HP15C LE vs. HP15C

[rkf]

Now the moment has come, when I put my HP15C LE finally into my collections display cabinet, since he only shows up Pr Error, no way. If I understood previous discussions to this "feature" correctly, it's due to an error by design, maybe correctable with a new firmware.

I bought this unit last October from thecalculatorstore, so its age is only about nine months

Now I'm using again my good old HP15C of 1987 U.S.A. (maybe on its first batteries?) 100 times slower, but nevertheless working like a charm

[Mark Hardman]

(07-03-2014 03:40 PM)rkf Wrote:  Now the moment has come, when I put my HP15C LE finally into my collections display cabinet, since he only shows up Pr Error, no way.

This is likely the result of low battery. It has been documented that the battery low (*) indicator does not come on before the minimum required voltage is reached.

Based on this issue and the issue reported in your other thread, you need to invest in a few CR2032 cells.

[rkf]

(07-03-2014 04:38 PM)Mark Hardman Wrote:  ... Based on this issue and the issue reported in your other thread, you need to invest in a few CR2032 cells.

Of course I understand this implication, but look:

HP15C LE after eight plus x months of virtually no use: empty batts.
HP15C after many years of normal use: still full batts.

You get the point ?

[kakima]

Ah, but you admitted the original 15C is a hundred times slower than the LE. So 8 months of LE is equivalent to about 800 months or more than 60 years of the original. I doubt your original 15C has been running quite that long on its original batteries :-)

[rkf]

(07-03-2014 09:03 PM)kakima Wrote:  Ah, but you admitted the original 15C is a hundred times slower than the LE. So 8 months of LE is equivalent to about 800 months or more than 60 years of the original. I doubt your original 15C has been running quite that long on its original batteries :-)

Now that's really an argument. I give up for now.

[Mike Morrow]

An emulation layer exists between the fast ARM hardware and the old original HP-15C firmware of the HP-15C LE. That emulation causes the HP-15C LE to run about one-eighth the speed compared to HP-15C LE firmware running directly on the ARM hardware.

Although the HP-15C LE is fast, it takes eight times longer (using eight times the battery energy) than the same problem would running directly on the ARM without emulation. Emulating a legacy calculator chip set on calculator ARM hardware in order to inefficiently (and in the case of the HP-15C LE...imperfectly) run old original calculator firmware is a foolish idea...it produced a cute gimmick machine that has very limited practical value.

The infamous HP-15C LE battery hog would see that problem largely resolved if 85 percent of its battery energy wasn't being wasted by emulation for every task it performs. The original HP-15C has no such handicap.

[Thomas Radtke]

I doubt that a reimplementation of the 15C functionality would have been possible at HP without introducing a lot of bugs. And there is no reason to believe that in this case the defect brown-out detection would have been working, as it is already written in native code. Without emulation, the 15C LE likely wouldn't exists at all.

[Mike Morrow]

(07-04-2014 08:48 AM)Thomas Radtke Wrote:  I doubt that a reimplementation of the 15C functionality would have been possible at HP without introducing a lot of bugs.

Some bugs are inevitable...but look at the HP 30b. It is a native-programmed ARM machine that was contemporary with the HP-15C LE. While not 100 percent bug-free, it is amazingly close.

Quote:And there is no reason to believe that in this case the defect brown-out detection would have been working, as it is already written in native code.

All that means is that the native code programming was defective. That could be detected and corrected in the development phase.

Quote:Without emulation, the 15C LE likely wouldn't exists at all.

That is the core of the truth of the matter. The nearly 30-year-old complex and sophisticated HP-15C firmware was able be used intact without further development effort, except for the task of emulator creation. Critics of the HP-15C LE should recognize the compromises that were required to quickly produce the HP-15C LE...a machine intended for a rather small specialty market during a limited marketing period. I'm glad HP went to the effort to produce the oft-criticized but still "cute" HP-15C LE, but I doubt that it was a wise and worthwhile business decision.

The basic ARM hardware and clock speeds of the 30b and the 15C LE are similar. The 30b allows programming the scientific-function-oriented Savage benchmark. That benchmark runs eight times faster while producing a result of much greater accuracy and precision, compared to the 15C LE.

