carteriii
Member
No, sorry, 1% is definitely not enough "room to charge". As I mentioned above, I didn't see the drop in frequency until I had 10-15% capacity, and even then the frequency was still ~62Hz (not standard 60Hz).
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Powerwall 2 + UPS Connundrum - and solution
- Thread starter NinjaVece
- Start date
-
- Tags
- ups 65 hertz
No, sorry, 1% is definitely not enough "room to charge". As I mentioned above, I didn't see the drop in frequency until I had 10-15% capacity, and even then the frequency was still ~62Hz (not standard 60Hz).
I just call the tesla energy phone number. 877 - 961 6752 is what I have as the energy support number. Thats what I called to get the specific issue addressed in this thread. I have never had it ring more than 4-5 times and could always talk to a human at that number.
Not sure what number you are calling, but try that one.
carteriii
Member
I think the post above might have accidentally mis-typed that number. From this page, I believe it ends is 7652: Powerwall Customer Support
Thanks guys!
I managed to get a knowledgeable person, who removed my old gateway. Within a week he said they will push the frequency change (62.7Hz).
Best to you all!
As a newcomer to this subforum, I'm reaching back a couple of years to the initial post in this thread. Tesla says the transition to backup power is "seamless", but as the OP and other owners I've heard say, the switch-over delay is long enough to cause a desktop computer to fail. This irks me because I currently have a 20-year-old solar system with DC-coupled backup using lead-acid batteries that switches in 20 milliseconds. It also avoids converting DC-AC-DC to charge the batteries from solar power. But my roof and my panels now need replacing. It seems to me that if Tesla is selling a truly integrated system of Solar Roof plus Powerwalls they could make a faster and more efficient DC-coupled system (there used to be DC Powerwalls).
But one comment in the OP's post is a question for me. If I understand the design correctly, the Gateway relay is normally closed and just has to open to disconnect the Powerwalls from the grid when the grid goes down. It should not be necessary for any relay to close (causing delay) for power from the Powerwalls to go to the loads. The Powerwalls just need to crank up their inverters, right? And that should be possible within a cycle or two.
aesculus
Still Trying to Figure This All Out
I totally agree but they are not there yet. Still selling independent components in old school AC.
And we should be able to charge the car directly from solar via DC too.
The delay is in that the Powerwalls have both charge, and discharge built in, and may not export any amps to the grid.
If the PW is currently full, and discharging, there will not likely be loss of power. If the PW is near empty, it needs to quickly switch from charging to discharging, perhaps needing to power up a large load immediately or not depending on load profile. This is where you may see loss of power. Powerwall did make some sacrifices compared to a DC coupled system. They gained flexibility and scalability with the AC coupled approach, but this is one of the downsides brief power interruptions.
I cannot say with any certainty how many times out of 100 outages that a home computer would die, but I suspect its less than 30.
How cow this thread is long! Here to report CyberPower ST425 does not work with our situation. I have 2 Powerwalls and 12.5 PV system. I am not supposed to be on yet so in my first few days testing, the solar was charging the 2 PWs to 100% before noon, to prevent excess going to the grid and SCE charging me since I am not supposed to be on yet, I flipped the main breaker to shut off the grid. Like many of you when Solar fully charged PWs and the hertz ramped up my CyberPower ST425 would not pass the power and my main computer crashed after a minute. I repeated this process 3 times thinking something wrong with my UPS, then I found this thread.
In the last 10 or so tests of killing the grid, none of my desktops (except the one on UPS) shutdown or rebooted, so for now I am connecting my main system directly to the outlet as the PWs seems fast enough in most cases. I know eventually I may have a situation where PW is not fast enough and the computer may crash.
This thread is years long, so am curious where most of you landed on this? Given we only need a UPS that faster than the PWs and is capable of 30 seconds running a computer, I don't need $$$ pure sine wave, just something the get over the hump that won't beep or complain about the ramp up in hertz.
