Pulse Oximeters for home use

Hello Bill,

Good to hear from you again.

NightHawkeye wrote:
Wiring it is pretty easy, Roger, so long as you don't mind soldering. The N-200 and N-395 use the same 15-pin interface. You only need to connect three wires to the computer 9-pin RS-232 interface.

Materials needed:
3-wire cable (I used 4-wire telephone cable.)
9-pin female DB-9 connector
15-pin male DB-15 connector

Hook-up
pin 2 of DB-9 connector to pin 3 of DB-15 connector
pin 3 of DB-9 connector to pin 2 of DB-15 connector
pin 5 of DB-9 connector to pin 5 of DB-15 connector

Note that pins 2 and 3 are reversed between connectors, just like a null modem cable. These are the RS-232 transmit and receive lines. Reversing them means that the transmit line from the N-395 connects to the receive line of your PC, and vice-versa. Pin 5 is simply the ground. One could shield the cable, but I'm using an eight foot cable which works fine without shielding.

Thanks for the pin information. It will save me the time spent trying to find my PC port information.

Nellcor’s Service Manual indicates a cable up to 25’ can be used without a problem, so I don’t an issue with your 8-footer. On my end, I have a spool of CAT-5 cable collecting dusk in the garage that will work great for an interface cable. To get started I’ll use an older copy of PROCOMM and then when I know things are working I’ll begin trying to grab data using a COMM library I have here. With the service manual indicating the machine uses software XON/XOFF that should make the process faster and simpler to handle.

NightHawkeye wrote:
The N-395 communicates at a speed of 19,200 baud. I don't think it is capable of other speeds, so the PC needs to be set appropriately.

Going into the Nellcor menu system shows most of the old popular modem speed with the highest being 19,200 baud.

NightHawkeye wrote:
Hmm . . ., maybe I've already done the hard part for you guys. OK, just kidding, I know you've still got to solve the problem of collecting the data bit by bit, but I seem to recall that that's explained in the oximeter manuals.

There is something in the Section-10 of the Nellcor N-395 that talks about how data is coming out, but I won’t be sure everything is there until it starts to collect in a file. Once a few files are collected I can look to see which rows are header rows and which are data records. According to the SVC-Man, a new record is created every 2-seconds. If this is just referring to the time the machine is turned on, a night of 8-hours of recording will create a text file with 1,920 records plus header and footer details if they are included. This isn’t much of a file to handle with today’s fast computer chips. Let’s see how this plays out.

NightHawkeye wrote: If you're really determined to get the Score software, one possible work-around is simply asking your physician's office to order it for you. (I realize that some might balk at such a request, but I suspect my sleep doc would be accommodating.)

Thanks for the offer. My MD isn’t into this sort of thing. His office is part of a larger family practice that is so busy I find it hard just to get in to see him with 3-weeks notice, and even then the nurse handles some visit details. If he wasn’t such a great MD and person I would be looking elsewhere, but for now I’m going to stick with him.

As for the software, if a used copy surfaces I’ll go that route. However, now that I’ve started down the path of rolling my own, I’ll continue with the current plan and try to keep expenses to a minimum. Already this stuff is consuming too much money, so expense control is a concern.

Of course, if you know a new version is only about $100 or so or even a used copy for a little less, then I can’t afford to spend my time on creating something from scratch as project. If a used copy appeared, I’d problem continue trying to roll my own to see if I could find a way to use the save Titration data from the REMStar’s database. This would allow us to overlay and synchronize the oximeter information to the titration data records giving us a parallel display. Seeing what might be causing a low saturation reading, or even know how a big mask leak affects the APAP treatment would take a lot issues off the table.

After writing that I began to wonder if it might not be time to go out and grab a copy of ALICE-4 (software used in sleep centers to score and create patient reports). Of course once you get the software then we’ll need to get the brain modules and wires. Yikes, talk about going off the deep end. Imagine having to tie yourself up with all those wires and electrode goop before hitting the sack. Whoa, did someone say keeping expenses down and sleep was important?

NightHawkeye wrote: For what it's worth, I'm not overly impressed with the Score software. It basically performs only one function and is not very flexible. If you guys can dump the data into Excel, I'd probably use that instead of Score in order to get the increased flexibility.

If would be interesting to hear what it really does and what you would like to see if do that it doesn’t even consider. If we succeed in rolling something useful and go beyond the grab the file level, tossing in graphics would be the next step and then it begins to make sense to think about useful features like, “And if you guys happen to create an oximeter analog to MyEncore . . ., that would be the ultimate.”.

