Project Logs

Collapse logs

• Some last pics

  Ole Andreas Utstumo • 02/15/2020 at 20:41 • 0 comments

  
  

  
  

• Shelvin'

  Ole Andreas Utstumo • 02/15/2020 at 19:45 • 0 comments

  I made sure the PLL cap was a NP0 type and added 10µF decoupling to the analog supply of the MAX30003 with no improvement. This project has been dragging on for too long, so it's time to cut the losses and move on. A little summary won't hurt, though:

  
  What works:

  • Sampling real time data to the microcontroller's RAM and, upon button
    push, store the entire RAM buffer to a file on the external flash file system
  • Using the STM32 with external flash as an USB mass storage device to access saved files (or any files!)
  • Runtime: approx a week on a single CR2032 cell
  • Silicone casing is smooth!

  What doesn't work:

  • The readings from the MAX30003 swing from rail to rail and makes no sense
  • The smoke escapes from the 3.3V LDO when there's no USB power... Seems like reverse conditions don't make these things happy!
  • The MCU is unsuccessful in recovering fully from deepest STOP mode

  Lessons learnt:

  • Tiny components in packed spaces are a bitch to patch and rework
  • Don't do hard to hand solder SMD components on the backside of PCBs in hobby projects with hobby tools
  • Include reset pin on debug connector if you plan on using sleep modes...
  • Segger has some truly excellent debugging software

  ... and lastly:
  

  • Ease down on scope on simple hobby projects, don't introduce too much new or tricky stuff or it's gonna be hard to keep up momentum, and as a result, motivation will suffer.

• Nonsensical data and an unlockable PLL

  Ole Andreas Utstumo • 02/10/2020 at 00:10 • 0 comments

  There is this slightly tiny issue where all the logged data makes no sense. A bit unfortunate for a data logger, yes, and one that's a bit harder to debug when dealing with a integrated circuit and not discrete components as previuously.
  

  After clearing away the metadata the logged data looks like this, electrodes attached or not:

  Just something that looks like a saturated signal going from rail to rail in a random fashion. I've verified the SPI transmissions with a logic analyzer so there's nothing wrong with the way data is handled, it's all fine there.
  
  The 1.8V supply is stable with only 3.3mV of digital ripple:
  
  Looking at the PCB design I cannot see any impedance or noise issues with the ground plane either. In addition, the MCU is basically sleeping when the front end is sampling data from the amplifiers..
  
  

  All registers are at default value, ECG channel is ON, sync register is cleared, the only thing reported in the logged metadata is FIFO END (but no overwritten values). Thinking it was some impedance or cross talk issue at the inputs I disconnected the inputs internally using the OPENP and OPENN bits in the CNFG_EMUX register, but with the same results..
  

  Reading from the MAX30003's status register might give us a slight clue, though: The PLL lock bit never clears from it's position, meaning the device never reaches a stable clock.
  

  
  

  The internal clock of the MAX30003 is generated from the incoming 32.768kHz clock from the STM32's crystal driven LSE clock:
  
  

  We can check the incoming clock for jitter:
  
  

  A closer look at one period's worth of jitter:
  So we have a period of ~30.5us +/- 0.2% which doesn't seem like too much, but I'm afraid my knowledge falls short here..
  
  A cause for the PLL issue might be the PLL capacitor which may or may not be a C0G / NP0.. If that's not it, I'm out of ideas, and to be honest, I'll also be out of motivation to dig further into the issue. It's been a cool little project, but to be honest there's nothing too interesting about it to warrant more hours sunk into it.. I'll order some new caps and give it one last shot before I shelve the project and move on.
  

• Moldin' 2

  Ole Andreas Utstumo • 01/26/2020 at 22:55 • 0 comments

  Didn't mix enough silicone and ended up with only half an enclosure.. Feels pretty damn fascinating, though.
  

  Letting gravity to its work on the silicone through that narrow opening in the mould proved to be impossible, so I resorted to the cordless drill to modify the resin printed mould a little...

  Whops. Brittle stuff, this is. Fret not, though, because my colleague Borut modified the cad files and my other colleague Börje printed a new mould with his cheap resin printer. The prints are slightly deformed but the level of detail and sharpness is on a whole other level from filament prints.
  

  Vaseline is again used to keep the mould from leaking.
  

  The negative filler..
  

  I used a pipette to push the silicone mix into the mould. Not ideal since air bubbles are easily introduced into the mix, but it'll work.
  

  
  

  Silicone doesn't take long to become unworkable.. I tried attaching a vibration motor to shake the bubbles out, but it didn't do much. A vacuum chamber would have given some solid results.

  Some rough edges, but it will be functional.
  

  Snug fit! As a bonus, the silicone works very well with the indicator LED.

  There will be a crack in the casing where the PCB slides in, though.
  

  
  
  

• Moldin'

  Ole Andreas Utstumo • 12/04/2019 at 23:41 • 0 comments

  Ok, no measurements have been done to verify the functionality, but that doesn't mean we cannot be allowed to get ahead of ourselves... Discussing the casing with my great colleague Borut we had some ideas on plastic casings and then soft casings made out of silicone. It turns out, of course it does, that there is a kind of silicone resin to be bought that is "medical grade". According to Wikipedia:
  

  Medical grade silicones are silicones tested for biocompatibility and are appropriate to be used for medical applications. (...) Medical grade silicones are generally grouped into three categories: non implantable, short term implantable, and long-term implantable.

  The list of examples given are

  • Tubing
  • Drains
  • Feeding tubes
  • Catheters
  • Implants for long and short term use
  • Seals and gaskets
  • Syringe pistons
  • Scar Treatment Silicone Sheets (FDA Class 1 Medical Device) and gels.
  • Condoms
  • Menstrual cups
  • Sex toys
  • Non-Stick Containers
  • Respiratory masks
    

  Some fun examples here :-)

  So Borut came back the next day with some pretty well thought out ideas on how the casing should work, how the edges should bevel with a shallow angle so the casing can integrate better with the skin of the user, how the PCB can be slid into the elastic casing with the USB port still accessible, how the mold would look in detail with a insert as negative space for the PCB and how the leads for the logger would also need to be printed on the insert. Few days later and he's made CAD models for the mold!
  

  
  Brilliant! There's even a small indentation for the tactile button to be felt on the outside of the soft casing :-)
  
  And what do you know, some other great colleague by the name Börje recently purchased a cheap SLA printer and a ultrasonic cleaner, and presented us with an amazingly sharp print of the mold over the weekend.
  
  I opted for some "food and skin grade" silicone resin (shure A 15, ebay link) and got down to business with syringes free from the pharmacy.
  
  To be continued..
  

• Consumin'

  Ole Andreas Utstumo • 10/25/2019 at 13:47 • 0 comments

  Did a crude test on current consumption during runtime. During runtime, the MAX ECG circuit is the only thing running constantly, waking the MCU up from Stop mode every 32 samples so it can receive these via SPI and store them in the RAM buffer.

  180 ish uA on average. On a 240 mAh coin cell battery that should provide well over a week of runtime. All components are selected to operate below the "knee" of the battery discharge profile.
  

• Loggin'

  Ole Andreas Utstumo • 09/07/2019 at 20:14 • 0 comments

  Fatfs is working with the new flash chip, the USB MSC services are running, the ECG front end is now also running smoothly via SPI and we're logging data to files.
  

  Binary data straight from the front end obviously won't look good in a text editor...
  

  And if you're working with STM32 and haven't tried out SEGGER's excellent software packages, I strongly urge you to. They provide by far the most handy tools I've used yet for embedded debugging. The real time variable graphing tool is damn cool:
  
  

  
  

• Blinkin'.

  Ole Andreas Utstumo • 09/01/2019 at 17:51 • 0 comments

  All good so far.

• Mounted.

  Ole Andreas Utstumo • 09/01/2019 at 17:48 • 0 comments

  Three cards were mounted a while back. Hand mounting SMD components on a PCB backside is as usual a pain in the arse... I will hopefully never fall for that one again.
  

• Design files uploaded

  Ole Andreas Utstumo • 06/29/2019 at 17:35 • 2 comments

  ... And can now be found on the main project page. I hope I didn't leave anything important out of the project folder :-)
  

View all 57 project logs