[CPAP] Realtime CPAP Metrics Possible?

[shikkic]

It sounds like an excellent project, and good luck and happy hacking!

Sleeprider's recommendations are spot on (unsurprisingly).  Yes, the Resmed Air10 series is the way to go, definitely.  The basic CPAP/APAP is the "Airsense 10 Autoset" (only get an Autoset, not an Elite or CPAP or Airstart model; see the wiki article), and the higher tiers of bi-level and ASV are in the Aircurve 10 series rather than Airsense.

There's no recourse available via the wireless facilities that have been provided.  The Air10 series has a built-in cellular modem (it can be  Resmed hasn't provided an on/off switch either hardwarily or softwarily), but that never transmits anything in real-time.  The firmware uses it to spew only compliance data (not full sleep data) once per day, an hour after the end of the sleep session, to a Resmed server.  The same goes for the optional, plug-in cellular modem for Resmed's previous generation, the S9 series.  I assume the same is true of the Philips Respironics machines, except that the Dreamstation (successor to the REMstar series) has an optional 802.11 module as an alternative to the optional cellular module, and I don't know what that's capable of.  But I seriously doubt that it can be used for anything deep.  Reports here over the last couple of years have said that it's used by the PR firmware only to do the usual upload of minimal compliance data to a Philips server, no different from the cellular modem.

For the at-home machines rather than the sleep-lab ones, the SD-card slot, via a Flash Air card, is it.  If it can't be done with that, then it's time for an open-source APAP hardware-and-firmware design (for various reasons, not just real-time data) ... a huge project.  But necessary, in the long run.  At the moment, the CPAP manufacturers have a lock on everything

Thanks for the information on the wireless facilities Fats Drywaller, it's helps put my mind at ease to at least know their capabilities.I would love nothing more than an OSS APAP, I got close to thinking about what it would take and if it was the right market to try and raise donations or capital for it. However medical equipment is probably very rigorously regulated and annoying to comply with. I still maintain that if Mark Watkins decided to try and create a company or initiate around it I'd jump on that next day.

---

Of course, one way of looking at it is that starting from the consumer-electronics (or patient-electronics) devices is the wrong approach.  Why not equip your own home sleep lab instead?

That's just the beginning of what you'll need ...

It's certainly an option, albeit an expensive and unfortunate one. Although I do like the sort of pirate attitude of saying screw you to medical equipment companies and DIY. It could also  make sense to try and collaborate with a sleep center in a big hospital and see if I can get more information or even equipment through them in some sort of partnership, as a "research"? I doubt it though...

[shikkic]

If you have a talent with hex data and you're interested in tearing apart machine data to help build the next version of Sleepyhead, you'd be welcome to  join that effort as well.  Nothing seems to be happening on the development front right now, but when JediMark awakens again, it could get busy.  I strongly recommend you try out a new machine and retire your 60 series to backup duty.  As you already know I have a strong bias for the Resmed technology and algorithms, but I think I come by that for some justifiable reasons.

I'm not sure I have much experience with using tool to inspect hex data but I'm certainly interested in helping. I wonder what JediMark's next project will be, I'd be very much down to work with him. He needs to get an account on Github though, no idea why he is still on source forge haha.

---

shikkic,
FYI: You can download the SleepyHead source code that includes the parser files for different CPAP's data, as well as Contec 50 series oximeter
data.  .  . if you get real industrious.  

Hey Crimson! As Fats Drywaller pointed out I already have created a custom parser using the SleepyHead source code, I made it so that it could attempt to parse the bits on the disk as they were being written. Now if only that SD card data was updated more frequently...

---

Sounds awesome, I'd personally love a knob to turn left or right that would immediately change the pressure setting instead of going through the hassle of entering the clinicians menu each time i want to switch from full face to nasal, sometime i need to change it during the night and id love just to set the correct pressure on the fly.

I've had this CPAP machine for years and I just realized yesterday I could access the clinical panel and change the pressure, I'd love to figure out how to issue those commands on the machine directly. It'd be a CPAP distributors worse nightmare though haha.

