Hand Hygiene

I made a histogram of all the trigger presses relative to the door entries. The histogram goes from -15 s to +15 s seconds after a door opening. Each trigger press is associated with the closest door event. The data show a clear correspondence between the trigger events and the door events:

Each box contains the number of trigger events for each trigger (in columns) for a particular door minder. For example, the box at (top) left is for door minder 96 and we can see that 3 of the 281 trigger events from trigger 127 came 4 seconds before a door minder event occurred.

From these figures, it is clear that the data are not uniformly distributed. The data in the columns highlighted in green are from triggers close to the door. Trigger events were much more likely to occur near a threshhold crossing when the trigger was close to the door. In three cases, 82-93% of the trigger events came at a time close to a threshhold crossing. One of the door minders does not follow this pattern. For doorminder 98, only 45% of the data were followed by a threshhold crossing.

The data highlighted in yellow are for triggers that are adjacent rooms to the door. These triggers are easily accessible to someone entering or leaving a room. In many cases nearby triggers have higher values than distant triggers, suggesting that some people used nearby triggers. Trigger 154 is often used close to other threshhold crossings. Perhaps it was in a particularly convenient spot for multiple room accesses.

To gain more insight into whether or not a trigger event was associated with a particular threshhold event, we need to allocate each trigger event to only one threshhold event, across the experiment, which will require another program, for another day.

On Monday around 7:30 am we set up 4 doorminders (96, 97, 98 and 99) and 4 triggers (127, 154, 161 and 162) in the peds unit.  Deepti tested each trigger and doorminder twice a day and collected them on Wednesday, December 1 around 7:30.  Her notes from the checks are as follows:

PM Monday – Doorminder 97 offcenter.
AM Tuesday – All good.  161×3 at 10:10 am
PM Tuesday – All good.
AM Wednesday – 138×5 at 8:50 am
PM Wednesday – Trigger 161 not flashing blue upon push.  138×6 at 6:45 pm.

I downloaded the data, stored a copy on vinci.cs.uiowa.edu in the doorminderTrigger.  Details after the bump.
(more…)

This is the graph from my previous post “Tilt sensor experiment graphs”. It shows that there is some drift in the measured angle. It starts out significantly higher than the expected angle, and by the end of the experiment, it is lower than the expected angle. We have a good idea what the cause of this is. The program measures the absolute voltage off the accelerometer. However, the accelerometer’s output is proportional to the supply voltage. The supply voltage will go down throughout the experiment as the battery voltage goes down.

The accelerometer being used has the part number ADXL103CE. Here is the digikey link http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=ADXL103CE-ND

The following link is to an openoffice spreadsheet of some data I collected concerning how the measured angle varies with the supply voltage when the tilt sensor is programmed with the current program.

https://docs.google.com/leaf?id=0B2GN0bxAc29yYTRjM2EwNDgtYmU2MS00NWEzLTljZGYtNWNkN2E3NGM5YmRi&hl=en&authkey=CIaH9r8P

The sensor’s expected angle was 45 degrees for all measurements. using an external power supply (abbreviated PS in the spreadsheet), I got a change of about 3 degrees for every 50mv change in power supply voltage. However, the readings from the external power supply didn’t seem to be entirely consistent with the data for when the tilt sensor was powered by usb or AA batteries.

Setting that aside, I think we can change the program and board design to make the tilt sensor work for a varying supply voltage. We can add a voltage divider between power and ground to the board and send it to another input pin on the mote. The mote will read the voltage, and be able to determine what the power supply voltage is in that moment. Since we already have an equation for the angle that works at 3v, we can multiply that equation by (supply voltage)/3 to get an accurate measure of the angle. This should hopefully solve the drift problem.

This first graph shows the expected vs measured (by the tilt sensor) angle for the time when I changed the angle every 40 seconds. It is fitted with a linear regression line. As you can see, the slope and R^2 value are both close to 1, indicating that it was a good fit. There is an average offset of about 6 degrees.

This second graph shows angle vs time for the weekend, when Phil changed the angle of the bed. As you can see, the offset seems to have changed during the experiment. At first the measured angle was consistently above the expected; later it was the opposite. I would assume this is due to the sensor moving during the experiment.

I just went over to the hospital to get the dimensions needed for the next generation of tilt sensor. I looked at three types of beds in the MICU: the sport, b-sport, and sizewise. The sport and b-sport are the most commonly used and have nearly the same dimensions. The sizewise bed is only used for patients over 500 pounds, which doesn’t happen often. I also looked at the bed used in the SICU: the zoom stryker. (for a total of 4 different bed types)

The following table summarizes the dimensions needed for the different beds and locations of attachment for the tilt sensor.

“end of head” means we would attach the sensor to the head of the bed on the short side. the “out” dimension is how far it would be able to stick out from the bed.

“side of head” means we would attach the sensor to the side of the bed near the head. Again, the “out” dimension is how far it would be able to stick out from the bed.

module: on the sport and b-sport beds, there is an empty slot for a device that the nurses said we could use.

Based on these results, I think the “end of head” option is the best. This would mean the limits for the dimensions will be 1.3″x1.3″x4.8″.  It will also mean that the accelerometer will have to be perpendicular to the length of the box.

Edit: the height will need to be slightly larger than 1.3″ to fit the mote (min. of  about 1.5″).  All this will mean is the box will hang off the beam slightly, which should be fine (although that still needs verifying).

Edit2: I went back to the hospital and found out a few more things. It turns out the magnets don’t stick to the “end of head” side of the beds for the sport (for whatever reason). We could work our way around this if there is another way to attach the sensor, or possibly with a stronger magnet. I also determined that a 2″ high sensor should be fine for the stryker bed. Also, the consensus in the MICU seemed to be that the sizewise bed is used so little that we don’t really need to worry about it.

Vacuum Forming Overview

Vacuum forming covers a variety of plastics thermoforming techniques.  Our equipment uses the “drape forming” technique, wherein a plastic sheet is heated and then pulled down over a mold while a vacuum is formed beneath.  There are many design considerations to account for when designing parts and tools for a drape forming process.  These considerations are outlined below.  Check the resources listed at the end of this document for more information on these and other thermoforming topics.

●   Draw Ratio

•     draw ratio =     ______surface area of mold and flange [m^2]_______
surface area of material within flange perimeter [m^2]
(more…)

Ted pointed me to an interesting announcement for a chip on a pill for wireless you can swallow.

The following are the results for the test described in the post “Tiltsensor experiment 1″ in the third entry of the table. Results seem to be consistently higher than the desired angle by 5 to 10 degrees. Since this offset is the same for all entries, it is probably a calibration error. The sensor could have been put on the bed at an angle, or our code could be calibrated wrong. Otherwise the results seem promising.

Each entry consists of the date, followed by the time on a 24 hour clock, followed by the angle in degrees.

–start at 45 degrees

11/04/10 15:06:00 53
11/04/10 15:06:10 53
11/04/10 15:06:20 55
11/04/10 15:06:30 43
11/04/10 15:06:40 24
(more…)

The bed angle sensor was connected to an empty bed in the MICU for several days. Phil periodically came and moved the bed position. Gregg and I wrote some scripts to analyze the data, which is after the fold.

Looks pretty good, except for one unexplained shift of 5 degrees.
(more…)

Herein lies the description of the toaster reflow oven description for using our toaster as a solder oven.

(more…)