The noise from the UPS is the power on self test - I think it's a common feature of APC as mine used to do that too - it will do it periodically as well - it switches over to battery to test if the battery can maintain the load and will alert if the state of charge drops too quickly.
Yeah seems fairly common on APCs to go into self-test after enabling the output. There is an option to stop it from doing that in the self-test schedule settings.
The EPO connector is for datacenters/machine rooms where there's an Emergency Power Off system in place (IE those big red buttons on the walls of datacenters). Hit the button and it kills power to the entire facility in the event of a fire. UPSes are required to have these connectors so that they will not continue to provide power in those EPO events. This is particularly important because the equipment attached to the UPS could be the source of the fire, also, it protects the emergency personnel who are dealing with whatever situation triggered the EPO.
I've dealt with the 3000VA model of these units at work in Australia, they have a superrr nasty hum at all times and will generally be the loudest thing in a rack. I opted for a 1500VA model at home and its practically silent, not to mention I dont have to worry about a 15A outlet. Oh, youll want to setup powerchute on any hypervisors you have as well for automatic graceful shutdown, you can also get a NUT addon for home assistant to get UPS data in there as well.
In case some people are wondering about the VA versus W and power factor. Electrical loads can be resistive, inductive (motor) or capacitive (switch mode power supply). Power factor basically means how well in sync or in phase the AC current is compared to the AC voltage. If you have a purely resistive load, say an incandescent light bulb or an electric heater, your current sine wave perfectly matches the voltage sine wave. Power factor equals 1. If you have a capacitive or an inductive load, the current sine wave will either be ahead or behind the voltage sine wave respectively. Power factor is less than 1.
UA-cam seems to be deleting my comments again.. Why is the UPS rated in VA instead of watts? The manufacturer knows exactly what it can deliver (voltage and current or VoltAmps) but does not know what equipment the user will have. It's up to the user because they know what equipment they're using.
If you have a power supply, it's power factor can be as good as 0.99 or it could be as worse as 0.7 or even lower. It usually depends on the load the power supply is under. If it's a fairly modern design with active power factor correction, you should expect it to climb above power factor 0.9 when it's loaded above 20% I'd say. If you have an UPS, you get watts by multiplying with the power factor. For example 2200VA and assuming 0.9 power factor is 1980W. If you have a power supply, you can divide watts with power factor to get VA. For example 500W and 0.9 power factor is 555VA.
@@Lagittaja The typical rule of thumb I was taught is to assume 0.6 until and unless a properly qualified engineer has done extremely extensive and expensive testing of the exact setup in question. Specifically because very little is actually THAT bad, so it's less likely (though not impossible) one will screw it up with their lack of said engineering qualifications.
This hits home. Just ordered a new UPS for my home rack. The old one is about 8 years old now and the second set of batteries has flaked out. Rather than do another battery replacement, I am going new because the UPS has not reported a single issue with the obviously bad battery. I can't trust a UPS that is blissfully unaware of its own issues. New it is.
Add an ATS to bypass the UPS for firmware updates and maintenance. I have an APC AP4421 and have been very happy with it, prevented downtime on multiple occasions.
Where I work we've purchased a number of the auto-transfer PDUs because we did a risk assessment and noticed that most telecom equipment that's vendor supplied does not have dual inputs. So in the event of a UPS failure or UPS maintenance the network circuits could go down.
That's a good shout - I've used ATSs before to feed single PSU kit in a rack with dual power feeds, never considered one for this use case. Also looks like you can get manual "UPS Bypass" transfer switches which might be an even simpler option. I'll probably keep an eye out on eBay and grab something if it shows up cheap.
EPO ON UPS - wire to an NC relay from the texecom. Fire causes the ups to shutdown when fire is in garage. So if the fire is caused from something on the rack, better chance to kill the fire than not.... Also stops other fires from starting with the heat
I'd like to add when trying to rebuild battery packs for APC units, they use some horrendously strong glue/tape to hold the batteries in. You may need to use a heat gun to get it to release or spend hours carefully prying up with a screwdriver being careful not to short the batteries or impale them.
