Hand Hygiene

First Test (Get parts done by June 25 – 8 working days left)
Full Test (Get parts done by July 19 – 24 working days left)

Anchors (pyramids) (2 by June 25 and 16 by July 19)
 Sent out board that Gregg made (60 pieces for $244)
 Ordered parts for Davie’s board (ordered 16 of each from Digikey, except for terminal block, only 5 available).  Also got 200 miniusb recepticals and 100 pin switches, which were very hard to find and used in several designs.
 Decided to rapid prototype two for the June deadline and manufacture 17 others for July.  Howard is working on a fixture for the odd angles.
 Ordered enough acrylic for a few, to make sure we’ve got the right stuff.  Should be here 6/16
 Cut acrylic
 Phil ordered screws today.
 Still need to send in order for prototyped parts.
 
Anchors (beds) (4 by June 24, 35 by July 19)
  Tim’s going to design the boxes based on the door minder design
   – Need to check on how that is going
  Had trouble with the in-hospital test yesterday — think it’s a wiring issue.  Will wait to buy antennas until it is resolved.  Parts ordered.
  Buy antenna adaptors – waiting on new antenna tests.
  Buy antennas
  Buy antennas wire
  Gregg found that we can get 5 wall->USB adaptors from the electronics shop for $4.95 each.  The electronics shop has ordered those and order 30 more from them.
  Gregg will check on power pack (3V) and a outlet->USB convertor.  We have a prototype battery pack built.  Need to order more black boxes to hold them.

Badges (20 by June 24, 100 by July 19)
 Secure 100 badge shells – Done.
 Find out what was wrong with the old badges – Checked broken badges over the weekend.  3 of 4 had broken wires, which I fixed.  We’ve soldered the wire to the boards which is not as good as a connection.  Not much room, but I’ll see if there is another way.
 Order boards (Main plus daughter)
 Order batteries (Replacement for Ipod mini, Model No EC003, 3.7VDC 850mA)
 Order parts:  Ordered 80 MAX1555 chips from Newark.  Have a parts list ready for digikey.  Just need to double check the list.
 Sherlock connectors, 3 pin, right angle (these are in the digikey order)
 Face plate.  Electronics shop (Tim?)
 Screws?
 Cut wires
 Make boards

Pucks (12 by 6/25, 75 by 7/19)
 Test latest design 14th and 15th
 redesign for manufacturing – Mostly done.  Howard is preparing to manufacture them when the plastic arrives.
 order machining raw stock – done
 order boards from external supply house – Need to look over board to see what we want to do.
 order parts from digikey for boards
 order hardware (screws and spacers)
 build boards
 machine parts
 assemble
 test
 
Door minders (4 by 6/25, 35 by 7/19)
 Test new boards 14th 15th
 Write program to write to flash
 Ordered 50 transmitter and receiver chips from Reynolds Electronics
 Ask Phil to make a couple more (maybe revisions)
 Resolve channel selector switch issue with Tim
 Make 4 with parts on hand and boards made by Phil
 Talk with Tim about the getting the bodies made
 ——– Enough for 4 prototypes ————-
 Send final board to board house (here by July 21)
 Figure out how to manufacture body (finish by July 21)
 Order components (all here by July 21)
 Assemble boards (perhaps 1/2 hour each)
  Through hole soldering
  Cut wires
  Assemble mechanical parts
  Test

Today, we are doing another trial implementation. There are badges on the counter for everybody in the lab, as well as one visitor. It would be great if everybody wore one throughout the day!

The “rules” are:

when you turn on the sink, push the puck

when you turn on the soldering iron, push the puck

when you enter the upstairs area, push the puck

–and every time you push the puck, please make a tally mark, so we can keep track of the data!

Thanks

I’ve successfully (I think) set up file sharing on the computer in the northwest corner. It uses Samba in the Windows ‘GROK’ domain. I noticed that Eric had previously set up a Samba account, but for the sake of starting new, I removed it and started a new account.

