@ChrisGammell,
Excellent idea, will do.
Testing MB85RS64V SPI FRAM with Bus Pirate. Very cool 64Kbit chip. It acts like flash memory. You can write to it and it retains the data between power off/on.
Oh yeah! FRAM is great. Itās finally starting to come down in price a little bit, it has been around a long time but the patent is up soon i think. I know TI started putting it into MSP430 chips.
I made a comment about this in the former build log ond posted the compared images.
(Perhaps this image will not be available to all)
Are these the settings you changed. If so do you know the old/new values?
OLD Pads Mask Clearance
Solder Mask Clearance 0.007874015748" = 0.2mm
Solder Mask Min Width 0
NEW
Solder Mask Clearance 0.002" = 0.0508mm
Solder Mask Min Width 0.004" = 0.1016mm
All of my older boards that worked fine
Solder Mask Clearance 0
Solder Mask Min Width 0
Here you go. In each photo, Left is the new and Right is the old. I havenāt done any soldering yet, but the mask looks the way I would expect now. Huge difference.
You can click on the images to get the full size original.
NOTE: On the Good board, you see a tiny reflection around the through holes. That isnāt an actually spot where the mask is missing. I had to use a flash where I am to help show the missing mask on the bad board.
Back
Front
I just noticed that the silkscreen is partially missing on the back of the bad board too. Thatās odd. I didnāt do anything to change that.
1 Like
If you think itās a manufacturing issue, be sure to let the OSHpark folks know. Theyāll at least make you another set of boards.
1 Like
Hi Brooke,
You can see by the screen shots, the only value I changed was the Solder Mask Clearance. Once I realised what my issue was, I went back on previous boards and saw that this value was also zero. I therefore set mine to zero and checked it carefully against the Gerbers and this worked out fine and has ever since.
The settings from the failed board
The settings from the corrected board
1 Like
I did report the issue 2 different times since it was happening on multiple boards and in both cases it was explained that my settings were the issue. At the time, I didnāt notice the silkscreen. And since changing my settings looks to have resolved the issue, I am hopeful just to move on.
Chris, do you have anyone you connect with to see if the KiCAD issue can be given a new look? It sounds like the current default isnāt a reasonable one and thus will likely cause issues for people new to creating PCBs. I am not sure how I would have handled it if the first PCBs I created from KiCAD cost me money and didnāt work, and I didnāt know how to figure out what was going on or who to talk to about how to fix it.
Hereās the issue I found: https://bugs.launchpad.net/kicad/+bug/1563744
1 Like
There are a couple developers over on the KiCad forums, might be a good idea to bring it up there. At the very least, it will give visibility to others as well, should the change take a long time.
I had time last night to work in the SensorBoard v2. I was able to laser cut the stencil, paste, place, reflow and visually inspect. Next stop will be to plug it in and test. During the fab downtime, I have been writing test applications for all the different sections. This should shorten my bring up time.
2 Likes
Thanks, Chris! Yes, we appreciate any issues being reported to support@oshpark.com, so we can work with the customer to resolve any problems. Our goal is always to produce perfect purple PCBs 
Thanks, Drew (drew@oshpark.com)
Sunday 6.4.2017
I tested each of the Arduino sketches with the SensorBoard v2 and they all worked. They covered both I2C temp sensors, the SPI FRAM, and the analog-read photoresistor ( values from about 20-1020 ).
The only chip I didnāt have a sketch written for yet was the SN74HC595 shift register. I am working on that sketch next.
Thought I would share a design mistake for the TeensyBreakout w/ SensorBoard v2. It turns out it is a good idea to keep in mind the physical design of your boards when you plan to stack them. I thought I had done pretty well, until I needed to press the reset button! Thatās right. It took some creative re-purposing of an ink pen by borrowing its refill to sneak in there and press the button. I am thinking the rev 3 SensorBoard needs to either have a hole that goes through every single āshieldā or bring a reset button up to at least the Teensybreakout.
1 Like
I wrote the test sketch for the SN74HC595 and it works. All sections of the SensorBoard v2 have tested as working. My next task will be to write some firmware that takes advantage of all 5 sections. Two temperature sensors, FRAM, photo resistor, and shift register with 8 LEDs.
3 Likes
I relaised this too - I will be exposing the ~RESET on the header in my next project.
Since the CE is general purpose the option would be either use one of its GPIO as the reset or to simply add an addition single Pin Header. The options of exposing it on the CE Header or, as you suggest, give some mechanical access to the Teency reset button do not cost anything extra.
Well, we can always flip the Teensy too for future versions of the teensy to CE header board.
Brooke, not sure if Iāve already asked but are you using NPTH for the mounting holes? It looks like it.
thanks.
1 Like
Hey @mikef,
Thatās correct, for the Teensy SensorBoard, I used Non-Plated Through Holes (NPTH). Why do you ask?
Thanks Brooke. Iām just curious about creating mounting holes. Iām not sure if I like the non plated look though.