Replacing a nRF8001 radio in a Wahoo Kickr indoor Cycle trainer

[erikbrenn]

Hi,

I'm trying to fix a sensor board for a Wahoo Kickr cycle trainer.  The sensor board has various inputs common on a cycle trainer, like rotary encoder (ir) for speed measurements, wattage sensor, etc.
On the board there are two Nordic radios for communication with phone/pc, one NRF8001 for BLE and one NRF24AP2 with an ANT protocol stack. In addition there is a MSP uController and a couple of smaller chips I haven't been able to identify.



Problem
My phone is able to detect the trainer and pair with it, which indicates that the BLE comms are somewhat ok. But the trainer does not acknowledge any instructions from me/my phone when I set the wattage.
I have testet the sensorboard on another working trainer, same problem, hence it must be this sensor board. And since I'm not really using the ANT features I recon the bluetooth comms are broken.

However, if I push hard  on the pedals the radio comms seem to work, at least the app then shows some wattage, which I assume it receives from the trainer. Altough I don't know if it's correct value.
 
Visual inspection of boards shows now damage. Measured all caps and no shorts. I have also taken diode measurements of all test points and compared with a board from  working trainer, they're all ok.

There are no replacement boards to be found for this rather expensive equipment. I want to try to replace the Nordic radios, they are quite cheap. But according to datasheet https://docs.rs-online.com/7e86/0900766b81429d60.pdf 
these chips have both common RAM and NVRAM to hold configuration for oscillator setup etc. And unfortunately the NVRAM is write-once only!



What are your thoughts on this? If configuration is stored in NVRAM, is there any way of reading it in order to program a new chip?  Is it likely that the radios are configured on-the-fly by the micro during startup? Probably not... 

I was considering ordering a couple of NRF8001 chips to test a drop-in replacement. Perhaps also a devkit if its possible to read the NVRAM ....

Cheers,
Erik

ps: in case you wonder about a couple of resistors that looks out of placement near the radios, they're like that from factory,

[thm_w]

So it can pair with bluetooth, it can read data back, it can take watt settings but does not apply them.

Wouldn't this indicate there is something up with whatever output it is using to set the wattage/resistance level?
Are there any other commands you can send that set parameters, that would tell you that writing data is or is not working?

[erikbrenn]

Thanks a lot for input!

Good point.  There are two connectors on the board, a 4-pin and a larger white connector, of which several pins are NC.
Both cables go down the "body" of the trainer as shown in pic below. The wire I'm holding has a infrared sensor in the end, which reads the encoded wheel you see on the trainer side.


Assuming that you are right, and that the 4-wire connection seen in the photo connects to the "wattage controller" board down/inside the flywheel, I might try to replace the two chips that seems to interface the microcontroller to this connector.  However I have not been able to identify them.  One is marked "DACL 1230 SVS",  while the other  is a TI chip marked 1333A or I333A Here is a closeup:



Any ideas what they could be?

[up8051]

I333A = TI INA333A

[cozza]

Have you checked the optical encoder is working properly? When faultfinding stuff like this, it pays to check the things which are likely to fail, the nordic radio chipsets are not really in this category unless ESD / or a regulator fails. A bad or low emission led in one of these arrangements wouldn't be unheard of.

Look at the LED with a phone camera, and check the photo transistor output with a scope for good clean switching on each black white transition. (you may need to shield ambient light from it)

If it pairs on BT and takes a wattage setting but doesn't apply it, the first thing to check would be all the feedback inputs to the controller first.

[amyk]

DACL is MCP4541-503E/MF

[erikbrenn]

DACL is MCP4541-503E/MF

Thanks a lot, how the heck did you determine that?  I've been googling all over the net

[erikbrenn]

> Have you checked the optical encoder is working properly?

I did check the infrared diode using my handhone and I don't see any signal. I wasn't sure abouit this however because I have experienced remote control leds which are not visible in certain mobile cameras. Will ask a friend tomorrow to do same check on his device, his trainer works.  I will also scope the diode to look for pulsetrains, should have done that of course. 
I do believe however that at faulty encoder does not explain the lacking wattage feedback.  The infrared emitter/diode reading the encoder is connected directly to the microcontroller via some resistors.  However, the wattage sensor is located inside the flywheel and is connected to the 4-pin connector on the shown board.  This 4-pin connector carries 3.3V, GND and a differential signal that is fed to the INA333 opamp D-/D+ pins. The output of the opamp then goes to the mikro. 

After looking at the voltage rail coming to this radio/sensor board I noticed it was very unstable. My multimeter showed anything between 3.8 and 5.5 volts.  When I scoped it I saw a very bad sawtooth like voltage rail, oscillating between 3 and 6V.   
Because of this I took apart some more of the trainer to locate the main board with the regulator and stuff.



Looking at the mainboard I noticed the LDO (IFX25001) had no filter-cap on the output. I checked its datasheet and it should defintetly have a cap of on its output. I added a 100uF cap and the output is now dead stable 3.3V.  The chip next to the connector and the adjacent tranny is a i2c temp sensor. The mosfet in the middle is used to activate the 12V output on the top connector, not sure where that's going, not able to follow the cable inside the trainer. The main 12V input to the trainer is coming in at the left connector (external switched psu)

A bit shabby soldering, but...


Unfortunately, after putting everything together things still doesn't work. 

I fear that something has been damaged by the bad voltage from the main board, the sensitive INA333  opamp or even worse the mikro.

 I have downloaded the datasheets for the MSP kontroller, the digital potmeter the Opamp etc.  Will continue investigations of those chips as well as the infrared sensor tomorrow. I have also done significant tracing of what is connected to what and will draw up everything in a diagram later. There are lots of trainers like this out there but I can't find any repair articles or YT rapair videos for Wahoo trainers, so this might be useful for others. 

Thanks a lot for your feedback so far!

[amyk]

DACL is MCP4541-503E/MF

Thanks a lot, how the heck did you determine that?  I've been googling all over the net

Famous Russian site.

[fzabkar]

More URLs for smd codes ...

http://www.smdmark.com/en-US/search/code?id=dacl
http://www.smdmark.com/en-US/search/code?id=i333a
https://www.ti.com/packaging/docs/partlookup.tsp

[BobRyan]

Have you checked the signal into and out of the op amp? If this is malfunctioning it could explain why you get a signal when producing more power.

[erikbrenn]

Have you checked the signal into and out of the op amp? If this is malfunctioning it could explain why you get a signal when producing more power.

Thanks, that makes sense, will try it out.   Haven't had time this weekend, I have to move my scope and stuff near the unit. Will post an update later, possibly after christmas