WS-12-250-DC hangs after hibernate

[foo00]

I have multiple WS-12-250-DC units setup in parallel to chargers/battery banks at AC sites such that they're pulling DC charger voltage during normal ops but have the "always on" connection to battery during a power outage event. Works fine during failures, but recently during longer outages they got down the hibernate voltage and the switches shutdown as they were set to. Now fast forward to the power coming back.. the charger lights up and starts recharging the batteries but the switch does -not- wake up. Again, this was on multiple units and separate sites on different days. It stays dead until I fully remove the DC connection, wait ~10 secs, and then reconnect it. It then boots and runs like a champ like it does every other day.

The first time is was on 1.4.7 so I upgraded them all thinking maybe it was an old glitch, but it happened twice more elsewhere -after- the 1.5.0 upgrade.

I thought I saw one thread warning us to not vary (like during bench testing) the input voltage too quickly as it could cause some system to freak out and misbehave. Could that be what's happening here? Power dies, we have the slow fall off all the way from 12V to 9V, go into hibernate, then an instant pop back up to 14+V when the charger gets power and goes straight into recharge mode (yes, this is the simplified version, I'm sure there's delays and in-rush protection, blah blah.. you get the picture).

If that indeed is potentially the culprit, is there any combo of settings on the Netonix to make it -not- misbehave? Or do I need to redesign the schematic and add a low-voltage disconnect -before- the Netonix as I've seen mention of in the FB groups. Then -it- worries about the dropping/raising voltage and the Netonix only sees a black and white 12V or 0V and either works or it doesn't?

[sirhc]

How many batteries are you running here?

Are the batteries in series or parallel?

Is your charger a smart multi stage charge capable of delivering 10A per battery to properly recover dead batteries?

Normally if you take a 12V battery below say 10V the depth of discharge exceeds normal recovery and requires 10A per battery to properly recover.

So for instance:
If 1 battery taken down to 9V than the depth of discharge is 25% which with some batteries "may" require 10A to properly recover. (Depends on battery specs)

If running 2 batteries in series to get 24V bank then the depth of discharge is 62% which definitely requires a charger that can produce 10A per battery to recover properly.

If your charger is not capable of 10A per battery recovery I would set hibernation voltage up to 10V x the number of batteries if in series. So if 2 batteries then hibernation should be set to 20V. You can allow the switch to discharge further in emergency but understand you may need to get on site with a proper charger.

What COULD be happening is the charger is not capable of recovering properly and the voltage is cyclic at first meaning the power cycles up and down fast and causes the power supply CPU/Program to get confused and stop. Like what happens if you plug and unplug any embedded device rapidly you often end up with a screwed up unit you need to unpower for a bit to fully discharge then apply power again cleanly.

This could be your charger causing this, all depends on how it ramps up power when AC power is restored.

[sirhc]

So example.

Your battery drops below 9V switch goes into hibernation mode.

AC power gets restored but the charger flickers on ramp up so the switch power supply sees 9V,15V,9V,15V very rapidly for a bit.

This could happen if your pushing your charging unit too hard as it can not provide 10A per battery.

Soon as the charger applies power the line voltage jumps but the batteries drag it down then up then down.

By the time you get there the batteries have recovered enough that when you power cycle the switch it see voltage above 9V and all is good but it was that flicker in the beginning that caused the problem. There is nothing in software we could do to fix this. That would be like asking for a SOHO router to deal with power on/off/on/off/on.

Take any soho AP and plug and unplug the power several times real fast before leaving power on and sometimes it will be just locked up as it lost power but before it could completely discharge and reset power is restored but the device is not fully down and things start to get confused. Firmware can not fix the way electronics behave under all circumstances.

I would look into the charger, maybe a different or larger/beffier one?

[sirhc]

Note: Some of the rants below may not be related to you, just ranting.

I guess we could try putting in a delay on WAKE UP.

Meaning the power supply must see voltage above WAKE UP voltage for say X minutes before it attempts to boot the switch.

Now this would only fix the issue if the switch is hooked directly to the battery bank because if powered by a power supply lord knows what is happening to the voltage fluctuations, could be dropping to 0V for milliseconds then back on then off then on which no delay in the world would fix this.

Some people incorrectly use the SMART DC switches and have them powered by a power supply.

SMART DC MODELS: WS-8-150-DC / WS-12-250-DC / WS-26-500-DC

What they use is some sort of Mean-well type power supply charger which means the Mean-well powers the switch and charges the batteries. THIS IS WRONG.

Simply hook the switch directly to the battery bank preferably with a breaker or at bare minimum an inline fuse so as to allow CLEAN power connects and disconnects and do not SPARK the connection which can mess up the power supply.

The reason why is the Mean-well charger may be causing the flicker. Sadly I get tech support from people powering a WS-12-250-DC with a power supply rated at 200 watts or far less. This is WRONG. The power supply in the switch is rated at 250 watts and the in-rush at power up can be much higher for milliseconds. So if your going to power our SMART switches with a power supply the make sure it is rated ABOVE the power supply rating of the switch.

Or better yet in these circumstances use the appropriate model which is meant to be powered by a 48V-54V power supply. I know Mean-well makes 24V and 48V versions.

DUMB Models: WS-12-DC / WS-26-DC

OR SIMPLY DO AS THE MANUFACTURER RECOMMENDS AND HOOK THE SWITCH DIRECTLY TO THE BATTERY BANK AND THE CHARGER TO THE BATTERY BANK SEPARATELY.

This just gets me every time. We tell people to use the equipment in a specific manor then they use it in another way and complain it does not work.

Honest to God I spent hours trying to help a guy with issues only to find out he was powering a WS-12-250-DC with like a 24V 1 Amp power brick.

Or people who put the switch 200 feet in the air and run 12 AWG wire up to it, this is not only BAD but DANGEROUS as the wire could melt and catch fire. Think about it, 250 watts draw at the switch and if the input voltage would drop to 9V that is 27 AMPS on 12 AWG at 200+ feet - DANGEROUS.

[sirhc]

So I need more info from you to even speculate on the cause.

How many batteries?
How are they configured (series or parallel)?
What is the charger?
Is the switch hooked DIRECTLY to the battery or batteries?

[ckoehler]

I am having the same issue it seems. Since this gentleman didn't answer your last question, I will. I am using 2 batteries in series (Expert Power EXP12330..33ah). I am using the MInn-Kota precision mk220pc. The switch is hooked up directly to the batteries. I have no issue setting the hibernate voltage to a higher value (i.e. 20v) but I did have mine set to 10 to see how long this setup could last. What is interesting is that I no longer received notification that the voltage was below the warning threshold (20) after it got down to 16 volts so I am not sure when it shut off. Is this match of battery to charger suitable?

[sirhc]

I am having the same issue it seems. Since this gentleman didn't answer your last question, I will. I am using 2 batteries in series (Expert Power EXP12330..33ah). I am using the MInn-Kota precision mk220pc. The switch is hooked up directly to the batteries. I have no issue setting the hibernate voltage to a higher value (i.e. 20v) but I did have mine set to 10 to see how long this setup could last. What is interesting is that I no longer received notification that the voltage was below the warning threshold (20) after it got down to 16 volts so I am not sure when it shut off. Is this match of battery to charger suitable?

The Minnkota chargers are good, I have no idea about that brand of batteries but I am sure they are fine.

What firmware version are you using?

"Almost" every WS-12-250-DC not all but most I get in for RMA are running v1.4.8 and v1.4.9.

We obviously lab these and they always work in the LAB?

[ckoehler]

1.5.1 is the firmware...I upgraded from the factory firmware the day I got it (Friday). After letting the batteries charge to full, I flipped the breaker off waited a few moments and flipped it back on. The unit still doesn't power up. I flipped the breaker off again and this time waited 10 mins but still no power led on the front although you can see a couple lights inside the unit come on. I am assuming I did not hard reset it correctly. Also on further reading, I did not receive the shutting down because of low voltage email. The last email was at 16 volts below the 20 warning but not yet to the 10 shutoff.

[sirhc]

1.5.1 is the firmware...I upgraded from the factory firmware the day I got it (Friday). After letting the batteries charge to full, I flipped the breaker off waited a few moments and flipped it back on. The unit still doesn't power up. I flipped the breaker off again and this time waited 10 mins but still no power led on the front although you can see a couple lights inside the unit come on. I am assuming I did not hard reset it correctly. Also on further reading, I did not receive the shutting down because of low voltage email. The last email was at 16 volts below the 20 warning but not yet to the 10 shutoff.

I have no idea then what's wrong with it?
Your saying it does not power on at all now?

[ckoehler]

That is correct.