Do you mean the dry cells glued together? I know APC and Tripp-Lite both do this. But i don't bother separating the battieries and just take them to recycling still bonded together. Usually batteries have been easy to remove and replace in every UPS I've used.
That's good to know! All the more reason to buy one that's already been rebuilt - cells are no longer glued in and therefore will be easy to replace in the future
@camerongray1515 at work we buy the generic APC batteries from CDW they are about half the price of the APC branded ones. I gave up on the rebuild that I had mentioned above after a few hours.
Couple years back, I was very lucky to get a new-in-box surplus Liebert GXT4 1500VA Online UPS for $200. I had to buy the network management card for $120 off ebay.. but still worth it. Only downside is that the fans are pretty loud (since the inverter is always running in normal operation). I have a bunch of Tripp-Lite line-interactive UPSes (SMART1500) and I have had *multiple* devices die due while connected to them after storms -- so not sure I fully trust their surge protection and their AC generation -- although I still use them because getting double-conversion UPSes for everything would be massively expensive and overkill.
Worth noting is that a great many UPSes have serial and/or USB ports even without the network card add-on. An older raspberry pi and Network UPS Tools can often replicate the functionality of those cards and then some for half the cost or less. Assuming the UPS in question is NUT-compatible, of course.
Was at Microcenter yesterday looking at UPS units and was surprised to see Liebert on the shelf alongside Cyberpower and APC. The Liebert was even priced competitively. IIRC 1500VA for like $230, new. Had no idea they were even in the consumer market. We use a lot of big iron Liebert stuff at my job.
We have some shorter APC PDUs at work. The serial port on them turned out to be RJ12. RJ11 has 4 pins, RJ12 has 6. Just a heads up in case you need a serial cable for it.
Yeah, the US tends to just use standard NEMA outlets - the reason simply being density. BS1363 and Schuko are just a bit large for PDUs. NEMA 5-15/20 is compact, even if it lacks any other redeeming features.
The IEC connectors are commonly used on what 240V UPSes exist in the US (and I assume Canada) as well. Tedious as they are to track down, they do exist here. Only reason to use them over more readily available 120V stuff is the slightly better power efficiency that most PSUs have at 240V.
I'm honestly surprised that more stuff isn't implementing USB-C PD as at least an option. PD4 can go to an insane 240 Watts now with a compatible cable, that could run the majority of network gear and even some servers outright. Like those N100 mini-PCs that the anti-Pi crowd keeps raving about for instance.
@@jamess1787 Some network gear has a 48V terminal-block-type input for the same reason. Side benefit being the ability to run them off of solar systems with the same efficiency increase.
34:20 Technically it's not the cross sectional area but the area as well as the length that matters. Assuming it's copper and 1 meter in length, running the full 16A across it at 230V, the voltage drop should be in the neighbourhood of 0.55V or so. Or in other words, under 9 watts lost as heat. Bumping it up to 1.5mm2 lowers the loss to under 6 watts. But as you said about how "well" the plug seats, I would "worry" more about that. If it's not confidence inspiring, then obviously you should change it for peace of mind.
Also going from a pvc cable to an H07RN-F flexible cable if possible so it flexes nicer. Another thing adding on to your point is the maximum output of the UPS is around 2kW so definitely the 1.0mm² pvc flex at about 30cm length could handle that. Personally I would never use anything less than 1.5mm² for a 16a connection, and often prefer 2.5mm² for any lengths over about 15m.
Alongside that I would have personally installed a 16a CEE socket instead of running of a 13a socket. Not that it matters for the current being drawn, but a 16a socket with a 3 pole isolator prevents being pulled out as easily. I can't remember exactly what he installed for the extra circuit, but if it was a 20a rcbo for the rack and alarm system you could technically install the 16a socket alongside but generally not recommend.
No chance of ever getting close to 13A, let alone 16! The UPS maxes out at under 10A and the cost of the power will become a limitation long before that!
