Lionel: I realize now (I'm a little slow ) that I didn't give your posts the attention they deserved.

Those adapters/cables would not be very useful for Chrisnet (it's a wildly different protocol than RS232), but they would be useful for an RS232 driver I've recently thought about writing for Punix. I could even transport IP packets over SLIP with the RS232 driver. The downsides are slower speeds (2400bps maximum) and only one calculator can be used on the network at a time. The benefit is a much simpler and less expensive connection (no hub is needed).

Edit: I shouldn't dismiss those cables for Chrisnet so quickly. I could use one for a Chrisnet hub, after all. They provide power and RS232 data signals that I could use for a hub. I wish they weren't as expensive, though (if they were under $10 I'd like to buy a few for various serial-related projects).

Edit 2: apparently the Nokia CA-53 data cable (and similar cables) is a USB to TTL RS-232 cable which goes for absurdly low on ebay and other places ($2-3). I might as well buy a handful of those right now.

So I learned my lesson on buying cables on ebay. A 99¢ cable from China advertised as a "Nokia CA-53" data cable is not going to work. I took mine apart and found that it's just a straight-through USB cable with no electronics in sight:



Maybe those crazy people thought they could get away with connecting the USB D+/D- wires to the RX/TX pins....

At least I didn't buy 10 of these junk cables!


Edit: It turns out that I read some incorrect information on the Internet (hard to believe, isn't it?). The CA-53 actually is a straight-through USB cable. The CA-42 cable contains electronics to convert USB to TTL-level RS232 serial data.

Edit 2: I just bought a CA-42 cable for $2.27. Hopefully that one will work.

Kerm: what's your preferred text representation of a CALCnet2.2 address?

1. ABCDEF0123 (all smushed together; I don't like this)
2. AB.CD.EF.01.23 (like an IPv4 address)
3. AB:CD:EF:01:23 (like a MAC or IPv6 address)
4. AB-CD-EF-01-23 (also like a MAC address)
5. AB/CD/EF/01/23 (like nothing I know)
6. some other format

I'm leaning toward #4 because it is the least ambiguous format with separators, and it resembles a MAC address (CN2.2 addresses are basically like MAC addresses in purpose and function).

I usually use #3, MAC-address style, mostly because CALCnet 2.2 addresses are hardware addresses. They're hard-coded into each calculator's certificate, so they most resemble MAC addresses in form and function to me.

That works for me (but I still prefer #4 because it's also used with MAC addresses and it's easier to type). I'm going to be using the same addresses with Chrisnet because they're a characteristic of the calculator and not specifically to CALCnet2.2 (as you pointed out), and I'm trying to avoid differences with Cn2.2 where such differences are not necessary.

My CA-42 cable came in! (Actually it arrived yesterday but my wife didn't tell me until late last night, so...)

I plugged it in and apparently it is a Prolific 2303, just like a real CA-42 cable. I tried shorting the RX and TX lines but got a bunch of garbage coming back in the terminal. Turns out that gpsd was sending data to it (curse you, ESR!!) which it promptly sent back to my terminal. After wrestling with gpsd for far too long, I tried it again and got a clean echo! Not bad for $2.27!

Here's the pinout for future reference (for myself or anyone else):

6 white RX 3.5V
7 green TX 0V (not pulled up)
8 blue GND

(Wire colors may vary.)

Nothing else was connected. I have not yet cracked open the USB case to see what signals are available directly from the circuit board (of course, I don't need any other signals for talking to a calculator). That's a project for another day. I started on this little project too late tonight as it is. Another project is to make a simple voltage divider to connect it to a TI link port (unless this cable can tolerate 5V+ coming from the link port).

Anyway, it's bedtime for me.

Kerm, in your experience with link ports on various calculators, have you found them to have a consistent pull-up resistor value? I want to make a voltage divider to convert 5V to about 3.3V with a single resistor if possible, but that depends on the pull-up resistance. I was thinking that if the pull-up resistor on the link port were 5k, I could use a 10k resistor between the link wire and ground, giving about 3.3V between the wire and ground. That should hopefully still register as a high value on the calculator (this matters only for the RX signal on the calculator).

Then again, the PL-2303 might not care if I use 5V directly. Do you think a small current at 5V would harm it?