Battery run competition, category 41

[Hans-Peter]

I invite you to a special race. The runners are our machines, the range is the program that drains the battery empty.

Here are the rules:
* Presentation of the candidate's inner values (electronic, retirement ...)
* Running program (hardware, listing)
* Finisher clip (hours)
* State after the finish (Memory Lost or normal status)
* Success factors
   
Here are my first four candidates, all with HP41CV, Time and IR module:
2xEneloop 1.2V / 800mAh, 5V regulator (30uA), Finnish with normal status, h 53:45
1xLipo 3.7V / 240mAh, 6V Regluator (18uA), finish with ML, h 32:00
1xLipo 3.7V / 300mAh, 6V regulator (18uA), finish with ML, h 27:00
1xLipo 3.7V / 450mAh, 6V Regluator (18uA), finish with ML, h 47:00
   
Good race
Hans Peter
   
PS: I'm still looking for a running program for the 97er category!

[Dieter]

(12-04-2016 01:11 PM)Hans-Peter Wrote:  I invite you to a special race. The runners are our machines, the range is the program that drains the battery empty.

Here are the rules:
* Presentation of the candidate's inner values (electronic, retirement ...)
* Running program (hardware, listing)
* Finisher clip (hours)
* State after the finish (Memory Lost or normal status)
* Success factors

Sorry, but I do not quite understand what you mean here. What is a "finisher clip" and what "retirement" do you refer to? Maybe you want to repost these rules in German ?-)

But looking at the pictures, the Li-Po battery built into the case of a 82120A battery sounds very promising. Seems to be charged using an USB connection. Can you say more on this? I only wonder why the 450 mAh battery yields 47 hours, compared to the 300 mAh version with merely 27 hours.

You say that all the Li-Po runs ended with a "Memory Lost". Looking at the program, the low battery flag is tested once an hour. Maybe you should change this to once a minute. Running up to an hour on a low battery simply is too much here, and whether the program fishishes correctly or a Memory Lost occurs depends on the BAT annunciator appearing at minute 59 (program finishes, calculator turns off) or at minute 01 (59 more minutes to go until the next test).

Dieter

[Hans-Peter]

(12-04-2016 08:13 PM)Dieter Wrote:  

(12-04-2016 01:11 PM)Hans-Peter Wrote:  I invite you to a special race. The runners are our machines, the range is the program that drains the battery empty.

Here are the rules:
* Presentation of the candidate's inner values (electronic, retirement ...)
* Running program (hardware, listing)
* Finisher clip (hours)
* State after the finish (Memory Lost or normal status)
* Success factors

Sorry, but I do not quite understand what you mean here. What is a "finisher clip" and what "retirement" do you refer to? Maybe you want to repost these rules in German ?-)

But looking at the pictures, the Li-Po battery built into the case of a 82120A battery sounds very promising. Seems to be charged using an USB connection. Can you say more on this? I only wonder why the 450 mAh battery yields 47 hours, compared to the 300 mAh version with merely 27 hours.

You say that all the Li-Po runs ended with a "Memory Lost". Looking at the program, the low battery flag is tested once an hour. Maybe you should change this to once a minute. Running up to an hour on a low battery simply is too much here, and whether the program fishishes correctly or a Memory Lost occurs depends on the BAT annunciator appearing at minute 59 (program finishes, calculator turns off) or at minute 01 (59 more minutes to go until the next test).

Dieter

The finisher clip is the small video, when you finish a run like this finisher clip (one of the yellow ones).
In our "calc world", it means a receipt of the reached time.

The retirement means the state of the calc (Memorylost or BAT). If you use a single cell Lipo, then you have a protection circuit witch drops the voltage imediatly, so the BAT flag has no chance to show low bat. The result is a ML.

The quality (real capaticity) of the lipos is sometime different. My actual 500mAh is now charging. In 3 days I will show more details.

The run with the eneloops shows the minutes and ends without a ML. The eneloops stopped by 1.02V each, when the booster went low.

Hans-Peter

[MeindertKuipers]

Hello Hans-Peter,

I do plan to submit my contribution, I have made my own battery booster, juts need to get my printer up and running again and prepare a webcam for a finisher clip. I just hope my computer runs longer without crashing than my calc ...

Regards,

[Nate Martin]

(12-04-2016 09:08 PM)Hans-Peter Wrote:  

(12-04-2016 08:13 PM)Dieter Wrote:  But looking at the pictures, the Li-Po battery built into the case of a 82120A battery sounds very promising. Seems to be charged using an USB connection. Can you say more on this? I only wonder why the 450 mAh battery yields 47 hours, compared to the 300 mAh version with merely 27 hours.