NUT tools for Home Assistant integration. I've just gone through the hassle of getting all my UPSs to report to HA via USB at 1st then network cards as I wanted them all to be the same (I know, OCD)
11:50 most of the 220-240V UPSes in the US use the same IEC connectors. My guess is it's for international market reasons: a unified component that they can just make ONE design for and then ship worldwide.
@@camerongray1515 The North American power implementation is weird, yeah. The "typical" configuration is that buildings get either 240V (house or small business) or 208V (apartments and smaller industrial) depending on wiring, which is then "split" into two 120/104V "split-phase" lines that go to /most/ of our everyday stuff. 240/208V remains as an option for stuff that needs or wants it, it just needs to be wired up as a special case. Get a few kilowatts of computer hardware, and the slightly better power efficiency makes said "special case wiring" a serious consideration.
@@quinnomega Actually, it's 120/208V - 120/240 is split phase, just a centre tapped single phase. 208V is two phases of a three phase supply, so the phase to neutral remains 120V. Then you get 'high leg delta', which just confuses matters by providing 240V phase to phase with a neutral between A and C supplying 120/240 split-phase, and phase B at 208V to neutral. That's just plain special.
@@Monkeh616 Yep, I was intentionally skipping over how convoluted it all gets. Split-phase /IS/ a 240V line coming in, with the center tap that makes it two 120V opposite-phases that power our everyday everything.
A 3-ph PDU shouldn't really be an issue. They still supply a single phase and neutral to the outlets, the internal power supply will just run from L1 + N. Just connect all three line inputs and you're off to the races. The only caveat is they may be doing phase to phase voltage monitoring for.. well, no reason at all. Not that I'm volunteering to spend a couple hundred quid to find out, mind you.
I was thinking that, but I couldn't find any concrete information on it so wanted to err on the side of caution for this video. If I ever come across a three phase UPS going free, I'd be tempted to try it out and see if it works.
Not a huge priority at the moment but I will eventually - in particular killing power to AV distribution kit in the rack in the event of a prolonged outage to preseve the UPS's run time.
I'd say do as much as you can through Network UPS Tools, and only bring Home Assistant into the mix for stuff that can't be shut down that way. One fewer element in the chain that can potentially go wrong.
I keep meaning to look into one although it's probably not something I'd use loads beyond printing pre-made models since I have no 3D modelling skills, but would be interesting to try out. I've also been toying with the idea of trying out a laser cutter to engrave labels on things such as light switches.etc
Have you got any information on the actual power consumption on those PDUs? Looking to get one for my rack and connect to the Eaton UPS. I have pondered to buy a bunch of smart plugs instead though… 😅
The noise appears to be vibration around the area of one of the transformers. I'll pop the UPS open at some point and take a look, wouldn't be surprised if the screws holding it in have just come a bit loose in transit.
I have 2 of that Exact Model UPS's ! Well 3 Actually & about a year to a year & a Half & 2of The 3 has went Bad(Not Battery's Either) I Have ben looking for a 4th But have not Found Any Good deals on one Lately ! I Have a New Battery Pack & a second NMC for one ! I Have about 4-5 Home Servers Running + my network Equipment a 1gb fiber Main Net work & a Private Offline 10Gb Network N a 1u Dell Running PfSense Then my Main High End Gaming PC & a 3+1 Monitor Set Up then a Couple Other Things But i Really Just Dont Like Running a Single UPS at 90-100% 24/7 its just Better for my set up to Use 2 Separate UPS's ! Dont Really Know Why Ive Had 2 Die with in a Year n a Half ish...a Bad Surge May have gotten the firs one but i dont have a Clue to what Happened to the Second One !
As I mentioned, the poor power efficiency of online UPSs made a line interactive UPS a better choice - power around here is generally very clean, I have no plans to run it off a generator, I have a whole house SPD and realistically, the value of equipment in the rack isn't really any greater than all of the other TVs, AV and computer equipment spread around the house. It's not as if I'm going to install an online UPS in practically every room. I also genuinely can't think of any time I've ever had a piece of equipment that has been damaged by bad mains power, nor do I know of anyone around this area who has.
