Mounting 2 boards under a 90 degree angle.

[free_electron]

Here is a problem i'm trying to find a nice solution for.
I have 2 circuit boards that need mounting under a 90 degree angle.

There is a large base board and a frontpanel mounted under 90 degrees.
Electrical connection between the two is done using angled pinheaders.

the board needs to be removable. so soldering angled constructions is out of the question.
i was thinking of soldreing brass standoffs on the vertical pcb and coming with a screw through the bottom board upward into the standoff.

the problem is positioning the standoffs and soldering them.
i cannot have anything extend over the edges either. the carrier board slides into a rack system and i need every millimeter to make this thing fit.. so left and right of the front nothing can be mounted.

The frame around the displays is actually another pcb so i can put copper structures there on the backside to solder something on.

it needs to be mechanically sturdy as people will be pushing those buttons...

idea's welcome

[alm]

Would screwing a small L bracket to both PCBs work? Or would that take too much PCB real estate?

[gregallenwarner]

Trying to obtain your source of strength and rigidity at the joint between those two boards is the most difficult option. The reason is because that's where stress will be the greatest. Use a header and socket for joining the two, but secure the vertical board in place with the casing. Perhaps two notches cut in the sides of your case that the vertical board slides downward into and mates with the connector on the bottom horizontal board. That way, the notches in the case grip the sides of the board and prevent it from moving forwards or backwards when you put pressure on the buttons. This way, the forces are not directed toward the 90 degree joint between boards.

Edit: I must've missed the part where you said you can't have anything on the edges. I take back my suggestion.

[DavidDLC]

Split the signals in two sets of DB-9 connectors that can be screwed, put the connector as close to the edges as possible.


David.

[tszaboo]

Something like this:
http://www.richco-inc.com/products/card-guides-and-pullers/98-card-guides-pullers-catalog.html?PartNum=&country=RICHCOUSA&grpcode=B1230
They send sample.

[gregallenwarner]

What if you used a short flex cable to connect the signals from the vertical board to the horizontal board? That way, the electrical connections aren't rigid, and therefore wouldn't be subjected to the stresses of any subtle movement.

[walshms]

Your best bet seems to be the idea of angle brackets, firmly attached on either side.  Don't rely on the PCB alone, or especially the pin headers, to provide the strength you need.

In fact, I'd be sure that the angle brackets were as close to the pushbuttons as I could get them.  They don't have to be particularly large, but they do need to take the stress off the boards.

[Jon Chandler]

Just had an idea that might work.  If you put edge-card connectors on each end of the vertical board and matching connectors on the horizontal board, 2 small circuit boards could connect the boards and provide support.  The drawing below illustrates what I'm trying to describe.

[free_electron]

Trying to obtain your source of strength and rigidity at the joint between those two boards is the most difficult option. The reason is because that's where stress will be the greatest. Use a header and socket for joining the two, but secure the vertical board in place with the casing. Perhaps two notches cut in the sides of your case that the vertical board slides downward into and mates with the connector on the bottom horizontal board. That way, the notches in the case grip the sides of the board and prevent it from moving forwards or backwards when you put pressure on the buttons. This way, the forces are not directed toward the 90 degree joint between boards.

Edit: I must've missed the part where you said you can't have anything on the edges. I take back my suggestion.

there is no casing. the main board slides in a VME case 19 inches wide holding 8 of these boards. 4 cases per rack, 22 racks in a room... this is a massive burn-in tester...

[free_electron]

Would screwing a small L bracket to both PCBs work? Or would that take too much PCB real estate?

i tried brackets. when you push buttons the front wobbles. not sturdy enough

[free_electron]

Split the signals in two sets of DB-9 connectors that can be screwed, put the connector as close to the edges as possible.


David.

no room behind the board to put such a tall connector. plus the solder joints on the connectors take the stress not reliable.

[free_electron]

Something like this:
http://www.richco-inc.com/products/card-guides-and-pullers/98-card-guides-pullers-catalog.html?PartNum=&country=RICHCOUSA&grpcode=B1230
They send sample.

close ! it is indeed something like this i am looking for.

[free_electron]

Just had an idea that might work.  If you put edge-card connectors on each end of the vertical board and matching connectors on the horizontal board, 2 small circuit boards could connect the boards and provide support.  The drawing below illustrates what I'm trying to describe.

there is no room behind the front panel for constructions like that. the thickest element i can put there is that dual row header. there is exactly 5.6 mm opening from the back of the PCB to the next module in the assembly. and i can't mechanically link to the assembly behind...

