Author Topic: EEVblog #1159 - World's Most Precise Pocket Calculator  (Read 21416 times)

0 Members and 1 Guest are viewing this topic.

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 16560
  • Country: 00
Re: EEVblog #1159 - World's Most Precise Pocket Calculator
« Reply #25 on: December 20, 2018, 10:06:41 am »
HP does the 15C (scientific) for Android:
For scientific I will keep my HP48 simulator instead, though. It does not replace my physical calculator, but it is darn convenient.

I like the shape/format of the Voyagers but my favorite in use is the HP20S. Not much to look at, but quite a beast under the covers. Plus: I can do hex<->decimal conversions and inches<->mm with a single (shifted) key.

Not RPN though, so HP purists will hate it.
 

Offline Grapsus

  • Regular Contributor
  • *
  • Posts: 242
  • Country: fr
Re: EEVblog #1159 - World's Most Precise Pocket Calculator
« Reply #26 on: December 20, 2018, 02:11:55 pm »
I'm genuinely curious, do you still have an actual use for a standalone calculator in your everyday life ?

I mean, I love calculators, and I always have a TI 86 (my favorite calculator of all times) on my desk. But with modern computers and high level programming languages I tend to always have a window open with interactive Python or Haskell. Those are so expressive and so handy that they tend to complete replace my use of the TI 86. Even when I do amateur electronics, the fact that I can copy paste from datasheets and write to files is a game changer. How about you ?
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 16560
  • Country: 00
Re: EEVblog #1159 - World's Most Precise Pocket Calculator
« Reply #27 on: December 20, 2018, 02:41:40 pm »
Used my HP20C about ten minutes ago.

5 Volts divided by 33 Ohms....  :popcorn:

(was feeling too lazy to use my slide rule)

Would a Python shell have been more convenient for that?
« Last Edit: December 20, 2018, 02:51:47 pm by Fungus »
 

Offline rsjsouza

  • Super Contributor
  • ***
  • Posts: 5980
  • Country: us
  • Eternally curious
    • Vbe - vídeo blog eletrônico
Re: EEVblog #1159 - World's Most Precise Pocket Calculator
« Reply #28 on: December 20, 2018, 03:09:52 pm »
Grapsus, I usually use my two calculators to do base conversions, arithmetic and, through the power of programming, do some Q arithmetic and floating point conversion. They sit quite conveniently on my desk.

At work I am frequently having to debug these types of data directly on a device's memory.
Vbe - vídeo blog eletrônico http://videos.vbeletronico.com

Oh, the "whys" of the datasheets... The information is there not to be an axiomatic truth, but instead each speck of data must be slowly inhaled while carefully performing a deep search inside oneself to find the true metaphysical sense...
 

Offline Grapsus

  • Regular Contributor
  • *
  • Posts: 242
  • Country: fr
Re: EEVblog #1159 - World's Most Precise Pocket Calculator
« Reply #29 on: December 20, 2018, 03:30:35 pm »
Used my HP20C about ten minutes ago.

5 Volts divided by 33 Ohms....  :popcorn:

(was feeling too lazy to use my slide rule)

Would a Python shell have been more convenient for that?

That might be scary but I would say yes if I'm at my computer when I need this computation.
I would bring the menu with the windows key, type "ipyt", enter, that would open the python shell and I would type "5./33", enter:

In [1]: 5./33
Out[1]: 0.15151515151515152


The whole operation literally took under two seconds.
 

Offline HKJ

  • Super Contributor
  • ***
  • Posts: 2899
  • Country: dk
    • Tests
Re: EEVblog #1159 - World's Most Precise Pocket Calculator
« Reply #30 on: December 20, 2018, 04:19:15 pm »
It looks like a very nice calculator, but I do not have any use for it. When out I use my phone and when home I use one on the computer I wrote many years ago (File date says 2012 but it is older), it works fine on newest Windows:

It is a bit more than just a calculator.
 

Offline 0xdeadbeef

  • Super Contributor
  • ***
  • Posts: 1570
  • Country: de
Re: EEVblog #1159 - World's Most Precise Pocket Calculator
« Reply #31 on: December 20, 2018, 04:39:00 pm »
the forward and reverse sin/cos/tan seem to work fine in the apps
so I went to try something on the app calc, 5000 nPr 5000
the answer given is 422.857 792 66 x10^16323
If "nPr" is supposed to mean "Nth Power of", this result seems to be totally wrong.
The correct result should be 0.7079811261048173e+18495.

so I hard reset it, and tried 6000 nPr 6000. this took 4.5s
26.839 997 657 x 10^20064
Which seems to be also wrong as the result should be 8.0816921258603719e+22668.


so I got curious, what can calculate 99999 nPr 99999? I tried the soft calc in windows. it produce error message and nearly jams the desktop.
for 10seconds worth, it is about 45000 n!
so which computer in this desktop world can calculate 99999 nPr 99999 in under 10seconds? intel i9? AMD ryzen?
On my  Core i7-6700K, the Java Calculator I wrote in 2004/2005 and updated now and then takes about a second to come up with 0.3678812805793781e+499995.
It uses a BigDecimal fractional representation with "endless" rational accuracy, so I would hope the result is correct. At least Wolfram Alpha seems to come up with the same result:
https://www.wolframalpha.com/input/?i=99999**99999
« Last Edit: December 20, 2018, 04:40:54 pm by 0xdeadbeef »
Trying is the first step towards failure - Homer J. Simpson
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 16560
  • Country: 00
Re: EEVblog #1159 - World's Most Precise Pocket Calculator
« Reply #32 on: December 20, 2018, 04:43:39 pm »
the forward and reverse sin/cos/tan seem to work fine in the apps
so I went to try something on the app calc, 5000 nPr 5000
the answer given is 422.857 792 66 x10^16323
If "nPr" is supposed to mean "Nth Power of", this result seems to be totally wrong.

Luckily it isn't, it's a permutation.
 

Offline Howardlong

  • Super Contributor
  • ***
  • Posts: 5315
  • Country: gb
Re: EEVblog #1159 - World's Most Precise Pocket Calculator
« Reply #33 on: December 20, 2018, 04:46:34 pm »
I also have a HP35S which is a nice compromise and not bad for a modern device.
:-+ :-+ :-+
Although a bit more cumbersome to do base conversion when compared to my 48GX, the 35S is quite nice. (I leave my 48GX at home as I am afraid it will be stolen).

The base conversion aspect is my biggest dislike on the 35S.

Some other comments...

I recently got around to getting a WP-34S, and I'm really liking it despite the fact that some of the common scientific functions aren't primary key operations. My only calculator with a direct Log base 2 function. It also has a satisfyingly complicated looking keyboard, a key requirement for any self respecting nerd.

I agree about the apps on phones, or indeed tablets or a PC, I find that the user experience is compromised badly.

Quote
do you still have an actual use for a standalone calculator in your everyday life

Yes, several times a day, both for programming and electronics. See my note above about user experience. It has to be sitting on the desk permanently, if it's kept in a drawer it's far less likely to be used.
 

Offline 0xdeadbeef

  • Super Contributor
  • ***
  • Posts: 1570
  • Country: de
Re: EEVblog #1159 - World's Most Precise Pocket Calculator
« Reply #34 on: December 20, 2018, 05:00:04 pm »
Oops. So you mean "n!/(n-k)!"? While I can't see the benefit of using that instead of n! which will give the same number obviously for n==k (as 0!=1).

fac(99999)/fac(99999-99999)
        out = 28.2422940796034787e+456567

Needs about a second.
Trying is the first step towards failure - Homer J. Simpson
 

Offline Grapsus

  • Regular Contributor
  • *
  • Posts: 242
  • Country: fr
Re: EEVblog #1159 - World's Most Precise Pocket Calculator
« Reply #35 on: December 20, 2018, 05:59:06 pm »
I agree about the apps on phones, or indeed tablets or a PC, I find that the user experience is compromised badly.

Could you elaborate on how a full keyboard and a 20+" color screen compromise the user experience ?

Meanwhile in GHCi:
Prelude> let n `nPr` r = product [n-r+1..n]
Prelude> 26 `nPr` 4
358800
Prelude> 6000 `nPr` 6000
2683999765726739596116316647462735512205018...

(it actually spits all the digits without rounding and it's instantaneous)
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 16560
  • Country: 00
Re: EEVblog #1159 - World's Most Precise Pocket Calculator
« Reply #36 on: December 20, 2018, 06:21:45 pm »
I agree about the apps on phones, or indeed tablets or a PC, I find that the user experience is compromised badly.

Could you elaborate on how a full keyboard and a 20+" color screen compromise the user experience ?

Meanwhile in GHCi:
Prelude> let n `nPr` r = product [n-r+1..n]
Prelude> 26 `nPr` 4
358800
Prelude> 6000 `nPr` 6000
2683999765726739596116316647462735512205018...

(it actually spits all the digits without rounding and it's instantaneous)

You had to type in all that " let n `nPr` r = product [n-r+1..n]" rubbish.

On my HP 20S it's a single button press.

200 nPr 200 = 7.88657867356e+374
 

Offline emece67

  • Frequent Contributor
  • **
  • !
  • Posts: 614
  • Country: 00
Re: EEVblog #1159 - World's Most Precise Pocket Calculator
« Reply #37 on: December 20, 2018, 06:59:18 pm »
.
« Last Edit: August 19, 2022, 02:06:03 pm by emece67 »
 
The following users thanked this post: jklasdf

Offline 0xdeadbeef

  • Super Contributor
  • ***
  • Posts: 1570
  • Country: de
Re: EEVblog #1159 - World's Most Precise Pocket Calculator
« Reply #38 on: December 20, 2018, 07:02:38 pm »
On my HP 20S it's a single button press.

200 nPr 200 = 7.88657867356e+374
Letting aside that this obviously more than one  button press, the result is actually wrong in the two least significant digits displayed. Or maybe this was a mistake when copying typing in the number.
Which is also something that is very convenient about PC based calculators: you can just copy/paste values.

For the record: the correct result is
788657867364790503552363213932185062295135977687173263294742533244359449963403342920304284011984623904177212138919638830257642790242637105061926...

So the last three digits should be 365 (rounding up) and not 356.
Trying is the first step towards failure - Homer J. Simpson
 

Offline 0xdeadbeef

  • Super Contributor
  • ***
  • Posts: 1570
  • Country: de
Re: EEVblog #1159 - World's Most Precise Pocket Calculator
« Reply #39 on: December 20, 2018, 07:05:28 pm »
The last time I looked at this, not all the firmware was open source. The manufacturers arranged things so that they supplied a closed source "OS" that can only run one application, a port of Thomas Okken's FreeHP42. Thus we have here an embedded system running GPL code but with not all of its firmware open.
There was a dispute this year but all the source code is available now since several months.
https://github.com/swissmicros/free42
Trying is the first step towards failure - Homer J. Simpson
 
The following users thanked this post: iMo

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 16560
  • Country: 00
Re: EEVblog #1159 - World's Most Precise Pocket Calculator
« Reply #40 on: December 21, 2018, 10:15:08 am »
On my HP 20S it's a single button press.

200 nPr 200 = 7.88657867356e+374
So the last three digits should be 365 (rounding up) and not 356.

Yep, I copied it wrong, it should be: 7.88657867365e+374
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 16560
  • Country: 00
Re: EEVblog #1159 - World's Most Precise Pocket Calculator
« Reply #41 on: December 21, 2018, 10:22:21 am »
now im confused, how did they do more than 200 digit precision on android?

The exact same way they do numbers with "single" precision - they write some code to do it.
 
The following users thanked this post: 3roomlab

Online iMo

  • Super Contributor
  • ***
  • Posts: 4674
  • Country: nr
  • It's important to try new things..
Re: EEVblog #1159 - World's Most Precise Pocket Calculator
« Reply #42 on: December 21, 2018, 01:06:08 pm »
Quote
now im confused, how did they do more than 200 digit precision on android?
It is called arbitrary precision.
For example with the "decNumber" math library (used in WP-34s) you may go for up to a billion digits of precision and 9-digit exponents (by changing a single parameter in the source).
The library was written by the IBM btw., not the Digital Equipment Corp. :)
 
The following users thanked this post: 3roomlab

Offline Poupsfamily

  • Newbie
  • Posts: 1
Re: EEVblog #1159 - World's Most Precise Pocket Calculator
« Reply #43 on: January 01, 2019, 06:17:17 pm »
Hi all and happy new year to everyone.
Thanks to accept me on this forum.
I have a DM42 too, and it's a great calculator, yes, but...
Mine have problem with battery consumption. I need to change the battery CR2032 every month (more or less).
Is any one has ever measure power consumption of the DM42. I'm using the last firmware but without improvement.
Any information on this problem ? Is it normal ?
Thanks
 

Offline Thomas

  • Regular Contributor
  • *
  • Posts: 117
  • Country: no
Re: EEVblog #1159 - World's Most Precise Pocket Calculator
« Reply #44 on: January 02, 2019, 01:54:31 pm »
My DM42 drain the batteries in a few weeks too.
I measured the current drain: 129µA when off, 160µA when on. Way too much when off, should be <10µA.
Mailed Swissmicros about the issue and got a reply saying it's probably C5 and C6 leaking.
I will replace the caps myself soon.
 

Offline SiliconWizard

  • Super Contributor
  • ***
  • Posts: 14297
  • Country: fr
Re: EEVblog #1159 - World's Most Precise Pocket Calculator
« Reply #45 on: January 02, 2019, 03:44:29 pm »
My DM42 drain the batteries in a few weeks too.
I measured the current drain: 129µA when off, 160µA when on. Way too much when off, should be <10µA.
Mailed Swissmicros about the issue and got a reply saying it's probably C5 and C6 leaking.
I will replace the caps myself soon.

I was also wondering about the screen displaying a still image when the calculator is off.
If I'm not mistaken, this is a Sharp memory LCD. I've already worked with smaller ones (B&W and color ones) and those are fantastic. But they still draw some current to keep displaying an image, even if not much. And If I remember well, this larger screen needs a 5V supply; I'm not sure how the power supplies are handled in the DM42 so not sure if it needs a step-up or SEPIC switching reg, which would further add to the power draw in "off" mode. My point is that some of the "off" current may be due to the screen not being off. Can this feature be disabled, and if so, have you tried measuring the current drain when the screen is completely off (again, not even sure this is possible)? Would have to check in the screen's datasheet as well for some figures. Will do later.
« Last Edit: January 02, 2019, 03:46:02 pm by SiliconWizard »
 

Offline Thomas

  • Regular Contributor
  • *
  • Posts: 117
  • Country: no
Re: EEVblog #1159 - World's Most Precise Pocket Calculator
« Reply #46 on: January 02, 2019, 07:38:21 pm »
My DM42 drain the batteries in a few weeks too.
I measured the current drain: 129µA when off, 160µA when on. Way too much when off, should be <10µA.
Mailed Swissmicros about the issue and got a reply saying it's probably C5 and C6 leaking.
I will replace the caps myself soon.

I was also wondering about the screen displaying a still image when the calculator is off.
If I'm not mistaken, this is a Sharp memory LCD. I've already worked with smaller ones (B&W and color ones) and those are fantastic. But they still draw some current to keep displaying an image, even if not much. And If I remember well, this larger screen needs a 5V supply; I'm not sure how the power supplies are handled in the DM42 so not sure if it needs a step-up or SEPIC switching reg, which would further add to the power draw in "off" mode. My point is that some of the "off" current may be due to the screen not being off. Can this feature be disabled, and if so, have you tried measuring the current drain when the screen is completely off (again, not even sure this is possible)? Would have to check in the screen's datasheet as well for some figures. Will do later.

I have now replaced C5 and C6. That helped a lot, the current consumption is now ~4.5uA when off.
A CR2032 cell is around 230mAh and will last around 6 years at that current.

I have not seen any possibility to turn the screen off, but I haven't looked hard either.
Anyway, I don't think it's necessary given the "calculator off" current consumption and the battery capacity.
 

Offline SiliconWizard

  • Super Contributor
  • ***
  • Posts: 14297
  • Country: fr
Re: EEVblog #1159 - World's Most Precise Pocket Calculator
« Reply #47 on: January 02, 2019, 08:05:00 pm »
I have now replaced C5 and C6. That helped a lot, the current consumption is now ~4.5uA when off.

That was quite a leakage! About 125µA... What kind of capacitors were they?

I have not seen any possibility to turn the screen off, but I haven't looked hard either.
Anyway, I don't think it's necessary given the "calculator off" current consumption and the battery capacity.

Yes, that's interesting actually. If I got it right, the screen must be a Sharp LS027B7DH01. From the datasheet, its nominal supply voltage is indeed 5V as I remembered. They state a typical static power draw of 50µW.

If the DM42 is powered by a single CR2032 cell (is it? or is this 2 cells?), then it must have a step-up regulator for powering the screen. The typical total static current draw on a single cell would probably be over 50µA (and assuming a very low power step-up reg). So I'm thinking they may have found out that the screen retained its image down to a much lower voltage than 5V, at a much lower power draw than the typical. I don't think I have seen anything about this in the datasheet which is kinda minimal. I would be interested in knowing more about this, as I've used this kind of screens but never thought of the possibility of leaving them on at all times (while on batteries) due to a bit too high static current draw (but I never bothered to fiddle with supply voltage and such).


 

Offline Thomas

  • Regular Contributor
  • *
  • Posts: 117
  • Country: no
Re: EEVblog #1159 - World's Most Precise Pocket Calculator
« Reply #48 on: January 03, 2019, 08:06:53 am »
I have now replaced C5 and C6. That helped a lot, the current consumption is now ~4.5uA when off.

That was quite a leakage! About 125µA... What kind of capacitors were they?

I have not seen any possibility to turn the screen off, but I haven't looked hard either.
Anyway, I don't think it's necessary given the "calculator off" current consumption and the battery capacity.

Yes, that's interesting actually. If I got it right, the screen must be a Sharp LS027B7DH01. From the datasheet, its nominal supply voltage is indeed 5V as I remembered. They state a typical static power draw of 50µW.

If the DM42 is powered by a single CR2032 cell (is it? or is this 2 cells?), then it must have a step-up regulator for powering the screen. The typical total static current draw on a single cell would probably be over 50µA (and assuming a very low power step-up reg). So I'm thinking they may have found out that the screen retained its image down to a much lower voltage than 5V, at a much lower power draw than the typical. I don't think I have seen anything about this in the datasheet which is kinda minimal. I would be interested in knowing more about this, as I've used this kind of screens but never thought of the possibility of leaving them on at all times (while on batteries) due to a bit too high static current draw (but I never bothered to fiddle with supply voltage and such).
C5 and C6 are ceramic capacitors, 100uF, 1210 size, probably X5R type. They don't leak 125µA DC (I tested after taking them out) but I think they have high dielectric losses.
I replaced them with 10uF X7R type, works fine.

The calculator is powered from a single CR2032 cell.
You are probably correct they are running the screen at a lower voltage when off, just for retaining the image.

For more information, check this thread over on HPmuseum.org:
http://www.hpmuseum.org/forum/archive/index.php?thread-10143.html
 

Online Kean

  • Supporter
  • ****
  • Posts: 2050
  • Country: au
  • Embedded systems & IT consultant
    • Kean Electronics
Re: EEVblog #1159 - World's Most Precise Pocket Calculator
« Reply #49 on: January 03, 2019, 10:43:10 am »
May not be leakage of the capacitors, but from flux residue - or a combination of the two.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf