Generator install - where to start

[WinterDeere]

One, see if you can adjust the input power acceptance settings on your UPS to allow a larger input power spread (+/-VAC, and a larger frequency spread).

Yes, all UPS's had already been set to widest possible tolerance range, or in APC parlance, "lowest sensitivity". But given how shady APC is with all of their specmanship and vague literature, I'm honestly not sure what parameters are being adjusted. Voltage bounds only? Frequency? THD? No idea...

Two, spend a little time with your Firman to make sure the frequency is set correctly (my suggestion would be to try for 61.5-62Hz unloaded), but try to see what works with your loads, and to dial the voltage up a bit to allow for the distance from the generator to your load and for the resulting voltage drop.

Being the classic engineering nerd, you just know that I adjusted the governor on my genny to hold 60 Hz. It as running 64 Hz out of the box, but the governor on this one actually works really well and holds steady. After adjusting it down a bit, I see it holds a near-perfect 59.5 - 60.5 Hz under all steady load conditions.

I'm not aware of any voltage adjustment mechanism. The onboard display reads 245 volts, so a hair high, but still in 240 ± 5% tolerance range. I have not checked balance on this one though, nor have I checked to see how equally both legs are loaded.

I also haven't gone so far as to hook up and check THD, spurrs, or noise, but I easily could with my o'scope. That said, hopefully I won't be relying on this stupid generator much longer.

So, I can't guarantee that your new generator will make the UPS happy, but possibly, even probably, as the new generator probably has a better voltage regulator, and probably a larger rotor, which adds to the stored energy, which helps keep voltage and frequency within specifications. Adjusting the UPS parameters will probably help a lot, and may even completely solve your issue of the UPS switching on/off line faster than it can charge.

Are any of the installed generators of the inverter type, or are they all old-school induction motor types?

I recommend Eaton UPS units. They aren't inexpensive, but they are in my experience rock solid, and have the ability to adjust a wide variety of parameters for the specific needs of a site.

I should look at them, when it comes time to replace the current fleet. I actually used to buy very large industrial 3-phase UPS's, so I'm quite familiar with the options and technologies out there. I used to really like the 3-phase inverter type UPS's made by JDSU, as your equipment was never connected to the mains, there was no switching and no possibility of transient pass-thru. But they cost like $25k each for small rack-mount units 25 years ago, I can't even imagine what they'd cost today.

I used to also buy 100,000 - 300,000 VA UPS's to run our test labs. They were the size of Euro refrigerators.

I went with APC here, because they were quick and cheap, when I was starting a new self-funded business. They get the job done, but maybe aren't the most robust option when trying to run off a dirty generator.

 

[WinterDeere]

Oh, and I learned that the power outage was caused by someone having a "medical emergency" behind the wheel. Took out a pole, flipped their car, and landed in the front yard of one of my daughter's former teacher and babysitter. A right big mess.

No word on the driver, hope they're okay. Looked like the car had about a million airbags deployed, as it sat on its roof in the ditch alongside the road.

 

[ning]

Yes. Most of my simulation runs are actually under 30 minutes, but optimizations or parameter sweeps taking hundreds or thousands of consective runs are common. Yesterday's chore was one such task, 234 iterations of a simulation task that was probably only 10 minutes.

It presently runs on a 32-core Xeon machine, which is a power hog, but with one big UPS for the PC and a few more smaller UPS's for networking hardware and NAS, I usually have at least 30 - 40 minutes runtime on the PC and many hours for network and NAS. This has always been plenty of time to set up my portable generator, such that there's no interruption.

That was with my old generator, which blew up last year. I replaced it with a Firman, only thing I could get in a pinch, and the UPSs just will not stay happy on that genny. They constantly toggle, and apparently lose charge faster than gaining it, to the point where the workstation shut down about 3 hours into yesterday's outage.

This is one of my biggest concerns with a $10k permanent generator install... will my UPS's be happy on it? I have four critical UPS's in the house, between workstation, networking hardware, and vector network analyzer, and several others that are less important. I can't have them toggling hundreds of times per hour between battery and generator source.


