HHC 2013: Day 1 Highlights



#8

http://edspi31415.blogspot.com/2013/09/hhc-2013-day-1-highlights.html


#9

In your blog, there's a sentence I've a query:

'During his talk a poll was taken and the majority of attendees preferred AA batteries over AAA and button sized batteries.'

Are AAA batteries less available in their countries? In Hong Kong, AA batteries are as common as AAA ones.

Edited: 22 Sept 2013, 4:52 a.m.


#10

I don't think an absolute question is appropriate in this topic. For pocket calculators, clearly the smallest reasonable battery size is the one to choose. Any connectable additional power source is appreciated if particular calculator states require more power - but I won't recommend loading big batteries in a low power calculator.

Just my 2 cents.

d:-)


#11

Richard Ottosen and I have used coin cells in all of our calculator prototypes to date. We originally used CR3430 cells, because they have twice the energy capacity of the more common CR2032 cells. Unfortunately they cost about five times as much. The latest prototype hardware uses a pair of CR2032 cells.

The rationale for using a pair of alkaline AAA or AA cells rather than the coin cells is that we cannot make the MicroSD card work reliably from coin cells (even the larger CR2430). An SD card (including MicroSD) is allowed by the SD specification to draw up to 100 mA at 3.3V. (In high speed modes, up to 150 mA, but we don't use high speed modes.) While not all SD cards draw the maximum current, there is no good way to select a low-current card; even cards sold by the same vendor with the same part number can vary considerably. Of the cards I've measured, most draw between 60 and 80 mA.

Even when two coin cells share the load, that is beyond the rating of the coin cell. With fresh coin cells, in practice everything works fine. However, when the cells are close to end of life, the internal resistance is very high (as much as 40 ohms), and if you try to draw a lot of current, such as by accessing an SD card, the voltage will drop substantially (to well under 1.0V), which will cause the processor to reset.

This is especially bad because the user is likely to want to write their programs and data to the MicroSD when they get a low battery warning.

Aside from that problem, abusing the coin cell by drawing high current from it reduces the effective capacity of the cell by about a factor of three.

Unfortunately there is no good way to determine at any given time whether using the SD will cause a problem with power from the coin cells.

Richard has attempted to solve this problem by designing a pair of 1F capacitors into the latest prototype. Unfortunately this has introduced other problems that are even more serious, so those capacitors have been removed.

Alkaline AA and AAA cells have much lower internal resistance than coin cells, even near end of life (around 0.3 ohms), so the voltage drop when using an SD card is not so much as to cause a reset.

Having been told about this problem, my poll of the conference attendees revealed that they were overwhelmingly in favor of having a wedge-shaped calculator with cylindrical cells rather than a flat calculator with coin cells.

The other alternative would be a calculator with coin cells but without the SD card. However, Richard and I do not have the resources to make two different models.


#12

Since you're designing a calculator mainly for us geeks (I think that's fair to assume), why not use AAAA cells that the user can cheaply obtain by ripping open any 9V alkaline battery. Cost would be comparable to CR2032, AA or AAA cells and they can easily supply the needed current. You'll have a lot more latitude in your physical design with AAAA cells than AA or even AAA.


#13

I didn't know AAAA batteries existed.


#14

They're often used in slim penlights. You can buy them individually with the standard button tops like these. Inside 9 volt alkaline batteries are 6 AAAA cells connected by welded tabs, which makes them a tiny bit shorted than the button top ones.


#15

I wouldn't really want to rip the 9V batteries apart to get to AAAA cells. Not because I find it too much hassle, but because they will leak easily when fully discharged. The manufacturer doesn't worry about this, as the 9V battery is sealed rather then the individual cells.

I would go with AAA batteries, too. Alternatively why not go down the route of installing a li-po battery with a cr2032 as a backup battery? The li-po could then be recharged via a USB port.


#16

Because we (Richard Ottosen and I) specifically do not want a calculator that has to be recharged.

