What a difference 8 thou/mil makes

[ruffrecords]

I designed the backplane PCB for the EZTubeMixer nearly a decade ago. It has gone through several variations and upgrades but the core layout is the same. One thing that has always bothered me is when modules sliding in card guides mate with the backplane mounted connectors the module moves up and to one side very slightly. It seems to make no difference to the performance or the force required to insert or extract a module. When I designed the backplane there was some ambiguity over the exact positioning of the connectors with respect to the fixing holes that attach the motherboard to the Eurorack extrusion. The centre line of the connector is 8 thousandths of an inch offset from the centre of the the connector's fixing holes. From the best connector and Eurorack data I had at the time I determined that the centre line of the connector should be aligned to the motherboard fixing holes. The connector fixing holes are therefore 8 thou/mil offset from the motherboard fixing holes.

I recently had cause to update the motherboard design so I revisited this question but this time I decided the fixing hole of the connector should be aligned with the motherboard fixing holes. This means the connector centre line is now offset by 8 thou/mil. I have just finsihed assembling a new motherbaord to this design into a Eurorack case and fitted the card guides. Now wehn I slide in a module it just goes smoothly striagh into the motherboard connector; no jogging or sideways movement of any kind. What a difference 8 thou/mil makes.

Cheers

Ian

 

[JohnRoberts]

when in doubt do it right...

I really appreciate how modern CAD allows you to check for fit and clearances inside a 'puter...

back in the bad old days we would make multiple generations of prototypes to tweak fit.

JR

 

[pvision]

Was it you, JR, who described how console channel strips were made to the correct width but the thickness of paint applied made them bind when inserted?

That really made me think about measurements & clearances

 

[cyrano]

One of the best solutions I came across, was connectors mounted on a spring, so they could slide a bit when being inserted.

This was for transmitter modules for automotive and military use. Standard D-connectors never lasted very long if mounted rigidly. Of course, there are more gorillas employed in military and automotive than in studios...

D-connectors were a lot cheaper than the military grade connectors used before. While these lasted a lot longer, they'd fail in the end too. And these were so hefty they went over the weight limit for aviation.

 

[JohnRoberts]

Was it you, JR, who described how console channel strips were made to the correct width but the thickness of paint applied made them bind when inserted?

Permission to share another old Peavey anecdote...? No paint involved by vinyl coated sheet metal.

The last big mixer I made was the AMR production series. The max configuration was 36 input channels, 24 submaster channels, 4 S/R strips, and a master section. For production we hard tooled a forming tool to bend the strips to a very precise width......BUT when we were trying to assemble the first prototype a day before the introductory trade show, the strips that were manually bent in the factory were not nearly as precise. Factory tolerances were +/- a few thousandths, but it seems these prototype strips were mostly on the plus side. A few thou here and there times 64x made it an extremely tight fit to shoehorn the last strip into the fully loaded chassis.

That really made me think about measurements & clearances

and accumulation of tolerance errors.

If the errors were truly random +/- they would average out, but I had no such luck. I guess the factory metal knockers figured it's better to be over than under, because you can always tweak an oversize down a little by hitting it again on the brake, but undersize are harder to unbend (my speculation). I never really got to the bottom of why they didn't fit, because we pulled an all nighter getting that prototype together and then drive it to the trade show in Atlanta (me and one other engineer who stayed up with us drove the van). We arrived just in time to carry it into the dealer pre-show that was already going on. The dealers thought that was a planned grand entrance... it wasn't. To add to our fun, our hotel rooms were not ready so we had to wander around like zombies for a few more hours.

JR