@@camerongray1515 have you calculated what the efficiency actually means in terms of actual cash ? You'll be surprised how small of a difference it makes on this scale.. it seems to me you selected the cheap version of something that literally does nothing to prevent the issue you're trying to avoid. An online ups will also do power factor corrections, frequency filtering/adjustments, and pretty much protect from all power irregularities. A line Interactive won't. You might have good quality power now, but that can change in an instant, its outside your control, but if you want to risk it, go ahead, up to you. I've had enough gear getting destroyed by random grid power issues in my experience so I'm going the safe route. And no, I'm not saying you should have a ups in every room, this was in relation to your homelab.. Just because you haven't experienced power quality issues doesn't mean you won't.
@@abn0rm1 At a load of 190W and 75% efficiency, about 60W of losses. 525kWh per year, or around £130 a year. Power factor doesn't matter. You're billed on real power and the power factor of your load has no bearing on the reliability of that load. Frequency doesn't matter - nothing in a homelab cares about it. This isn't rotating machinery. Not that it would matter if it were, because it's quite carefully controlled. About the only argument is transient protection.
£130 a year is definitely a considerable amount to just be wasting, and that's only going to get worse once I add more kit to the rack and increase the power consumption further. Additionally "the issue I'm trying to avoid" is short power interruptions causing kit to lose power/reboot unexpectedly and the associated inconvenience and data loss. A line interactive UPS does perfectly well. The financial value of the equipment in the rack isn't particularly high in comparison to devices I have in other parts of the house, if I was particularly concerned about protecting devices from bad power, there's other, more valuable kit elsewhere that would be the same/higher priority to protect.
Some information for you to know and understand as it relates to both PDU / UPS. Do not upgrade the firmware on either if they are not under APC Warranty! There is a greater than 50% of bricking a perfectly functioning unit while doing so. ☝️ If you ignore the above information and want to pursue this anyways. Read and fully understand all the requirements and prerequisites in the firmware procedure. In enterprise there is always a exit strategy / recovery process in place whether it be a hot spare. You don’t have either so understand how the recovery process works using X-Modem. Other important factors to consider is (IF) The firmware is below a major revision that you try your best to obtain each major version instead of jumping from say 5.XX ~> 8.XX! So it would be 5 -> 6.XX -> 7.XX -> 8.XX firmware. The use of the proprietary wired serial cable must be used! Do not attempt to update either hardware using NMC (Network) or USB!
I was going to say, with the UPS, I've never updated APC UPS firmware due to horrible warnings, but I have regularly updated the NMC which have had a lot of CVEs in the past. On one occasion the NMC firmware update bricked the card, and I had to brush up on my Xmodem skills to successfully recover it.
It would be nice if a UPS could have several USB-C PD ports (and if more rack devices could be powered by USBC). That would avoid AC-DC conversion losses
The noise from the UPS is the power on self test - I think it's a common feature of APC as mine used to do that too - it will do it periodically as well - it switches over to battery to test if the battery can maintain the load and will alert if the state of charge drops too quickly.
Yeah seems fairly common on APCs to go into self-test after enabling the output. There is an option to stop it from doing that in the self-test schedule settings.
The EPO connector is for datacenters/machine rooms where there's an Emergency Power Off system in place (IE those big red buttons on the walls of datacenters). Hit the button and it kills power to the entire facility in the event of a fire. UPSes are required to have these connectors so that they will not continue to provide power in those EPO events. This is particularly important because the equipment attached to the UPS could be the source of the fire, also, it protects the emergency personnel who are dealing with whatever situation triggered the EPO.
I've dealt with the 3000VA model of these units at work in Australia, they have a superrr nasty hum at all times and will generally be the loudest thing in a rack. I opted for a 1500VA model at home and its practically silent, not to mention I dont have to worry about a 15A outlet.
Oh, youll want to setup powerchute on any hypervisors you have as well for automatic graceful shutdown, you can also get a NUT addon for home assistant to get UPS data in there as well.
In case some people are wondering about the VA versus W and power factor.
Electrical loads can be resistive, inductive (motor) or capacitive (switch mode power supply).