A thought experiment: how well would the 15C have sold if it needed to be recharged every few weeks (or more frequently if used heavily)?


#17

Technology was very different back then. All rechargeable batteries had a high rate of self discharge.
And I don't think you will manage to build a calculator with a power consumption anywhere close to what the 15C had (and at the same time meet all the other design goals).

Don't get me wrong, I don't have a problem with AAA cells, but where is the problem in recharging the calculator once a month rather than replacing AAA cells every two month?

My WP34s with li-po battery lasts me several month without recharging.

#18

Hi Katie,

Having not opening up a 9v battery for many years, I never realized that they were composed of six AAAA cells. I remember the old carbon 9v batteries having six square cells. Shows how long it's been since I looked inside one.

Just for fun, I opened up an Energizer 9v Alkaline battery tonight, and the six cells just fell out. They did not have welded tabs. They were just pressure fitted. They have a clear plastic sleeve on each and a button for the positive end. I don't have a regular AAAA, so not sure if they are any shorter.

So the Energizer makes it really easy to get a set of six batteries.

Bill


#19

I've only opened a few 9v batteries, so was just reporting from the ones I've seen. I guess that each manufacturer has their own way of packaging these. The old carbon-zinc batteries were quite a mess and would often leak in old transistor radios since the "stuff" was just under the outer shell. I've never seen an alkaline 9v battery leak through the outer shell, but I guess it could happen.

I would also guess that Harold is right, the individual AAAA cells used inside a 9V batteries might be more prone to leaking since the manufacturer has the outer shell as a second line of defense -- they could skimp on the inner shell quality.

I like the idea of a li-po rechargeable cell with a CR2032 as a usable backup (without being about the write to an SD card).

Edited: 23 Sept 2013, 5:46 p.m.


#20

Quote:
I've never seen an alkaline 9v battery leak through the outer shell, but I guess it could happen

They sure can leak outside the shell. I have a homemade sprinkler system for my lawn which uses these small valve timers that run on 9v batteries. Last year I forgot to remove the batteries when I put the timers away. When I opened them up this past spring to put new batteries in them, they had broke open and ruined the battery contacts in the timers. The battery compartment is sealed with gaskets, and I didn't see water damage, so I'm thinking they may have froze during the winter and broke open. I store them in an unheated shed.

I've also seen a few others that have corroded in place. I've seen them with the two contacts completely corroded off the 9v battery. While it's not common, it can happen.

Bill

#21

Aside from the problems mentioned in other replies in this thread, AAAA cells have low capacity and high series resistance, so they don't really solve the problem.

One of the major alkaline battery makers once announced a rectangular AA-ish battery for MP3 players, but it sank without a trace. The press release was still online for a long time, but I can't find it now.


#22

As Eric was at the conference, this is mainly for those who were not.

The AAA vs. AA is about several things.

Size. Yes, AA are a good bit thicker.

Cost. AA's are much cheaper than AAA, especially when you consider...

Capacity. AA's last a lot longer.

So cheaper and longer. Over the life of the machine, this reduces the cost to operate, which is why I suspect the AA's won the poll at the conference.

The thickness was not as big an issue to attendees (it did not seem).

FWIW.


#23

Well, if it is to be a pocket calculator, then I would favour AAA size to reduce the thickness. My Microsoft Arc Touch mouse uses these batteries and when folded flat it is still slim enough to fit in a shirt pocket without being bulky. That is more or less the same thickness as my HP-15c calculator which I consider to be the gold-standard for pocketability.

Regards

Nick

Edited: 26 Sept 2013, 5:37 a.m.

#24

might i make a desperate plea for using a full-sized SD card? the micro-SD cards are so small, that i've seen them accidentally thrown out with the product packaging. i can certainly see the rationale in a product where space is at a premium (like a hearing aid or a cellphone that slips behind your eyeball), but for a product that runs on AA cells surely there is no such shortage of space?

rob :-)


#25

Quote:
i can certainly see the rationale in a product where space is at a premium (like a hearing aid or a cellphone that slips behind your eyeball), but for a product that runs on AA cells surely there is no such shortage of space?

