Symmetra 16kVA Bypass after Power Failure?

[Anonymous user]

Hey Guys,

Had an issue with our Symmetra 16kVA last month and have been trying to get an answer to what happened and so far nothing has come up, looking to the forum here to see if anyone can help.

Long story short we had a power outage at our data center on December 10th and forced all non-UPS loads to fall over to UPS backup power. All of our UPS's reacted fine (we have 4 other SM 6000's) however our 16kVA unit immediately went into bypass/overload and dropped it's load. This UPS only feeds dual power supplied servers, with one side on UPS power, the other on the interruptable hydro feed.

Logs from the UPS from the incident:

12/10/2013

15:05:20

UPS: In bypass due to an overload cleared.

12/10/2013

15:04:18

UPS: In bypass due to an overload.

12/10/2013

15:04:18

UPS: In bypass due to an overload cleared.

12/10/2013

15:03:46

UPS: Redundancy below alarm threshold cleared.

12/10/2013

15:03:46

UPS: Redundancy returned.

12/10/2013

15:03:32

UPS: In bypass due to an overload.

12/10/2013

15:03:26

UPS: Redundancy below alarm threshold.

12/10/2013

15:03:26

UPS: Redundancy lost.

Since everything attached to the UPS is dual feed and the current usage is ~45% with the transfer from hydro it would have only put the UPS at 85-90% capacity.

Currently specs on the UPS are:

	Describes utility power status
	Input Voltage:	199.2	VAC
	Input Frequency:	60.01	Hz
	Maximum Input Voltage:	199.2	VAC
	Minimum Input Voltage:	199.2	VAC
_
	Describes output power status
	Output Voltage:	198.3	VAC
	Output Frequency:	60.03	Hz
	Output Current:	21.3	Amps
	Output Watts at n+0:	044	%
	Output Watts at n+1:	044	%
	Output VA at n+0:	030	%
	Output VA at n+1:	030	%
_
	Fault tolerance parameters
	Redundancy:	n+3
	Present kVA Capacity:	16.0	kVA
_
	Describes battery status
	Battery Capacity:	100.0	%
	Runtime Remaining:	0029	Minutes
	Nominal Battery Voltage:	120.0	VDC
	Actual Battery Bus Voltage:	136.9	VDC
	Number of Batteries:	006
	Number of Bad Batteries:	000

Not sure if there is something I am missing here and we need to relieve some load from the UPS itself, or if I've managed to get some config wrong. But any help would be much appreciated.

Thanks!

[Anonymous user]

A couple if things/questions come to mind on the resiliency of your redundancy setup and the readings:

A. Are you sure your servers  are set for power usage within the servers config to balance the load on each of the two A and B power supplies? IE they are not drawing primarily off one power supply while the other is warm-standby ?

B. To back up what I said in A., Have you ever done a live measurement of the "b-source" power drawn by the servers on the non-UPS interruptible hydro feed ? Eg If they are drawing 25 to 30 amps at 200 volts the ups could overload if the hydro side fails.

C. Also noticed your incoming voltage is a little low @200 volts if you are on a 208 volt or 240 volt source that may slightly reduce available power from the UPS when its online.

If this is a Symmetra LX, the loading should be reported for both the phases L1 and L2....but I'm guessing this is an older Masterframe Symmetra instead ?

[Anonymous user · Accepted solution]

Thanks for getting back guys,

A: They are set to balance the load were they can. Each UPS/Hydro feed does have an APC Switched PDU on the line so we can measure it's Amperage. In the cases were they are not balancing we're making sure that they do not exceed the recommended 12Amps combined across both A (UPS) and B (Hydro) PDU's.

B: I have done some calculations based on what each of the supplies is currently drawing, it does appear to match up with what the UPS is showing on it's panel, and it does appear that the Hydro side of things does have a bit more load then the UPS currently. will need to do an actual measurement to confirm.

	Feed	Voltage	Amps	Watts			Watts	Percent of Total Usage
pdu7a	UPS	120	6	720		UPS	4929	44%
pdu7b	Hydro	120	1	120		Hydro	5789	52%
pdu8a	UPS	120	1.4	168		Total:	10718	96%
pdu8b	Hydro	120	1.7	204
pdu9a	UPS	120	7	840		UPS MAX (W)	11200
pdu9b	Hydro	120	6	720
pdu10	UPS	120	0	0
pdu11a	UPS	120	6.2	744
pdu11b	Hydro	120	5.8	696
bc1-1	Hydro	208	8.2	1712
bc1-2	UPS	208	6.9	1442
bc2-1	Hydro	208	6.9	1444
bc2-2	UPS	208	4.4	908
bc3-1	Hydro	208	4.3	893
bc3-2	UPS	208	0.5	107

C: Yeah, I see what you mean, if we are drawing less on the voltage end we could be hitting about the 96% when it does transfer. Unfortunately this is an older Masterframe so we can't see the load per phase.

So I guess the main problem here is that even if we did give some wiggle room, when the load does transfer to the UPS we could theoretically be looking at ~95% load, I guess with a UPS as old as this one we won't be able to necessarily guarantee that this thing is 100% operational and will transfer to battery correctly....might need to shift some load off of it.