UPS, part 2

[Migizi]

gpez points out some important consideration with the different type's of UPS models.
Another very important consideration is the type of output waveform.

the less expensive UPS models output a modified square wave instead of a true sinusoidal waveform.
less tolerant devices may suffer from non sinusoidal inputs

Here is a good link with examples:

Database of UPS inverter output waveforms

 

[MarqueePhillip]

FYI Tesla lowered my hz to 64.5 or 63.5(cant remember what he said) and all is good. Took them about a week to do but it worked!

 

[h2ofun]

FYI Tesla lowered my hz to 64.5 or 63.5(cant remember what he said) and all is good. Took them about a week to do but it worked!

I asked them to go to 62. How can one check to see what it is changed to?

 

[jboy210]

I asked them to go to 62. How can one check to see what it is changed to?

Good question. I just waited until my PWs were full, the sun was out, and dropped the grid. PWs took over powering the house, then solar topped off the battery, and finally, the gateway shut down solar as I monitored frequency with meter and it went close 62 Hz. I watched and made sure it did that a cycled a couple of times. I am sure there is a better way to do this, but that was very realistic.

 

[BGbreeder]

I asked them to go to 62. How can one check to see what it is changed to?

Since you have a generator as well, I recommend a frequency meter. $10 or so from Amazon or Ebay. You could also use a DVM set to Hz.

All the best,

BG

 

[jjrandorin]

Good question. I just waited until my PWs were full, the sun was out, and dropped the grid. PWs took over powering the house, then solar topped off the battery, and finally, the gateway shut down solar as I monitored frequency with meter and it went close 62 Hz. I watched and made sure it did that a cycled a couple of times. I am sure there is a better way to do this, but that was very realistic.

Thats what I did, as well, but I did the monitoring with a kill a watt type device. I would link the one I bought (from amazon) but the price seems to have close to doubled since I bought it in feb of 2020.

 

[jboy210]

Thats what I did, as well, but I did the monitoring with a kill a watt type device. I would link the one I bought (from amazon) but the price seems to have close to doubled since I bought it in feb of 2020.

I needed an excuse to upgrade my DMV, so having frequency measurement to test the Powerwall Gateway was my justification. BTW, it turns out even the cheap DMMs on Amazon have frequency measurement. Unless you use it for work and the boss is paying, no need to plunk down $200+ for a Fluke.

 

[h2ofun]

Thats what I did, as well, but I did the monitoring with a kill a watt type device. I would link the one I bought (from amazon) but the price seems to have close to doubled since I bought it in feb of 2020.

I just ordered this

https://www.amazon.com/Poniie-PN1500-Electricity-Electrical-Consumption/dp/B07VPTN8FZ/

 

[gpez]

Many of the "mid grade" residential UPSes have a display that includes input frequency. My CyberPower LCDs (BRG1500AVRLCD and CP1000AVRLCD) do and while I'm not sure about accuracy the precision is down to the 0.1hz. The home PC software doesn't show it but it does appear on the "enterprise" ones do.

 

[jjrandorin]

I just ordered this

https://www.amazon.com/Poniie-PN1500-Electricity-Electrical-Consumption/dp/B07VPTN8FZ/

That actually is the same brand and similar price that I paid before.. it looks like the newer version of the one I bought, so it should work fine for what you want

 

[dc81]

Is this frequency shift still applicable to Tesla's inverter? I'm still waiting for PTO so haven't been able to test it out yet.

 

[eml2]

You can find the frequency from the Web API. First, use your browser to login, then go to <powerwall ip>/api/meter/aggregates, look for battery.frequency.

 

[dc81]

You can find the frequency from the Web API. First, use your browser to login, then go to <powerwall ip>/api/meter/aggregates, look for battery.frequency.

s/meter/meters/


Doesn't that show the current frequency, not the frequency it would shift to in order to cut off solar.

 

[gpez]

s/meter/meters/


Doesn't that show the current frequency, not the frequency it would shift to in order to cut off solar.

