[Crepuscular]

From what I can see the software does not have any idea of how many records the device has until after it downloads them. (OneTouch Software Ver2.2). When sniffing the serial port I see the status bar sit at the beginning 2 bars until after all of the HR records are read. It makes a substantial jump after that. It doesn't seem to track the actual download process.

BTW what bits indicate the record type? It turns out the two bytes I thought were reserved for GL are multi-use depending on the record type.

[Crepuscular]

Looks like it might be the high order bit of the last byte for GL. Is that what you are using?

[Crepuscular]

What polynomial are you using for the CRC-7? I tried (x**7 + x**3 + 1) same as that used for MMC cards but it does not result in the correct CRC.

[Crepuscular]

I have decoded the Pill info and the Insulin info. Is there some place I can post it? I have also decoded most of the health info and all of the exercise info. Do you want that also?

[JasonJayhawk]

Wow, good job!! Out of curiosity, what was your method for figuring it out?

Do you have any algorithms that you could post here, so that we may all marval in your coding?

[Crepuscular]

I have an office doc that I would like to post. It just describes the data format. The trouble is that I added tables tomake it easier to understand but tables don't translate well if I were to post it to this forum. If there Is a repository where I could put it It would be great.

[Crepuscular]

I found some errors in my doc so I better go through it with a fine tooth comb before I post it.
In the mean time here are a couple of the record types I have decoded.
I will post the rest when I verify them.

BTW if condensr could post his code for encoding the CRC7 it would be much appreciated as I can't seem to get mine to work. Has anyone found out what the last byte of data is for? It almost looks as if they have a second crc on the payload????
Have fun!!!

Medication Record type 0x28
Uses first byte of data after the time stamp.
Bits 0 – 3 Quantity in half pill increments.
Bits 4 – 7 Pill type.
1 = Pill A
2 = Pill B
3 = Pill C
4 = Pill D
5 = Pill E
Note: the record does not contain the name of the meds if the user has changed them. This info has to be extracted out of another type of record.
************************************************
Insulin Record type 0x2c

Uses the first 3 bytes of data after the time stamp.
Bits 0 – 2 Insulin reading mod 8
Bits 3-4 unused ??
Bits 5-7 Insulin type
Byte 2 Insulin reading low byte div 8
Byte 3 Insulin reading high byte div 8

Insulin type:
1 = Rapid
3 = Lente
5 = Ultralen

Example: A reading of 0x21 0x38 0x01

0x21 = 00100001
First three bits = 001. add 1 to final result.
Last five bits = 00001 = Rapid

Bytes 1 and 2 = 0x38 0x01 reversed = 0x0138
0x138 = 312 decimal
312 * 8 = 2496
Add 1 from the mod bits = 2497
Insulin reading = 2497 / 10 = 249.7

Keytones type 0x78
Uses first byte of data after the time stamp.
Bits 0-4 unused ???
Bits 5-7 Keytone reading
Keytone values
0 = Negative
1 = Trace
2 = Small
3 = Moderate
4 = Large

HbA1c Record type = 0x7C
Uses first byte of data after the time stamp.
Reading is in tenths of a percent so divide by 10

Exercise Record type 0xB4
Uses the first two bytes following the time stamp
Bits 0-5 time value mod 64
bits 6-7 Type
Byte 2 time value div 64

Type:
1 = mild
2 = Moderate
3 = Hard

Value represents number of 5 minute intervals

To calculate time in minutes:
Multiply second byte by 64 and add first 6 bits of Byte 1 then multiple by 5

[condensr]

Check your PM.

[Mica]

Hi,

for some time now I'm used to evaluating the BG levels of our 6 yr old daughter, that stem from an OTU/OTU2, with a custom application, as the representation of the data in DMS Pro is very poor. Then I got an OTUS hoping to facilitate my recordings of the meals taken, the corresponding insulin dosages and the phases of physical strain. It was very disappointing to realize, that LifeScan changed policy and no longer provides customers with sufficient information about their products. My notes aren't copyrighted material that have to be protected from being read by unauthorized, illegitimate pieces of software. So what the **** is going on in the heads of those guys? Making a lot of money with the test strips doesn't seem to be enough. Any ideas what their long-term strategic goals are?

[JasonJayhawk]

Lifescan (J&J) implies they do not want any company or person from making money by creating a product better than their own software or better than the companies that have licensed the protocol from them.

Thus, by forcing a license in order to download the data, they can control who makes any potential money.

Off topic, but related -- CVS Pharmacy, for example, has in their privacy/usage policy that anyone who prints photos through their company must adhere to their rules of allowing CVS to use the said photos, license-free, in any medium that they desire. People don't protest such policies, so it allows them to get away with what they want -- until people protest loudly enough.

[Mica]

Sure, an innovative product would allow them to dictate the rules. But I wonder, what kind of innovation they believe to offer here for licensing. AFAICS an exotic CRC method, nothing else. The coding of the user data is standard at best. Compared to the profits they make with their test strips the selling of software might never run up to a reasonable amount. DBMS Pro is anything but a killer application, at least in a technical sense . So wouldn't it be better for them to attract software developers in order to push the hardware into the market flanked by a diversified portfolio of applications?

[AirJordan]

In a few days I have the description of the communication protocoll for the OTUS.

[someone]

Now someone needs to decode the protocol for the Guardian RT. I tried sniffing the port, but I would have no idea how to decode that!

If anyone is interested, I can send them the serial port communications and the data table from the carelink site.

[BriOnH]

Originally Posted by someone

Now someone needs to decode the protocol for the Guardian RT. I tried sniffing the port, but I would have no idea how to decode that!

If anyone is interested, I can send them the serial port communications and the data table from the carelink site.

Then send the decoded information in XML format using SOAP to Diabuddies .

[someone]

By the way, has anyone here tried to use Jad to decompile parts of the CareLink uploader? After all.. it supports most meters.