You say that all the Li-Po runs ended with a "Memory Lost". Looking at the program, the low battery flag is tested once an hour. Maybe you should change this to once a minute. Running up to an hour on a low battery simply is too much here, and whether the program fishishes correctly or a Memory Lost occurs depends on the BAT annunciator appearing at minute 59 (program finishes, calculator turns off) or at minute 01 (59 more minutes to go until the next test).

Dieter

The retirement means the state of the calc (Memorylost or BAT). If you use a single cell Lipo, then you have a protection circuit witch drops the voltage imediatly, so the BAT flag has no chance to show low bat. The result is a ML.

The quality (real capaticity) of the lipos is sometime different. My actual 500mAh is now charging. In 3 days I will show more details.

Hans-Peter

I tried several different LiPO batteries and 5V circuits, always ran into two issues: 1 was the low voltage cutoff as mentioned above, not really a issue if we could get long battery life, but 2 was the way the circuits operated caused a much higher drain then expected and the ~500mah batteries only lasted about 3 days as the 5V circuit was always on. I found Nimh N-cells much easier overall for a recharagble solution.

[MeindertKuipers]

I just started my test with 2* 1.2V NiMH batteries, advertised as 1000 mAH each, using a 5V boost converter. HP41CX has a PPC ROM, and is running a user program calling NP, calculating primes.
I have a webcam and a timelapse program (taking a picture every 60 seconds), so no printer connected, the first snapshot is attached.

Regards,

[Hans-Peter]

[/quote]

I tried several different LiPO batteries and 5V circuits, always ran into two issues: 1 was the low voltage cutoff as mentioned above, not really a issue if we could get long battery life, but 2 was the way the circuits operated caused a much higher drain then expected and the ~500mah batteries only lasted about 3 days as the 5V circuit was always on. I found Nimh N-cells much easier overall for a recharagble solution.
[/quote]

Hello Martin
The big deal with the current drain is the quiescent power of the used chip on the booster. Your board needs in 3 days 500mA. That are 7mA quiescent current. My best one takes 18uA, thats 390 time less.

For the next weekend, I publish my table with all the tested chips and boards.

Hans-Peter

[Nate Martin]

Hans-peter

The boards I was using had very low specification for quiescent current in the ~x0uA range. But a freshly charged LiPO with the 41C(CV/CX) in standby only lasted about 3 days. What kind of life are you getting with a 41C in standby versus operating?

Specifically I used a:
1. SparkFun Power Cell - LiPo Charger/Booster (~50uA)
2. Adafruit Mini Lipo w/Mini-B USB Jack (~25uA)
3. Converter salvaged from a Battery stick
All of these should have been several orders of magnitude less then what I actually saw which was batteries dead at low voltage cutoff after about 3 days. I used several different batteries and calculators to test. If I leave the batteries out of a calculator that last as long as I have been testing them (about half a year with only minor voltage drop from full charge).

I am a Chem Eng, not a EE, so this is frustrating to me and I do not know that I am doing something wrong.

[MeindertKuipers]

My setup is still running, more than 60 hrs now. Standby is also excellent, this is better than 5-6 weeks.
I will post details of the boost converter and batteries when my test run is ready

[MeindertKuipers]

The setup stopped this afternoon. Total run time 66 hrs (plus or minus 2 minutes), on 2* 1000 mAH AAA batteries, see picture.
This was a bare HP41CX with only the PPC ROM plugged, running a small user program generating primes with the NP routine.
Boost converter used is one from Polulu https://www.pololu.com/product/2560, set at an output voltage of 5.2V.
It started at prime 409 (needed some time to start the sequence), last prime seen was 74357.
I have a timelapse movie to prove, need to edit it a bit!

Regards,

[emece67]

(12-16-2016 06:01 PM)MeindertKuipers Wrote:  The setup stopped this afternoon. Total run time 66 hrs (plus or minus 2 minutes), on 2* 1000 mAH AAA batteries, see picture.
This was a bare HP41CX with only the PPC ROM plugged, running a small user program generating primes with the NP routine.
Boost converter used is one from Polulu https://www.pololu.com/product/2560, set at an output voltage of 5.2V.

Being the converter regulable, perhaps you can get a little more run time slightly lowering the output voltage, say to 4.3 V or so (slightly above the low bat indicator). The internal voltage converter inside my 41 seems to have better efficiency around such voltage. On the other side, the efficiency of the Boost converter you use (it seems to be a TI TPS61200) may be slightly less at lower voltages, in part compensating the better efficiency of the 41's converter, but according to my measures/calculations such lower input voltage can give you around 3 hours more.

As a side effect, the battery life with the machine turned off will also rise, as the current consumption when off (in my 41) at 4.3 V is half of that at 5.5 V. The TPS61200 has a much greater quiescent current, though, so the overall effect is not such high.