Samba is sharing /home/groksmb under the ‘groksmb’ user name. You shouldn’t need to authenticate from any of the computers, but in case you do, use that user and the password is written on the whiteboard upstairs. Each computer upstairs has been renamed according to cardinal directions (north, southwest, etc). I have set up a persistent share on each of the Windows computers so that /home/groksmb is mapped to S:. This is where we should save project files from now on. I’ve set up a directories in /home/groksmb for ProE and Eagle (as well as Eagle part library). Deepti/Geb/Gregg — I don’t know if you guys want to put all of your code there, but if you do, create a new directory. Should be full RW permissions. I talked to Gregg about setting up the backup — we should just be able to tell the software that he has been using to back up from /home/groksmb. Gregg — is there anything other than these types of files that will need to be backed up?

The Unix computers that use Ubuntu have a samba client and Nautilus samba extensions installed. To get to the network shares, you should just be able to go to Places > Network > Windows Network > GROK > GROK-SERVER….. Printers should work the same way, although I’m not sure Linux will play well with PNP driver propagation (probably will just print PostScript?).

Some notes on that Samba installation on grok-server:

-The Samba daemon binaries are located in /usr/sbin (smbd, nmbd)

-The configuration files are located in /etc/samba

-Not sure why, but there is an additional configuration file in /usr/share/samba — this is unused

-To have samba rehash its config files use SIGHUP (sudo killall -HUP smbd|nmbd)

-To restart the daemon (usually not necessary) manually kill all smbd,nmbd processes without SIGHUP

NOTE: I noticed that the southeast computer upstairs, all the computers downstairs, and the unused ethernet cable coming UP from behind the support column next to the treadmill are all connected to the downstairs router (which is in turned connected to the upstairs router) on 192.168.1.0/24 rather than 192.168.0.0/24 — this means that the upstairs SE computer will not ’see’ the GROK workgroup and will not be able to fileshare. I’m not sure if there is some way to reconfigure the router downstairs (maybe if that router is on the same subnet and its DHCP assignments are also on the same subnet?) so that it’s computers can see that workgroup, or if a switch needs to be in place of that router, or what….

Geb and I had a brief discussion the other day about organizing computer usage and files on the computers. The general idea is to ensure that each computer (especially upstairs) acts as a workstation, so that it doesn’t matter which computer someone uses — the files s/he needs to get to will be accessible and persistent from any computer (with exception to OS-specific software). In other words, there should be one centralized storage location for project files such as code, Pro/E files, Eagle files, etc. Thus, there should only be one version (with a trail of older versions) of each file, rather than different versions on different computers. As far as design goes, this will become more useful as time goes on if we build part libraries with components such as mote boards (with mounting holes), bottles, baskets, etc as well as components that are accurate to the specific things we find in the electronics shop (accurate drawings of fasteners, battery clips, etc). It should be relatively easy to configure software to recognize this type of file storage. Both Pro/E and Eagle use specified ‘home’ directories where it will start looking for projects. It will also be important to use one persistent Eagle part library, whose location can be specified in the Eagle config…. The next step is to back up this file storage on to the external HD (Gregg is working on this, but his work may need to be adapted if we change the way files are stored).

This can be kind of complicated with both *nix and Windows computers. I think that using the grok-server computer on the northwest wall as a file server with a Samba server is the way to do this, but I’m not sure. This article, (found on the Samba website) seems like all the guide we will need to set this up. I’ve started by renaming each of the computers based on cardinal directions (east, southeast, etc) so we know who is who. The next step is to make sure the grok-server can see everybody else, and then to get an installation of Samba running. I know that I’ve played with Samba on that computer in the past (and I think Geb has as well). It may be easier to wipe the configs and start from scratch?

