I’m certainly still finding a footing with these projects and making videos. While I shot a bunch of footage, no single project has come together in two weeks. Unfortunately publish or perish is real.
However I figured this would be a good time to update people on what I’m doing.
Maelstrom Controller
Since stopping the ads and the traffic from Ray’s week in review has effectively stopped I’m left with a number of pre-pre-orders. I was behind for a bit responding to questions but I am caught up. I do note that if something said “UK please” or “EU please” I did not email back. Instead I put a FAQ up. However if someone said “will you ship to the UK/EU/etc” I did reply. The explanation was a bit canned / copy pasted but I did get around 50 of these.
The big thing I revealed in the video is that I haven’t made up my mind. The reason has to do with Certification quotes. Essentially when you’re small they don’t really respond to you or you end up in a phone system from the underworld. While pre-order numbers are more than I expected, I show the math for 1000 units. I allude to an alternative method to launch. There are actually several. However they fall into the category of closer to opensource or kickstarter. More to come shortly.
I finally got the CSA information back after multiple attempts. I still have nothing from UL yet, and I’ve yet to try ETL. However, certifying outside of Canada / US is a more complicated manner and I’m trying to find support for something that could do all. Considering that the requests are split all over the world, this might be the death of it. No single market might be able to support the idea.
In more positive news, I have found that there are IEC (like the desktop computer sockets) that are reverse style that can go to receptacles! This is awesome. It would add costs to the hardware but could be worth it since there would be only one SKU to manufacture and certify. It also allows a fuse holder with replaceable fuse.
Strain Gage Board
I’ve been slowly redesigning this so that people can actually buy and use it. The main problem with the design was that while a nordic nrf51/52 can have any pin be an interrupt, a real arduino based on the atmega series cannot. In fact it varies from board to board. So in order to ensure compatibility I have to check all the boards I want to support.
It’s almost done, I’m just debating on some 1st order classic differential filtering. For a lot of applications it’s not needed, but for some it helps a bit, and some very needed.
DI2 Hacking
I’ve gotten a decent library together that is much much much simpler than what is supplied with the nrf52. I’ve mentioned it before but I will say it again: I hate how the Nordic guys lay out their SDK. It’s just overly complicated and convoluted. Why are there upwards of 40 files referencing each other to do what I literally did in one. I’m not joking! Sure it is more versatile as having individual page encodes and decodes help, but it’s a mess of pointers everywhere and hard to follow. It’s layers on layers on layers so by the time you figure out what the command actually does you’re 8 files deep.
I have short and long presses working reliably but the double pressing has kicked my butt for days. I get so close but then it all falls apart. Effectively I can get double press working once and then nothing, or it works but throws a single press afterwards no matter what. I’m starting to understand what it is, but it’s annoying. All this for making buttons more configurable than what the ANT+ profiles allow. And other reasons…..
Future Stuff
Main things I’m working on while waiting on all the UL and ETL information are small. The small USB DI2 charger inspired me to start work on a EPS version. More portable! Also part of the reason is also cell imbalance, though I have to test this. None of the electronic groupset manufacturers seem to be doing cell balancing. It’s a shame. I’ve talked to a few shops. Some of these batteries that won’t hold a charge are likely just imbalanced. The cells end up with different effective capacities and that will eventually result in imbalances and what seems like premature death. They all have great protection but all setup for 1 – 1.5 hour charges. That’s actually pretty quick. However they stop charging in one of a few reasons. Battery voltage is reached and current dropped, a timer runs out, OR a single cell went a *smidge* over 4.2v. That means if one cell starts to overcharge then the lights go off and you think the battery is charged. It does not resume it does not try again, it just stops.
Anecdotally I’ve heard slower charging keeps things more balanced. There are numerous videos where people say to plug in their Shimano charger to a computer to charge it. The reason they don’t explain is a lot of ports won’t support PD so they will limit to 500ma usually (technically should be 100ma, but… ya… reasons). On the boosted to 8.4v Shimano side this will effectively limit you to about half of the rated current. So at 1A in it’s 0.4a out, so a computer would limit its output to 0.2A or 200ma. The battery is around 600mah taking over 3 hours to charge with losses.
So the same could have positive impact on EPS batteries, but more importantly a tiny little pocket module about 1/2 the size of the Shimano one to charge my EPS system would be so convenient when travelling.
General EPS Hacks?
With the talk of EPS I’m also debating some “how to hacks” about bikes. Starting with my Wilier and how I put the interface on the seat and mounted an internal only electronic groupset to a 1980’s frame that only has external routing. A friend is doing a new build and transplanting his EPS and will need some hacking too. No holes for FD cable? No problem! We got that.