In the last 10 or so tests of killing the grid, none of my desktops (except the one on UPS) shutdown or rebooted, so for now I am connecting my main system directly to the outlet as the PWs seems fast enough in most cases. I know eventually I may have a situation where PW is not fast enough and the computer may crash.
This thread is years long, so am curious where most of you landed on this? Given we only need a UPS that faster than the PWs and is capable of 30 seconds running a computer, I don't need $$$ pure sine wave, just something the get over the hump that won't beep or complain about the ramp up in hertz.
(There is even a part 2 thread with more discussion because of the length of this one.)
The best solution is to contact Tesla (probably after PTO - they may not support you yet) and request that they reduce the frequency from 65 Hz to more like 62. (They may only agree to go to 62.5 or 63 Hz.) If they ask, you can tell them that you have a UPS that cannot handle 65 Hz (and you can check what its specs are - I am guessing it will be 60 Hz +/- 3 Hz) and provide them the model info. That solution avoids the need to replace the UPS you have and any other issues related to 65 Hz operation. Then yes, you should only need smaller UPS devices to guard against a temporary power loss during cut-over, as well as to provide surge protection.
UPS, part 2
There is a really long UPS thread from a while ago, and what I am hoping for in this one is to get a listing of UPS models that work seamlessly with Powerwalls. I have an APC Back-UPS NS 650M1 that goes into battery mode whenever the grid goes out and Powerwalls kick in even though we had Tesla...
teslamotorsclub.com
We are using CyberPower units, specifically the 850PFCLCD. They have worked fine on the 8 outages and PW takeovers we have had since November as documented in the Tesla app's Backup History.
I have 4 CyberPower units, 5 if you count the one in my closet that I haven't needed and doesn't have batteries (I promise I'm not a shill). After I got Tesla to reduce my max frequency to 62hz they have been working perfectly during both actual and simulated outages.
wbhokie
Member
I replaced mine with Eaton units. Keep in mind, they don't really need to perform a UPS function if they are on a circuit backed up by a powerwall, they only need enough battery capacity to properly fail over until the power switch. Eaton units are a little more expensive but they all worked without any changes to the existing system. 5S700LCD, 1500, 3S550 are the ones I used. My primary computer (gaming) needed the 1500, but the rest of my devices didn't need the higher wattage rating.
Same here with going with Eaton units. Yes, they're a little more but good quality UPSs (from my experience and what I've heard) and don't have to worry about contacting Tesla or other changes that might come with firmware updates. I mostly have 5S700LCDs but just picked up a Eaton 3S 350 VA recently.
@wbhokie and @thegdgteer I can't stress this enough: do not use the Eaton UPSes without asking Tesla to lower your maximum frequency.
The Intel ATX standard and subsequent power supply design guidelines state that the maximum input frequency is 63hz: Power Supply Design Guide for Desktop Platform Form Factors (intel.com) (page 14). Exceeding that limit could cause all sorts of problems, from crashes all the way up to equipment damage. Of course some PSUs may have higher tolerances but that will vary by manufacturer and model. The Eaton UPSes simply pass that 65hz waveform through to your PC putting it at risk.
Additionally the 65hz frequency is bad for many appliances, especially those with a motor or compressor (garage door opener, microwaves, furnace blower, A/C, etc). Your best approach by far is to get Tesla to reduce the maximum frequency of your system below 63hz with 62hz being a good value. Getting a UPS that is tolerant of high frequencies is a band aid and not a very good one.
Tesla does not change the min/max frequencies with firmware updates. This is a one-and-done exercise.
Well, keeping the line frequency at 63Hz or below is fine. And non-ECM motors are run cooler at their design frequency. So will power supplies that use a line frequency transformer. But any computer supply made in the last couple of decades is an AC to DC converter with a transformer that runs at >20kHz. AC to DC converters that incorporate power factor correction circuits have in principle a limited frequency range, but the range is broad compared to 60Hz or 65Hz. Converters that are small enough to not need PFC will operate even with a DC input. The Intel paper is describing the minimum input frequency range needed for ATX compliance; actual power supplies built today significantly exceed the range.
Similar threads
