HP Forums
82182A Time Module question - Printable Version

+- HP Forums (https://archived.hpcalc.org/museumforum)
+-- Forum: HP Museum Forums (https://archived.hpcalc.org/museumforum/forum-1.html)
+--- Forum: Old HP Forum Archives (https://archived.hpcalc.org/museumforum/forum-2.html)
+--- Thread: 82182A Time Module question (/thread-209145.html)



82182A Time Module question - Monte Dalrymple - 01-14-2012

The Time module includes hardware and functions to adjust the accuracy factor for the time module clock. How important are these features?


The hardware required is almost exactly 10% of the total, and with this hardware included it is probably impossible to fit the logic into a low-power programmable logic device. Hence my question.

Monte


Re: 82182A Time Module question - Massimo Gnerucci (Italy) - 01-14-2012

What a pity... I've always found the accuracy factor management to be a real helper: I missed it in the newer calcs.

Greetings,
Massimo


Re: 82182A Time Module question - Håkan Thörngren - 01-14-2012

I think it is very useful, as it tends to make the clock quite accurate. I can never remember if it works with daylight savings hour (T+X). Hence, I tend to correct it once and live quite happily with it after that not trying to correct it again.

I would say it is important. The whole Time module is a piece of very well designed part of the HP-41 system. I am very especially impressed by the programming of it..


Re: 82182A Time Module question - Eric Smith - 01-14-2012

It does the right thing with T+X. In other words, it knows that T+X is not an adjustment for an error, so it changes both the current time, and the record of when the time was set, such that the effective elapsed time since setting remains the same.


Re: 82182A Time Module question - Garth Wilson - 01-14-2012

I suppose it depends on the cost versus accuracy of available crystals. Do you have room for a 3mm variable capacitor to trim the crystal frequency? I must have gotten lucky. My 41cx (time module is not separate) keeps very accurate time without ever applying a correction factor.


Re: 82182A Time Module question - Eric Smith - 01-14-2012

I've got the accuracy factor implemented in the firmware for the DIY4, though I couldn't do it the same way the timer chip hardware does. The way I've done it has the same long-term effect, but bunches the correction into a periodic "leap-centisecond" or missing centisecond. While having the correction occur closer to continuously would be better, this seemed like a reasonable compromise. Maybe something similar can be done in hardware with less logic required?


Re: 82182A Time Module question - Howard Owen - 01-14-2012

Perhaps you could condition the clock in software? How about NTP/HP-IL? (0.5 :))


Re: 82182A Time Module question - Monte Dalrymple - 01-14-2012

Hmm... back to the drawing board. If I restrict the use to the 41CL only perhaps I can pull enough of the functionality into the logic on the 41CL board to make everything fit.


Re: 82182A Time Module question - Monte Dalrymple - 01-14-2012

But that requires periodic communication with something outside the calculator. The advantage of the accuracy factor correction circuitry is that it is self-contained.


Re: 82182A Time Module question - Geir Isene - 01-14-2012

Needed.


Re: 82182A Time Module question - Eric Smith - 01-14-2012

Quote:
Needed.

That depends on a lot of factors, and without considering those factors I don't think it's fair to simply say that it is needed.

It is possible to get higher-stability 32.768 kHz oscillators than the 82182A/41CX used, which would reduce the need for the AF.


Re: 82182A Time Module question - Kerem Kapkin (Silicon Valley, CA) - 01-16-2012

Nice to have, but not necessary especially if it will make the time module form factor or system integration with 41CL much more difficult. Also I think with the current xtal and components the time keeping precision of the time module would be as good as a regular wrist watch.

In addition, personally I have no problem of the time precision of the WP-34S. If the same was achieved on 41CL with a time module, personally I don't see the need to use the AF adjustment.

Thank you again for making the 41CL a reality and investing your time and resources into the time module. Very much appreciated!

-Cheers, Kerem


Re: 82182A Time Module question - BobVA - 01-16-2012

I think a good solution would be to integrate as much of the timer functionality in the 41CL as you can and give the user the option of disabling it entirely in favor of an actual time module if they wish.

If the omitted functions are critical for a particular user they could then install an actual module and go with that.

The impact of no accuracy adjustment could be ameliorated by (as noted above) using a higher accuracy crystal, incorporating a physical trimmer or a even a TCXO.

I'd say it's well worth the effort of including. I know it's not a 100% solution, but for me it would be 95%. Given the price of used time modules, that's a significant accomplishment.

Bob


Edited: 16 Jan 2012, 12:00 p.m.


Re: 82182A Time Module question - Ángel Martin - 01-16-2012

Well said, that´s the perfect compromise I concur.

I´ve never used the accuracy adjustments on the CX - but having said that I have noticed significant errors in the TIME keeping using the TIME module on a CL, so it may well be more needed there.




Re: 82182A Time Module question - Monte Dalrymple - 01-16-2012

As far as the Time Module is concerned there shouldn't be anything specific to the 41CL that would affect the time-keeping accuracy. The bus still runs normally even during Turbo operation, and all of the Time Module code runs at 1x speed. Signal edge rates are probably different, but that should not affect the time. I suppose that noise on the power lines might have some effect...


As far as a TXCO, they draw a lot more current, like 220uA, so that is not really an option.