Another 15C LE speed test



#2

I know that you've all been waiting for my pi calculation test so here it is (scroll to the bottom). :) 306 decimal digits of pi in a bit over 16 minutes. This works out to about 160 times as fast as the original 15C, nice!

Edited: 22 Sept 2011, 12:47 a.m.


#3

thanks Katie. I'll put my 15C LE to the torture test tonight and share the results here.


#4

You all should measure the battery voltage before and after the test. It would be interesting to know how much we can get out of the batteries with heavy use. @20mA and a rating of 220mAh per cell it will not be more then 22 hours.


#5

Quote:
@20mA and a rating of 220mAh per cell it will not be more then 22 hours.

Actually it's much, much less than this.

The 240mAh (Energizer's) rating is based on a continuous current draw of 0.19mA. At higher current the capacity drops off very rapidly. They give a pulse rating of 200mAh based on 12 2-second long pulses per day of 6.9mAh. The pi program is a 960 second long "pulse" at 3 times this current (but drawn from 2 cells, so really 1.5x this current). This simply destroys the cells. I'd estimate only a couple of hours running long programs on the calculator at best.

Edited: 23 Sept 2011, 2:14 a.m. after one or more responses were posted


#6

Hope they address this with an update.

~ btw, where can the flash cables be purchased and are they USB or serial interface?


#7

Gene has a few left apparently. He has asked for a moratorium on requests until after HHC. He sends them out for the cost of shipping and I don't know what is going to happen when they are all gone.

They are serial cables and they are finicky about the USB-Serial converter you use (if you have to use one). Someone else will be better able to tell you which converter or adapter to use.

Hope this helps a little.

Cheers,

-Marwan


#8

Quote:
Gene has a few left apparently. He has asked for a moratorium on requests until after HHC. He sends them out for the cost of shipping and I don't know what is going to happen when they are all gone.

They are serial cables and they are finicky about the USB-Serial converter you use (if you have to use one). Someone else will be better able to tell you which converter or adapter to use.

Hope this helps a little.

Cheers,

-Marwan



Thanks for this info, hopefully there may be some news of an update that will come out of the conference. I'm on a mac, it only has USB/Firewire but could access a pc with a serial port if necessary for the update.

Seems if the cables are finicky and scarce perhaps HP will establish update centers or maybe a forum member may offer the service. Since HP has lots of uncertainty at the moment this is most certainly a low priority for them now. Appears we're just lucky they released the 15C LE, thanks to some dedicated folks within.

Fortunately, I've got my original 15C for daily use so my 15C LE is a secondary unit which will help maintain it's battery life until a solution unfolds.

#9

The current draw is a direct function of speed. So the 20mA while a program is running cannot be addressed by a firmware update without severely lowering the speed.


#10

Quote:
The current draw is a direct function of speed. So the 20mA while a program is running cannot be addressed by a firmware update without severely lowering the speed.

The 20ma during a key press is certainly fixable in firmware, the 12C+ uses only 0.5ma when a key is held down.

The current draw when running a program is nearly perfectly linear with clock speed. Indeed the "L" in the chip's part number (AT91SAM7L128) is there to indicate that the current draw is 0.5mA/MHz. The maximum advertised clock rate is 36MHz and there's not much else in the calculator that consumes appreciable power. So the 15C LE must be running at maximum speed.

They could keep the clock at full speed for say, the first second or two of running a program or function and throttle is back to 5 Mhz or less and still advertise the 100x speed up, I think.

Better still would be a self test-mode function (number 4, for example) that would allow the use to toggle between full clock speed and 1/10 speed (or so). Running at 4MHz translates to 1ma/cell and would give you around 300 hours of run time on a set of batteries.

Marcus, you've got a SLOW/FAST toggle in the 34s, what are the corresponding clock rates? I think all ARM-based HP calculators should have this toggle.


#11

Quote:
Indeed the "L" in the chip's part number (AT91SAM7L128) is there to indicate that the current draw is 0.5mA/MHz.

It's more in the range of 1mA/MHz according to my measurements.

#12

Quote:
It's more in the range of 1mA/MHz according to my measurements.

That's pretty far out of spec from what I can tell. Even if all the peripherals are on it shouldn't be that high. Maybe HP got a deal on out-of-spec raw chips.

In looking at the original schematic in the hp-20b SDK I see that they apparently planned to run this on a single 2016 cells with a second one optional, ha! They must have been thinking at running this at a much lower clock rate originally.


#13

I obviously haven't recalled my own measurements correctly. At full speed, WP 34S consumes 20mA. See here. It's not running at 40 MHz though so 0.5mA/MHz is a bit on the optimistic side.

#14

Quote:
Indeed the "L" in the chip's part number (AT91SAM7L128) is there to indicate that the current draw is 0.5mA/MHz.

I thought (wrongly) that the "L" just indicates that it has on-chip LCD display driver circuitry. The AT91SAM7S128 is almost identical except that it lacks such circuitry. NOTE: Edited to correct wrong information.

Quote:
They could keep the clock at full speed for say, the first second or two of running a program or function and throttle is back to 5 Mhz or less and still advertise the 100x speed up, I think.

That would then constitute a dishonest claim, IMHO.

A program that takes five seconds on the 15C-LE would take ten minutes on the original 15C. Such programs, if they are to run at all, should not be slowed after an arbitrary time elapse. And if such programs will never be run, then there is no real battery life issue.

I want the SAM7L in my 15C-LE running at maximum capacity. That is, after all, the only characteristic that is an improvement over the original HP-15C.

Quote:
Better still would be a self test-mode function (number 4, for example) that would allow the use to toggle between full clock speed and 1/10 speed (or so).

Excellent idea!

Edited: 23 Sept 2011, 2:02 p.m. after one or more responses were posted


#15

The L stands for the low power modes. The chip can power down each and every single component and can even reduce the kernel voltage below 1.8V if set to a lower speed. It has wake up pins which enable the "being dead but looking alive" mode used in the latest HP calculators. The chip is powered down except the LCD and some wake-up logic tied to the keyboard while the calculator is waiting for user input.


#16

I was wrong drawing assumptions from the block diagrams of the SAM7L vs. the SAM7S.

Thanks. My apologies!


Possibly Related Threads...
Thread Author Replies Views Last Post
  48G vs 49G+ User RPL Speed Comparison John Colvin 7 524 11-16-2013, 10:07 PM
Last Post: Han
  HP-42S ESD self-test Yriarte 2 301 10-24-2013, 09:08 AM
Last Post: Yriarte
  OT - A lucky find - Busicom Handy-LE LE-120A Cristian Arezzini 2 341 09-26-2013, 04:43 AM
Last Post: Cristian Arezzini
  So I took a test with my Prime today ... kris223 12 821 09-25-2013, 07:19 PM
Last Post: kris223
  Low power warning for HP-15C LE and batteries Nick_S 1 306 09-16-2013, 09:34 AM
Last Post: Hans Brueggemann
  WP-34S: Speed of y^x Marcel Samek 1 255 09-14-2013, 07:31 PM
Last Post: Paul Dale
  WP-34S function execution speed ? Gene Wright 4 367 09-04-2013, 05:40 PM
Last Post: Paul Dale
  Concern about Voyager keyboard test Matt Agajanian 2 366 08-30-2013, 07:56 PM
Last Post: Matt Agajanian
  HP 15C-LE replacement still available? Borja 16 1,141 08-22-2013, 11:16 AM
Last Post: Michael de Estrada
  JTAG on HP-12C and HP-15C LE Ingo 5 596 07-01-2013, 06:37 PM
Last Post: Paul Dale

Forum Jump: