Compatability of Selected UPS Device to Handle Inductive Loads

Related to this forums area of expertise, I have a question I would like to ask. Any input would be greatly appreciated!

First, some background:
My church's bells are rung by three separate Rabbit 2000 Embedded Processors. The combined wattage output of the processors, when not activating solenoids to ring the bells, is only six watts. Two of the bells are rung with a single solenoid which, at activation, draws a peak current of 250 watts. The big bell uses two solenoids which has a peak current of 500 watts. The activation periods for the solenoids is between 60 and 90 milliseconds. The only situation in which more than one bell is rung at the same time is when regular AC current is available and then manually by pushing a button. Lately, our town has had a slew of power outages. During the last outage, one of the processor's flash drives was corrupted (i.e. power surges, other current irregularities). We will have the system designer reload the program and plan on purchasing a good surge protector designed for electronics. As a result of the outages, we are considering the purchase of a line interactive, pure sine wave UPS that is rated at 1500 VA and 1050 watts. I have been told and have read that lower priced simulated sine wave UPS will usually overload when they encounter inductive loads. Based on our system requirements, do you believe the selected device should be adequate to handle the relatively infrequent inductive loads (bell ringing cycles) without overloading? Of course, if we know the power is out, we will have the church office delete any programmed rings until the power is restored. That way, the UPS will only have to handle the very small power load of the powered up processors.

Again, any input is greatly appreciated.

Sincerely,

Steve

Read more: http://www.physicsforums.com

I would recommend isolating the electronics from the bell solenoid loads first, In other words, is the goal really to ring the bells when the power is out or to keep the electronics from being damaged by the outage? I would use solid state relays to drive the bells - these are isolated and protected from line transients and so forth. something like this should workhttp://www.mouser.com/ProductDetail/Omron-Automation-and-Safety/G3NA-210B-DC5-24/?qs=sGAEpiMZZMvF1DT/iDLzVhcn0ZUkc0tE92YyBQP+6OQ=
then make sure you have TVS protection on the input AC line and simply run the bell circuits right off the AC line and the UPS and protection circuitry is used for the electronics - that way you dont have to worry about the load on the UPS.

its a good idea to make sure you have enough bulk capacitance as well as some .1 uF caps to clip transients when they occur. As you say a good AC line filter and TVS protection on the AC line cant hurt. Dont forget fuses in all of this for safety. you did not mention if the power supply for the electronics is a switching supply or linear supply? can you let us know that please? A switcher can usually be supplied by just about any UPS

I would use car-batteries to drive the solenoids. They are excellent for that use, refer to starting your car....
Use an relay to isolate the batteries from power when ringing! And optocoupler isolating uP from solenoids will safeguard your Rabbits .
You have protection from power failure and a no impact on your powerline .
The only highpower (expensive) things are safety diods and solenoids drivers.
This way you wont notice a minor power failure. If the embedded processors have a ups.
Good luck

First of all, I'm curious: Why one Rabbit per bell? It seems to me that one Rabbit could easily handle all three bells. I mean, the thing has 40 I/O lines, far more than enough to handle 4 solenoids.

Next, if the bells are not to be run during a power outage, I would only battery backup the low voltage supply to the processors and let power fail on the solenoid circuits (i.e. have your designer add a battery to the VDD line with a battery charging system [can be as simple as a resistor to a lead/acid gelcell], or even better, use an UltraCap [see description below]. Then have your designer use one of those 40 I/Os to monitor the mains voltage, and when it drops below a set-point (i.e. when the power fails), it goes into a power down routine that gracefully shuts down the processor(s) -- fairly easy to do and really they just have to be put to sleep--then,even if the backup power fails [i.e. the battery/UltraCap runs down] the processor(s) won't be "compromised"--they'll merely stop running.

An added plus of this design [low-voltage battery back-up with automatic shutdown] is that the bell ringing schedule won't have to be altered -- let them ring--or try to ring. Nothing will happen, even if the processor(s) try to energize the solenoids--because, there will be no power available to the solenoids.

As for whether a lower priced UPS will have problems with an inductive load, I would ask the manufacturer. And, if the UPS is only used to keep the processor(s) powered up, then no inductive load is involved, even if the little dickens attempts to ring the bells -- with no power to the bells, no inductance is energized or de-energized. (unless relays are being used to switch the solenoids--in which case, also remove power to the relays, during a power outage).

UltraCap low voltage backup system: Allowing 100ms for the Rabbit(s) to shut down [should be plenty of time], and using a 2.7V UltraCap with one of the amazing Pololu step up regulators (see below) to boost the UltraCap voltage (even when it gets down to 0.5V!) to whatever VDD is on those Rabbits, and allowing the UltraCap to only charge to 2.5V, and assuming worst case efficiency on the Pololu power converter as 70% [for a 5V output device], apply the following math to calculate the size of the UltraCap:

1. E = PT/e = (6W * 100ms)/.7 = 0.86W-s ['e' is the efficiency of the Pololu converter and 'E' is the energy required to keep the Rabits running until they shutdown/go to sleep]

2. E = 0.5CV² ; and since the cap voltage will only apply until it drops to 0.5V (min. voltage on the Pololu converter): E = 0.5C(V - 0.5)² = 0.5C(2.5 - 0.5)² = 2C: C = E/2

3. C = 0.86/2 = 0.43F

4. Keep the UltraCap topped off with a current limited 2.5V regulator, or just a 2.5 volt regulator and a 250Ω 1/4W series resistor (which will fully charge the capacitor in around 540 seconds). Or a 100Ω 1/2W resistor for 215sec, etc.

5. Use a pair of Schottky diodes to switch in the backup power upon power failure [drawback: you lose 0.3 to .05 volts to diode forward drop], or just run the rabbits off the backup system (but, increase the capacitance, so it can charge while powering the Rabbits -- drawback, of this system is: there will be a delay in start up as the Cap charges from 0V to 0.5V [i.e. until the Pololu converter kicks in]

So, a low cost backup system, that is far more reliable than a battery system, is possible -- and could even be used to keep the RTCC osc running (if such is used), thus preventing the need to reconfigure the time and date.

URL to Pololu StepUp regulators: http://www.pololu.com/category/132/step-up-voltage-regulators