EUC World - some alarm delay when if both non-peak and peak current alarms are set.

[rcgldr]

Update - the issue is an apparent delay in reporting data, which may be specific to the V8F. The issue may be that when the non-peak current alarm is set, it is one more packet of data that the V8F reports to EUC World. I set non-peak current alarm to 5 amps, and peak current alarm to 7 amps. Battery was about 60% charge. I started each test at about 5 mph, going up a slight incline around 1.5 degrees, leaning very gradually forwards until I either reached about 10 mph or got an alarm. If I didn't get an alarm, I gradually slowed to about 5 mph and repeated the test. If I got an alarm, I gradually slowed and stopped, and checked the stats.

In one of my tests, I was already braking after accelerating before I got a peak-current alarm, so a noticeable delay on that one test. I wasn't able to duplicate this again.

With non-peak current set to off, I managed to get an alarm and a peak of 7.1 amps, but that was a fluke. Most of the time after getting an alarm, the peak current was 8.0 to 8.3 amps, whether or not non-peak current was off or set to 5 amps. These tests were done with very gradual movements, and I was anticipating the alarm, so I was reacting faster than normal.

I also did a test where I triggered two alarms on the same run, and looking at the speed | current graph, the first peak current was 8.3 amps, the second one 8.0 amps, so the peak current alarm will trigger when the peak current alarm setting is exceeded, regardless of prior peaks.

My assumption is that the PWM from the controller board in effect increases current while voltage remains near constant or sags as speed increases during acceleration, so a delay in the alarm is probably what caused the relatively higher peak currents when I tested with non-peak current at 10 amps and peak current at 15 amps. I was accelerating harder and at higher speed, 8 mph to around 15 mph runs. Maybe coincidence, but doing 3 tests with non-peak current off, I got alarm and peak current was around 17.5 amps, and another 3 tests with non-peak current at 10 amps, where I got alarm and peak current was around 20 amps. Assuming this was all a delay issue, then with non-peak alarm off, if alarm was triggered around 15 mph, then with non-peak current set to 10 amp, delay resulted in about 14% faster speed, around 17 mph, except that voltage sag would be greater, so the speed difference would be less.

Edited January 25, 2022 by rcgldr

[Planemo]

The non peak setting simply adds a duration variable afaik, and it isn't tied to the peak setting. Neither should influence each other.

The peak setting should still have triggered an alarm though if you had it set to 15A and the wheel peaked at 20A. Seems a bit odd...

[rcgldr]

47 minutes ago, Planemo said:

The non peak setting simply adds a duration variable afaik, and it isn't tied to the peak setting. Neither should influence each other.

The peak setting should still have triggered an alarm though if you had it set to 15A and the wheel peaked at 20A. Seems a bit odd...

I agree that is the way the non-peak setting should work, but that's not what happened when I tested it. The wheel peaked at 20A, but part of that is reaction time and having to accelerate to balance the rider, except in this case I'm anticipating the alarm, and did a few tests while riding, each test just a bit more lean than prior test until I got an alarm or got the impression I needed to lower the settings and test again. So the alarm may have gone off at around 18A, but still well above 15A.  

After turning off the non-peak setting, I lowered peak setting to 7 amps, to confirm I would get multiple instances of alarms on the same run, instead of just a single alarm event that would require me to exit and restart EUC World. Most of my rides have peak amps below 18 amps, unless I'm deliberately testing peak current alarm. I tested a 22 peak amp setting, and I think I got a 23 peak amp reading. My vendor thought that 25 amps would be safe, but I want a bit more margin so setting it 20 amps for now, since my reaction would be a bit slower since I wouldn't be anticipating an alarm.

[Planemo]

I think you meant to say 'that is the way the peak setting should work...?'

I think @Seba or one of the EUCW experts would need to comment on this. I can't see why a peak alarm wouldn't trigger when it sees a peak, irrespective of what the non peak is set to.

[rcgldr]

5 hours ago, Planemo said:

I think you meant to say 'that is the way the peak setting should work...?'

Peak setting is working as I would expect it to work as long as the non-peak setting is off. It seem that enabling the non-peak setting is disabling the peak setting, resulting in peak current much higher than either setting.