Power factor basically means how well in sync or in phase the AC current is compared to the AC voltage.
If you have a purely resistive load, say an incandescent light bulb or an electric heater, your current sine wave perfectly matches the voltage sine wave. Power factor equals 1.
If you have a capacitive or an inductive load, the current sine wave will either be ahead or behind the voltage sine wave respectively. Power factor is less than 1.
UA-cam seems to be deleting my comments again..
Why is the UPS rated in VA instead of watts?
The manufacturer knows exactly what it can deliver (voltage and current or VoltAmps) but does not know what equipment the user will have.
It's up to the user because they know what equipment they're using.
If you have a power supply, it's power factor can be as good as 0.99 or it could be as worse as 0.7 or even lower.
It usually depends on the load the power supply is under.
If it's a fairly modern design with active power factor correction, you should expect it to climb above power factor 0.9 when it's loaded above 20% I'd say.
If you have an UPS, you get watts by multiplying with the power factor. For example 2200VA and assuming 0.9 power factor is 1980W.
If you have a power supply, you can divide watts with power factor to get VA. For example 500W and 0.9 power factor is 555VA.
@@Lagittaja The typical rule of thumb I was taught is to assume 0.6 until and unless a properly qualified engineer has done extremely extensive and expensive testing of the exact setup in question. Specifically because very little is actually THAT bad, so it's less likely (though not impossible) one will screw it up with their lack of said engineering qualifications.
This hits home. Just ordered a new UPS for my home rack. The old one is about 8 years old now and the second set of batteries has flaked out. Rather than do another battery replacement, I am going new because the UPS has not reported a single issue with the obviously bad battery. I can't trust a UPS that is blissfully unaware of its own issues. New it is.
Add an ATS to bypass the UPS for firmware updates and maintenance. I have an APC AP4421 and have been very happy with it, prevented downtime on multiple occasions.
Where I work we've purchased a number of the auto-transfer PDUs because we did a risk assessment and noticed that most telecom equipment that's vendor supplied does not have dual inputs. So in the event of a UPS failure or UPS maintenance the network circuits could go down.
That's a good shout - I've used ATSs before to feed single PSU kit in a rack with dual power feeds, never considered one for this use case. Also looks like you can get manual "UPS Bypass" transfer switches which might be an even simpler option. I'll probably keep an eye out on eBay and grab something if it shows up cheap.
EPO ON UPS - wire to an NC relay from the texecom. Fire causes the ups to shutdown when fire is in garage. So if the fire is caused from something on the rack, better chance to kill the fire than not.... Also stops other fires from starting with the heat
I'd like to add when trying to rebuild battery packs for APC units, they use some horrendously strong glue/tape to hold the batteries in. You may need to use a heat gun to get it to release or spend hours carefully prying up with a screwdriver being careful not to short the batteries or impale them.
Hours? I have done that in less than 10 minutes with only a screwdriver and without damaging anything.
Do you mean the dry cells glued together? I know APC and Tripp-Lite both do this. But i don't bother separating the battieries and just take them to recycling still bonded together. Usually batteries have been easy to remove and replace in every UPS I've used.
That's good to know! All the more reason to buy one that's already been rebuilt - cells are no longer glued in and therefore will be easy to replace in the future
@camerongray1515 at work we buy the generic APC batteries from CDW they are about half the price of the APC branded ones. I gave up on the rebuild that I had mentioned above after a few hours.
Couple years back, I was very lucky to get a new-in-box surplus Liebert GXT4 1500VA Online UPS for $200. I had to buy the network management card for $120 off ebay.. but still worth it. Only downside is that the fans are pretty loud (since the inverter is always running in normal operation). I have a bunch of Tripp-Lite line-interactive UPSes (SMART1500) and I have had *multiple* devices die due while connected to them after storms -- so not sure I fully trust their surge protection and their AC generation -- although I still use them because getting double-conversion UPSes for everything would be massively expensive and overkill.