The batteries affect the height of the case. The SD card and USB affect the width of the case. There is barely enough case width as it is.


-- Richard


#26

i've not seen the layout - is it available to view anywhere?

to save real estate around the sides you could use a dual-ring 3.5mm socket for usb, as is used on some apple ipod nano products. this has an added advantage of being much more robust than mini/micro usb, along with being quite narrow.

regarding super-caps: how did you implement using them? trick is a simple cc charger, then a step-up converter on the output to keep a constant voltage as the cap is discharged (plus a schottky diode and p-channel mosfet for supply switching). i worked on the design of a product that did this, with the caps able to keep a whole win ce machine (3a consumption) running for around 30 seconds after power loss.

rob :-)


#27

Quote:
i've not seen the layout - is it available to view anywhere?

Not really. Maybe Eric can take a couple of quick pictures of the prototype and post them here.

Quote:
to save real estate around the sides you could use a dual-ring 3.5mm socket for usb, as is used on some apple ipod nano products. this has an added advantage of being much more robust than mini/micro usb, along with being quite narrow.

I think we need to stick with a standard micro-USB connector. It has the advantage that it is low profile. The width is tolerable.


Quote:
regarding super-caps: how did you implement using them? trick is a simple cc charger, then a step-up converter on the output to keep a constant voltage as the cap is discharged (plus a schottky diode and p-channel mosfet for supply switching). i worked on the design of a product that did this, with the caps able to keep a whole win ce machine (3a consumption) running for around 30 seconds after power loss.

The step-up converter and diodes are already there.

I did a simple resistor to limit charge current. I have considered a constant current charge but I don't think it gains much. CC charging does have the advantage of a lower peak charge current for a faster charge rate. However, it is going to be a bit tricky to do on a single 3-volt lithium coin cell while maintaining extreemly low current consumption.

I expect to run hours to _days_ on the cap if the SD card is not used and for seconds when it is.

#28

Quote:
Richard has attempted to solve this problem by designing a pair of 1F capacitors into the latest prototype. Unfortunately this has introduced other problems that are even more serious, so those capacitors have been removed.

While this is true, I still have hopes of putting the huge caps back into the design.


-- Richard


#29

What sort of problem did you run into with the caps?


#30

Quote:
What sort of problem did you run into with the caps?

There were/are 2 potential problems.

The first, is that the only way to turn off the power for doing hardware modifications is to short out the caps. That is why I removed the caps on the prototype.

Second, it takes a _real_ long time to charge the caps the first time a battery is installed. I am worried that this will either harm the battery or take so long that the power from the caps won't be available the first time the SD card is used.


-- Richard


#31

Quote:
The first, is that the only way to turn off the power for doing hardware modifications is to short out the caps. That is why I removed the caps on the prototype.

That makes sense for the prototype.

Quote:
Second, it takes a _real_ long time to charge the caps the first time a battery is installed. I am worried that this will either harm the battery or take so long that the power from the caps won't be available the first time the SD card is used.

Hm, that is a tricky one ;) Unless of course you used AA cells ;)

I am really looking forward to see the calculator!!

#32

AA's give several times as much energy per dollar as AAA's give. I just bought 100 alkaline AA's for the many things around the house that use them, for $16 at Fry's. That's about eight cents per amp-hour. And while I occasionally see AAAA's, most battery displays in the stores do not have them.

I definitely would not want a battery type, rechargeable or otherwise, that won't be available indefinitely. My HP-41cx has 27 years of service on it and seems to be going strong. I use it every day. Getting N cells is becoming more challenting, but we can still get them. If HP had used a custom battery and had discontinued it 15 years ago, I would be rather upset.