To perform this task, the 15C LE depleted its two CR2032 cells eight times what the 30b did. The gross waste of battery energy (not speed) caused by emulation is the HP-15C LE's most serious adverse characteristic...one shared by all emulated machines, including the HP 50G.

[Tugdual]

Suprisingly I don't have this battery issue with my 15CLE but to be fair I don't use it everyday. I find the LE more comfortable and though I found its design much cheaper than the original one, I finally adopted it for those reasons:
- much faster and much more comfortable to use with instant key stroke reaction.
- the original 15C keyboard is probably made of rubber or something that became extremely stiff and I can't really use it without pressing very hard on key with the risk to deteriorate it. Wonder if there is a way to fix that BTW...

[JimP]

This presents an interesting question -- of all the calculators available on the market today (including HP and "the rest" eg TI, Sharp, Casio, Canon, etc.) -- maybe a question answerable by an expert collector -- when compared head to head for battery longevity, which brand comes out the best?

I suppose there are potentials for lots of categories -- graphing, business, scientific, dual power (ie solar/battery) so it's not an uncomplicated question. There is also the type of battery -- alkaline vs. flat Li type. Maybe a silly one, though...

[Steve Simpkin]

(07-05-2014 02:24 AM)JimP Wrote:  This presents an interesting question -- of all the calculators available on the market today (including HP and "the rest" eg TI, Sharp, Casio, Canon, etc.) -- maybe a question answerable by an expert collector -- when compared head to head for battery longevity, which brand comes out the best?

Excepting solar powered calculators, I don't know any models from any manufacturer, past or present, that could last as long on a set of batteries as the original Voyager series. 22+ years is an amazing lifespan for batteries sitting new in the package let alone for a set that powers a calculator in regular use AND keeps continuous memory intact 24/7. Like landing a man on the moon, I doubt we will ever see anything like that again in our lifetimes.

[jebem]

(07-03-2014 08:38 PM)rkf Wrote:  

(07-03-2014 04:38 PM)Mark Hardman Wrote:  ... Based on this issue and the issue reported in your other thread, you need to invest in a few CR2032 cells.

Of course I understand this implication, but look:

HP15C LE after eight plus x months of virtually no use: empty batts.
HP15C after many years of normal use: still full batts.

You get the point ?

I have just acquired a vintage HP-15C to join my modest collection.

Concerning the batteries subject, according to the original HP-15C Owner's Guide:

"(...) In typical use, the HP-15C has been designed to operate 6 months or more on a set of alkaline batteries. The batteries supplied with the calculator are alkaline, but silver-oxide batteries (which should last twice as long) can also be used.
A set of three fresh alkaline batteries will provide at least 60 hours of continuous program running (...)
(...) If the calculator is being used to perform operations other than running programs, it uses much less power. (...)

(...) If the calculator remains turned off, a set of fresh batteries will preserve the contents of Continuous memory for as long as the batteries would last outside of the calculator - at least 1 1/2 years for alkaline batteries or at least 2 years for silver-oxide batteries. (...)"

As I see it, this was good design engineering, taking into account the chosen battery characteristics.
As for the HP-15C LE, more processing power really consumes more current, despite all the energy optimizations available in the processor power management circuits.

[Katie Wasserman]

(07-05-2014 07:17 AM)Steve Simpkin Wrote:  Like landing a man on the moon, I doubt we will ever see anything like that again in our lifetimes.

You're probably right about that. The Voyagers used SOS (Silicon on Sapphire) CMOS technology that HP had developed just a couple of years earlier. Using a sapphire insulating base for the silicon made for ultra-low current leakage transistors. SOS (and similar tech) is still used for some wireless technology chips, but it's more expensive than bulk silicon. The real problem is/was that it's not that widely used and that economy of scale meant that most "fabs" (semiconductor fabrication plants) just don't/didn't produce chips using it and, of course, there are many fewer fabs than there used to be.

[Thomas Radtke]

My 15C LE still runs on its first set of batteries. I guess energy consumption when idling and running is reasonable and probably much more efficient than with e.g. the Clamshells. The defect brown-out detection is the problem here, and I'd even pay for a firmware that corrects for it.

[jebem]

One (big) difference between HP-15C and the 15C LE is the total battery capacity:

- The HP-15C works with three 1.5Volt batteries in Series, for a total of 4.5Volt providing just the equivalent capacity of a single battery (150mA/h for Alkaline or 200mA/h for Silver-Oxide types).

- The HP-15C LE works with two 3Volt batteries in Parallel, for a total of 3Volt providing a total capacity equivalent to two batteries (total of 450mA/h for CR2032 Lithium-manganese dioxide type).

[Katie Wasserman]

(07-06-2014 12:37 PM)jebem Wrote:  One (big) difference between HP-15C and the 15C LE is the total battery capacity:

- The HP-15C works with three 1.5Volt batteries in Series, for a total of 4.5Volt providing just the equivalent capacity of a single battery (150mA/h for Alkaline or 200mA/h for Silver-Oxide types).

- The HP-15C LE works with two 3Volt batteries in Parallel, for a total of 3Volt providing a total capacity equivalent to two batteries (total of 450mA/h for CR2032 Lithium-manganese dioxide type).

You're conclusion is correct, but just to be explicit about this .....

Using Energizer specs sheets:

Total energy available for three 357 (silver oxide) cells is nominally 3 x 150mAh x 1.55 volts = 697.5 mWh.

Total energy available for two CR2032 is nominally 2 x 240mAh x 3.0 volts = 1440mWh.

[Katie Wasserman]

(07-06-2014 08:23 AM)Thomas Radtke Wrote:  My 15C LE still runs on its first set of batteries. I guess energy consumption when idling and running is reasonable and probably much more efficient than with e.g. the Clamshells.

The clam shells use alkaline N cells which have a relatively high self-discharge. But they also have built in RTC (real time clocks) and that is the main battery drain even when off.

The 15CLE's power consumption in the off state is about 4uA, that translates to about 13 years of battery life when off. It's 20mA when pressing a key, that's just a few hours if a key is held down continuously.

Quote:The defect brown-out detection is the problem here, and I'd even pay for a firmware that corrects for it.

I agree, this, the PSE (pause) problem and the possibility of corrupting memory when the old self-tests are run are the biggest issues with 15CLE. It sure would be nice to have new firmware available for download.

[Tugdual]

(07-06-2014 04:27 PM)Katie Wasserman Wrote:  I agree, this, the PSE (pause) problem and the possibility of corrupting memory when the old self-tests are run are the biggest issues with 15CLE. It sure would be nice to have new firmware available for download.

I stopped dreaming about that. Also I read somewhere that unlike the 30b, the 15CLE cannot be flashed though it has the same connector.

[Mike Morrow]

(07-06-2014 04:27 PM)Katie Wasserman Wrote:  The clam shells use alkaline N cells which have a relatively high self-discharge. But they also have built in RTC (real time clocks) and that is the main battery drain even when off.

I never owned any of the unfortunate clamshells except the HP 28C and HP 28S. These do not have RTCs. I never had much issue with their battery life, except for the battery door area breakage that occurred, especially in the S model, if batteries were left installed for a long time. :-)

Quote:...this, the PSE (pause) problem and the possibility of corrupting memory when the old self-tests are run are the biggest issues with 15CLE.

That defective brown-out detection discussed here creates the very serious problem of memory corruption that often occurs when the 15C LE turns off due to battery exhaustion. It would be only a minor nuisance if all that this defect caused was the machine unexpectedly turning off. This is IMHO the major flaw in the HP 15C LE that the original HP 15C did not have.

[Katie Wasserman]

(07-06-2014 05:18 PM)Tugdual Wrote:  Also I read somewhere that unlike the 30b, the 15CLE cannot be flashed though it has the same connector.

What you read is absolutely wrong!

The 15CLE uses the same hardware as the 12C+ (also 12CAE) and they both flash just fine via the 6 point connection.