09/10/2018 at 02:11 •
September 9, 2018: I bet the followers are shocked to see an update on this project. It's definitely not dead. In fact, it's been quite busy, but I didn't really know what I should say. I'll hit some highlights and say that you should expect some more updates in the future.
On the technical front, we've had several batches of flex circuit assemblies made at contract manufacturers, and they've been installed in footballs, car tires and bike tires. Everyone who has used them has been very impressed with the overall accuracy and ease of use. It's pretty easy to get people excited about the technology.
We did a pretty major overhaul on the firmware in the tag that will enable us to operate with iOS devices. When we first started, iOS had no NFC support. About a year later, they added NFC reads, but no NFC writes. Our original tag architecture needed the reader to write to the tag to request a measurement. However, for various reasons, we overhauled the firmware, and we should be able to support iOS sometime in the next couple months.
The app is mostly the same, but we now upload all the data to a web server. This allows users to monitor their sensor's readings in an internet-connected database. This is useful for compliance management (making sure your employees are actually checking the pressure), and fleet management (is the whole fleet of trucks, cars, footballs, rafts, etc. running well?).
From the business front, we made it into another accelerator, and made some good contacts. I have to say that I am done with accelerators though. There's so much emphasis on pitching and taking on investors. For those of us who have a more complicated or risky product (e.g. not just "Uber for hairstylists"), it can be impossible to make any real headway in the actual competition. The real benefit of the accelerators appears to be the introductions. If I could do it all over again, I would have leveraged those a bit harder. It can be hard when you have a real full time job though. And that was another thing that floored me about the accelerators. We'd look around and realize that the various startups were everyone else's full time job, where it was just a hobby or pet project for us. We're not to the point yet where we could quit our real jobs to do this full time.
Which brings me to the final part of the update. We are running a beta trial right now with fat tire bike riders as part of a test run of a business to consumer model where we'd sell sensors to bicyclists to monitor their pressure, starting with the fat tire bikers, who run low pressure at varying temperatures on some pretty expensive tubeless tires. It's easy to get the sensor in, and it works well. I think we'll learn some interesting stuff from this beta and we'll step and repeat from there. It continues to be exciting to work on this project, and I expect it will continue to be for some time.
07/25/2016 at 18:45 •
July 25, 2016: First off, I finally got the design back to where I wanted it to be to stock it on Tindie, so there's a handful in stock. I plan to build more soon, because those first ones are probably going to sell out quickly.
Second, I was able to buy inexpensive flexible Kapton circuits from Itead, and it looks like they did a great job building them. I only built one so far, and it works great. I will be building more soon for mounting in tires and balls.
Speaking of tires and balls, I tried two different types of tags in the first tires and balls. When I got the samples back, one type of tag had died due to the vulcanization process. In both cases, it appeared as though the "gold" pads on the Kapton had corroded. I suspect what actually happened was that the gold plating didn't completely cover the copper, and the copper corroded due to the heat and bloomed through the gold, breaking the trace. These traces were the ones that go to the antenna, so when they broke, the sensor became unresponsive. The type of tag that survived was not a sensor tag - it just served a URL. The fact that it survived is good because it tells me that sensors can survive the vulcanization process, but it's bad because these tags aren't sensors, so their survival doesn't tell me as much as it might have if they were sensors.
So now the most recent version of sensors (that still have a rigid board) are getting ready for mounting into footballs. I just ordered a solder stencil for the flex circuit array so I can make those 8 at a time. It should be here within a week or so and I'll knock out a couple dozen of the flexible circuits for more mounting into tires and balls.
Hopefully I'll be having these produced at a contract manufacturer instead of hand assembling them. It will take a fairly big commitment on my part to ramp up production though. I'm not sure I can sell enough to pay for that. I'm working to do some customer discovery to understand my value proposition and so on to determine if this is something that I can sell in any sort of quantity. I'll keep you posted.
06/03/2016 at 13:37 •
June 3, 2016: Wow, a lot of things have been going on in the past month. First, we were selected as finalists in the MassChallenge startup accelerator program in Boston. This accelerator provides office space, mentorship, in-kind support, and the ability to compete for zero-equity funding at the end of the program. We're very honored to have been selected, and the judge's feedback has given us a slightly new focus. The feedback was extremely positive. I think they're as excited to have us in the program as we are. The only thing we have to figure out is how we're going to get to Boston, since we're in Minneapolis.
