EEVblog  

I have the same calculator and I just tried some really weird roots to give it a work out and it took about the same amount of time as 69!. I also think if it had a look up table it would be instant instead of having that small delay. I also think these ASICs have the memory hard programmed into the silicon so it would make much more sense to go for the route without the huge float array if you had the processing capabilities. Factorial can be programmed with a simple float factorial(int n){ float answer = 1; for(n,n>1,n--) answer*=n; return answer; } (in C) which is way less programming space than 69 floats.

Mason - I think you are right that it is going through the calculations base on the update time. I like your function call but would the assembled code really be less space? It has a ten digit plus exponent display that could be stored in simple BCD in just 6 bytes. 6 x 70 = 420 bytes plus the lookup code. How much for floats? If I had the code space I would use the table just to get the speed but it seems clear that they did not.

I haven't had much too much experience in assembly unfortunately and the size of the function will depend a lot on the microcode and instructions they've implemented into their ASIC. But I'm assuming it wouldn't take even 20 words, let alone 420. I tried to compile it with arduino real quick just to see how many bytes it would take to compile that function but it said 444 before and after the code was inserted. And if I assume that max displayable digits are what is stored in memory (10 digits) and 2 for the exponent, log2(9999999999) + log2(99) ~ 40 bits or 5 bytes per number in the lookup table. With 69! being the biggest number we can show on the calculator with 2 digits for exponents that leaves us with 345 bytes. The lookup code would just take 2 or 3 bytes to say have a starting point address and then add the factorial to it and then go to that lookup address. So it would probably take 350 or so bytes in total. Since I'm assuming they role their own chips they would just require more memory location be built in as long as they didn't need to add an extra bit to the memory address. My guess is that they didn't see it as necessary to build or spare the extra memory since their ASIC is clearly fast enough.

Mason - Close but you forgot to store zero so it is 70 numbers. And adding log2(digits) + log2(exponent) isn't quite right as you are adding the fractions but they are separate numbers. log2(9,999,999,999) = 33.22 but .22 of a bit is a bit (ha!) hard to store in a non-analog processor! ;-) 2^33 = 8,589,934,592 so you need the 34th bit. And for the exponent log2(99)=6.63. If we round both of them up we have 34+7=41 bits to store but also the hassle of converting it all back to a format for display which is likely 5 + 1 = 6 bytes. You can do it but will it take less bytes to code it than what the 6 byte wide table takes? Also, there is a need to test that it is an integer and not negative in both versions so that adds a bit more to the code size. That's interesting that the Arduino code size didn't change. Must be a minimum amount or something? Weird. But a big reason to use the table is because 69! seems to be a common speed test. "Hey my calculator is faster than yours!" Increased sales!

oh right, you got me there, I did forget 0. I added them as seperate number because computers represent floating point numbers as the mentesa (the decimal) and then the power of 2. I assumed if you were going to do a lookup table that you would just represent it as a power of 10 because that is how it is displayed and there is no reason to add extra complexity. I just wanted a decent estimate as they may just be representing it as a standard (long) floating point number, 8 bytes, or try to save some space. And I rounded down the 33 because I'm not actually sure the calculator goes up to 9.999999999, and they wouldn't even need to in the lookup table seeing as how the largest is 69! (1.71*10^98). I did round up the 6.63 because you do need the full range there. So I still think my best case scenario for memory storage required is still decent. So the range would be 350 bits to 560 bits. I am still very much convinced that a talented programmer could easily program a factorial function with around 20 bytes (which corresponds to around 20 assembly instructions!) which is an order of magnitude less than the lookup table. Yeah I'm not really sure why the arduino IDE did that. I really should load up the PIC c compiler but I was feeling lazy. I don't think that is a major selling point of this calculator, I think it is just meant to be a budget calculator that never runs out of juice. But I think we might be worrying too much about the specific details of a random calculator lol

He used up all his flux caps on 121GW multimeters

If you can source the thiotimoline for the electrolyte, he'd do it.

from the solar road

Or they just haven't seen carbon...

JJayzX I'm registered blind does that count? 😀 Although even I know it's plastic.

Or because stupid tripod companies market their bollocks as "x69! layers" and claim the "stripes increase rigidity", which gullible people fall for...

I like SHARP calculators, because they need very few keypresses for any operation and you get hex even on cheapest models. But i wish they still made EL-506P though :( I used to have a Tozai-made clone of that, but it has worn off. It's a metal faceplate model from the 80ies.

I use a fx-991MS, an 991ES PLUS , and an 991EX, and the free42 app on the phone. I also have a fx-992VB (1994), but the display doesn't display all segments. The machine is still working though! I know more about Casio calculators than I am supposed to know.

I'm tempted, too. Really nice for the price.

What kind of features would you like in such a calculator?

I think retroreflective coating is a bad idea, because your eyes don't actually emit light, and i think wearing a head lamp just to be able to read a calculator is stupid. But what isn't such a terrible idea is EL film. It's a thin film backlight that is normally white, but emits azure glow when powered with something like 100V but there's very little current going through it. I wonder how small you can make an EL step-up circuit? After all, i've seen wrist watches do it.

In practice it actually is not such a bad idea, you don't need a light behind you to see the numbers (because the coating is white), just some ambient light and in many situations it's better than the regular old-school reflector since in lower light conditions they don't reflect that well because you need to see the reflector from the correct angle plus the correct angle for the LCD and they are often not the same. On the other side, the retro-reflective coating scatters better the light, the only problem may be using it with a bright light behind you like the sun. It may be a bit too reflective, but you could use some sort of matte filter. There are many types of retro-reflective materials, like tapes, stickers, films, paints and even fabrics so there's quite bit of options to try.

voltage clamping. red leds start conducting at ~1.5v and really take off at 2v Really handy if your moding cheap lcd electronics to be solar powered.

What purpose would that serve?

MrMegaPussyPlayer I think there was room, if looking closely at the placement of those ribs.

Not enough for the 1F one ( 7:34 )

I was thinking that maybe it is just a different board without that button connected...

Or maybe c120 is responsible for this setting

Valentino Saitz Or maybe they just omit a part from the button mechanism itself. For example if these are really tactile domes behind the buttons, they can just omit the metal dome itself.

I really doubt they use a cap for that, it could be a bond wire on the die, it could be whatever, but a cap? Odds are slim...

Don't forget smasnug and nichocn capacitors!

Err, yeah, doh.

I'll have to keep an eye out for that. Alan has some good videos and they keep getting better.

Correct, I stand corrected.

She'l no stand the strain harald, she'l blow for sure.

Cliff Matthews Depends on the country or exam system. Some places allow bringing as many books and notes as you can carry, so a few extra kilobytes of calculator memory is not a problem. Others still focus on testing rote learning of rarely used formulas so they insist on not letting students bring any way of storing extra notes or preprogrammed formulas. Bans on phones and other ways to communicate makes more sense as they want to be sure they are not getting results from someone else, inside or outside the room.

@@jameshewlett7798 You own all Casios?🙄?? Assuming pockets, let's count: fx-10, 15, 19, 21, 102, 201P, 39, 140, 180P, 3400P... 310, 350, 510, 570, 370, 361, 81, 82, 85, ... I stop here (those were only a fraction of the 70s and 80s!).

@@jameshewlett7798 Great. And i have one Casio from 1994 fx-992VB broken, and few from 2000s - fx-260 solar, fx-991MS, 991ES PLUS, 991EX, and basics : a credit card size SL-760LB, SX-300P and a dual leaf SX-220. A micro / nano museum? Do you have any of these? 😀

Yes, but the big "to do" was when Energizer came out with the ads. That backfired on them when Duracell "re-introduced" their bunny, and everyone promptly forgot about Energizers bunny campaign.

EEVblog Except people keep mentioning the "Energizer bunny" out of context, so much that having never actually seen an Energizer branded bunny commercial, I just thought Duracell and Energizer in America was the same company as Duracell in Europe and they just edited the text in the commercials for different markets.

Could well do.

Why couldn't a configuration trace/pad hide right under the epoxy blob? Or simply different PCBs. I'm sure they reuse silicon a LOT, but reusing PCB layouts completely doesn't really save you much cost at volume, unless you have to order batches that are too large to consume and have to store them.

It's at the end.

EEVBlog #473

Watch my previous video, it works on 20-30 lux.

KX36 some propellers and hobbyist model aircraft parts are in fact carbon fiber

Huh? I had a tempered glass phone screen protector that was fairly obviously glass. It became obvious because it shattered a few weeks in, and it didn't just break in one spot, when it shattered, it shattered across the whole surface, but it was really not possible to scratch it up. I got it for a couple dollars from China too.

KX36 .... Tempered glass screen protectors meet the MOHs hardness rating that corresponds to glass. You may have purchased some "fake" tempered glass.

I've bought hundreds of tempered glass screen protectors over the years they have all been glass however the quality varies significantly! One I bought was amazingly thin, about a quarter of the thickness of a typical tempered glass screen protector you could curl it round on itself and it wouldn't break however eventually it did break cracked all over the place.

Shape, design elements and probes resembles Rigol, but I can't find this particular model on their website. Something very new perhaps? I think Dave hid the model number very intentional.

I wastn't paying attention but I think Dave was looking at the voltage pump output necessary for the LCD. See his videos on LCDs.

uhmm... no

zerpBot I haven’t got one to test but it might be enough to power a l.e.d 😐

Adding a boost converter does not increase the power available, it just increases the voltage. Adding a boost converter will only lower the overall power because it's not 100% efficient.

Mason Daub it’s worth a try, but might not be very bright? 🤔🤷‍♂️

It would be useful if the voltage coming of the nuclear battery is too low for the thing you're trying to power. The monocrystaline solar panel Nurd Rage used produces a high enough voltage. With a solar cell it depends on the wavelength of light that's hitting it and the energy it takes to free the electron from the substrate. I believe this is called the work function. Anyways with this type of solar panel and the energy in each photon of visible light the individual electrons have enough potential energy to make the voltage required to run the calculator. The problem is that there isn't a high enough number of photons, and thus electrons, produced by the decay of the tritium vials to provide enough power to the ASIC, so it can't provide enough current. (current is basically the flow rate of electrons through the wire and power is the flow rate * the potential of each electron. This gets you Watts which is J/s. Increasing the voltage and decreasing the flow rate keeps the power the same. Remember, energy is a conserved quantity)