May Progress

A lot got done this month!

Workshop Progress

This month I set about reworking the mud room in my house. It was originally used to store inventory from my mom's ecommerce shop, but going forward I want it to be the ThinkMods workshop, where all of my manufacturing equipment is located, components and inventory are stored, and a full desk exists for designing new mods. The full detailed progress of that is being shared as it happens over in the #progress channel of the ThinkMods Discord server.

To briefly summarize for this post, I've cleared out everything that needs to go, I've built a new shelf for housing the 3d printers, and converted its top rack into a dual-printer enclosure. The printer is just waiting on a wired internet connection before it's fully functional. The necessary equipment for that, along with a second 3d printer, arrive tomorrow (5/14), so I'll be showing that off in Discord when it happens.

Unfortunately, while I know exactly what desk+cabinet I wanted to buy for the second half of the workshop, IKEA is completely shut down in the US due to COVID. I'm going to wait until one of the stores near me reopens for online orders, but if that takes too long I'll just set up a folding table in the meantime.

Chassis Development

I have prototyped a large number of different designs over the past month, including ones using the material that I plan to use for the final revision, black PETG. Some of these have been coming out great, however one of the key problems is print time. I am going to be sending both 34 and 54 variants with every single order; the PCB will only be installed in the chassis you selected when ordering, but you can easily unscrew it and move it over to the other chassis later down the line for use in a different system. This is unfortunately the cleanest way I was able to design to solve that problem.

That means that print time is measured as how long it takes to print a pair of chassis; on a single printer, the estimated time to print 7 pairs (the maximum number I can fit on the print bed) was around 20 hours. That's honestly pretty lackluster; I would be running the printer nearly 24/7, and it would be running while I slept, which is really not great in terms of safety. This, combined with the fact that the original chassis didn't look great inside a T430 (that I had access to thanks to my brother sheltering with me from COVID), told me I needed a better solution, so I set about redesigning the chassis to be easier to print.

For comparison, this is what the chassis looked like beforehand:

And here's what the redesign looks like now:

It still needs some work; for instance, the LED cutout is just a placeholder and needs more closely reflect the original design, and there aren't yet holes for the screws/mounting nut. But overall, this design is simpler, but looks far better on the outside of the ThinkPad where it really matters, and is far easier to print. While 7 pairs of the original would take 20 hours, 7 pairs of this version take just 10.

With 2 printers, that means I can print at least 14 pairs a day. I would be able to print every chassis I need to complete every order I have so far in a little under a month, which I would say is definitely progress.

PCB Design

Unfortunately, before I even got to manufacturing the Prototype 2 boards, I realized a key issue; I made a rookie mistake and hooked up an LDO regulator to convert the 3.3V power rail to the necessary 5V for the USB chip. For those without electrical engineering experience, an LDO regulator very basically reduces voltage by shedding excess volts as heat. It can only reduce voltage; it isn't a buck converter. I basically set it up backwards, feeding its output voltage and expecting it to give me the necessary input voltage.

I can't believe I made such a rookie mistake at the time, but that doesn't really matter now. I've redesigned the PCB around a proper buck converter now, and I'll be placing the order for that in a day or two. I've also gone ahead and revisited other parts of the board; I've changed the LED shutoff from a jumper + header to an actual switch, I've swapped the LED mounting style to SMT rather than TPH, and I've fixed an issue where the LED woiuld be enabled at all times, regardless of activity - for some reason I connected it to the 3.3V power rail instead of the SSD's LED line, because I guess rookie mistakes are my forte. I've also decided to switch to a different mounting nut for the SSD, which necessitates a different hole size on the PCB.

All of this is to say that Prototype 2 isn't really going to work, and I'm going to need a Prototype 3, which will be the design that I ship to testers, partially because of the changes needed and partially because I still haven't received the USB chips from my supplier; COVID has really taken a toll on shipping.

Sticker Orders

As a quick update on this, I mentioned last update how I would be sending these out as soon as I was able to get stamps. Unfortunately it took a long time for USPS to reopen its store; fortunately, I placed that order and they should be here any day now. I've already got envelopes, so I'll be shipping those orders as soon as they get here.

Looking Forward

Needless to say, due to both COVID and a couple of rookie mistakes on the engineering side (though mostly the former), I am in no way going to meet the May-June shipping window given in the original campaign. I could probably start shipping in June if I didn't test any further, but there's a number of issues that I really want to iron out before I ship my first order. I want to make sure the USB chip's idle power consumption isn't too major (online estimates are 400mW, which is too much IMO). I want to make sure everything truly fits together. I also want to test power consumption, heat, and performance under a number of different situations, in a number of different ThinkPads.

Due to all of the aforementioned issues, my best-case estimate at this point for shipping final orders to backers is (very roughly) late July. There are a number of factors at play here; it takes 30 days to get any order out of China, which means that each 'cycle' of prototyping has a 30-day lead time. I've placed orders for all of my changed components, so a little less than 30 days from now I can actually start testing them. At that point, if I place an order immediately for the quantities needed for final production, it will be another 30 days before I receive those and can possibly start manufacturing.

In the meantime, I am going to try shipping out a few Prototype 2's to testers without the USB chip or LEDs, just to make sure the core functionality works in most systems.

I know it's a frustrating norm to have delays in an Indiegogo campaign, but I am going to do my absolute best to make sure it's worth the wait.