Suggestions/Observation/Bug Reporting

Forum for announcements and discussion of beta firmware.
Paul.Allen
Site Admin
Posts: 98
Joined: Tue Aug 06, 2013 5:33 am
Location: Utah, United States

Suggestions/Observation/Bug Reporting

Post by Paul.Allen » Wed Sep 11, 2013 7:39 pm

Version: 0509

Suggestions:
  • If two batteries are inserted and one is determined to be already full then the Green light flashes. However once the other battery is full and both are Green the Green light still flashes for the first. I think they should both go solid.
  • It would be nice if the very first line printed to the SD card and on serial out was a line that named each column and what info they contain like Voltage or Current.
Bug:
  • Mark, I know you already are aware of this, but I thought I would put it here for reference. If two batteries are inserted then after pressing the button there is still sometimes that not both are detected.
Observation:
  • Feel free to delete this if it is all wrong, but i think I may have had a few batteries terminate early because of a bump right when it begins charging. If you don't already I think we should have a min voltage threshold that the battery must be above before we consider it charged. The other thought is that this bump only seems to happen with the combination of new batteries or highly abused batteries and at high charge current. So the other option would be if we could guess if this might happen based on the impedance check and then easy into the full charge current over a period of maybe 5-10min or something to help avoid this bump.

Mark
Site Admin
Posts: 418
Joined: Sun Jul 28, 2013 6:47 am
Location: Brisbane, Australia

Re: Suggestions/Observation/Bug Reporting

Post by Mark » Thu Sep 12, 2013 12:03 am

Paul.Allen wrote:Version: 0509

Suggestions:
  • If two batteries are inserted and one is determined to be already full then the Green light flashes. However once the other battery is full and both are Green the Green light still flashes for the first. I think they should both go solid.
  • It would be nice if the very first line printed to the SD card and on serial out was a line that named each column and what info they contain like Voltage or Current.
It would be a bit tricky to do the first one, although it should be doable in theory at least. Personally, I think it makes sense to leave it flashing so that it's obvious that the cell wasn't charged at all since it was inserted. There has to be a window where a cell will be rejected as fully charged even though it's only mostly charged, so it might be at only 95% whilst the other is at 100%, so I think it's useful for the user to be aware that one cell wasn't actually charged. It probably doesn't matter too much in the end since the flashing green happens straight away, I think most users would just take the battery back out straight away rather than leaving it in there while the other charged.

It would be interesting to see what others think about this though?

I did have a header line being added to the start of the log, but I disabled it in the quest to save some space for other program enhancements. I might just re-enable it though...
Bug:
  • Mark, I know you already are aware of this, but I thought I would put it here for reference. If two batteries are inserted then after pressing the button there is still sometimes that not both are detected.
Darn - I thought that it was fixed with the latest version that I sent to you!

I was checking a few times for it last night but didn't see it happen. What percentage of the time do you see it?
Observation:
  • Feel free to delete this if it is all wrong, but i think I may have had a few batteries terminate early because of a bump right when it begins charging. If you don't already I think we should have a min voltage threshold that the battery must be above before we consider it charged. The other thought is that this bump only seems to happen with the combination of new batteries or highly abused batteries and at high charge current. So the other option would be if we could guess if this might happen based on the impedance check and then easy into the full charge current over a period of maybe 5-10min or something to help avoid this bump.
The battery does need to hit 1.44V resting before my inflection termination method will kick in, so that's generally already pretty full. Can you send me the log file from where you're seeing this problem?

Mark
Site Admin
Posts: 418
Joined: Sun Jul 28, 2013 6:47 am
Location: Brisbane, Australia

Re: Suggestions/Observation/Bug Reporting

Post by Mark » Thu Sep 12, 2013 2:20 am

I've just finished testing the push button literally 50 times and I couldn't get it to fail once!

Assuming that it's happening reasonably regularly for you still, I might have to get you to do testing for me if I can't reproduce the problem here...

