diy solar

diy solar

UPS Transfer Time - longer?

DIYrich

Solar Wizard
Joined
Mar 6, 2022
Messages
3,165
Location
New England, USA
The Sol-ark 15k has a "UPS Time" setting for transfer to battery when grid goes down. The minimum is 4ms (any setting below 4ms is 4ms).

Why would you want a longer transfer time?

Minimize nuisance disconnects from grid in areas with dirty power?

I could see a longer time that doesn't require keeping the inverter energized (eco mode) but it doesn't offer that.
 
Wow...4ms is fast. Extremely fast.
I am left wondering whether that is even possible, most equipment has a bit longer duration switchover.

Example: My Multiplus has a spec of less than 20ms. No adjustments are possible that I know of. I notice nothing blinking or any behavior to indicate it has made a switchover.

To your question I don't believe that setting has an effect on nuisance disconnects.
 
Wow...4ms is fast. Extremely fast.
We do for certain have a default 4ms transfer time. The adjustment was added because some users wanted to be able to tell if there was a transfer happening as it was happening so they could change processes. Now if those changes were happening physically or some sort of operations change when batteries were being used I'm not sure. I have not seen anyone use it myself though.

Believe it or not there are some servers and sensitive electronics where 4ms still isnt fast enough. For instance, some 911 call centers use us to back up their systems in the event of an outage. Most of them experienced no issues, however, I specifically talked to one site where their equipment was still picking up on the transfer and even though they wouldn't lose power, the software would reset. The only solution was to integrate an even faster UPS system fo them to filter out the transfer blip. Bit of an edge case for us.
 
Because in some instances you want power to bounce equipment. Say you have a full N+1 or N+2 system where you have multiple reduncancies. If power is lost you might want 1 system to stay up with 4ms but the other one to wait a full second to come back online. This way the equipment fails over but the host is still online incase of a fault.

For instance if you have 2 server racks with identical load balancers and your runtime is 1 hour for both running, if you bounce 1 rack and its just sitting idle then you can increase your runtime close to 2 hours. If there's an issue you can still manually run them.

On top of this there could be devices on the line that cause false positives. We have manufacturing clients and lots of large equipment can cause voltage to drop for a few ms and constantly trip UPS's.
 
We do for certain have a default 4ms transfer time. The adjustment was added because some users wanted to be able to tell if there was a transfer happening as it was happening so they could change processes. Now if those changes were happening physically or some sort of operations change when batteries were being used I'm not sure. I have not seen anyone use it myself though.

Believe it or not there are some servers and sensitive electronics where 4ms still isnt fast enough. For instance, some 911 call centers use us to back up their systems in the event of an outage. Most of them experienced no issues, however, I specifically talked to one site where their equipment was still picking up on the transfer and even though they wouldn't lose power, the software would reset. The only solution was to integrate an even faster UPS system fo them to filter out the transfer blip. Bit of an edge case for us.
Any thoughts on offering an "eco" mode where you don't keep the battery inverter energized. There is a constant 15 watt draw (0.38kWh per day). How long a transfer time would it need to energize the inverter?
 
Any thoughts on offering an "eco" mode where you don't keep the battery inverter energized. There is a constant 15 watt draw (0.38kWh per day). How long a transfer time would it need to energize the inverter?
Not sure what you mean by Eco mode in this instance so let me know if Im not understanding. Grid following to grid forming happens all on the AC side so even if there is a set transfer delay the DC side would stay active the whole time. I don't think it would be practical to turn the DC side off during the inbetween period, especially since most delays would only be for a few seconds.
 
Clean cut vs. brownout also makes a difference on the speed of transfer. Brownout ends up being a much cleaner and faster transfer because the Sol-ark "sees" it coming, where a clean cut such as breaker flipping causes a longer transfer time.

Friend of mine tested this with oscilliscope. Brownout scenario ends up with hardly even any glitch! I'll see if I can get a screenshot from him! Normal grid outage will be slight brownout, so very clean transfer to battery/ inverter!
 
Back
Top