Yuwell CPAP Data: YH-450 to YH-920

Yuwell CPAP machines write their data in the company's own .bys format, and none of them opens in OSCAR as you would download it. CPAP Clarity reads all three families directly in your browser: the YH-450, the YH-680 and YH-690, and the YH-920 bilevel. This guide covers what each one records, what we show you, and what we deliberately leave blank rather than guess.

Which Yuwell Do You Have

The model number is on the label underneath the machine, and it decides everything about what your card holds. Three families read today.

YH-450 (YH450A). An auto-adjusting CPAP. Writes one summary file per night. This is the model with the longest track record here. One firmware variant was checked against a real card; a second, which writes seconds into its timestamps, was worked out from the community spec and checked only against a card we built.

YH-680 and YH-690 (BreathCare III). Usually auto-adjusting, though the format also carries a fixed-pressure mode. They write a folder per session containing two or three files: a summary, a minute-by-minute record, and on the YH-690 an optional flow file. Both have an oximeter built into the machine, so your card carries oxygen saturation and pulse alongside the therapy data.

YH-920 (BreathCare, bilevel). A bilevel machine running Auto-S mode, so it has separate breathing-in and breathing-out pressures rather than one. It writes files named the same way as the YH-680 and YH-690 but laid out differently inside, which CPAP Clarity works out by inspecting the card rather than trusting the filename.

If your machine is a YH-550, YH-560, YH-580 or YH-830, CPAP Clarity cannot read it yet and will say so plainly instead of showing you a blank dashboard.

Why OSCAR Cannot Read Any of Them

OSCAR only reads devices someone has written a loader for, and mainline OSCAR has no Yuwell loader at all. A community fork adds one for some models, but that is a separate project you would have to build yourself.

This matters for how much you should trust the numbers here, so it is worth being direct about it. For most machines we check our figures against OSCAR reading the same card. For Yuwell there is no OSCAR answer to check against, so we used two other methods, described further down.

What Each Model Records

YH-450

One sample every minute: your pressure in cmH2O, your mask leak in liters per minute, and a flag for any obstructive apnea or hypopnea in that minute. Each night also carries a header with your event totals, average leak, pressure range, ramp and humidity.

CPAP Clarity shows your AHI, your per-minute pressure curve with the peak and typical pressures, your average leak, your settings, and your night-to-night trends. The headline AHI matches the Yuwell app exactly, including the way the app rounds it down. The per-type breakdown underneath can differ from the app by a tenth, because the app's rule for splitting the index is not something we have been able to reproduce, and the central-apnea byte on this model is unvalidated.

YH-680 and YH-690

Minute-resolution pressure, and obstructive apneas, central apneas and hypopneas as separate events. The format carries oxygen saturation and pulse from the built-in oximeter, and where those readings pass our validity checks they land on the same charts an external oximeter would feed.

On a YH-690, if the card carries the optional flow file, CPAP Clarity also renders the breath-by-breath flow trace from its 10 Hz samples. The YH-680 has never been seen writing one, so we do not prompt you to go looking.

YH-920

Because this is a bilevel machine, the headline pressure is your IPAP, the breathing-in pressure, with the EPAP and IPAP envelope behind it. You also get minute ventilation, your inhale-to-exhale ratio, the oximeter data, and your configured Auto-S limits and pressure support read off the card as settings.

Events on this family are counted as the machine counts them: a run of flagged minutes is one event, not one per minute. That is a real difference from the YH-680 and YH-690, and it is deliberate on both sides rather than an inconsistency to report.

What We Leave Blank, and Why

This is the part most worth reading, because a blank field is a decision, not a bug.

Leak, on the YH-680, YH-690 and YH-920. The card carries a leak number. What we cannot establish is whether it means total leak (including the air your mask vents on purpose) or only the unintended part, and those two readings call for completely different reactions from you. Showing the wrong one is worse than showing none, so leak stays hidden on these families until a real card and an app screenshot settle it. The YH-450's leak IS shown, because it was matched against that model's own app.

Overnight oxygen desaturation index, on every model with the built-in oximeter. The oximeter records once a minute, and our scoring needs a reading every one to four seconds. At a one-minute cadence it cannot tell one long dip from a run of separate ones, so it closes and reopens the event every minute: a four-minute desaturation would be counted as four. We measured the inflation at roughly two to eight times, the factor being the length of the dip in minutes. That errs toward alarm rather than reassurance, because an inflated index can cross the line that drives the verdict on your dashboard. So it is refused outright rather than approximated. You still get the oxygen trend, the minimum and the average, which stay honest at any cadence.

