HP Forum Archive 14

[ Return to Index | Top of Index ]

possible explanation for 49g+ missed keystrokes Message #1 Posted by Eric Lundgren on 22 Feb 2005, 12:18 a.m. Just got off the phone with a helpful HP support tech. He pointed out that displaying the clock would contribute to missed keystrokes, after fiddling with the machine further, it appears that I never miss keystrokes (pressing four '4's very rapidly generates '4444' on the stack) if the stack is clean. Once the stack gets 3 levels deep, I begin getting 44 instead of the inteded 444, and so forth. So could 'overclocking' the machine resolve this (as VPN suggested earlier)? If you read this, VPN, could you please help me understand how to accomplish 'overclocking' it? Also, will it cause damage over time to the hardware? I can't imagine clearing the stack every few calculations.. Eric

Re: possible explanation for 49g+ missed keystrokes Message #2 Posted by Raymond Del Tondo on 22 Feb 2005, 2:44 a.m., in response to message #1 by Eric Lundgren Hi, I think the only chance to get a somewhat stable entry mode, is to experiment with ->KEYTIME . There were several threads discussing this design flaw. Of course you have an alternative if input reliability is important: Just use an HP-48, because this machine normally doesn't miss key strokes;-) Raymond

Re: possible explanation for 49g+ missed keystrokes Message #3 Posted by bob on 22 Feb 2005, 3:16 a.m., in response to message #2 by Raymond Del Tondo ->KEYTIME has nothing to do with it. Overclocking will not help either. When the user strikes a key, it triggers an ARM interrupt. The ARM emulation layer must then fake a saturn interrupt. The problem is sometimes this takes too long (the OS has several tasks to juggle) and the key is no longer registering by the time the Saturn keyboard handler is called. It is more complicated then this, but that gives you the general idea.

Re: possible explanation for 49g+ missed keystrokes Message #4 Posted by Raymond Del Tondo on 22 Feb 2005, 5:42 a.m., in response to message #3 by bob Yes, I remember a discussion about that, and that there was a problem recognizing keys when the ARM is in low power state. I also think to remember that ->KEYTIME has some influence to the key registering even on the 49g+, but I won't search for the articles, because I don't care about the 49g+. Thanks god I sold my 49g+ ! After all, there are immanent design flaws, and the problems won't be resolved. So my suggestion is still active: Use an HP-48 instead;-) Raymond

Re: possible explanation for 49g+ missed keystrokes Message #5 Posted by Erik Ehrling (Sweden) on 23 Feb 2005, 9:57 a.m., in response to message #3 by bob Quote: ->KEYTIME has nothing to do with it. Overclocking will not help either. When the user strikes a key, it triggers an ARM interrupt. The ARM emulation layer must then fake a saturn interrupt. The problem is sometimes this takes too long (the OS has several tasks to juggle) and the key is no longer registering by the time the Saturn keyboard handler is called. It is more complicated then this, but that gives you the general idea. Hm... interesting. I have been running Hrastprogrammer's HP-42X on my HP-49G+ for over a half year now without any missed keystrokes. Maybe the "inner loop" of HP-42X is much lighter than that of the HP-49G itself? Best regards, Erik Ehrling (Sweden)

Re: possible explanation for 49g+ missed keystrokes Message #6 Posted by Veli-Pekka Nousiainen on 23 Feb 2005, 2:28 p.m., in response to message #5 by Erik Ehrling (Sweden) Maybe that program never allows the calc to go to Slow Mode (12MHz) during key entry or operator hesitating? I though the problem happens when ARM ramps up to 75MHz and user presses a key and clock updates the LCD all the same tick. [VPN]

Re: possible explanation for 49g+ missed keystrokes Message #7 Posted by Veli-Pekka Nousiainen on 22 Feb 2005, 6:41 a.m., in response to message #1 by Eric Lundgren Actually the 49g+ is underclocked, so overclocking does no damage (sue me:) ClockSpeed Adjustment Tools - Overclock your 49G+ 2.00: http://www.hpcalc.org/details.php?id=6081 Try running your calc at a fixed speed all the time I only overclock for some long programs (and go back at the end) Just avoid slow mode and displaying the clock [VPN]

[ Return to Index | Top of Index ]

Go back to the main exhibit hall