulate and prove as the best, and the idiots just tossed it out the window because they are too stupid to know that metallic form lead is not a environmental toxicity hazard. The entire industry has suffered and even irreplaceable non-serviceable items like satellites have suffered as a result. Dumbest thing the euro dopes ever did to this day.
Some plastics might get leaky... particularly in a high humidity place... say Singapore. :^)
George H.
Have something to add? Share your thoughts — no account required.
Ask the community — no account required