[Seba]

On 1/22/2022 at 1:35 PM, Planemo said:

I think @Seba or one of the EUCW experts would need to comment on this. I can't see why a peak alarm wouldn't trigger when it sees a peak, irrespective of what the non peak is set to.

I've made several tests to verify what @rcgldr reported and didn't found any issue. Current alarms are working as intended.

There are two separate current alarms that works in independent way. First one is "current alarm", second one is "peak current alarm". First is based on a current value that is averaged over period of 1-2 seconds, so all spikes are cancelled out. Second one works on momentary (non-filtered) current values, so will react on any spikes that are reported by the wheel. First alarm is intended to avoid wiring, battery and mainboard circuitry overload, while second alarm is intended more to avoid MOSFET failures. This means that current alarm should be always set to a lower value than peak current alarm. It also means that even if you see a certain current value in wheel parameteres, current alarm will react with some delay. It's intended, as current values one can see in the app are momentary, not averaged.

So for example if you set your current alarm to 30 A and peak current alarm to 60 A, dynamic acceleration causing the current to rise above 30 A (but below 60 A) won't activate current alarm immediately after passing 30 A threshold. Depending of the acceleration dynamics, it will activate somewhere between 30 A and 60 A. But as soon as the current value goes above 60 A, alarm will trigger.

Of course you have keep in mind that there may be additional delay caused by Bluetooth data link limitations. Depending on wheel model, data is sent with certain repetition rate and there is always some lag. Also, radioelectric interferences or noise may lead to dropping some data frames. More, if you have lot of apps installed on your phone, some of them may work in background, consuming processing resources and thus slowing your phone and EUC World. Unfortunately, communication protocol used by older Inmotion wheels isn't great. It's possible that wheel parameteres are send with interval of one second or even mode. Wheels like V11, V12, King Song and Begodes are much better in this area. Still, there is always a 0.2 - 0.5 sec lag, even in perfect conditions.

[Patrick Robert]

I was thinking about this also. 

On my S18, I set a current alarm on 22A (7.3A x 3p) and a peak of 30A (for curiosity). Seems to fit with was Seba just described.

On my Sherman, I used 60A for current and 100A for peak. I've gotten the 60A alarm, never the 100A one. 

Edited January 24, 2022 by Patrick Robert

[rcgldr]

8 hours ago, Seba said:

I've made several tests to verify what @rcgldr reported and didn't found any issue. Current alarms are working as intended.

I updated my original post to note this is a delay issue.

Edited January 25, 2022 by rcgldr

[rcgldr]

20 hours ago, Seba said:

Unfortunately, communication protocol used by older Inmotion wheels isn't great. It's possible that wheel parameteres are send with interval of one second or even more. Wheels like V11, V12, King Song and Begodes are much better in this area. Still, there is always a 0.2 - 0.5 sec lag, even in perfect conditions.

Which might explain the apparent delay on my V8F. Also I probably had non-peak current alarm too close to peak current alarm, 10 | 15 amps on one test, 5 | 7 amps on another test. Unless it was coincidence, I seemed to be getting somewhat longer delays if I had non-peak current alarm enabled.

I'm also using the EUC World option to use the V8F speaker for alarm, since the V8F speaker is much louder than my phone, which helps if I'm riding into a headwind. I'm using a Bell Super 3R Mips helmet with a lot of holes in it, which is probably increasing wind noise. Using the V8F speaker is probably adding an additional delay.

However, the first time I enabled the non-peak current alarm with higher settings, the issue didn't seem to be a delay. I already had peak current set to 20 amps and had tested it, getting a peak current around 22 amps due to delay. On my first test, I set non-peak current at 10 amps, leaving peak current at 20 amps. I did a few test runs and gave up on triggering alarm since I felt the acceleration and speed should have been enough to trigger peak current alarm. It turned out that peak current was around 24 amps, but this did not trigger an alarm. I didn't feel comfortable trying to repeat this test, so I retested at lower values as noted in my prior posts. This could be some sort of issue with the V8F. For now, I have non-peak current alarm off and peak current alarm set back at 20 amps.

Edited January 25, 2022 by rcgldr