The software handles distributed computing and MPA, more for splitting between multiple large-scale workstations, than any small Raspberry Pi clusters. But it adds about $10k per node per year to the leasing cost, and usually the benefit is negligible... sometimes even negative. The trouble is the overhead required to split the calculation task onto multiple nodes is high enough to undo the benefit of parallelization larger than that already built into a reasonable 24 - 64 core Xeon machine.

Some of the solvers also support GPU. In fact, I ran it for many years on two workstations (distributed), each workstation having Tesla K80 GPU's. This was back in the days when the best solvers ran memory-bound processes that could take 100 hours to solve on the best multi-core x multi-socketed CPU's. Then it was worth the processor and networking overhead to split a solver process onto multiple machines, and even port to GPU's.

But today's solvers are more efficient, and the latest Xeon CPU's are so much faster, I don't really see much need for all of that overhead anymore. Oh, and starting this year, the software lease comes with many hours of cloud computing, which allows me to port the rare very-large-scale task onto their cloud. I haven't had need for that, in the short time it's been available, single solver runs that take long enough (> 2 hours?) to require that, are fairly rare.

Seems like you may be well-served by a partial house battery backup system (ie, your office/LAN/NAS), which itself may be recharged with a generator. Could be easier to count that as business expense as well, separated from the rest of the house.

 

[JRH02]

What's the problem? We've been off for 16 years. The electronics isn't the problem. It's the rest of your life's consumption. It's that getting in the way of your work.

 

[Larry Caldwell]

Seems like you may be well-served by a partial house battery backup system (ie, your office/LAN/NAS), which itself may be recharged with a generator. Could be easier to count that as business expense as well, separated from the rest of the house.

PG&E is distributing residential 1--13 kwh battery backup systems to vulnerable users. If the system is in mass production, it might be worth looking at even without the subsidy.

Residential Storage Initiative | PG&E

Learn about the Residential Storage Initiative (RSI) pilot program to see if you qualify to get a home battery storage system installed for free.

www.pge.com

 

[WinterDeere]

Thanks, guys. I see storage as a huge benefit, not only for the uninteruptibility aspect, but for surge loads. Although base load on the house rarely gets much above 55 amps, just one of our larger HVAC units pulls 60 amps all on its own, for a few seconds on startup. Same issue with some of my shop machinery, although use of that is more discretionary, not really needed during a typical outage.

It sure would be nice to install a generator < 100 amps capacity, and rely on storage for surge loads, than to have to install a generator large enough to handle every possible surge load combination.

Spent all day in the shop today, packing up a large order that I thought would only take me a few hours. Potential vendors were already closed, by the time I got to sit down at my desk. I'm spending tonight doing what should've been done last week, and tomorrow doing what should've been done today... that's just how it's been, the last several weeks.

 

[GrandpaSmurf825]

Happy Generac whole house user here. I had mine installed by a local electrical contractor that is an authorized Generac dealer/installer. I'm on natural gas and that knocks the power output down to 19,500 from 22,000 watts. It is NOT brushless as I initially stated. grsthegreat pointed that out and i looked at the parts break down and confirmed I was wrong. I have had it for not quite 2-1/2 years. It's not logged a lot of time working. After a Derecho went through just 3 or so miles north in 2023 and all the carnage that brought with some people being without power for up to 2 weeks I felt it was a good investment. It runs I believe 30 seconds after start up before the power transfers. Upon power restoration it's seamless and instantaneous on return to line power. It exercises every 2 weeks for I think 5 minutes. It sets outside my shop on a pad and is very quiet. According to Generac 67 decibels under full load.

 

[grsthegreat]

Happy Generac whole house user here. I had mine installed by a local electrical contractor that is an authorized Generac dealer/installer. I'm on natural gas and that knocks the power output down to 19,500 from 22,000 watts. It is BRUSHLESS as info. I have had it for not quite 2-1/2 years. It's not logged a lot of time working. After a Derecho went through just 3 or so miles north in 2023 and all the carnage that brought with some people being without power for up to 2 weeks I felt it was a good investment. It runs I believe 30 seconds after start up before the power transfers. Upon power restoration it's seamless and instantaneous on return to line power. It exercises every 2 weeks for I think 5 minutes. It sets outside my shop on a pad and is very quiet. According to Generac 67 decibels under full load.

