These are intense weeks at the office with lots and lots of things to do; Lots of work but also lots of fun, this is a very exciting time! Unfortunately, we are now predicting a delay in the production. Read on to see why and how we also have an unexpected bonus added in the hardware.
Second revision has arrived
We have the second revision of the camera now and we’re right in the process of verifying it! As we wrote in the previous hardware update, we had some issues on the last PCB that prevented us from verifying all of the functionality, which was due to mis-documentation from one of the chip vendors. Basically it meant the SDRAM didn’t work, which effectively stopped booting any OS larger than what would fit in the internal, very small RAM or reading out photos. That is fixed now though and we’re resuming the verification work, but it did delay us a couple of weeks.
Sample photos coming soon!
We also got an update of the camera modules last week, and since the RAM is now working, we can read out images and expect to be able to show some representative mobile photos very soon now in the next hardware update.
GPS issues and predicted production delay
Everything in the PCB-hardware seems to be right on track now again but we do have one remaining issue regarding the GPS reception signal strength. We can see satellites using the GPS receiver, but with too weak a signal to get a position fix. Normally, antennas in small portable products do require some iterations of placement and PCB layout to get right, so this was as expected. But in combination with the SDRAM problem delay over Christmas, it does look like we have to shift forward the full production batches until the antenna is adjusted.
The consequence is that we are predicting that the final production will be delayed. The exact delay is difficult to estimate but since we believe that the GPS is an important part of the Memoto camera, we really want to get this right before we ship the first batches. This means the batches will not be produced in February as was originally planned but more probably in the end of March / beginning of April. Currently, the antenna manufacturer has our PCB’s and is measuring them to give us feedback for the necessary adjustments. We have the best people on top of this issue and we want to be completely open with the development process.
Bonus: added magnetometer / compass functionality
When we had to do a new revision of the camera, we did manage to insert one component that we didn’t have access to when we finalized the electronics design a few months ago but is surely interesting to have in the Memoto camera – a magnetometer / compass.
The compass is not verified together with our steel clip yet but our supplier thinks it will work fine. Together with the accelerometer, it will enable us to store the 3D-direction the camera was pointing at the time of each photo snap (with an accelerometer, we only get the inclination/pitch of the camera).
The magnetometer we are evaluating is the tiny chip on the lower left PCB.
To be clear, the addition of this sensor does not delay production. Instead, we are able to include it due to the fact that we had to wait for a second hardware revision.
What can a compass be used for?
A compass is basically another sensor that you can do cool stuff with. One of the ideas would be to allow you to create something like your own Google Street View since the app knows in what direction each photos is taken. We can also give you statistics of in what directions you are facing during the day and we are sure that some of you will come up with very cool applications of that. The compass data will be available through our API.
On the software side
On the software everything is coming together very nicely and we are actually a bit ahead of schedule. We will have both iPhone and Android apps ready, in their first versions, by the end of the month and the backend is already up and running and able to process all the photos we can throw at it.
Videos from the pre-production process at Note in Lund, Sweden
When we produced the latest PCB batch we managed to get some movies that illustrate the solder paste application and robot mounting process for your entertainment. These were taken by our good friends at Note Labs in Lund, our local, very helpful PCB assembly house who is doing prototype and pre-production PCB assemblies for us.
In the video below, the solder paste is applied to the bare PCBs using a solder mask stencil lasercut through very thin steel. The solder paste is a sticky, granulated version of normal tin solder, which makes the components that are placed on the PCB afterwards stick to the PCB. The solder mask stencil has openings for each pad on the PCB where a component pad should be attached. The pads can be very small, around 0.2 mm in diameter, so the entire process is very sensitive and actually the equipment itself adjusts the solder mask and PCB using computer vision, even if it looks a bit crude in the video.
In the next two videos, the PCBs with solder paste on all component pads, are attached to a slider in the component mounting machine, which can move the PCB in the y-direction while a component dispatcher head which moves in the x-direction is placing components using a vacuum system. The components are fetched from feeders attached to wheels or trays of components not visible in the video. One side of the PCB is mounted first, then the PCB’s are turned upside down (the solder paste is sticky and so the components already placed will not fall off) and the second side is mounted. After this, the mounted PCB’s are sent into an oven to be heated so the solder paste melts and creates a final permanent attachment.
Currently, our hardware engineers are testing these newly manufactured modules for proper functionality. We’ll keep you posted! And we love to hear from you, get in touch with us on http://memoto.com.