[DavidDLC]

Split the signals in two sets of DB-9 connectors that can be screwed, put the connector as close to the edges as possible.


David.

no room behind the board to put such a tall connector. plus the solder joints on the connectors take the stress not reliable.

There are connectors that you screw to the boards and also screw between them, but anyhow there is no space.

[david77]

Can you fit some kind of 5,6mm spacer behind your front panel pcb that physically touches the assemply behind it? That would absorb the stress of people pushing the buttons.
Maybe a 5,6mm brass stand off soldered to the back of the front panel. There should be enough give in the pin header so that it doesn't stress the solder joints too much, I'd think.

[free_electron]

The other problem is that this frontpanel must be easily removable to reach the module behind it...

I think i have a solution. instead of trying to solder the standoffs i am going to use double standoffs screwed into the motherboard. these will be placed with a gap of exactly 1.62mm between them (board thickness) the board will slide between them vertically.

There is only pressure around the buttons so i will put double standoffs there

All the standoffs shown will be screwed into the motherboard. they are not soldered to the frontpanel.
I added a second angled pinheader that serves no electrical purposes but serves purely as sideways anchor.
The extra standoff next to absorbs any pressure in the middle of the panel

What is not shown are the screws with standoffs that bind the front pcb , through the displays , to the back pcb. So that is a rigid inflexible assembly.

to remove the frontpanel assembly , simply lift it upward out of the pin sockets.

[walshms]

If you're going to use tall standoffs like that, then you can eliminate the ones in front; they're not providing any support unless someone is going to pull on the front panel.

I think that's a pretty creative solution, though.  Nicely done.

[free_electron]

pulling  could accidentally happen ... so i will leave them in.
i need to see the actual display and take a look at how the backlight is formed. i may stick an additional post in the opening between the displays.
( these are displays with rgb backlight. illuminated text on black background. i switch them to blue during boot , green during run and red when a fault is detected in that channel )

[baljemmett]

If you're going to use tall standoffs like that, then you can eliminate the ones in front; they're not providing any support unless someone is going to pull on the front panel.

They might, however, help with alignment when slotting the front panel into place -- and also prevent it from flopping forward slightly if there's any give in the headers.  Not sure how likely that is with dual-row headers, though!

[DavidDLC]

Use round standoffs, otherwise you may need to align the flat face in all of them.

David.

[walshms]

pulling  could accidentally happen ... so i will leave them in.
i need to see the actual display and take a look at how the backlight is formed. i may stick an additional post in the opening between the displays.
( these are displays with rgb backlight. illuminated text on black background. i switch them to blue during boot , green during run and red when a fault is detected in that channel )

Now I'm back to what you said before -- this is going to go into a backplane, I think I took from what you said.  So in terms of pulling the board out, you will face the friction in the assembly and the connectors.  If this is something you think may be removed or reinserted regularly, you might want to consider a light aluminium backplate for the main board; then you can just put levers or hoops on that to make insertion and removal easy without stressing the panel.  Would also give you a potential thermal path, and make the whole thing more robust.

[free_electron]

The board has a mechanical reinforcement structure underneath it that prevents it from warping. This structure has threaded studs that fit in holes in the board. The structure extends in front of the board and doubles as a handle to pull the board out. So nobody will grab the display assembly and try to yank the board out.

[pickle9000]

Hex standoffs could allow you to have some adjustment (by rotating them), elongated mounting holes would do the same. If there is space a couple flat washers will help stiffen up the pcb.

[Magicmushroom666]

PEM make some nice solder in nuts etc which may help with mounting those:
http://www.pemnet.com/fastening_products/circuit_board.html

[ElectroIrradiator]

i tried brackets. when you push buttons the front wobbles. not sturdy enough

Then you are in trouble. Brackets are about as good as it gets, as by using them you would be limited by the mechanical rigidity of the PCB material to limit any flexing. Using fancy connectors or anything like that would hardly make things any better.

I'm speculating your problem is a lack of mechanical support at the top of the front panel assembly. If you could somehow arrange for that, then the torque applied to the angle bracket(s) would be hugely diminished, limiting bending and flexing.