The Rogue HP-12c

[Surfin]

(10-11-2024 05:43 PM)AnnoyedOne Wrote:  

(10-11-2024 05:29 PM)Surfin Wrote:  Just to check if now we have the cable connection.

If there is one probably not serial or USB.

I guess it´s I2C, also serial but almost sure it´s not the same protocol as used before.

(10-11-2024 05:43 PM)AnnoyedOne Wrote:  

(10-11-2024 04:23 PM)Idnarn Wrote:  In my units, the display shows 1.L 2.C 3.h 4.CPu...

What does "4.CPu" report? The GPLB31A version says "3886160 h2". The "rogue"?

A1

For record, my Prestige reports "4029814 h2"
And also:

Quote:ChE--564Ah
Ver 02
C 2004 hP

But the older Platinum* model, all silver and no Undo button:

Quote:CHE--0EdFH
VEr 01

no '2004 hP' at the end, there is also no 4.CPu, just "1.LCD 2.CHE" options.

[AnnoyedOne]

(10-11-2024 06:56 PM)Surfin Wrote:  I guess it´s I2C, also serial but almost sure it´s not the same protocol as used before.

Maybe so. My 2022 12C+ looks the same. USB (HP-15C CE cable). Confirmed. I've sucked and saved the firmware,

A1

[cheesy]

The 4. CPU value should change each time you run it, pretty sure it’s just reporting the clock speed

[RPNerd]

(10-11-2024 04:59 PM)Idnarn Wrote:  An example of a monochrome LCD that I used recently which was impressive is the one on the Garmin eTrex SE GPS receiver. I believe it's a Sharp memory LCD: https://sharpdevices.com/memory-lcd/
Perhaps some of the Swiss Micro flagship calculators use these displays too.

They do. This is exactly what the DM42 and similar use.

[AnnoyedOne]

(10-12-2024 04:02 AM)cheesy Wrote:  The 4. CPU value should change each time you run it, pretty sure it’s just reporting the clock speed

Could be!

From the GPLB31A datasheet I have

Quote:Max. operating speed: 4.0MHz @ 2.4V - 3.6

...

6.14. Mask Options

6.14.1. 32768 oscillator

1). X’TAL
2). R-oscillator

From the GPL833F datasheet

Quote:Built-in X’tal 32.768KHz oscillator circuit for real time clock function and system operating clock.

Built-in PLL 2.048MHZ/4.096MHZ/8.192MHZ for system operating clock.

So the latter can operate at twice the speed (8MHz vs 4MHz).

If a GPLB31A based HP-12C Platinum/Prestige does show the system clock speed with the 4.CPu item the numbers for them would make sense (around 4,000,000). By contrast a GPL833F one would display around 8,000,000.

I've no idea what the "h2" means though. Could be a chip rev ID. I'd expect the GPL833F to be different if so.

A1

PS: 4,000,000 / 32768 ==> 122 for the PLL (phase locked loop) multiplier.
8,192,000 / 32768 ==> 250

32kHz crystals are usually accurate to ppm (parts per million).

It is possible that the firmware "throttles" the clock speed to save power,

[cheesy]

(10-12-2024 12:42 PM)AnnoyedOne Wrote:  I've no idea what the "h2" means though. Could be a chip rev ID. I'd expect the GPL833F to be different if so.

A1

I'd hazard a guess that it's Hz

[AnnoyedOne]

(10-12-2024 01:24 PM)cheesy Wrote:  I'd hazard a guess that it's Hz

Makes you wish for a Swissmicros clone with a dot-matrix LCD vs a 7-segment one.

A1

[dm319]

(10-11-2024 06:56 PM)Surfin Wrote:  no '2004 hP' at the end, there is also no 4.CPu, just "1.LCD 2.CHE" options.

Same here with the rogue - no 4th option.

[AnnoyedOne]

(10-13-2024 09:34 PM)dm319 Wrote:  Same here with the rogue - no 4th option.

Yeah. That's to be expected. The "4.CPu" menu option is only present in the HP-12C Platinum/Prestige v2 firmware from 2004 or so. The early "silver" Platinum/Prestige models have v1 firmware and thus don't have that function.

Since you've already verified that the "rogue" doesn't have any hidden extra Platinum functions present (e.g. g x^2) that option being absent isn't really surprising. Unless pressing '4' in the menu works anyway (unlikely).

Nor is it a HP NUT emulation per my HP-12C+. You've proven that also.

So what is it? The firmware probably can't be "sucked" (no programmer and/or it is "secured") so still a mystery. At least for now.

A1

[lvt]

(10-10-2024 06:58 PM)dm319 Wrote:  I am so confused now! I thought Platinums were made in China? Does anyone fancy getting hold of one of these and we can scratch our collective heads over another 12c?

My Platinum was made in the Philippines.

[Idnarn]

@dm319 I came across the following today:

https://www.msn.com/en-us/money/news/the...r-AA1qWIEy
https://www.theregister.com/2024/09/21/h...r_mystery/
https://dm319.github.io/pages/2024_09_09...comma.html

Congratulations on the media coverage.

I want to ask you about something in your GitHub article:

Quote:

Quote:What interest rate would balance a payment of $50 upfront, returning you $30 each year for 10 years, with $400 to be paid at the end?

This would be a very unusual arrangement, and fittingly, there is an unusual answer in that there are two possible interest rates which would satisfy this. One is around 14.4% and the other 53.2%. No HP financial calculator (or many others) will attempt to return a result on this, but interestingly in a rare win, the rogue HP-12c gives us 14.44… and an accuracy of 8.7. Not bad.

Is this correct? My newly acquired Moravia HP 12C and a recent Moravia HP 12C Platinum model both solve it with i=63.31%, the new 12C solving for i almost 10 times faster than the 12C Platinum. I've assumed that the interest is compounded annually which is not explicitly mentioned, and $400 returned at end.

[Idnarn]

(11-16-2024 01:12 PM)Idnarn Wrote:  @dm319 I came across the following today:

https://www.msn.com/en-us/money/news/the...r-AA1qWIEy
https://www.theregister.com/2024/09/21/h...r_mystery/
https://dm319.github.io/pages/2024_09_09...comma.html

Congratulations on the media coverage.

I want to ask you about something in your GitHub article:

Quote:This would be a very unusual arrangement, and fittingly, there is an unusual answer in that there are two possible interest rates which would satisfy this. One is around 14.4% and the other 53.2%. No HP financial calculator (or many others) will attempt to return a result on this, but interestingly in a rare win, the rogue HP-12c gives us 14.44… and an accuracy of 8.7. Not bad.

Is this correct? My newly acquired Moravia HP 12C and a recent Moravia HP 12C Platinum model both solve it with i=63.31%, the new 12C solving for i almost 10 times faster than the 12C Platinum. I've assumed that the interest is compounded annually which is not explicitly mentioned, and $400 returned at end.

Solving manually in a spreadsheet, I think I misunderstood the problem. $50 + $400 is paid out. So the two solutions you've offered look fine.

[dm319]

(11-16-2024 01:12 PM)Idnarn Wrote:  @dm319 I came across the following today:

https://www.msn.com/en-us/money/news/the...r-AA1qWIEy
https://www.theregister.com/2024/09/21/h...r_mystery/
https://dm319.github.io/pages/2024_09_09...comma.html

Congratulations on the media coverage.

A bit of fame!

(11-16-2024 02:10 PM)Idnarn Wrote:  Solving manually in a spreadsheet, I think I misunderstood the problem. $50 + $400 is paid out. So the two solutions you've offered look fine.

Yes it's an odd one - originally posed by Dieter back here in 2012 when suggesting improvements to the TVM solver for the WP34s:

https://www.hpmuseum.org/cgi-sys/cgiwrap...ead=234439

AFAIK no HP TVM solvers solve this, other than the rogue 12c.

The plot is quite interesting:

[dm319]

I had a little brainwave yesterday - the rogue HP-12c exhibited some quite specific behaviour when manipulating numbers at the limit of the digit precision (16) of the device. I was curious if this behaviour was similar to the latest Casio calculators (checking on an fx-85gt cw), but it seems like the Casio calculators easily handle calculations going to 17 digits. I.e. 1e16-3, then ANS-1e16 gives -3.

So my idea that it might be based on a similar platform to a Casio is probably not the case.

[Albert Chan]

(11-17-2024 09:52 AM)dm319 Wrote:  [Negated NPV] plot is quite interesting

It had plotted rate below -100%, which does not exist.
Rate search domain can be reduced further, using edge rates.

PMT/FV = 30/-400 = -0.075
PMT/-PV = 30/50 = 0.60

Also, we should dampen compounding effect to straighten curve.
FYI, this is the curve used by Plus42 to solve rate.

npmt = NPMT/n = NPV * r/(1-1/(1+r)^n)

plot (((-400-50)/((1+r)^10-1)-50)*r+30), r = -0.075 .. 0.60

(11-17-2024 10:07 AM)dm319 Wrote:  Casio calculators easily handle calculations going to 17 digits. I.e. 1e16-3, then ANS-1e16 gives -3.

That's only 16 digits. For 17 digits test, try 1e16 + 3 - 1e16

[lvt]

The firmware test on my 12C shows

ChE - - 4141h
2015-01-05
C 2008 hP

What does the second line mean, was it the date when the chip has been programmed? Tks.

[RPNerd]

(11-19-2024 04:57 AM)lvt Wrote:  The firmware test on my 12C shows

ChE - - 4141h
2015-01-05
C 2008 hP

What does the second line mean, was it the date when the chip has been programmed? Tks.

I would expect it to be the build date of the binary.

[lvt]

(11-19-2024 06:31 AM)RPNerd Wrote:  

(11-19-2024 04:57 AM)lvt Wrote:  The firmware test on my 12C shows

ChE - - 4141h
2015-01-05
C 2008 hP

What does the second line mean, was it the date when the chip has been programmed? Tks.

I would expect it to be the build date of the binary.

It could be, my HP12C was made in the 4th week of 2015.

[lvt]

BTW there is a bug in the firmware that has been loaded in my 12C.

While checking firmware info (test mode?), if I don't exit the test mode (for example by pressing ON button twice), after a while the keyboard is not responsive anymore and the calculator will be stuck "forever" until I open the battery compartment and use the Reset button.