I am thrilled to announce we have overcome the single greatest obstacle to our forward progress: custom thermocouple connectors. There are a few manufacturing consistency issues to work out, but the design is sound and the connectors work.

We previously discussed the terrifyingly bad samples that were rushed out of the factory and the proper samples that came soon after. The newer sample worked quite well except for the gap that the male thermocouple fits into; it was a bit too large. I decided to decrease the size of that gap and to change the insertion flanges from thirty degrees to forty-five degrees. We ordered the tooling modifications and the wait began.

connector test

Footprint spacing test

While waiting we designed a connector spacing test PCB to help design the perfect PCB footprint. The board is a series of contact pairs with a range of spacing between them in 0.5 mm steps. We then made a second test PCB to get the individual footprint just right. We nudged around the slot size and position in increments of 0.1 mm.

Screen Shot 2015-08-06 at 17.52.00

Detailed footprint test

We always use DirtyPCBs.com and recommend them highly. It’s run by Ian from Dangerous Prototypes and they’re the fastest and cheapest place around.


Externally constrained thermocouple connector

We tested two configurations: constraining the male connector on the outside of its contacts and constraining on the inside. The T400 will have the male connector constrained from the outside, but a future single channel device may find a junction temperature sensor directly between the contacts. Drop a glob of thermally conductive silicone on top and you have the best thermal contact you can get.

If you would like to use this connector in your own designs, you can download our Eagle libraries on github and purchase thermocouple contact in our store once we have the production batch. For now they are packaged loose, but we would like to get them on reels for pick and placing.

Yesterday we received samples from the updated tooling and were a little disappointed in the factory quality. I know they can do better, but we will have to babysit them. All the problems basically come down to manufacturing consistency. From the previous samples we know the factory can produce good work. These issues should be resolved in the next few days.

Here are the images we sent to the factory to explain what is wrong with the samples.

entry bend angleshort edge anglelong edge parallelbottom parallelshoft edge shift