[JasonJayhawk]

Originally Posted by someone

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

I think someone was working on that in another thread in this same forum. If you Google it, you'll find some other people are working on it, too. I suspect that they actually encrypt their data to prevent third-party companies from making use of it. (The XML data had at least one RSA key in it.)

[someone]

Originally Posted by JasonJayhawk

I think someone was working on that in another thread in this same forum. If you Google it, you'll find some other people are working on it, too. I suspect that they actually encrypt their data to prevent third-party companies from making use of it. (The XML data had at least one RSA key in it.)

I was able to decompile the ultrasmart classes, but I didn't go any further than that.

[condensr]

They send the raw data from the meter to their web service. The web service handles all data parsing etc. server side.

[JasonJayhawk]

That makes complete sense, then! They must have decided it was easier to encrypt on the device for easy web transmission.

[someone]

Originally Posted by condensr

They send the raw data from the meter to their web service. The web service handles all data parsing etc. server side.

Awe too bad. That didn't occur to me.

[meffertf]

Hi,

I've read this thread with great interest. I have 2 diabetics in my household, both use UltraSmarts. I write software for a living, and it's very frustrating to me that there is no common transfer protocol or data format for all meters, but this is the world we live in.

Well, all I'm trying to accomplish is to automate the bi-weekly data sends from both our UltraSmarts to the doctor's fax machine. Being an engineer, it seems that I should be able (from a 15000 foot level) to automate using a Serial to Bluetooth adapter to transmit data from the meter to a collection point (a authorized desktop/laptop/smartphone near the meter), store the data, parse out the information I need, use that to generate a graph on a fax, then connect to the fax machine and send it. Or email, whichever is preferred. Or send it using Diabuddies. Etc. Etc. Basically do an automated graph send every x days/weeks. A little .NET app imo.

None of this should be that hard, I've done each (except the US data collection part and I've just started reading about Diabuddies) in one form or another in the past.

Anyway, I've been trying to connect to my son's UltraSmart for a couple days now via HyperTerminal using 38.4k, StopBits 1, Parity None, DataBits 8. Sending a DM? just to get the sw version, etc. doesn't work.
I Wrote a fast .NET class to try it using the System.IO.Ports.SerialPort namespace, but no go there also. I know that class works 'cause I can connect to a normal OneTouch Ultra2 and OneTouch Ultra using it (at lower baud rates, prefixing it with \x11 and \x0d, etc.). Could it have something to do with the USB/Serial cable I'm using? Shouldn't, that's how I connect to the other OneTouch meters.

I'm about to hook up a sniffer to start disassembling the data, but I figured I'd give this group a try and see if I can get the work (code) that's already been done for the UltraSmart instead of reinventing the wheel. Maybe I could help move things forward even?

Replies would be appreciated.

Sorry about the rambling lol,

Thanks
Frank
(not a diabetic, but a parent of 2)

[JasonJayhawk]

Hopefully, Bryon might know how to help you with that. The Ultrasmart is difficult. I've contacted the company with no success, other than a "licensing agreement." I think someone might have figured out how to talk to it -- I've been busy with twins -- but hopefully someone can help you!

[DeanM]

Know what you mean about the protocols not being standardised - it's a pain in the arse. I'm writing my own application too, and whilst I can't help you with yours since mine only talks to the Roche Accu-chek Integra for the moment (for that's the meter I bought), I think the best thing is to put on a serial monitor and pull it apart. At least that way you can see what's going on.

[lar3ry]

Originally Posted by meffertf

Hi,

I Wrote a fast .NET class to try it using the System.IO.Ports.SerialPort namespace, but no go there also. I know that class works 'cause I can connect to a normal OneTouch Ultra2 and OneTouch Ultra using it (at lower baud rates, prefixing it with \x11 and \x0d, etc.). Could it have something to do with the USB/Serial cable I'm using? Shouldn't, that's how I connect to the other OneTouch meters.

I'm about to hook up a sniffer to start disassembling the data, but I figured I'd give this group a try and see if I can get the work (code) that's already been done for the UltraSmart instead of reinventing the wheel. Maybe I could help move things forward even?