It does, so the suggestion would be to simulate an outage then query the API to see what the local frequency is.

I did notice during my annual system test last month that the Load Frequency and Solar Frequency lagged the Battery Frequency by about 4 seconds so either be sure to check both or give Load Frequency a few moments to catch up.

 

[wbhokie]

There are 3 types of UPSs, in order of cost: standby, line interactive, and online. Standby provides no regulation until the battery kicks in. Line interactive can boost or reduce voltage but not frequency. Only online models, the most expensive, provide output regulation of both voltage and frequency when not on battery. Eaton makes online UPSes, but they're very pricy, such as the 9SX Double conversion UPS | Network UPS | 9SX | Eaton. A quick search shows that a 1000 VA 9SX model is ~$1300 retail.

The Eaton 3S UPS which many here have suggested is standby (Eaton 3S UPS | Home & Office Equipment | Eaton) and the Eaton 5S UPS is line interactive (Eaton 5S UPS (550-1500 VA) 230V). Neither provide output frequency regulation until the battery kicks in.

Eaton has a nice FAQ for this: Uninterruptible Power Supply (UPS) FAQs | Eaton.

tl;dr those are the $$$ UPSes

Not to be repetitive, but I can confirm all of these models work without any changes or calls to Tesla. 5S700, 5S1500, 3S550. They are a bit pricey but you don't really need any of the normal UPS features since the Powerwall itself is what is your real battery backup.

 

[zƬesla]

They are a bit pricey but you don't really need any of the normal UPS features since the Powerwall itself is what is your real battery backup.

You do need the primary UPS feature as otherwise a power blip will shut off a server.

We recently had a third PW installed to address an unrelated issue. One of the test cases I ran yesterday was grid out while in full solar production and PWs full. There was a loss of power for a few seconds which would have caused my server to shutdown without a UPS.

Since Tesla lowered the frequency to 62Hz last month, the APC UPS noticed the brief outage then reconnected. Prior to the latest Tesla change (it was at 63Hz before), the UPS kept discharging even with the PWs supplying power

 

[wbhokie]

You do need the primary UPS feature as otherwise a power blip will shut off a server.

That's what I mean. Nothing other than than the ability to account for 10s or so max downtime. No other features are necessary.

 

[Rapidrezults]

I recently had a system installed with 2 PWs and the new Tesla inverter and all of the simulated outages that I've had (turning off the main breaker) have been almost seamless. I noticed, not even a flicker but a slight dim in one of the lights but all of of my routers (mesh) and modem were still intact and online. Not one clock in the house was even tripped for reset. I was thoroughly impressed. Unless of course this version of a simulated outage is not actually simulating a real outage?

 

[wbhokie]

I recently had a system installed with 2 PWs and the new Tesla inverter and all of the simulated outages that I've had (turning off the main breaker) have been almost seamless. I noticed, not even a flicker but a slight dim in one of the lights but all of of my routers (mesh) and modem were still intact and online. Not one clock in the house was even tripped for reset. I was thoroughly impressed. Unless of course this version of a simulated outage is not actually simulating a real outage?

That is how most are. The time of the transition depends on how much load being placed on the powerwall. The majority of them will be exactly as you described but they can be longer.

 

[jjrandorin]

I recently had a system installed with 2 PWs and the new Tesla inverter and all of the simulated outages that I've had (turning off the main breaker) have been almost seamless. I noticed, not even a flicker but a slight dim in one of the lights but all of of my routers (mesh) and modem were still intact and online. Not one clock in the house was even tripped for reset. I was thoroughly impressed. Unless of course this version of a simulated outage is not actually simulating a real outage?

Whether you notice it or not depends on a few different factors. There was a thread recently that had a decent amount of discussion around when you might notice it and when you might not, and why.

Here is a link to that thread: How come power changeover was not seamless during power outage blip?

I recommend skipping / ignoring any posts in that thread that devolve into the meaning of the word "seamless", if you choose to review that thread.