Paul.Allen
Site Admin
Posts: 98
Joined: Tue Aug 06, 2013 5:33 am
Location: Utah, United States

Re: Suggestions/Observation/Bug Reporting

Post by Paul.Allen » Thu Sep 12, 2013 7:24 am

It is not very often at all, but I did get it to do it tonight. I will send you the data in the morning (my morning). Are you using an interrupt or a loop? If you use an interrupt and don't enable interupts until after you have checked both then I don't know how it would be possible? Same with a loop.

Also after your explanation on the green light I agree with you. Keep it flashing so people know it was not charged.

Mark
Site Admin
Posts: 418
Joined: Sun Jul 28, 2013 6:47 am
Location: Brisbane, Australia

Re: Suggestions/Observation/Bug Reporting

Post by Mark » Thu Sep 12, 2013 2:12 pm

Paul.Allen wrote:It is not very often at all, but I did get it to do it tonight. I will send you the data in the morning (my morning). Are you using an interrupt or a loop? If you use an interrupt and don't enable interupts until after you have checked both then I don't know how it would be possible? Same with a loop.
I've got the current code working off an interrupt that is triggered by both the rising and falling edge of the pin connected to the switch, although I'm thinking about moving the code run on a regular basis from the Timer ISR that currently handles flashing the LEDs - this should save a few bytes from the compiled size.

When a button press is detected, it sets a flag for each channel, so if it works for one, it should work for the other.

I have got some code in there to do software debouncing on the switch. I can imagine a scenario that if the switch is bouncing a lot (i.e. never settling down for long enough), that could cause a problem with the way that the code is working.

Can you try with a different charger and see if another one also exhibits the same problem - it might just be a dodgy switch on the one that you're testing on. I'll also make some changes to the code to work around this type of scenario.

JGSchubert
Posts: 25
Joined: Sun Jul 28, 2013 7:59 pm

Re: Suggestions/Observation/Bug Reporting

Post by JGSchubert » Thu Sep 12, 2013 11:35 pm

Am I missing something, or did Mark get a new board with a button and not the rest of us? Also, I seem to have read earlier on a post here that if I download the newest code, it will work fine on my non-button board right? Do I have to change some #defines?

Thanks...

- Geoffrey

Mark
Site Admin
Posts: 418
Joined: Sun Jul 28, 2013 6:47 am
Location: Brisbane, Australia

Re: Suggestions/Observation/Bug Reporting

Post by Mark » Fri Sep 13, 2013 1:09 am

I'm not sure about Paul's intentions with regards to sending out boards to the rest of you. I'd say that he sent me the revision 6 board so that I could finish adjusting the code for the final version. He might be intending to send boards to the rest of you from the first batch.

The next version of the code will still support the older versions but I'll probably remove that support in the following version just to clean up the code to make it easier to understand for the masses. There is a #define at the top of charger.h which sets which version of the hardware that you're compiling for.

Paul.Allen
Site Admin
Posts: 98
Joined: Tue Aug 06, 2013 5:33 am
Location: Utah, United States

Re: Suggestions/Observation/Bug Reporting

Post by Paul.Allen » Fri Sep 13, 2013 3:54 am

I sent Mark a Rev. 6 Board because it takes a long time to get to him. However there were a few small things I didn't like about the Rev. 6 board so I didn't send them to the rest of you. Instead I decided I would just wait and send you the Final black boards which I have now. I will send them out hopefully tomorrow.

JGSchubert
Posts: 25
Joined: Sun Jul 28, 2013 7:59 pm

Re: Suggestions/Observation/Bug Reporting

Post by JGSchubert » Sun Sep 15, 2013 3:42 am

Paul,

Works for me! (sending out final boards - yesterday now as I write this) Just wanted to make sure I did not miss anything. Lots going on with this project (which is great)!

-- Geoffrey

Mark
Site Admin
Posts: 418
Joined: Sun Jul 28, 2013 6:47 am
Location: Brisbane, Australia

Re: Suggestions/Observation/Bug Reporting

Post by Mark » Sun Sep 15, 2013 4:22 am

Let me know when you get the board - assuming that you get it before I release version 0.510 of the firmware.

The next version already has quite a few changes but I've been holding off on releasing it since most of the changes are for the version with the button.

Post Reply