NightHawkeye wrote: I also think that some here would benefit from using an oximeter. For example, I get somewhat better oxygen saturation when using the Swift nasal interface than my other interfaces. I don't really know all the reasons why, but the effect is real and is not only not obvious in the Encore Pro data, it is not even there. The Encore Pro data simply doesn't show any real differences between my nasal interfaces.

I’ve suspected some mask do a better job of providing treatment than others. For example, when I use my RESMED Ultra Mirage Full Face mask, it leaks like a sieve, but I find I can breathe so much easier. I loose large portions of the REMStar’s treatment data with that mask because the leaks are so large the REMStar thinks the mask has fallen off and stops recording information. Without data, I can’t tell if the treatment is working or not. This is where I can get some incremental feedback from the oximeter to see how the mask helps or hinders oxygen saturation. It might be the flow and pressure in this leaking mask might still be giving good treatment. When I wear it I don’t wake from apnea events, but I don’t know if the oxygen saturation stays above the 85% level. So far my Mirage Active has been my workhorse mask, but I’ve notice more than a few oximeter recorded blips where is shows the oxygen saturation drops down to 87%. That value doesn’t seem like a bad reading, but without some of other data I don’t have a baseline to work from. Time, testing and some study should help me close the ignorance gap.

I have been tied up for the past few days ' also for the next 2 weeks so may be slow responding.

Bill, thanks for the cable info - that does help a lot.

But, one point you made is very interesting to me & one reason I want to do some SpO2 experiments. This being that different mask interfaces may result in different levels of SpO2 sat. I have suspected this to be true & want to prove it (unless someother lab rat here can do so 1st).

I believe the Adams Circuit nasal mask)s0 I have, will result in me finding hgher levels of SpO2 sat.

Cheers

DSM

Roger... wrote:If would be interesting to hear what it really does and what you would like to see if do that it doesn’t even consider. If we succeed in rolling something useful and go beyond the grab the file level, tossing in graphics would be the next step and then it begins to make sense to think about useful features like, “And if you guys happen to create an oximeter analog to MyEncore . . ., that would be the ultimate.”.

Score plots Pulse Rate and Oxygen Saturation vs Time where the choices for time are:
- full study (usually 48 hours of data since that is its memory storage capability)
- 24 hours
- 4 hours
- 2 hours
- 1 hour
- 30 min
- 15 min
- 5 min

Note that 8 hours or 10 hours is not one of the selections, although it would be the most useful for me. (When I've posted a full night's chart in the past it was by splicing two 4-hour charts together.)

Score also finds the following events:
- desaturations (default of 4% below previous peak, somewhat selectable)
- motion events

Score also has histograms which are essentially worthless because the bins are so broad, and the software allows no customization of the histograms.

Roger... wrote:I’ve notice more than a few oximeter recorded blips where is shows the oxygen saturation drops down to 87%. That value doesn’t seem like a bad reading, but without some of other data I don’t have a baseline to work from. Time, testing and some study should help me close the ignorance gap.

One tid-bit I've learned from oximeter use is to distinguish between quick desaturations vs apnea induced desaturations which seem to require anywhere from 30 sec to several minutes before bottoming out. I can get quick desaturations (10 sec or so) down to the mid-80% range by hand position. I wonder if this is related to low blood pressure for me. The only time it seems to happen is right after going to bed and the oximeter will sometimes alarm on this, but changing hand position resolves it immediately. The desaturation is over in a matter of seconds. At first I thought it was false alarms, but after seeing it a few times decided the oximeter is recording a real phenomenon (but I don't believe it to be apnea related). Don't know if you might be seeing something like this, but thought I'd mention it.

The thing I've most missed with the Score software is the ability to look at events from multiple nights as MyEncore does with AHI for example.

Just my $0.02 worth.

Regards,
Bill

This is good feedback Bill. I’ll keep your notes handy when I start pulling the process together. Earlier today I poked around at the COMM routines to be sure everything I need is there. So far, it seems fairly complete, but only actual coding will tell the real story. I still haven’t found a 15-pin connector locally, so I’ll need to track one of those down soon. I’ll put more time into this after the 15th of April is behind us.

My desaturations were mostly small period events, but a few covered a broader period. I think the short spikes are caused when the finger probe moves to a poor location as I toss and turn. I don’t know about the broader drops so having some meaningful analysis view of the data should help sort out the machine tracking records to what the REMStar is recording. I have some Disposable Oximax Sensors I can try when things begin to come together with the software to see if a more stable finger probe helps to remove some of the spike readings.

One of the things I did with my N-395 is to reset the default Alarm levels to where I think they should be for this kind of work. Factory default has the Low Adult Alarm sound if saturation drops to 85% or less. I’ve dropped this down to 80% to keep the noise this thing makes to a minimum. In addition, I set the Probe-Off alarm to 120 seconds because small bumps were causing alarm triggers. I also went in and stopped the Probe-Off Reminder so it would not cause more noise when I deliberately paused the tracking during excursion from the bed. Dropping the Alarm volume down to a 1 level has helped to stop the bedroom complaints. The first night this finger probe came off the alarm volume level was set to a 7 when it went off making a racket that had the bride on the ceiling instantly. When that happened I thought my bride was going to grab the phone, bean me and then call for an ambulance. Now when it goes off, the sound is so low it doesn’t scare her and have her scare me.

Roger... wrote:I still haven’t found a 15-pin connector locally, so I’ll need to track one of those down soon.

I was surprised to find that Radio Shack didn't carry the DB-15. It's a common connector, nonetheless. It is used in the PC for the "gameport/joystick" connection, although not all PC's seem to have that anymore. I got my DB-15 connector by robbing it from an old joystick. If you choose this route, be advised that some are better for scavenging than others. Molded ones are the worst to use (cosmetically), but any should work so long as you are able to solder to the pins.

Roger... wrote:I’ll put more time into this after the 15th of April is behind us.

I wish you hadn't brought that up. I'm in denial today, but that's gonna consume a lot of my free time for the next couple of weeks too.

Regards,
Bill

NightHawkeye wrote: I was surprised to find that Radio Shack didn't carry the DB-15. It's a common connector, nonetheless. It is used in the PC for the "gameport/joystick" connection, although not all PC's seem to have that anymore. I got my DB-15 connector by robbing it from an old joystick. If you choose this route, be advised that some are better for scavenging than others. Molded ones are the worst to use (cosmetically), but any should work so long as you are able to solder to the pins.

After checking with a few more Radio Shack stores, I finally thought about asking them to look in their catalog to see if they still carried the old standard DB-15 connector. Sure enough, it isn’t listed any longer. In its place is the new style High-Density DB-15 connector. This new connector is the same size as the DB-9 and uses the same size hood housing. Not even our local Fry’s Electronic outlet store carries the old style DB-15 connector any longer.

It looks like I’m going to need to go to the computer surplus store and see if they have any connectors, or cables that aren’t of the molded variety.

On a positive note, I’ve been able to cobble a simple COMM program together that should work at being able to talk to the Nellcor and accept its data into a file it saves on the disk. Once I get the cable working, I’ll begin looking at how to parse the records in the Nellcor file.

If anyone has a Nellcor file they can share, I’ll be able to look at the data sooner and not let the cable connection stop the entire project from making progress.

More as it happens,

It sure is confusing when the forum randomly decides to not log me in. It shows my name and password, but it still requires me to click on the logon button even though I have the auto-logon option enabled.

In any event, the above post is mine.

Roger wrote:It looks like I’m going to need to go to the computer surplus store and see if they have any connectors, or cables that aren’t of the molded variety.

That's probably your quickest bet, Roger. In case that doesn't pan out, though, here's a place that sells an entire db-15 to db-15 cable cheap. (Shipping's reasonable, too.) You could lop off the connector from one end and replace it with a db-9.
http://www.winfordeng.com/products/ext15.php
Entire cost should be less than buying a simple RS-232 cable from RS. Of course, there's still the issue of waiting another week or two for the cable to arrive.

Roger wrote:If anyone has a Nellcor file they can share, I’ll be able to look at the data sooner and not let the cable connection stop the entire project from making progress.

My files wouldn't do you any good, Roger. They are saved as binary data in rather small files, not ascii.

Regards,
Bill

Well I finally located a DB-15 connector and hood at an R&D Electronic store here in the area. However, using a metal hood for the housing of the connector says I'll need to get creative with the hood housing so the connector will fit into the opening on the back of the case. Boy, if it isn't one problem, it certainly will be another. Oh well, that is part of what this is about I guess.

Bill,
What is the image you are using for an Avatar?
It certainly looks scary, so I hope it isn't how you are feeling.

I was also able to make connection with the Encore Pro database and now have a listing of some of the tables. My next step is to discover the fields and their purpose. That will come after I get the new COMM program downloading data from the Nellcor.

More as it happens,

Roger... wrote:However, using a metal hood for the housing of the connector says I'll need to get creative with the hood housing so the connector will fit into the opening on the back of the case.

Never did like those metal hoods . . . The most important thing you need is to make sure you have strain relief on the cable so that your connections stay connected. Also, make sure you put some insulating tape between the cable connections and the hood, otherwise, one day the wires will short to the hood and then you could have even more problems.

Roger... wrote:What is the image you are using for an Avatar?
It certainly looks scary, so I hope it isn't how you are feeling.

I thought everyone would recognize it. It's the AHI screen from MyEncore. . I tried an Encore Pro screen for the avator, but it just didn't look right. I thought the MyEncore version looked good, but judging from your reaction . . ., maybe not. . Guess I need to keep looking for an avatar

Regards,
Bill

Bill,
What confused me about the avatar image was how bad it looked. In fact after seeing it I had to go take a nap because it felt so bad.

My confusion comes from the left side of the image looking at the screen. Those peaks look like you've got two graphs blended together. At least that is the reaction from a MyEncore 1.4 user. Maybe Derek did something in 1.5 that puts more information on the screen, but even if he did, to see that many untreated Apneas is making me tired again.

After finding out about the connector housing interference with the back panel of the Nellcor, I ordered the cable you referenced to see if I can make a molded cable work. I wanted to stay away from the molded cable because of what you mentioned earlier, which is why I chased the connector component and housing, but if I've got to fiddle to make it work it might be best to give a molded connector a go and see what life will be like.

In your avatar image, what did you do to calm the oscillations?

Roger... wrote:What confused me about the avatar image was how bad it looked. In fact after seeing it I had to go take a nap because it felt so bad.

Gee, Roger, I'm proud of it. Despite what you see on the chart, I felt better the very first night! Here's the full size chart:

Roger... wrote:My confusion comes from the left side of the image looking at the screen. Those peaks look like you've got two graphs blended together. At least that is the reaction from a MyEncore 1.4 user. Maybe Derek did something in 1.5 that puts more information on the screen, but even if he did, to see that many untreated Apneas is making me tired again.

Roger... wrote:In your avatar image, what did you do to calm the oscillations?

I had to do my own titration, Roger, but ran into a few obstacles along the way. Here's a short synopsis:
1. Began therapy with 4 cm CPAP just to start. Observed AHI around 15.
2. Auto function with BiPAP and small pressure increases didn't improve things.
3. Decided to try APAP, but AHI still around 15.
4. Tightened mask to eliminate leaks at higher pressures and tried to tough it out. AHI jumped to between 20 and 30.
5. After more than a week of high AHI, I suspected tightening mask had increased apnea by pushing jaw back. (DUH! )
6. Started using nasal mask and AHI dropped back down to around 15 for a few days.
7. Tried APAP again, but this time with nasal mask and AHI dropped consistently to less than 5 with very few apneas.
8. The jump to AHI of 10 in late March was caused by sleeping on soft pillows while vacationing. (The lost day right before it also had AHI around 10, but I only have this data on my laptop.)
9. Firmed up pillow and AHI dropped back down to consistently less than 5 again. (No doubt this is all jaw related.)

So, anyway, that's where things stand now. Apnea count is very low now. The residual apneas usually occur faster than the APAP can respond to them, but I may just have to live with that since increasing minimum pressure causes me grievous aerophagia. I'm not sure what to think about the high residual hypopneas, but suspect sinus related constrictions.

Roger... wrote:After finding out about the connector housing interference with the back panel of the Nellcor, I ordered the cable you referenced to see if I can make a molded cable work. I wanted to stay away from the molded cable because of what you mentioned earlier, which is why I chased the connector component and housing, but if I've got to fiddle to make it work it might be best to give a molded connector a go and see what life will be like.

If you're not used to working with hardware, the slow and cautious approach may be best. Putting the cable together was a multi-hour project for me too, mostly trying to figure out which parts of the interface were being used. Initially, I even neglected to hook it up as a null modem cable and wondered why it didn't work. (I experience the DUH! factor a lot. )

Regards,
Bill

Has anyone devoped software to produce useful charts from the oximetry data, once it's downloaded to the computer?

It's a shame the ProFox software is apparently priced only for labs (around $500). There is probably something else out there that would do.

Guys n Gals, If anyone can't get the 15 Pin Sub-D tail plugs (solder wires onto the pins) & Backshells, I can get them here in Sydney from a Tandy equivalent for approx $4 US complete. The postage might be about $6 US so if anyone can't get a complete one & can spare $10 (via Paypal to me) I can buy the plugs & send them to you. Just PM me. But, I am sure you can get them there .

Re software - I hope to be able to look at the programming side after Easter.

Cheers

DSM

Congratulations, DSM.

As I read your last post, your total post count is now 1000!!!

Regards,
Bill