The generac 22KW is not brushless.

 

[GrandpaSmurf825]

The generac 22KW is not brushless.

You're right. After looking closer it does have a brushes. That doesn't change my satisfaction level and if it were to become necessary I can change the brushes if they wore out. However I doubt that will be an issue. They should be good for at least 1500 hours and it doesn't even have 10 on it yet. Original post is corrected as well.

 

[ponytug]

Yes, all UPS's had already been set to widest possible tolerance range, or in APC parlance, "lowest sensitivity". But given how shady APC is with all of their specmanship and vague literature, I'm honestly not sure what parameters are being adjusted. Voltage bounds only? Frequency? THD? No idea...


Being the classic engineering nerd, you just know that I adjusted the governor on my genny to hold 60 Hz. It as running 64 Hz out of the box, but the governor on this one actually works really well and holds steady. After adjusting it down a bit, I see it holds a near-perfect 59.5 - 60.5 Hz under all steady load conditions.

I'm not aware of any voltage adjustment mechanism. The onboard display reads 245 volts, so a hair high, but still in 240 ± 5% tolerance range. I have not checked balance on this one though, nor have I checked to see how equally both legs are loaded.

I also haven't gone so far as to hook up and check THD, spurrs, or noise, but I easily could with my o'scope. That said, hopefully I won't be relying on this stupid generator much longer.


Are any of the installed generators of the inverter type, or are they all old-school induction motor types?


I should look at them, when it comes time to replace the current fleet. I actually used to buy very large industrial 3-phase UPS's, so I'm quite familiar with the options and technologies out there. I used to really like the 3-phase inverter type UPS's made by JDSU, as your equipment was never connected to the mains, there was no switching and no possibility of transient pass-thru. But they cost like $25k each for small rack-mount units 25 years ago, I can't even imagine what they'd cost today.

I used to also buy 100,000 - 300,000 VA UPS's to run our test labs. They were the size of Euro refrigerators.

I went with APC here, because they were quick and cheap, when I was starting a new self-funded business. They get the job done, but maybe aren't the most robust option when trying to run off a dirty generator.

I should have guessed that you would have set the generator by the numbers. My apologies. 245VAC should be on the money, and if the Firman is holding +/-0.5Hz, no load to load that is amazing. Can it do that if you get 0 to 80% load in an instant?

Given the generator numbers, I rather doubt it is the source of the issue. Is it possible that the APC units are close to maximum load and that is causing them to be dropping off due to internal heating/over temperature? My other thought is that either the load, or something else in the house is feeding noise, or causing frequent surges.

I hadn't thought about JDS/Uniphase in ages, but yes they made some monster, and wonderful, UPS units.

Most of the the backup generators, ok almost all of the backup generators that I know of are not inverter. To me there is a logic in avoiding inverters for permanent backup: there are more parts to fail. Keep It Simple. Larger generators often surpass local grids for power quality, but that is a trust but verify item.

If you go with battery storage, I would choose the storage before choosing the generator, as not all battery systems support generators, and those that do support generators may only support certain model/brands. I personally like the Enphase battery setup, because it is built around multiple microinverters, like their solar systems, so if an inverter fails, you only lose a percentage of your battery storage system, not the whole thing. Belt and suspenders. I know too many folks who have had an inverter fail when it was important. I would rather lose a little rather than all of the battery.

All the best,

Peter

 

[grsthegreat]

You're right. After looking closer it does have a brushes. That doesn't change my satisfaction level and if it were to become necessary I can change the brushes if they wore out. However I doubt that will be an issue. They should be good for at least 1500 hours and it doesn't even have 10 on it yet. Original post is corrected as well.

Your right. I service these units and i very rarely have issues with the brushes.

 

[WinterDeere]

I should have guessed that you would have set the generator by the numbers. My apologies. 245VAC should be on the money, and if the Firman is holding +/-0.5Hz, no load to load that is amazing. Can it do that if you get 0 to 80% load in an instant?