Worth noting is that a great many UPSes have serial and/or USB ports even without the network card add-on. An older raspberry pi and Network UPS Tools can often replicate the functionality of those cards and then some for half the cost or less. Assuming the UPS in question is NUT-compatible, of course.
Was at Microcenter yesterday looking at UPS units and was surprised to see Liebert on the shelf alongside Cyberpower and APC. The Liebert was even priced competitively. IIRC 1500VA for like $230, new. Had no idea they were even in the consumer market. We use a lot of big iron Liebert stuff at my job.
We have some shorter APC PDUs at work. The serial port on them turned out to be RJ12. RJ11 has 4 pins, RJ12 has 6. Just a heads up in case you need a serial cable for it.
Yeah, the US tends to just use standard NEMA outlets - the reason simply being density. BS1363 and Schuko are just a bit large for PDUs. NEMA 5-15/20 is compact, even if it lacks any other redeeming features.
That feels mildly terrifying lmao
and NEMA 6/5-XX to IEC cables are a lot cheaper and more available than IEC TO IEC ones
The IEC connectors are commonly used on what 240V UPSes exist in the US (and I assume Canada) as well. Tedious as they are to track down, they do exist here. Only reason to use them over more readily available 120V stuff is the slightly better power efficiency that most PSUs have at 240V.
Regarding the CPC cable, iirc 1.0mm² is rated higher when the cable is fitted with a moulded plug, but I agree 1.5mm² would be better
It would be nice if a UPS could have several USB-C PD ports (and if rack devices could be powered by USBC). That would avoid AC-DC conversion losses
I'm honestly surprised that more stuff isn't implementing USB-C PD as at least an option. PD4 can go to an insane 240 Watts now with a compatible cable, that could run the majority of network gear and even some servers outright. Like those N100 mini-PCs that the anti-Pi crowd keeps raving about for instance.
That's why some UPS' have DC output, tie the -48vdc equipment "directly" from the batteries in the UPS.
@@jamess1787 Some network gear has a 48V terminal-block-type input for the same reason. Side benefit being the ability to run them off of solar systems with the same efficiency increase.
34:20 Technically it's not the cross sectional area but the area as well as the length that matters.
Assuming it's copper and 1 meter in length, running the full 16A across it at 230V, the voltage drop should be in the neighbourhood of 0.55V or so. Or in other words, under 9 watts lost as heat.
Bumping it up to 1.5mm2 lowers the loss to under 6 watts.
But as you said about how "well" the plug seats, I would "worry" more about that.
If it's not confidence inspiring, then obviously you should change it for peace of mind.
Also going from a pvc cable to an H07RN-F flexible cable if possible so it flexes nicer.
Another thing adding on to your point is the maximum output of the UPS is around 2kW so definitely the 1.0mm² pvc flex at about 30cm length could handle that. Personally I would never use anything less than 1.5mm² for a 16a connection, and often prefer 2.5mm² for any lengths over about 15m.
Alongside that I would have personally installed a 16a CEE socket instead of running of a 13a socket. Not that it matters for the current being drawn, but a 16a socket with a 3 pole isolator prevents being pulled out as easily. I can't remember exactly what he installed for the extra circuit, but if it was a 20a rcbo for the rack and alarm system you could technically install the 16a socket alongside but generally not recommend.
Next you need a dedicated 16amp socket on the wall..... for when you get close to overloading that 13amp standard UK outlet.
No chance of ever getting close to 13A, let alone 16! The UPS maxes out at under 10A and the cost of the power will become a limitation long before that!
OpenSCAD is fine for creating 3D mounting brackets, adaptors, etc.
Same UPS ai got on eBay for £65 few weeks back. Great piece of kit.
NUT tools for Home Assistant integration. I've just gone through the hassle of getting all my UPSs to report to HA via USB at 1st then network cards as I wanted them all to be the same (I know, OCD)
Epo port is good to connect to a fire alarm so if smoke is detected it turns off
11:50 most of the 220-240V UPSes in the US use the same IEC connectors. My guess is it's for international market reasons: a unified component that they can just make ONE design for and then ship worldwide.
