HP 50g/HP 49g+/HP 39gs Battle of the beams -- Increasing IR communication range

[Robert VM]

I've been rather lucky obviously to get a secondhand 49g+ and 50g for a reasonable price. I couldn't resist taking a look inside the 49g+ especially at the IR tranceiver. See if there wasn't an easy way to increase IR LED drive current. There seemed little hope as I couldn't find any reports of such modification while it's been around for more than 10 years.
Other people found out how to open: 49g+ cut the heads off the rivets in the battery compartment; 50g and 39gs screen cover has to be lifted to access the two screws underneath (a suction cup from e.g. a satnav device will do). Guide a guitarpick or other such implement between the bottom half and the side ridge to pry open the case.
The IR transceiver is a six pin device.
   
Pc board tracks give some hints as to pin function. Leftmost pin is connected to a large pcb area, the only wider track is connected to the next pin. Probably GND and supply positive.
I found three brands of IRDA transceiver with 6 pins, all having the same pinout consistent with pcb track layout. Sharp, zilog and Vishay.
   
Finally a resistor connected in series with the rightmost pin, kathode of IRED according to the datasheets, was enough hint to do a little experiment with the series resistor. To encourage the experimenter the series resistor is labeled R1 on the pcb (49g+). The HP 49g+ would serve as the guinea pig, 50g was next if everything went well.
   
(This is a 7 pin device (as used in the 39gs?) almost same pinout)
The resistor is EIA coded 30B (2KOhm). Supply is 3.3V, IRED volt. about 1.8V, gives 0.75 mA IRED current.
Considering the currents mentioned in the datasheets, I thought it safe to increase the current to 7.5 ma.
   
First experiment was with a 330 ohm resistor, was later changed to 200. The change was rather noticable visually (with a camera). While originally nothing would show on the camera during printscreen, IRED output is very visible now. (example fotograph on website: http://www.site2241.net/october2008.htm).
Increasing the current further would be more of a problem for the 3.3V PSU than for the IRED I guess.
IR communication range of the pair (49g+/50g) is now about 30 cm, near halogen desk lamp. (Both devices need to be modified for 2-way comm)
HP 50g: resistor is in slightly different place, same value. In the picture the resistor has been replaced.
   
HP 39gs: 7 pin IRDA transceiver, almost the same pinout (IRED39-01). I haven't been able to test this device except visually. I think this device is easier to open, starting from the top. While the 50g can do very well without the two screws, the 39gs opens too easily without them.
(39gs pic in next post)
Robert

[Robert VM]

hp 39gs IRDA transceiver:
   

Robert

[Dimitri]

Interesting. If the calculator is IrDA compliance, the wavelength of the infrared diode should be between 850 nm and 900 nm. Hoping the 50G, the 49G+ and all other calculators use the same wavelength to maximize the communication range.
The HP48S(X) anf HP48G(X) use a 940 nm wavelenth.
Another problem is the emplacement of the emitter and the receiver IR diode. They are inverted between the 48S/G and the 50G.

[rprosperi]

(01-15-2018 07:10 PM)Dimitri Wrote:  Hoping the 50G, the 49G+ and all other calculators use the same wavelength to maximize the communication range.

While I have no exact numbers, I know HP "tweaked" the IR on some late model machines (perhaps all post 48-series?) to limit IR range, as many schools objected that this could be used for students to exchange data. This is the same reason the Prime has no IR at all.

[Robert VM]

(01-15-2018 07:10 PM)Dimitri Wrote:  The HP48S(X) anf HP48G(X) use a 940 nm wavelenth.
Another problem is the emplacement of the emitter and the receiver IR diode. They are inverted between the 48S/G and the 50G.

A little test revealed that indeed rx/tx are switched (39gs <> 48G).
Differently placed emitter and receiver will be less of a problem though with increased range.
While difference in wavelength may shorten the range the main problem will rather be IRDA handshaking/negotiating which afaik the 48 does not do.

Robert

[Robert VM]

Another thread discussing range and wavelength.

And a quote from the 82240B IR printer Techn. interfacing guide:

The transmitting element is an infrared (IR) light-emitting diode. The optimum optical wavelength for the printer is 940 nanometers (nm), but wavelengths as short as 850 nm and as long as 1000 nm can be used with only minor reductions in the distance over which data can be sent to the printer.

I don't own an IR printer so haven't been able to do any test with it.

Robert

[Robert VM]

I haven't really tried to find max distance for printing but printing goes well at 30cm with printer 82240B.