Tidal volume and respiratory rate, on the YH-920. Both are decoded and both sit in the file. Neither has ever been confirmed against the machine's own app, and the way we validate this family cannot catch a mislabeled byte, so they stay out of sight.

Per-breath detail. On the YH-450 the format genuinely has no flow, snore or flow-limitation channel, so that one is the machine rather than us. The YH-680 has never been seen writing a flow file. The YH-920 may write one, and if a card arrived carrying it we do not decode it today, which makes that a limit of the tool.

How These Numbers Were Checked

Being straight about this matters more here than on any other brand we read.

The YH-450 was validated against the Yuwell app's own Detailed Data report across several real nights from a real contributor, including the way the app rounds the AHI down. Those values are pinned in our test suite so the mapping cannot drift.

The YH-680, YH-690 and YH-920 have been validated a different way, because no real card for them has reached us yet. We built a synthetic card for each format and checked our decoder against the community OSCAR fork's own reader, sample by sample. The fork's author confirmed several of those channels against real hardware. That is a strong check on the decoding and it is not the same as a real card, which is why we withhold anything the check cannot see.

If you own one of these machines, importing your card helps directly, and if something looks wrong, tell us. We use what you send only to build support for your device.

Importing Your Card

Power the machine off and take out the SD card.
Put it in your computer or phone. Any SD reader works.
Open CPAP Clarity and drop the card in. The whole card is the safest choice, though the model folder or even a single session folder also works. The format is detected from the files themselves, so you never have to tell us your model.
Read your data. It appears in seconds, in the page, on your device. Nothing is uploaded.

If your card writes its sessions in a layout we have not seen before, the import refuses and says so instead of guessing. Some real BreathCare cards do this today, and a copy of one is where adding them starts.

Reading Your Numbers

AHI. Breathing disruptions per hour of use. An AHI under 5.0 is the range the AASM associates with well-controlled CPAP, a figure that comes from its in-lab titration guideline rather than a home therapy target. Consistently above that, or a clear change from your own normal, is worth raising with the clinician who manages your therapy. What is AHI? covers the categories.

Pressure. On the YH-450, and on a YH-680 or YH-690 running in auto mode, you are watching the algorithm move inside your prescribed range. On the YH-920 you are watching two pressures and the gap between them.

Usage. Medicare's coverage rule defines adherence as at least 4 hours a night on 70 percent of nights across a consecutive 30-day period in the first three months. That is a coverage threshold, not a clinical target. See the Medicare rule.

Your pressure settings are a prescription. If the data raises a question, bring the data, not a conclusion.

Frequently Asked Questions

Can OSCAR read a Yuwell CPAP?

Mainline OSCAR has no Yuwell loader, so no. A community fork adds one for some models, but you would have to build it yourself. CPAP Clarity reads the YH-450, YH-680, YH-690 and YH-920 in your browser with nothing to install.

Why does my Yuwell show no leak figure?

On the YH-680, YH-690 and YH-920 we cannot yet establish whether the card's leak number means total leak or only the unintended part, and the two call for opposite reactions. We leave it out rather than show you the wrong one. The YH-450 does show leak, because it was matched against that model's app.

Why is there no ODI even though my machine records oxygen?

The built-in oximeter samples once a minute, and the scoring needs a reading every few seconds. At that cadence one long dip is broken into a separate count for each minute, so the index would run roughly two to eight times too high and could push your dashboard toward a warning you do not deserve. We show the oxygen trend, minimum and average, which stay accurate at any cadence, and refuse the index.

Which Yuwell models are supported?

The YH-450, the YH-680 and YH-690, and the YH-920. The YH-550, YH-560, YH-580 and YH-830 are not readable yet, and the site says so rather than failing quietly.

Has a real YH-920 card been tested?

Not yet. That family, and the YH-680 and YH-690, were validated against a synthetic card checked sample by sample against a community OSCAR fork's reader, whose author confirmed several channels on real hardware. The YH-450 is the one Yuwell validated against real contributor nights.

Primary Sources

Kushida CA, Chediak A, Berry RB, et al. Clinical Guidelines for the Manual Titration of Positive Airway Pressure in Patients with Obstructive Sleep Apnea. J Clin Sleep Med. 2008;4(2):157-171. PubMed 18468315 (opens in new tab). The AASM guideline behind the under-5.0 figure.
CMS Local Coverage Determination L33718, Positive Airway Pressure (PAP) Devices for the Treatment of Obstructive Sleep Apnea (opens in new tab). Source for the 4 hours a night on 70 percent of nights adherence definition.