[Fats Drywaller]

It could also  make sense to try and collaborate with a sleep center in a big hospital and see if I can get more information or even equipment through them in some sort of partnership, as a "research"? I doubt it though...

A medical degree would be one prerequisite, I imagine. Medicine, or "health care" as it's called these days, is a club, no doubt about it. A very exclusive club, with high barriers to entry.

Actually, I think you've pretty much explored the possibilities of working with the consumer-electronics machines and you know what's what. The SD-card slot is the best resource we patients (outsiders) have. I haven't tried to find the prices of the components of that Resmed S9 Tx sleep-lab system, but I'm sure the sticker shock would be extreme. Clinics have deep pockets when it comes to buying equipment. So a homebrewing approach of copying the data from the Air10's card every time it's updated, and working from that with the Sleepyhead source code as your reference, is probably the best you can do. But there's no guarantee that it'll be recent data, because the A10 has lots of flash RAM other than the SD card. I'd say get an A10, try it and see, and if it turns out to be no good for real-time or anything close to that, at least you can just use the machine, which has some advantages over the REMstar.

[shikkic]

One thought I had was that if it's not possible to get real time metrics from the machine itself, would it not be possible to use a flow meter of some kind and actually derive the metrics from the characteristics of air flow. I'm no Dr and certainly not an hardware engineer, but CPAPs seem fairly one dimensional in how they derive all of their metrics, it's only from one source, the air it pushes and the resistance it faces when you breathe out. Now I'm sure it's a bit more complicated to actually derive the algorithms that look at things like the volume of air in the lungs. However whats really awesome is that this might work for any CPAP because it relies exclusively on the flow rate of the air. So you could open source the hardware that is the generic flow meter and the algorithms designed and you've got a neat little system that can sit on top of any CPAP without being treated as medical equipment necessarily. Could even be a first step in an open source CPAP machine ya know?

---

It could also  make sense to try and collaborate with a sleep center in a big hospital and see if I can get more information or even equipment through them in some sort of partnership, as a "research"? I doubt it though...

A medical degree would be one prerequisite, I imagine.  Medicine, or "health care" as it's called these days, is a club, no doubt about it.  A very exclusive club, with high barriers to entry.

Actually, I think you've pretty much explored the possibilities of working with the consumer-electronics machines and you know what's what.  The SD-card slot is the best resource we patients (outsiders) have.  I haven't tried to find the prices of the components of that Resmed S9 Tx sleep-lab system, but I'm sure the sticker shock would be extreme.  Clinics have deep pockets when it comes to buying equipment.  So a homebrewing approach of copying the data from the Air10's card every time it's updated, and working from that with the Sleepyhead source code as your reference, is probably the best you can do.  But there's no guarantee that it'll be recent data, because the A10 has lots of flash RAM other than the SD card.  I'd say get an A10, try it and see, and if it turns out to be no good for real-time or anything close to that, at least you can just use the machine, which has some advantages over the REMstar.

I will be definitely doing that and happy to report back if I find anything useful.

[pholynyk]

>>>He needs to get an account on Github though, no idea why he is still on source forge haha.<<<

Actually, JediMark has been using GitLab, although he may be using a private server for his latest work (if any).

[shikkic]

>>>He needs to get an account on Github though, no idea why he is still on source forge haha.<<<

Actually, JediMark has been using GitLab, although he may be using a private server for his latest work (if any).

Ah that would explain it. Thanks for the heads up.

[Sleeprider]

Start here https://github.com/jedimark/sleepyhead

When Mark gets busy, he has been using Discord as the collaboration platform. Here is the non-developer side https://discord.gg/zBcqcdR If you want to get into the development, I'm sure there is room for you.

[Dog Slobber]

One way to get real-time access to the data would be to use a FlashAir WiFi SD Card. As the device is writing to the the SD Card your app can be interrogating the data through the API.

https://www.flashair-developers.com/en/documents/api/