May 13, 2015 at 8:06 pm #37652
I don’t have a good IDE for editing the OS cpp files or I would have this done already. Is it possible to have the durations moved out of the program structure and stored on an SD card? This eliminates the need for encoding and decoding the times, while not changing much elsewhere in the core files. The other idea would be to store programs on the SD card entirely (if it is present of course). I see that a few functions suggest this might be an option in the future, but it would be nice to see soon. I have worked out everything for ET based watering except the program to store the 1mm coverage times, which is then multiplied by the ET value.
The problem is most of my times are under 3 minutes and then multiplied by a typical 7.5 or so. This makes for rather erratic watering. 75 seconds becomes 60, which results in a near 2 minute difference toward under watering.
If this isn’t possible or reasonable can the encoding and decoding of times be adjusted, I know this was discussed a while back, but I don’t see a reason to allow a single station to run for more than 6 hours when 4 starttimes are available. This gives 10 more values under the current scheme, then swapping seconds and minutes resolution seems to be more reasonable. Second resolution to 179(3min), minute resolution from 4 to 75, then hours from 2 to 6, leaving 255 for special situations. This seems to provide greater accuracy to the times where resolution counts more, and less as it moves up.
I would like to release the script for use and testing, but don’t feel comfortable doing so while the script is doomed to kill your lawn by loss of resolution during ‘compression’ of duration values.May 18, 2015 at 3:51 pm #37772
This is certainly possible. The only reason we haven’t changed the firmware to do so is for backward compatibility — we only started including microSD card from OS 2.3. For previous versions, one would need to install a microSD card to use the firmware. We don’t want to make this a required step.
We’ve done a survey earlier on the forum about compressing the water times, and I think the decision was unanimous that granularity of minutes is sufficient. Intuitively, the longer the water time, the less accuracy it needs.
You can feel free to modify the firmware to support higher precision. The compression is transparent to the UI — the firmware always decompresses the value when sending the data to UI, so the compression remains local to the firmware and the UI is not aware of it.May 18, 2015 at 8:24 pm #37790
I honestly never even thought about backward compatibility issues, as I picked up my OS right when data logging was made available, and essentially installed an SD card on day one, before I set anything else up. I could see how this would be a problem for you guys, and is of little consequence to me as a single user. I’ll just go forward with the updates for myself. I suppose I could get it figured out and fork the project so others could mess with it as well, if they want to take on the additional task of installing an SD and updating the firmware, without affecting the general user.
The time compression thing just seemed odd to me, but again I’m a single user, so I’ve already modified my own to meet my needs. I was just wondering why the granularity wasn’t linear, and I didn’t really see an answer in the forum thread about it. If it was there and I missed it, I apologize, but I like understanding why decisions are made the way they are, perhaps to the dismay of the designers and programmers.
Thanks for the reply and insight, your product and continued support is quite incredible. I used to run a major commercial irrigation system, and I am far more impressed with the capabilities of your device and staff.
Anyway, back to playing with my sprinklers 😀 Thanks again,
ShawnMay 19, 2015 at 10:36 pm #37823
Regarding ‘why the granularity wasn’t linear’, I think you probably have got the answer from my brief reply. Just in case it’s still not clear: the reason is that typically as the water time increases, the demand for accuracy also decreases. For example, if the water time is only 1 minute, then a few seconds of difference would matter a lot. However, if you are watering for an hour, a few seconds of difference typically wouldn’t matter as much. Hope this makes sense.
- You must be logged in to reply to this topic.