Low Voltage Action

[yahel]

We have an issue with the way WISP-switch handles set low-voltage thresholds.
Here's a common scenario:
A solar-powered backup site with two power hungry Air-Fibers.
We set the "power down" voltage to 48.5v for both Air-Fibers, and should the voltage go that low, the first Air-Fiber (the lower port number) gets shut, which results in voltage increase and hence the 2nd Air-Fiber does not get shut.
This sequential behaviour of shutting ports is counter productive as we're left with one useless Air-Fiber operational which just continues to drain the battery, while not passing any traffic (since the 1st Air-Fiber is down).
We wish that all designated ports would be shut at the same time upon voltage threshold hit, as not to be affected by voltage fluctuations that results from shutting of load.
Is there anyone who can think of a reason why this should not be the normal behaviour of the switch, I would love to hear.
Current firmware version 1.4.7RC4 (but this was always the behaviour of the switch from early days, and have not changed in recent versions).

Thanks,

Yahel.

[Dave]

Yahel

How the code works is that it takes a voltage measurement, and then checks all port limits, and will turn off all associated ports, before making another measurement cycle.

We designed it to do exactly as you are requesting, and was verified to work this way in our lab.

What is the power back up voltage level you would set to, assuming you set it to 48.5 volts for power down?

Dave

[Dave]

Yahel

Give me until tomorrow for this issue....it might be possible we have things staggered for shutdown, depending on port type.

We will do more testing tomorrow & confirm.

Dave

[yahel]

Guaranteed that shutdown is indeed staggered, and it should not be.
Hope to get a fix before this winter is over (although this would again become relevant next winter ;-) )...

Thanks,

Yahel.

[Dave]

Yahel

We will get a firmware RC done this week for this.

Thanks for your feedback.

Dave

[Julian]

Yahel;

I setup a ws-12-250-DC in-house with 4 WS-6-minis on it, powered via 48vh. Configured shutdown setpoint at 47.5v, and turn-on setpoint at 48.2v for ports 1-4. I started my adjustable power supply at 49v, and allowed everything to boot, and then made the following observations:

I adjusted input voltage to 47.3v, and allowed it to remain for 5 seconds or so, until the first 6-mini shut down.
At this point, I turned input voltage quickly to 48.3v, and the remaining 3 6-minis never actually shut off.


per my conversation with dave, the best case for your battery system is:

1) Voltage sags below setpoint for 3 seconds =>
2) All ports configured for shutdown at that setpoint shut down
3) Once a port is scheduled for shutdown, it shuts down.
4) Shutdown ports do not turn on again until voltage stays at or above user configured setpoint for 3 seconds or more,
5) all ports scheduled for turn-on at a given setpoint turn on at once

Obviously, this requires some intelligence when setting shutdown/turnon voltages; if the port shuts down, and voltage rises above turnon due to decreased load, then you've made a loop, basically.
Work proceeds apace, please bear with :)

[lligetfa]

I think with an adjustable power supply, it will work to regulate the output as the load changes. A battery bank on the other hand is unregulated so YMMV.

The other YMMV is the voltage drop due to wiring.

[yahel]

Julian, you miss the point...
When one load gets shutdown the voltage goes up beyond the shutdown set-point..
Obviously, it never reaches the turn-on set-point, which would indeed be a loop (that would be an unwise setting).
(my turn-on set points are in the 56v area, while my shut-down set points are at 48.5v or so).

The problem, as Dave acknowledged and will fix soon, is that ports are being shutdown one after the other, and due to the rise in voltage after the first shutdown, the next port would never shut.
All ports with the same settings should be all shut together at the same time to avoid that problem.

As for your test itself, lligetfa's point is valid... your bench equipment is unlikely to exhibit what real batteries would...

[Julian]

Yahel

The staggered shutdown is indeed the issue, I'm afraid I may have been unclear in conveying my understanding of that as an issue, I apologize.

I had no trouble duplicating the issue you described with a regulated power supply, since it's constant voltage adjustable, it approximates (although not millisecond accurately) the behavior of your battery bank with manual adjustment.



I will be emailing you the fixed software image shortly.

[yahel]

Thanks, and no rush (at least not on an hours time scale).... I'm in India, going to sleep (the switch is in California)...

Thanks!

Yahel.