Replies would be appreciated.

Sorry about the rambling lol,

Thanks
Frank
(not a diabetic, but a parent of 2)

Frank,
I just downloaded the pdf file for the Ultra2 from the LifeScan site. I have been trying (using TeraTerm) to communicate with the meter, but have't yet managed to even wake it up. I followed their documentation, making a file using a hex editor, with the contents being (in Hex) 11 0d 0a 44 4D 40, and sending it at 9600 baud, 8N1.

I see that the protocol file is dated 2006, and am wondering if something has changed.

Any idea you could give me would be much appreciated.

[lar3ry]

Originally Posted by lar3ry

Frank,
I just downloaded the pdf file for the Ultra2 from the LifeScan site. I have been trying (using TeraTerm) to communicate with the meter, but have't yet managed to even wake it up. I followed their documentation, making a file using a hex editor, with the contents being (in Hex) 11 0d 0a 44 4D 40, and sending it at 9600 baud, 8N1.

I see that the protocol file is dated 2006, and am wondering if something has changed.

I ended up hanging the logic analyzer on it, and figured it out. They lied about the contents of a packet. They also don't tell about all the message types (like "clear meter"). I am using SniffUSB to figure them out.

[drummerAL]

I just bought a onetouch ultrasmart, but I have a mac. I am considering buying the cable and software from Lifescan as I have access to a PC, but I'd rather not. I was thinking of writing an app for the mac, but I don't have the protocol and didn't get any info from Lifescan when I called them.

Anyone have the ultrasmart protocol?

Thanks

[dsimms]

Originally Posted by lar3ry

I ended up hanging the logic analyzer on it, and figured it out. They lied about the contents of a packet. They also don't tell about all the message types (like "clear meter"). I am using SniffUSB to figure them out.

By the way, using USB sniffer to accompany the protocol document is a good strategy. It's frustrating, but the Ultra2 protocol description is not very clear about the meter acknowledgment. (One might go so far as to call it a lie, but I can't figure out why J&J would do it purposefully.) If you follow the recipe using HyperTerm, it works but only because HyperTerm quietly does some CRLF -> CR conversion. No one has mentioned here yet, but the Ultra 2 meter acknowledgment is hex 11, 0d on the wire.

I might mention that when looking at the OneTouch Ultra, I found that misc.health.diabetes FAQ:

diabetes FAQ: bg monitoring (part 2 of 5)

to be helpful since it points out a non-lifescan description of earlier versions of the protocol:

ftp://vic.cc.purdue.edu/pub/

. Despite that Vic said that he faithfully copied the protocol description from a Lifescan memo, I think his is still much clearer.

Last, I'll ask: Does anyone have a published reference for the UltraSmart? It seems to look an awful lot like the Ultra 2 protocol so I don't know why it wouldn't be publishable.

[gnickm]

Originally Posted by dsimms

It's frustrating, but the Ultra2 protocol description is not very clear about the meter acknowledgment. (One might go so far as to call it a lie, but I can't figure out why J&J would do it purposefully.) If you follow the recipe using HyperTerm, it works but only because HyperTerm quietly does some CRLF -> CR conversion. No one has mentioned here yet, but the Ultra 2 meter acknowledgment is hex 11, 0d on the wire.

Wow, big ups dsimms, I have been pounding my head on my desk for the past two days trying to connect to my Ultra 2. Changed the ack from 11 0a 0d to 11 0d, and it worked like a champ!

Ugh, who wrote this spec? I'd like to introduce them to my middle finger.

Anyway, thanks again! Now to see if the bluetooth cradle will work...

[jjames]

I still can't believe that the US2 is still not "hacked" . . . kinda funny watching this thread!

[someone]

Originally Posted by jjames

I still can't believe that the US2 is still not "hacked" . . . kinda funny watching this thread!

I think someone here has successfully reverse engineered the protocol for the Ultra 2 but I guess they aren't going to give out the info. Too bad since I would love to see more open source diabetes software. If you want the protocol for the original Ultra, have a look at Glucomodul.