I looked at replacing the manual test switch on the FSR circuit board (the switch that forces the interrupt on the mote to go low) with an on-off-momentary switch that would also replace the power switch. I found that this would not work because Vout (from the comparator) is not the same signal as Vcc — the on-off-mom switches I found require the on/momentary signals to be the same. Geb asked why I was bothering, and I said that it would be nice to combine the two switches because the SPDT momentary switch we have now is the largest footprint on the board (other than the battery). We found that we could use a smaller SPST momentary switch (this one) if we connected it between V+ on the comparator and ground — where the old switch simply forced Vout to ground, the new switch forced the comparison V+ to certainly be less than V-, implying Vout goes to ground. I also replaced the 2 FSR pins with a 4-circuit ZIF connector (more information in ‘FSR distributor‘ post; 2-circuit connectors have a minimum order of 10,000 everywhere). These changes required a reconfiguration of board layout, and I found that the changes made to the switch were very difficult (if not impossible?) to lay out. So for now, I’ve decided to keep the larger switch and relayout the board with the ZIF connector for the sake of ease. We do have the smaller switches on order and if we have time we should try again to make a layout with this switch. It may require a two sided board.

EDIT: Lies! I’ve got the circuit board laid out with the new switch and ZIF connector…. figured it out. The Eagle files are on the southwest computer

From the TFlash program, I found the PAGE_LENGTH variable to be set to 256.

256 bytes * (1 message / 6 bytes) = about 43 messages would fill a page

43 messages * (7 seconds / message) = 301 seconds = 5 minutes

So, using only two motes, it should take about 5 minutes to fill a page of flash.

The Ubuntu machine mcgwire was acting up today, in that its network connection died.  I found I could get it going again by telling them to get a new IP address with the command:

sudo dhclient eth0

Overview:
This is a 24 hour run-through using only pagers and anchors. Three anchors were placed around the lab, and everybody had an assigned pager to wear. Pagers were worn throughout the day, or, when not in use, they were kept on in the pager area near the sink. At the end of the 24 hours, the pagers were collected and the memory was downloaded to the computer. Except for two parts, the mote flash was successfully downloaded.

Notes:
(1) Pagers were on all day regardless of whether or not they were being worn.
(2) Pager 218 and Anchor 42 did not download successfully. **Still trying to find where this problem comes from.
(3) Start Time: Tue, 01 Jun 2010 08:30:10 // Stop Time: Wed, 02 Jun 2010 02:42:08
(4) Only data from 65518 seconds were available or 18.2 hours. **Probably because of data types
(5) We still don’t have a way of looking at the data.

Using:
8 Pagers
-Geb (217), Tim (227), Mike (221), Greg (218), me (225), Phil (219), visitor1 (220), visitor2 (228)
3 Blackbox Anchors
-Upstairs: middle of room (42)
-Downstairs: near sink (46), on black countertop (41)

Process:
Pager/Anchor flash cleared and reprogrammed. (8AM 6/1/10)
Anchors placed around lab. (8AM 6/1/10)
Pagers downstairs, near sink, turned on, ready for pick-up. (8:30AM 6/1/10)
Pagers worn throughout day and returned to same area near sink.
Pagers turned off, Pagers/Anchors brought upstairs. (9AM 6/2/10)
Pager/Anchor flash transferred to computer. (9AM 6/2/10)

Geb and I did some calculations to see how many hours of messages we could get from the flash memory on the motes.

The flash memory can hold 1024 kilobytes of memory. Each message is 6 bits.

[ 1024 kilobytes * 1000 (bits/kilobyte) ] / 6 bits gives approximately 170,000 messages.

Using a previous experiment, we calculated that there are approximately .5 messages per second.

[ 170000 messages / (.5 messages/second) ] / 3600 (seconds/hour) gives us a grand total of 94 hours of storage, which amounts to almost four days of storage per mote.

We also learned that the messages will go up to 65536 seconds in mote time and then the mote time will reset.

All plants have sprouted except for 17, 13, 8, 9, 4 (all plants that are closest to the middle). The LED plants are less mature than the rest (~10mm vs ~5mm)