The HP 82240 uses a somewhat distinctive IR protocol developed in the 80's. I don't know what you would classify it as -- at first I thought it was 'biphase level', but that was a misinterpretation on my part. It does have some similarities to that, but the pulses are actually for 1/4 bit time, rather than 1/2 bit time. Another way of looking at it is that a marking signal is actually 1/4 bit time marking, and 1/4 bit time spacing.

I can appreciate the motivation for that: marks always have transition to detect their boundary, so a 0 followed by a 1 will still have a detectable transition, and because detecting the otherwise-illegal start sequence of 3 half-bit marks is easier. But, this does have a consequence: the pulse bursts are unusually short ('unusually' relative to the myriad of other IR protocols).

My hope is that I can use one of the Vishay TSOP IR detectors. This would be a boon for many reasons:

* cheap

* reduced component count (the whole thing is in a single 3-pin package with demodulated signal out)

* sophisticated: the unit includes amps, filters, AGC all in the unit

otherwise I will have to build the IR detector/demodulator from discretes&op-amps.

As mentioned, the major problem is that this IR protocol is unusual in the carrier frequency (32768 Hz -- wonder where that value came from? haha), and the timing involves comparatively short bursts followed by short quiescence.

The first part is not that big a deal because the parts typically have a fairly broad bandwidth -- I picked up the signal just fine on a part I have on-hand that has a center frequency of 36 KHz (TSOP4836), for example. The second part is trickier, though. On that same part the pulses were so stretched that the signal did indeed look like proper biphase level signalling. This cause me confusion for about a day when I was trying to do the reverse: to emit a signal to the printer. I couldn't see any significant difference between what I was generating, and my control signal generates by an HP-28C. But my signal would not print and the calculator's did. After several hours, I broke out a phototransistor and crufted a crappy detector together so I could see the raw modulated signal, and then it was clear that the TSOP was pulse-stretching. And it was all there in the spec, but you know how the brain and cognitive dissonance go, and I misinterpreted it as a standard line encoding that I already knew.

Anyway, Vishay has a ton of different parts, so I started rummaging though all the specs. The frequency part was trivial, they have a unit centered at 33 KHz, so that's solved. The timing was the tricky one. There are a couple more obscure IR protocols, such as RC-MM and RECS-80 and some Sharp that have bursts/quiescence of similar durations, and at length I found a part Vishay number TSOP4133 that just might work. I couldn't find any on eBay, so I ordered them from Mouser. Mouser will let you order just one, but you will be paying for shipping, so I usually order 10 of anything small like this. Who knows what I'll use the others for -- Christmas gifts, perhaps. Imagine mom's delight! Anyway, the upside of ordering from Mouser is that they ship quickly. I chose 'economy' shipping for this one, but I still expect that it will be here in the next week. Then I can test reception. I reeeeeally hope this unit works out..