The second big thing is that we have collaborated with a partner in China to put prototype sensors into bike tires and footballs. We expect to have these back in hand in the next few weeks. It is pretty difficult to seal the large holes you need to make in an inflatable in order to get the sensor installed inside. By installing them before the tire or ball is vulcanized, we will never have to worry about sealing these huge holes.
Next, we paid the Android developer fee and we're working on getting an app up on the store. Right now we have it installing to our phones from the sandbox, but nobody else can see it yet. We hope to change that in the not too distant future.
We have to make more sensors. There have been some patient customers on Tindie that are waiting for us to make a new batch. This will probably happen in June. Those sensors will work with the app that will be on the store at that time.
Finally, we're planning to shift our near term focus a bit and focus on balls instead of tires. The MassChallenge judges were excited about our work partially because it aligns very well with the whole deflategate thing going on. Since they're in Boston, I'm sure they've heard even more than we have about the Patriots and Tom Brady as it pertains to air pressure. They were excited that we could measure the air pressure in a football without having to use a leaky air gauge and stick a needle into the football. So we plan to shift temporarily to balls from tires. We will come back to tires after we get a good ball design.
We don't yet know how we're going to try to market this. We were thinking of doing a Kickstarter again, since we were successful with the BooSTick, but there may be a better way. Hopefully MassChallenge can help us in that area.
04/26/2016 at 13:21 •
April 26, 2016: Since the unrelated flat that I got in the front tire shook the original sensor loose from its attachment, I've been trying to test with a back tire with little success. Last night I had a moderate success that still ended in failure. The issue is that I have been testing flexible sensors for durability. Last night I put a flexible sensor in the original tube. Basically, the tube has a large hole in it as pictured in the gallery, and I put the flexible sensor over this hole and covered it with a large rubber patch as I had before. This effectively put the sensor inside the tube, and no longer was it pinched between the tube and tire. I figured this might help with the sensor cracking issue I was seeing previously.
I checked the sensor periodically, and it worked happily up to about 5 miles. Then the tire went flat. I wasn't that surprised because I hadn't been able to get the patch to seal properly. Unfortunately, the brief time I was riding on the flat tire without realizing it killed the sensor, so I won't know how long it would have lasted. It appears as though a single component on the back wore through the patch and created a pretty major leak.
I have one more flexible patch I can try - I think this time I would plan to cut a slit in the tube and try to adhere the patch to the inside of the tube and then patch over the slit. I would remove the single (and unnecessary) component from the back of the sensor so it wouldn't wear through. One issue I'd have to overcome is that tubes are coated on the inside with a powder that makes it hard to stick anything to it. It's there so the tube interior doesn't stick to itself.
Otherwise, I need to wait for the sensors I shipped to various Chinese tire companies to reach them. It's been over a month so far and none of my three shipments have made it. Maybe I shouldn't have tried to save money by shipping them without tracking!
04/19/2016 at 13:56 •
April 19, 2016: I have a lot of new updates on this project. Strap in, because this is probably going to be a long one.
First, I got a flat in the tire that had the sensor in it after about 200 miles of riding. The sensor was in my front tire, which generally gets flats less frequently. I suspected the sensor may have caused the flat, but I was pleased to find that a tiny rock had punctured my tire. I probably need a new front tire. I fixed the tube and the sensor was still functional. However, after one ride, the sensor was not working anymore! I soon discovered that it was because the sensor had come loose from the tire because of all the jostling around when the tire was flat. That bent the patch out of shape a bit and it fell loose inside the tube. I pulled the sensor out and it was still fully functional. This is a hard sensor to kill! I have not mounted it back in the tire yet because of the other updates below.
The second thing I tried to do was to mount three different sensors in my tire to test for durability. However (and I believe this is key), I mounted the sensors between the tube and the tire. This means the sensors were constantly being pushed on by ~100 psi, and they were probably also constantly rubbing between the tire and tube. Not one of the three sensors lasted for the ~15 mile test ride. Two of them cracked and one of them had the wire antenna wear completely through (all pictured below). At first I convinced myself that the units that cracked did so because I had put epoxy over the components to protect them from the forces of the tire. I figured the epoxy made it too rigid and gave the sensor an opportunity to crack rather than flex. I tried another of the TI Kapton sensors without epoxy and it cracked as well, so I knew it wasn't the epoxy. The sensor that wore the antenna through gave me a few more clues. In that picture, you can see a faint vertical line that goes right through the failed point on the antenna. That line corresponds to a parting line on the tire tube, that had some extra rubber flash. The flash appears to have worn through the antenna, even though it was protected by a cover layer not present in the picture. This fact and the fact that the successful tag had a very similar Kapton antenna tell me that the real issue here is that the tags are between the tire and the tube, not that the flexible design in general is flawed. More updates below the pictures.
The final update is that I have sent some of the very flexible designs to a few tire manufacturers for test mounting inside tubes. The sensors will have to withstand ~200 C temperatures for the vulcanization process, but if they do, it will answer tons of questions for me. Even if they don't survive, I plan to do some beta testing with these tubes for a few hundred miles and then cut them open to see how the sensor substrate physically fared. However, the tire manufacturers are in China, and none of them have received the sensors yet because I tried to send them via USPS to save money. Seriously, it's $1.20 to send an envelope to China through USPS, but jumps to $60 when you add tracking. Problem is, nobody has received the sensors yet and I have absolutely no idea where they are. They will probably show back up in my mailbox in a month or so as undeliverable for some reason.
So that's where I am. I am just coming off the successful Kickstarter for the BooSTick: https://hackaday.io/project/7050-boostick-small-aa-voltage-booster, so I am starting to look at this one a lot more to figure out how to proceed. I can tell you it is pretty cool to have that sensor in your tire, and I have used it quite a bit more than I would have guessed.
03/12/2016 at 20:24 •
March 12, 2016: TI got wind of my project and was kind enough to send me some flexible patches that they had made. I finally got around to hand wiring a pressure sensor and voltage regulator onto one of them, and I'm pretty proud of my work (pictures below). I did this one quite a bit different. There's no differential amplifier, so my signal is very tiny AND it's biased toward the high rail, so I can't use the gain stages inside the IC. The pressure signal is measurable though. I don't think I would do away with the op amp if it weren't such a pain to hand wire in. I also used a capless voltage regulator to provide me with a more predictable voltage than the RF430 can. This way, I don't have to spend ADC time (or more hand wiring) measuring the pressure bridge's reference voltage. This one is going to get the components epoxied and it's going into my BACK tire. It's about to have a hard life. We'll see how it turns out.
On a related note, the original sensor is still going strong after 90 miles in the front tire. Still no leaks, still no problems.
02/28/2016 at 20:24 •
February 28, 2016: The patch has remained in my tire all winter, and the tire has held at least some air all winter. The weather was nice yesterday so I took a 15 mile ride and the patch is still holding up great. After the ride I was able to see that my pressure was 83 psi. Everything has gone very well with this first design. I am still trying to figure out what to do with this, since the flexible patch would cost so much to make.
02/19/2016 at 15:03 •
February 19, 2016: I created a flexible patch design that is suitable for lower pressures, and had some prototypes built. These prototypes consist of a Kapton flexible circuit on a PET carrier. The antenna is wound with magnet wire. These work very nicely and they give me a great deal of confidence that I could have higher pressure units created for a low cost and still achieve excellent pressure performance. I would like to make these new patches in higher volume, but the quantity order is too high. I have thought of turning to crowdfunding, but I'd need to know first that people were interested. I posted the original design on Tindie (https://www.tindie.com/products/CaptMcAllister/rfid-pressure-sensor/) for people to sign up for a waitlist for the original revision. If you're interested, please head over and sign up. If I get enough interest, I will look to crowdfund the flexible version.
12/03/2015 at 03:08 •
December 2, 2015: Files have been uploaded to Hackaday project page. There are native EAGLE files for the schematic and board, a CSV file for the BOM, and Gerber files for the PCB.
11/15/2015 at 15:58 •
November 15, 2015: I road tested the tire yesterday. A short 12 mile ride along my typical route - I didn't go easy on the tire. Just normal riding conditions.
The first thing I (didn't) notice was the ride. There was no difference in the ride quality. Some have asked me if the huge patch was going to create a wobble or a bounce in the tire. I didn't notice that at all. I think 100 psi forced the patch smooth up against the inside of the tire. The tire itself also has a lot more structural integrity than the tube.
Second, even with pressure on it, I noticed no leaking. The tire came back with the same pressure it went out with.
Finally, and most importantly, the pressure sensor worked at the end of the ride. The RFID still reported the pressure very accurately at the end of the ride. It was a very successful road test. I'm running out of weather to do any more testing this year, but so far it's a good sign. I think I can make it much more durable too, so that will be a focus over the winter.