That's why I said "any steady load", or something to that effect. I'm sure that when the well pump or our largest and oldest HVAC compressor kicks on, there's a momentary dip... just no way to avoid that. But in standing and watching the thing for awhile, with my wife and kids doing whatever they regularly do in the house on a cold winter day, I never saw it budge more than ±0.5 Hz.

Given the generator numbers, I rather doubt it is the source of the issue. Is it possible that the APC units are close to maximum load and that is causing them to be dropping off due to internal heating/over temperature? My other thought is that either the load, or something else in the house is feeding noise, or causing frequent surges.

Not in this particular case. Two of my UPSs are right here in my office, so I can easily watch them. One was loaded only 4%, the other 35% - 40%, and both were randomly toggling between genny and battery.

Most of the the backup generators, ok almost all of the backup generators that I know of are not inverter. To me there is a logic in avoiding inverters for permanent backup: there are more parts to fail. Keep It Simple. Larger generators often surpass local grids for power quality, but that is a trust but verify item.

Good to know.

If you go with battery storage, I would choose the storage before choosing the generator, as not all battery systems support generators, and those that do support generators may only support certain model/brands.

I'm so swamped with work right now, with no let-up in sight, that I'm going to be looking for a turn-key solution. Hand them a purchase order or credit card, and just let them handle the whole thing, start to finish. That said, I'll get a few quotes and only work with the vendor who can offer the best solution, which I strongly suspect will involve some level of battery storage.

I have a lot to learn, and not the amount of time I usually like to devote to a purchase like this. So I'll be relying on weeding through the recommendations of the few vendors I manage to call, and weighing them against one-another. I'm probably going to be one of those customers they hate, nature of being an engineer, but that's where we are.

I personally like the Enphase battery setup, because it is built around multiple microinverters, like their solar systems, so if an inverter fails, you only lose a percentage of your battery storage system, not the whole thing. Belt and suspenders. I know too many folks who have had an inverter fail when it was important. I would rather lose a little rather than all of the battery.

I'll check that out, and be sure to ask vendors about failure mode of the battery rig. I agree, we don't want battery storage with a single point of failure taking down the whole works. Batteries fail, especially when you have a permanent installation that can be ignored a decade at a time, and then only tested in emergency situations.

 

[ponytug]

@WinterDeere I'm glad you have the work, but yes, it has its downsides. PM if you want any specific help.

Given the loads on your UPS units, I suspect that, yes, it is your HVAC, or well pump kicking in and throwing the APCs offline, just by the combined voltage and probably frequency droop, but whatever is actual the issue driving the APCs offline, it is happening frequently enough to keep the APCs offline enough of the time to not be charging. That's a lot of the time, especially for the 4% load unit. Short term, I'd try some new surge/EMI protector strips, e.g. from Tripplite (Isobar or medical versions) or Belkin, to see if they can attenuate enough of the "whatever" to keep the APCs happy until they are replaced. Dollars to donuts, a ferroresonant voltage stabilizer would help, but they are big, heavy, energy hogs, and usually pretty loud.
I've used these for years;

Amazon.com

Isobar Surge Protector, 8 Outlet, 3840 Joules, Metal Case | Eaton (Pricey!)
I put them on the grid side of the UPS units.

One way to avoid not noticing that the batteries are not working is to have them cycle 80-60-80% every day, so you pick up the failures more quickly and hopefully before outages happens. That's what I do. Modern battery storage systems pretty much all have apps and alarms to make it easy.

All the best,

Peter

 

[grsthegreat]

On buildings with alot of ups systems, like call centers, the generators are like 4x rated need. Helps with the ups devices.

 

[ponytug]

Great point, although I haven't seen the 4x number before, but there is certainly redundancy nine ways to Sunday in a modern data center.

Personally, I think that data centers are a power quality nightmare. A hundred thousand identical devices mean there is the huge need to stagger load balancing across power circuits, as well as in time, and for intermediate transformers and AC/DC/AC inversion for better power quality control. The generator side is pretty fascinating as well as those get interconnected, synchronized, and load balanced in what I think are some interesting ways. Basically rather large grid complexity in a small area with identical consumers. Plus all of the HVAC refinements and control. There are lots of redundancies and capacity across the web of systems.

I think that the design and operation are quite interesting, if like me, you find that sort of thing of interest.

Definitely not a "throw the main breaker on" sort of place.

The flip side is also true. If the grid operator screws up the voltage levels as they did in Virginia a while back, then you have pretty much every data center in the area simultaneously disconnecting and switching to their backup power in a microsecond. That amount of load causes a big voltage surge for the remainder of the utility's grid. That means equipment losses up and down the grid, as well as safety shutdowns rippling through the grid. So utilities are finding their feet being held to the fire in meaningful ways ($$$ losses, and upset customers) to actually meet their power quality standards in ways that homeowners and small industries never have. E.g.

https://www.wsj.com/business/energy-oil/a-new-threat-to-power-grids-data-centers-unplugging-at-once-741f1bda

Coming from a background of setting up computers on lousy power grids, to me this is/was a 100% foreseeable issue. I am interested in how various utilities decide to deal with it, lest it turn into a Portugal/Spain grid issue.

All the best,

Peter

 

[ning]

Thanks, guys. I see storage as a huge benefit, not only for the uninteruptibility aspect, but for surge loads.

My experience with battery backup (Enphase) is that typically *I* don't notice any switchover from grid to battery, but occasionally I do still see a flicker in the lights.

A computer system running plugged in directly would still fail with the switchover onto battery - there's still a non-zero interruption, it's just much shorter; you'd still want UPS though it wouldn't need much reserve capacity any more.

 

[ericm979]

You can put soft starts on the A/C and maybe also the well pump.

I have solar a Sol-Ark inverter and Briggs SimpliPHI batteries. When the power goes out the lights barely flicker. It's a 5ms cut over. Most of our computers are laptops but the desk top, NVR etc aren't on UPS and didn't reboot.

The solar and batteries were expensive. We had lots of outages in California. The utility there is terrible but also the mountains are steep and the trees are tall. The utility here is seeing what California utilities get away with and copying some of those ideas. One of those is special breakers that cut power very aggressively. It's to prevent fires, which is good, but its mostly a bandaid that allow skimping on maintenance. The summer before last our property in Oregon had more outages than our place in California. When I saw that I wanted the whole house backup.

 

[ning]

The summer before last our property in Oregon had more outages than our place in California. When I saw that I wanted the whole house backup.

We had a zillion power outage - short, unplanned (quick trip breakers, cars crashing into poles, and very occasional tree vs power line) as well as planned (pge's "psps") - after the slew of bad fires in the later twenty-teens, which pushed me to get the whole-house backup. Since then, so far (knock on wood) no nearby fires, no psps (presumably lots of line work - nonstop it seems nearby with almost quarterly visits on our land (!) - have made PSPS only affect more remote areas any more), only the previously typical mini-fails.

I'm still glad I've got the battery backup for certain... sounds like if your utility is looking to emulate PGE's (hopefully former, but probably permanent) ways though, yours will get more and more of a workout...

 

[ericm979]

... sounds like if your utility is looking to emulate PGE's (hopefully former, but probably permanent) ways though, yours will get more and more of a workout...

They call the special breakers ESS which is Enhanced Safety Setting or something like that. Normal breakers have a recloser. If they were triggered by something that gets fried and falls away from the wires, like a falling branch or a squirrel, the power glitches and then comes back. The ESS is a more sensitive breaker with no recloser.

We were in the same outage block as the nearest town. The summer the power kept going out, people there were extremely agitated. The mayor had a bunch of meetings with the utility and the utility promised to turn the settings down and split the block up into smaller units. Exactly what PG&E was forced to do though they took longer to do it.

 

[ponytug]

@ericm979 I think that they are also called sectionalizers. They generally reduce the amount of energy transferred into a short below the ignition point of tree limbs and such. I'm not saying always, but often.

All the best,

Peter