Sorry, I gave Ann some bad advice I want to correct. Constant voltage transformers use 'ferroresonent technology', a phrase which nagged at me. In checking it out, they in fact use resonance at exactly 60Hz as part of the control scheme. A generator may vary output frequency significantly for variable loads resulting in poor performance, thus generators are not a good match for these power conditioners.
After confirming this with Sola/Hevi-Duty tech support, they recommended 'double conversion' UPS's for mission critical electronics with standby generators. These, as I understand it, eliminate any 'switch' between line power and battery, instead the output inverter runs continuously on the battery supply voltage, and the incoming power only charges the battery. Somehwere in the description of a UPS you are interested in, you might also look for 'compatible with generators', before purchase.
We also did a bit of experimenting in my plant, and found that sequenced power-ups seem to have eliminated our instability problems. (We tested by hand, and will automate it shortly. We also did everything we could with software in the UPS's to reduce sensitivity to volts and frequency, as cvandewater suggested.
Appended are links to a couple of technical white papers from Cummins. The first may make you despair of ever using a generator with a UPS, but take heart, and go for it with eyes wide open. The second one has some mild contradictions with the first, but both are the best reads I found, if you are trying to spec a back-up generator for electronics.
bob from maine, we've looked at UPS in the past and the ones we saw are insanely expensive and don't last long enough for the home user who only needs an occasional 1-2 days during a power outage. The ones targeting clean computer power, at least what we have found, are also hard to buy. Meaning, if you're not a specialist in power, trying to translate all the terminology into the features and specs *you* want is too big of a headache. The buyer should not be having to do so much work! I'm surprised that an industry that's been around this long, selling stuff to non-techie consumers for this length of time, is so bad at communicating well enough for potential buyers to buy something. Even though I used to work in high-end UPS 30 years ago in marketing, my eyes glaze over and my mind goes blank every time a salesman starts talking to me about joules.
bdcst, thanks for the redefinition--I agree. Yet more reasons why we stopped looking at "UPS."
I hate what the industry calls an "UPS", uninteruptable power system as most are not truly uninterruptable and should be instead be called an SPS (standby power system). These so called UPS's are line interactive devices lying in wait to pounce on a loss of a couple of cycles of AC power and fire up their inverters whilst switching the load from the power line to the inverter's output. This does not happen instantaneously. The result is a glitch to the equipment being protected. Depending upon the energy storage capacity of the protected equipment's own power supplies this might be acceptable, or not!
I've seen too many instances where "Smart" UPS's, (without naming names :-), fail to switch rapidly enough if at all when a power loss isn't a draconian cut but rather a slow brown-out to nothing. Then there are the units that auto test their batteries weekly by switching the load to the inverter for a few seconds. Some of these have been responsible for odd seemingly random equipment lockups. Worse yet is the test that fails due to a battery with a shorted cell such that the DC power drops low enough to cause the controller to lose it memory of the failed test. So, the bad battery indicator light doesn't stay on long enough to warn you that a test in your absence had failed! How smart is that?
I've been replacing all of my so called UPS's with true double conversion units which are always powering the load from its batteries and inverter. The one big drawback might be MTBF for the inverter since its always running. But the benefit is truly uninterruptable power that is fully isolated from power line transients. In addition I specify external automatic bypass switches to restore power around a failed UPS and to permit routine swapping of UPS's without having to power down critical loads.
CVT's typically are power hogs as they run at full saturation with 60Hz resonant windings. They can run hot and add a lot of acoustic noise. But they do the the job well of isolating output powe from line transients. Just remember that there are two types. One delivers fairly dirty distorted AC while the other provides a fairly clean sine wave.
A fairly new technology in inexpensive standby power is the gasolene driven DC generator/AC inverter combination. Honda has several models. The inverter produces a true clean 60 Hz sine wave without regard to generator speed! And that's the key. The motor can idle when the energy demand is low and yet the generator will still put out 120 VAC at 60Hz.
Honda has lightweight 1-2 kW had carried generators up to a dolly mounted 6 kW 120/230 VAC model that'll also run your well pump using the same inverter technology.
There are also a bunch of far less expensive versions now on the market though I'm not sure about ultimate quality and reliability.
I assume the Honda inverter circuitry has built in regulation control and fail safe shut-down. It does control motor throttle when it needs more energy from the DC generator/alternator. At any rate it kept my fridge cold, my home entertainment center running, computers and Internet working and my ham station on the air during hurricane Irene with no smoke emitted.
I've installed back-up power for public safety departments and the best way to absolutely guarantee power never, ever goes down is to install a high-end battery powered UPS which powers the radios and computers all the time, the batteries are charged from whatever power comes into the building; the UPS output is sine-wave, voltage and frequency regulated. Generator output power quality is typically directly proportional to cost. A good 4 pole 1800 RPM generator that is electronically governed/regulated and provides noise-free AC costs up to 10X a (relatively) cheap 3600 RPM 2 pole portable generator. If you want good, clean, uninterruptable, reliable power, go to a professional, talk it over and evaluate your options. Inverters, batteries, wind or solar or diesel or gasoline, flywheel/motor/generator sets all are designed to meet a particular end-user need. Single and dual cylinder Briggs, Honda, Kohler, Kawasaki (etc.) engines are providing excellent, relibable power (when maintained) for 1500 to 10,000 hours with no issues, diesel engines frequently run for 12 to 15,000 hours between overhauls. You pretty much get what you pay for.
Typically in a setup for a computer room, the generator does not feed the loads directly, but the genset (as well as the incoming grid) power the UPS'es which will provide uninterrupted power to the computer equipment.
So, a bad genset connection will fry UPS'es (and other non-critical loads that are on raw power, such as lights) but should not affect the computers.
Regarding the load transfer of the UPS'es: indeed, if all UPS'es are of the same brand and type then they are likely to all detect good power at the same moment and according to Murphy, go for the maximum damage (all switch for max load and battery charging at the same moment, to increase the results of their synchronization).
What I would suggest is to review the voltage and frequency limits that the UPS'es accept - often they are set very critical to make the UPS'es look like they are needed a lot, while they can actually work with a much larger variation of voltage and frequency.
If the delay of the individual UPS'es to "jump" on the generator cannot be staggered to spread the load changes, then I suggest to configure the individual UPS'es with slightly different parameters when they consider power "good" again, so they will switch at different times. For example if power is nominally 230 +/- 10% then set 4 UPS'es to 207; 208; 209 and 210V so they likely will sequence their switch from battery backup to generator accordingly. Also try to set the frequency they accept as wide as possible so they don't trip off when another UPS is jumping on the genset, or you get an instable system as you have experienced. Success.
..Interesting, as just recently I converted a motor-home type generator to stationary/portable use for my brother, who lives in a remote area subject to outages.
The unit is a 5KW Onan..2 cyl opposed gasoline. Fairly decent sine-wave output from 4-pole wound stator and slip-ring rotor. The speed control is a mechanical unit on camshaft..appears to maintain 2-3% @ 1,800 RPM. No electronics in this unit to fail. Small fluctuations in speed affect frequency much more than the output voltage.
The output is 2, 120 volt windings that can be connected at junction box, to series or parallel..50 amps @120, 25 amps @240. The output/s are through 2 30 amp circuit breakers. The generator ground (green) is not connected to the neutral. The two slip-ring 120 rotor outputs are completely isolated from each other, or the frame ground. The start winding is a heavy copper, brushed commutator, on the same rotor as the 120 output, and that start winding is also used to generate the rotating field current. A relay works from the oil pressure, to switch this winding from 'start' to 'run'.
I'd guess some of the...um...low-cost units would use a (permanent magnet?) alternator for the generator, and 'speed control' maybe just by adding some load, or the cooling air-vane flap attached to the carb..?..?
I am amazed that anybody would connect a generator without first checking the output voltage. Universal-motor appliances will work well with anywhere between 80m and 140 volts, and any noisy waveform that you can imagine. The ultimate protection is indeed a "constant voltage" transformer of the resonant-saturated-core type. Not only do they hold the voltage fairly constant, they also stop some of the generator harmonics. Unfortunately they demand that the frequency be held very close to the design target value. Also, they are fairly expensive, and not "wonderfully efficient".
Another option would be to use a good quality DC powered inverter and have a generator charging batteries to power the inverter. THat would be a lot more complicated, but it would have the big advantage of not needing to run the generator constantly. IN addition, using an automotive style regulated charging system and commecially available inverters allows a great deal of flexibility, and the option of switching to backup power with very little effort. In fact, using the DC charging system to hold up UPS batteries could be a very cost effective system that would also avoid the need for a large transfer switch. The main downside to this approach would be the need for heavier conductor wiring to the various UPS packages.
Ken, thanks for the detailed feedback--it's very helpful. I know computers run off generators, the problem is how to do it at home with usually lower-end, non-server-farm-quality equipment, especially when running other stuff off the generator, as you point out. We will check out the CVT--sounds like just what we need.
In an age of globalization and rapid changes through scientific progress, two of our societies' (and economies') main concerns are to satisfy the needs and wishes of the individual and to save precious resources. Cloud computing caters to both of these.
For industrial control applications, or even a simple assembly line, that machine can go almost 24/7 without a break. But what happens when the task is a little more complex? That’s where the “smart” machine would come in. The smart machine is one that has some simple (or complex in some cases) processing capability to be able to adapt to changing conditions. Such machines are suited for a host of applications, including automotive, aerospace, defense, medical, computers and electronics, telecommunications, consumer goods, and so on. This discussion will examine what’s possible with smart machines, and what tradeoffs need to be made to implement such a solution.