TPMS... DIY and Design Anatomy

[srt8-in-largo]

If I were designing my own TPMS, these are the useability features I'd want:

- System… must read pressure and temperature and have user selectable units for both
- System… must be able to obtain live readings when the bike is parked
- Display… must be weatherproof
- Display… must be backlit for night viewing
- Display… must be hard wired to the bike instead of being battery powered only
- Display… must show pressure and temperature by default and not require a button press to see this
- Sensors… must report frequently; at least once every 1 minute
- Sensors… must have a “fast pressure loss” mode that reports immediately when this happens
- Sensors… must have reasonably long battery life; at least two or three years… ideally 5 to 7 years

- Nice to have... low sensor battery indicator (maybe this should be a requirement)
- Nice to have... since the display is hardwired to the battery, battery voltage can be displayed
- Nice to have... battery option for the display instead of requiring hard-wiring


I don't like dash clutter so I'd see about getting a panel-mount display that installs in the panel near the shelter keyhole. This is an out of the way location but is at least somewhat visible at a glance. Since the panel isn't in the riders field of view, you could install a small but BRIGHT dash mounted LED to alert you if any pressure or temperature thresholds have been crossed.

[srt8-in-largo]

The OEM Goldwing TPMS is what I would call a simplified system that is typical of those made specifically for motorcycles. The OEM system is illustrated below and consists of three parts... two sensors/transmitters in the tires and a controller that monitors the transmitted signal and produces a low pressure warning when appropriate.

TPMS sensors have three function blocks: 1) a sensor that measures pressure and temperature, 2) an RF transmitter that sends the measured data to the controller, and 3) an LF receiver that accepts commands and issues them to the RF transmitter. A "simplified" system is one that doesn't have a built-in way to issue commands to the LF receiver in the sensors.

Why would you want to issue commands to the sensor? Read on...

The second picture below is an exploded view of the Goldwing sensor; this is typical of any other sensor on the market. The three function blocks mentioned above are implemented with electronics on a tiny PCB housed in an enclosure with a battery. Function 1, the sensor, is a tiny analog IC that continuously reads pressure and temperature. Function 2, the RF transmitter, is a programmable microcontroller that reads the analog pressure and temperature data, converts it to digital, and transmits it to the controller on a 315 or 433 MHz carrier. And function 3, the LF receiver, is an LC tank circuit tuned to pick up 125 kHz signals and send them to the microcontroller.

The OEM sensor uses built-in accelerometers to send motion data to the microcontroller. When the bike is parked and the microcontroller does not sense motion for 7 minutes, it will not transmit data and therefore you cannot read tire pressure after the bike has been parked. This is done to save battery life of the sensor. When you get on and ride faster than 9 mph for 20 seconds, the microcontroller senses this and begins transmitting data every 1 minute. This is a cheap usability feature IMO because the most important time to check pressures is in the morning BEFORE you ride.

If the bike had an LF command module it could be configured to detect a key on, and when it sees the key is on it can send a single transmit command to the sensor... and then allow the sensor to go back into sleep mode until the accelerometers take over. With this setup you could read pressure whenever you wanted and not really affect sensor battery life.

So where can you get an LF command module that will do this? That's the complicated part... do-able but complicated.

[srt8-in-largo]

After taking a good hard look into all of this... I've decided to move forward and make my own TPMS system.

I'm taking a slightly different route than what I pictured above and I'm going with this setup pictured below. The location of the Honda plaque is just big enough to be replaced with a 2-line 8-character display that is enough to display pressure and temps for both tires simultaneously... and this location is easier to glance at than the side panel. I've already found an LCD that matches the Honda OEM display and so it'll look pretty good, almost OEM. I still plan to wire a glaringly noticeable LED that will alert the rider of pressure warnings or rapid decompression events and have this mounted near the gauge cluster.

I've already prototyped some hardware for the pressure sensor using Freescale's FXTH87 TPMS chip; proto board is shown below. I'm currently getting the programming sorted and then will move onto developing the receiver/display module.

[Steve 0080]

Impressive..... we have a lot of talent on this board !!!!!

[Spanky]

sitting on a million bucks! great job! I can see a Hellcat in your future! now don't forget all those on the board that support and encourage you ( who support you enough to pay cost+ 10%)! remember make it easy to install! :-)

[srt8-in-largo]

Thanks guys!

The hardest part of the install will be pulling the shelter cover to get to the underside of the Honda plaque and routing the alert LED. I'm hoping to develop the low frequency (LF) communication so that riders can get a fresh reading just by turning the key to ACC. These LF communication modules will need to be mounted near the tires... and I haven't yet worked out how all that will come together.

I plan to use an in-tire sensor that will mount to the rim on the opposite side of the air valve. This creates a better tire balance and, more importantly, allows you to use tire sealer like Rideon and not worry about that stuff clogging the sensor. Like this picture from the Mobiletron manual.

[Phantom]

[srt8-in-largo]

Thanks Phantom!

I built a test chamber with a pressure gauge, thermocouple, and a schrader valve to let me test out the sensor board. PVC in general is NOT a good material to pressurize, especially the type that says "not for pressure", but it did ok to 40 psi. The sensors seem to be working fine.

The part of the system that I've been concerned with was the antenna circuitry; currently it has no filtering and a first pass design of the matching network. To test how good (or bad) it is I downloaded HDSDR and bought a $20 antenna to go with it. The video below shows the sensor transmitting at a distance of about 4ft from the antenna. From what I can see, I think Freescale designed a helluva RF transmitter on the chip! Even with this non-optimized proto-board signal strength is fantastic.

For test purposes I have the transmitter configured to send 16 consecutive packets... these are the pulses and beeps you see. Each packet contains pressure and temperature data measured by the chip. I plan to have the software setup where it'll take measurements every 0.5 or 1 second but only send the packets every 20 or 30 seconds in order to conserve battery power. I might even extend the transmission to 60 seconds or more... but I have yet to create a power budget. Like I said above, I want a sensor with long battery life.

Sending consecutive packets is done for redundancy; if one packet gets lost there are multiple others available and therefore the data monitoring can be consistent and continuous. When I look at the power budget and do some link testing I'll probably reduce the number of packets to 2 or 4 to, again, optimize battery life.

In addition to the above software configuration, I plan to create a comparison loop where if a significant pressure drop is detected the chip will transmit immediately and more frequently.

I've also been working on the receiver module. Next steps are to get the transmit-receive link operational and then work on incorporating the LCD display. Here's a picture from the vendor website; I think this is with the white backlight turned on.

[hiflyer]

Thanks Phantom!

I've also been working ......

srt8, I think that's an understatement!!! You are serious about this thing!! Get er done!!

[srt8-in-largo]

Haha thanks Flyer; I hope I can get it all together before long.

To illustrate the transmission options, I reduced the packet count to 4 in this video and maxed out the time between each of the four packets. I have the program compiler in debug mode and you can hear me clicking the keyboard to initiate each transmission; when the chip is programmed it'll do this automatically.