Another week dominated by Museum in a Box second batch production. Again it wasn't supposed to, but at least my work has been on the software side rather than hardware—a change is as good as a rest right?

In the end it seems to have been third-party hardware that's let us down. This time a batch of dodgy SD cards giving a combination of failing after a couple of okay boots and corrupting some of the files on disk. Tracking down the problems when I'm a couple of hundred miles from the troublesome cards has been tricky, but has meant that some of our plans for better remote debugging capability have been brought forward and partly implemented.

The RFID code also got a good looking over, because given our experiences with the hardware in this batch they were in the frame as possible culprits for the problems we were seeing. I've now got some stress testing code that writes tags and verifies what's written many, many times. With zero failures after two thousand writes it was clear it wasn't the RFID side.

It's been nice to get reacquainted with the Museum in a Box code. The pandemic has meant that all my time on Museum in a Box has been on customer support and various box production tasks, and my muscle-memory for how it all fits together had faded a little. It feels good to have stretched that muscle and realise it hadn't atrophied as much as I'd feared. Could've done without the added pressure of debugging strange behaviour to an overdue deadline, mind...

There was more picking up skills I haven't used in a while outside the Musuem in a Box work too. This was more of a re-learning though. Following the Fusion360 CAD work I've been doing recently it was time to move onto the CAM side and machine some of the design. I have used the DoES Liverpool CNC router before, but only once and that was over a year ago. Luckily Russ, who gave me my induction initially, was up for giving me a refresher session. Things mostly went to plan, but we didn't have the right diameter bit in stock for some of it and the MDF test part didn't survive our improvising. I've got a new, correct size bit on order and I could pick Russ' brain on maybe a better technique to adopt with it too. He also pointed out how I could tweak the toolpaths to reduce the machining time by about a third. Tacit knowledge FTW!

The only other notable point from the week was a small but interesting conversation about my plans for epxanding the company. It's something I was pondering in my decade-and-a-half-notes and have arguably been on the brink of for a few years. The company has been a bit too precarious for it before, but is on a firmer footing now—as ever, a bit more work in the incoming pipeline would be better...—and this week an opportunity presented itself that isn't quite what I thought I needed but might be a better fit. That's crystallized things so this week I need to get into the details of our finances. There are more discussions to have, but it feels like a good step forwards.