andrew@earwax.pd.uwa.oz.au (Andrew Williams) (10/18/90)
Hi- I'm trying to build a receiver for the infra-red printer output on my HP28S. The data seems to be RS232 protocol, but getting the frequency right is a bit tricky. I assume thats its a carrier of around 40kHz, with a couple of signal tones for 1 and 0 around a few hundred Hertz. Does anyone out there know the exact frequencies, for both carrier and signal? (I dont have a storage CRO, and its pretty tricky triggering properly without one). With a bit of luck, I'll be able to make a 'Infra-red to RS232' converter, and keep up with the 48 owners (at least partially)! Andrew Williams, Physics Department, University of Western Australia. (andrew@earwax.uwa.oz.au)
chuck@mitlns.mit.edu (10/19/90)
-Message-Text-Follows- In article <andrew.656252611@earwax.pd.uwa.oz.au>, andrew@earwax.pd.uwa.oz.au (Andrew Williams) writes... >Hi- I'm trying to build a receiver for the infra-red printer output on my >HP28S. The data seems to be RS232 protocol, but getting the frequency >right is a bit tricky. I assume thats its a carrier of around 40kHz, with >a couple of signal tones for 1 and 0 around a few hundred Hertz. Does >anyone out there know the exact frequencies, for both carrier and signal? >(I dont have a storage CRO, and its pretty tricky triggering properly >without one). With a bit of luck, I'll be able to make a 'Infra-red to >RS232' converter, and keep up with the 48 owners (at least partially)! > Andrew Williams, > Physics Department, > University of Western Australia. > (andrew@earwax.uwa.oz.au) HP has a news letter that they publish, in it several years ago they gave a complete spec. I did it from a scope and it was damn hard but kind of fun to crack. It's a great way to get data out of your microprocessor project. Get the newsletter from HP for the real dope, but if you can't wait I'll follow with a bit of what I remeber, don't count on it being absolutly accurate I don't remeber so well after 2 years. This is what I remeber. Tone 32,768 Khz I'm pretty sure, I remember the striking match to a power of 2. I assumed it was to match up with the cheap crystals they use in watches. since I got it off a scope I only had it to 1% or so 8 bit ascii data with 4 error correction bits, also a 1 1/2 bit introduction bit. The correction bits come first. To generate the error correction bits you need to figure out 4 1 byte numbers. I did it by printing out the alphabet and figuring it out, HP has it in there news letter article. Basically each ECC bit is: BIT_i = PARITY(CODE_i.AND.ASCII) At least thats what I did, I'm sure ther are other ways of calculating it that look quite different but give the same result. The ones and zeros are sent as follows (S=silience T=carrier) tone, what is important is the is the transition between time slices 2 and 3, does it go tone to silence or silence to tone? 1 = SSTS 0 = STSS The tone bursts a 7 cycles long I think Chuck@mitlns.mit.edu