That makes sense, I hadn't considered the fact that the US uses both 110v and 230v UPSs.
@@camerongray1515 The North American power implementation is weird, yeah. The "typical" configuration is that buildings get either 240V (house or small business) or 208V (apartments and smaller industrial) depending on wiring, which is then "split" into two 120/104V "split-phase" lines that go to /most/ of our everyday stuff. 240/208V remains as an option for stuff that needs or wants it, it just needs to be wired up as a special case. Get a few kilowatts of computer hardware, and the slightly better power efficiency makes said "special case wiring" a serious consideration.
@@quinnomega Actually, it's 120/208V - 120/240 is split phase, just a centre tapped single phase. 208V is two phases of a three phase supply, so the phase to neutral remains 120V. Then you get 'high leg delta', which just confuses matters by providing 240V phase to phase with a neutral between A and C supplying 120/240 split-phase, and phase B at 208V to neutral. That's just plain special.
@@Monkeh616 Yep, I was intentionally skipping over how convoluted it all gets. Split-phase /IS/ a 240V line coming in, with the center tap that makes it two 120V opposite-phases that power our everyday everything.
@@quinnomega Yes, indeed, my point was that 120/208 is not 104/208 and doesn't involve a split phase.
Here in New Zealand all the ones I have use wall style outlets on them.
Biggest issue with home rack ups is depth. Most home racks are Comms racks and won’t fit the ups you’re reviewing.
A 3-ph PDU shouldn't really be an issue. They still supply a single phase and neutral to the outlets, the internal power supply will just run from L1 + N. Just connect all three line inputs and you're off to the races. The only caveat is they may be doing phase to phase voltage monitoring for.. well, no reason at all.
Not that I'm volunteering to spend a couple hundred quid to find out, mind you.
I was thinking that, but I couldn't find any concrete information on it so wanted to err on the side of caution for this video. If I ever come across a three phase UPS going free, I'd be tempted to try it out and see if it works.
any plans to use HA to automate graceful shutdown of services when the UPS kicks in?
Not a huge priority at the moment but I will eventually - in particular killing power to AV distribution kit in the rack in the event of a prolonged outage to preseve the UPS's run time.
I'd say do as much as you can through Network UPS Tools, and only bring Home Assistant into the mix for stuff that can't be shut down that way. One fewer element in the chain that can potentially go wrong.
Do you own a 3d printer? I feel like someone like you would make excellent use of one, creating mounting brackets and other nifty solutions.
I keep meaning to look into one although it's probably not something I'd use loads beyond printing pre-made models since I have no 3D modelling skills, but would be interesting to try out. I've also been toying with the idea of trying out a laser cutter to engrave labels on things such as light switches.etc
Have you got any information on the actual power consumption on those PDUs? Looking to get one for my rack and connect to the Eaton UPS. I have pondered to buy a bunch of smart plugs instead though… 😅
Are there power conditioning options that you could use between a generator and your equipment? And in your opinion, would it be worth looking into?
As soon as you're using a generator, you'd want to be using an online UPS, that'll provide all of the power conditioning you'd need.
It will kick in for the voltage frequency going too far out of spec.
soho lead acid batteries don't last long and cost as much as the UPS. And if you're lucky enough to find one.
46:14 is the rack served by a type B or F RCD?
It's Type A (16A B Curve RCBO), specifically a Hager ADA316G.
I think your UPS is wonky or has lose mounting, my old APC 3000 VA UPSs made some noise, but nothing like that.
The noise appears to be vibration around the area of one of the transformers. I'll pop the UPS open at some point and take a look, wouldn't be surprised if the screws holding it in have just come a bit loose in transit.
@@camerongray1515 agree
I have 2 of that Exact Model UPS's ! Well 3 Actually & about a year to a year & a Half & 2of The 3 has went Bad(Not Battery's Either) I Have ben looking for a 4th But have not Found Any Good deals on one Lately ! I Have a New Battery Pack & a second NMC for one ! I Have about 4-5 Home Servers Running + my network Equipment a 1gb fiber Main Net work & a Private Offline 10Gb Network N a 1u Dell Running PfSense Then my Main High End Gaming PC & a 3+1 Monitor Set Up then a Couple Other Things But i Really Just Dont Like Running a Single UPS at 90-100% 24/7 its just Better for my set up to Use 2 Separate UPS's ! Dont Really Know Why Ive Had 2 Die with in a Year n a Half ish...a Bad Surge May have gotten the firs one but i dont have a Clue to what Happened to the Second One !