I keep hearing about "pocket calculators;" but I would not consider any of these valuable calculators "pocket" calculators. If you put it in your shirt pocket and bend down to get to a lower file cabinet drawer, what happens? The calculator falls out on the floor. If you put it in a pants pocket and sit down, what happens? You break it. And even if you don't do either of those, you get lint into the keyboards. No, good calculators never go in a pocket.

My first choice would be the AA's, followed by AAA's. I'm not convinced AAAA's will always be available, short of disecting new 9V batteries which is not a practical option.

If you know of a good web page describing how to use SD cards, with their many types of file systems (FAT12, FAT16, FAT16B, FAT32, FAT32X, NTFS, exFAT, HFS Plus, UFS, Ext2, Ext3, Ext4, btrfs, HFS Plus, ReiserFS), partition types, the two or three different CRCs, three different block sizes, different densities, etc., in SPI mode, for dummies, please share it with me. I find SD to be daunting.

Edited: 24 Sept 2013, 10:49 p.m.


#33

Hey, I realize that this is non HP related but in a class I am currently working on SD SPI interface (on the pic microcontroller) I guess you can have a look around and see what is going on. I mostly use Application note 1045 (AN1045) from Microchip, if you are good at C, you can quite understand the functions' "underground" jobs.

I successfully used this application note to write sensor datas from a PIC24fj256gb206 to uSD card.

There is also This website that provides libraries for most Microcontrollers that may be useful to you


Hope it can help you a little.

P.E.F.

P.S. Someday, when I have time, I plan to use the IR communication from a 48g to communicate to a PIC and then log datas using USB to the computer or then communicate to other IR calculators calculator to parallelize calculations.


FatFs - Generic FAT File System Module


Edited: 30 Sept 2013, 2:56 p.m.

#34

The poll was an informal one. I expect that batteries of all types would be available worldwide. Here in the States AA, AAA, and coin size batteries are readily available.

#35

In your blog, there's the tip:

Quote:
Tip: Do not plug a Woodstock calculator (HP 21 to HP 29, 1970s) unless it had a good battery. Otherwise, the battery may get fried.

Shouldn't that read "Otherwise, the calculator may get fried."

Thanks for the blog about the HHC.

Bill


#36

Yes, that's true. Unfortunately I have an HP-25 that suffered that fate. It fried the main processor chip. Of course if I want to cannibalize one of my HP-97's....NO!


Possibly Related Threads...
Thread Author Replies Views Last Post
  An amazing day: Giving a talk at HP about their calculators Geir Isene 9 683 12-16-2013, 06:14 PM
Last Post: aurelio
  HHC 2013 RPN Programming Challenge - Final Thought Jeff O. 7 394 10-17-2013, 11:02 AM
Last Post: Jeff O.
  HP Prime 2013 8 13. Rev:5106 bluesun08 2 210 09-28-2013, 11:05 AM
Last Post: debrouxl
  HHC 2013 programming contest winners Brian Walsh 52 2,025 09-26-2013, 11:39 PM
Last Post: David Hayden
  Box left at HHC Tim Wessman 1 224 09-26-2013, 11:05 PM
Last Post: Wlodek Mier-Jedrzejowicz
  HHC 2013 speakers' files Joe Horn 2 230 09-26-2013, 02:51 AM
Last Post: Geoff Quickfall
  HHC 2013 Day 2 Highlights Eddie W. Shore 6 438 09-23-2013, 04:03 PM
Last Post: Kimberly Thompson
  HHC / HP Museum Programming Contest for RPN and RPL machines Gene Wright 18 818 09-22-2013, 09:39 AM
Last Post: Miguel Toro
  HHC 2013 room numbers David Hayden 2 213 09-20-2013, 05:34 PM
Last Post: sjthomas
  Flash Flood Warning: 9/16/2013 (One Week from HHC13) Eddie W. Shore 8 485 09-17-2013, 09:20 PM
Last Post: Craig Ruff

Forum Jump: