How does ResMed determine the intentional leak rate?

[palerider]

Is that labeled "intentional leak" or just "leaks?" We know that unintentional leak is calculated from actual flow data. I'm discussing the value used for "intentional" leak rate.

https://www.youtube.com/watch?v=1eeDtO- ... e=youtu.be

at appx 23 seconds, I created an intentional leak by loosening my mask, at about 40 seconds, I tightened it again, and then at about 1:05 I took the mask off.

[palerider]

I spent a lot of time looking at pictures of the circuit boards and it definitely looks like a 3 phase brushless DC drive system, based on the number and type of driver chips you can see in the pictures.

I've got high res images of the front and back of an elite board, and a vpap adapt board, also, if you're interested.

[archangle]

I spent a lot of time looking at pictures of the circuit boards and it definitely looks like a 3 phase brushless DC drive system, based on the number and type of driver chips you can see in the pictures.

I've got high res images of the front and back of an elite board, and a vpap adapt board, also, if you're interested.

Yes, post them somewhere, please. I suspect a number of us would like to see them.

Is that labeled "intentional leak" or just "leaks?" We know that unintentional leak is calculated from actual flow data. I'm discussing the value used for "intentional" leak rate.

https://www.youtube.com/watch?v=1eeDtO- ... e=youtu.be

at appx 23 seconds, I created an intentional leak by loosening my mask, at about 40 seconds, I tightened it again, and then at about 1:05 I took the mask off.

Interesting. My AutoSet doesn't have that screen, or I can't figure out how to get to it.

Do you have a mask with an oxygen port or some way to generate a small additional leak? If the machine calculates a running measured intentional leak rate, it should zero out such an additional "intentional" leak. You could also do something like putting something small under the edge of the mask to create a constant small leak.

I used a Quattro air, and opened the oxygen port. On SleepyHead, it showed a fairly constant leak of about 8 l/m. It didn't drift down the way I'd expect it to if it had a "real time" intentional leak calculation. I also blocked off about half the vent on a mask with my thumb and started the machine and let it run for a few minutes. If it was doing "real time" intentional leak calculation, it should have lowered it's estimate of intentional leak. Then I unblocked the vent, which would make the total leak jump up. It didn't show up on SH.

I guess I could sacrifice an old mask and start drilling additional holes in the vent and see if the AutoSet compensates for the leaks. If it's "real time" tracking the intentional leak rate, it should still stay at zero.

[aytikvjo]

Since there are no wires for stator position sensors (usually hall effect) it's either a sensorless BLDC algorithm or as i previously suggested with onboard commutation and the three wires are simply VCC, VDD, and duty cycle

Anywho back to the matter at hand:

On SleepyHead, it showed a fairly constant leak of about 8 l/m

This is interesting.
Were you breathing through it at the time? For how long?

[archangle]

Since there are no wires for stator position sensors (usually hall effect) it's either a sensorless BLDC algorithm or as i previously suggested with onboard commutation and the three wires are simply VCC, VDD, and duty cycle

I just assumed it's functioning basically as a stepper motor and the CPU just drives the wires at the RPM it wants and assumes the motor turns at that rate.

Anywho back to the matter at hand:

On SleepyHead, it showed a fairly constant leak of about 8 l/m

This is interesting.
Were you breathing through it at the time? For how long?

I was using it for a couple of minutes. And breathing at the time.

It was late last night and I didn't take a lot of notes, so I'm gathering better data now. I'm using a Quattro Air, but set the mask type as "nasal" and I'm seeing what it registers in terms of leak.

[palerider]

I used a Quattro air, and opened the oxygen port. On SleepyHead, it showed a fairly constant leak of about 8 l/m. It didn't drift down the way I'd expect it to if it had a "real time" intentional leak calculation.

that is illogical.

once the machine fires up and calculates a baseline leak, it makes no sense for it, at a set pressure, to compensate for an additional leak.

what makes sense is for it to determine that 10lpm is the baseline for that session, anything larger in it's realtime calculation is therefore a leak, and it would report it as such.

for it to then assume that your exstra 8lpm was 'normal' would be.... well, I'd call it a bug, if not a fatal flaw.

try opening the port and starting the session from that point. if your theory about it assuming a vent rate for the mask is right, you'll start off with a 8lpm leak.

[palerider]

Yes, post them somewhere, please. I suspect a number of us would like to see them.

your wish, my command, etc.

http://imgur.com/a/s8nn7#0

don't forget the little gear in the upper right corner of the images to 'view full size'.

[archangle]

I used a Quattro air, and opened the oxygen port. On SleepyHead, it showed a fairly constant leak of about 8 l/m. It didn't drift down the way I'd expect it to if it had a "real time" intentional leak calculation.

that is illogical.

once the machine fires up and calculates a baseline leak, it makes no sense for it, at a set pressure, to compensate for an additional leak.

what makes sense is for it to determine that 10lpm is the baseline for that session, anything larger in it's realtime calculation is therefore a leak, and it would report it as such.

for it to then assume that your exstra 8lpm was 'normal' would be.... well, I'd call it a bug, if not a fatal flaw.

try opening the port and starting the session from that point. if your theory about it assuming a vent rate for the mask is right, you'll start off with a 8lpm leak.

The port was open before I turned the machine on.

I don't know what you're suggesting. Do you think it calculates the intentional leak rate based on the measured airflow shortly after you turn the machine on, or do I misunderstand what you're saying? I thought you were suggesting it gradually determines the intentional leak rate in the first minute or so after you turn the machine on.

[palerider]

The port was open before I turned the machine on.

I don't know what you're suggesting. Do you think it calculates the intentional leak rate based on the measured airflow shortly after you turn the machine on, or do I misunderstand what you're saying? I thought you were suggesting it gradually determines the intentional leak rate in the first minute or so after you turn the machine on.

you watched the leak readout fluctuate in the video, yes? that's how long it looks to have established a baseline.

so, you had the port open, and sh reported a 8lpm leak without you doing anything else?

[archangle]

your wish, my command, etc.

Very good. Now send me the winning powerball numbers for Wednesday's jackpot sometime in the next 24 hours. Via PM.

your wish, my command, etc.

http://imgur.com/a/s8nn7#0

don't forget the little gear in the upper right corner of the images to 'view full size'.

Neat pics. Nice touch whiting out the serial numbers. I love a fellow paranoid.

[archangle]

The port was open before I turned the machine on.

I don't know what you're suggesting. Do you think it calculates the intentional leak rate based on the measured airflow shortly after you turn the machine on, or do I misunderstand what you're saying? I thought you were suggesting it gradually determines the intentional leak rate in the first minute or so after you turn the machine on.

you watched the leak readout fluctuate in the video, yes? that's how long it looks to have established a baseline.

so, you had the port open, and sh reported a 8lpm leak without you doing anything else?

Yes, although I need to run a few more tests with better record keeping. It was too late last night.

I think this is the correct data. Note that it has a somewhat constant "unintentional" leak rate of 8 from start to finish. If it auto-adjusted, I would think it would have set this baseline to zero. I'm not quite sure I had the CPAP set to FFM or pillow at that time. Better data later.



The session in question starts around 22:42:10 or so.

[palerider]

Do you have a mask with an oxygen port or some way to generate a small additional leak? If the machine calculates a running measured intentional leak rate, it should zero out such an additional "intentional" leak. You could also do something like putting something small under the edge of the mask to create a constant small leak.

*sigh* the things I do for science. I'd blocked off the o2 ports in my mask because the little plug thingy kept falling off at night, so, anyway, I opened them up.

removing one of the two plugs resulted in about a 8-9cm leak:

full size

removing one plug, waiting a few minutes, then starting up, the leak rate settled at zero, I waited a bit, and then removed the other:


full size

then put it back in.

the fact that it set a baseline rate at two different leak rates on the same mask tells me that it's calculating the leak rate the way aytikvjo described, and setting a baseline on startup.

[palerider]

Neat pics. Nice touch whiting out the serial numbers. I love a fellow paranoid.

I'm only paranoid if they're NOT out to get me! *tightens strap on tinfoil hat*

[palerider]

I think this is the correct data. Note that it has a somewhat constant "unintentional" leak rate of 8 from start to finish. If it auto-adjusted, I would think it would have set this baseline to zero.

I had the same thing happen to me, I was running without an extra leak, I stopped, unplugged a port and turned it right back on, it settled with a 8lpm leak.

I turned it off, waited a couple of minutes, and started up again.... this time it settled to 0.

I believe that if it's a really brief interruption, the machine assumes, perhaps, that, for instance, the mask slipped off, the (in my case) smart-start stopped the machine, I put the mask back on again, and it resumed the session without a new calculation.

[The Latinist]

If it's setting a baseline at startup, how is it accounting for changes in the leak rate when pressure changes? Does it recalculate the baseline every time the pressure changes? And why doesn't this show up in the 1 Hz leak data as temporary jumps in leak rate each time the pressure decreases? Does it go back and subtract the new, lower leak rate from the start of each pressure change?