I cant belive that I have a better UPS than someone like you.... Makes no sense.
I have an SMX3000RMLV2U 4u
It's not as if I have an unlimited budget for this sort of thing, realistically this UPS is more than enough for my needs.
Never use a line interactive ups if your gear is worth anything, never ever do that. Online ups or bust.
As I mentioned, the poor power efficiency of online UPSs made a line interactive UPS a better choice - power around here is generally very clean, I have no plans to run it off a generator, I have a whole house SPD and realistically, the value of equipment in the rack isn't really any greater than all of the other TVs, AV and computer equipment spread around the house. It's not as if I'm going to install an online UPS in practically every room. I also genuinely can't think of any time I've ever had a piece of equipment that has been damaged by bad mains power, nor do I know of anyone around this area who has.
@@camerongray1515 have you calculated what the efficiency actually means in terms of actual cash ? You'll be surprised how small of a difference it makes on this scale..
it seems to me you selected the cheap version of something that literally does nothing to prevent the issue you're trying to avoid. An online ups will also do power factor corrections, frequency filtering/adjustments, and pretty much protect from all power irregularities. A line Interactive won't.
You might have good quality power now, but that can change in an instant, its outside your control, but if you want to risk it, go ahead, up to you.
I've had enough gear getting destroyed by random grid power issues in my experience so I'm going the safe route.
And no, I'm not saying you should have a ups in every room, this was in relation to your homelab..
Just because you haven't experienced power quality issues doesn't mean you won't.
@@abn0rm1 At a load of 190W and 75% efficiency, about 60W of losses. 525kWh per year, or around £130 a year.
Power factor doesn't matter. You're billed on real power and the power factor of your load has no bearing on the reliability of that load.
Frequency doesn't matter - nothing in a homelab cares about it. This isn't rotating machinery. Not that it would matter if it were, because it's quite carefully controlled.
About the only argument is transient protection.
£130 a year is definitely a considerable amount to just be wasting, and that's only going to get worse once I add more kit to the rack and increase the power consumption further. Additionally "the issue I'm trying to avoid" is short power interruptions causing kit to lose power/reboot unexpectedly and the associated inconvenience and data loss. A line interactive UPS does perfectly well. The financial value of the equipment in the rack isn't particularly high in comparison to devices I have in other parts of the house, if I was particularly concerned about protecting devices from bad power, there's other, more valuable kit elsewhere that would be the same/higher priority to protect.
Some information for you to know and understand as it relates to both PDU / UPS.
Do not upgrade the firmware on either if they are not under APC Warranty!
There is a greater than 50% of bricking a perfectly functioning unit while doing so. ☝️
If you ignore the above information and want to pursue this anyways. Read and fully understand all the requirements and prerequisites in the firmware procedure.
In enterprise there is always a exit strategy / recovery process in place whether it be a hot spare.
You don’t have either so understand how the recovery process works using X-Modem.
Other important factors to consider is (IF) The firmware is below a major revision that you try your best to obtain each major version instead of jumping from say 5.XX ~> 8.XX!
So it would be 5 -> 6.XX -> 7.XX -> 8.XX firmware.
The use of the proprietary wired serial cable must be used!
Do not attempt to update either hardware using NMC (Network) or USB!
I was going to say, with the UPS, I've never updated APC UPS firmware due to horrible warnings, but I have regularly updated the NMC which have had a lot of CVEs in the past. On one occasion the NMC firmware update bricked the card, and I had to brush up on my Xmodem skills to successfully recover it.
It would be nice if a UPS could have several USB-C PD ports (and if more rack devices could be powered by USBC). That would avoid AC-DC conversion losses