EMERSON (Liebert) Himod issues

[coolhandlew]

Hi all,

I've recently installed a few Liebert Himod M47 DX units to a data centre. Instead of the standard Liebert controller these have Trend fitted. On studying the units and drawings etc there appears to be a few flaws in their design.

1. The LP switch is in the compressor contactor coil circuit with no relay and thus does not create an alarm in the event of a trip. This is made even worse as there is a timer which shorts out the LP switch for 3 minutes each time the compressor starts (to allow it to run on low pressure at start up in certain situations). This means in the event of refrigerant loss the compressor will run for 3 minutes without refrigerant on each cooling signal.

2. The HP switch, upon tripping does energise the alarm relay which is nice. This obviously is then flagged by the Trend and in my case is logged on the BMS.
Thing is, (and I know it doesn't really matter) when the HP switch goes open the HP relay drops out. The alarm relay then energises via a NC contact on the HP relay. The alarm relay gets it's supply from the cooling signal from the Trend, therefore if there is no cooling demand there appears to be no alarm.

I wondered if anyone had come across these problems, and also if anyone had done any mods to improve.

Cheers

[nike123]

1. The LP switch is in the compressor contactor coil circuit with no relay and thus does not create an alarm in the event of a trip. This is made even worse as there is a timer which shorts out the LP switch for 3 minutes each time the compressor starts (to allow it to run on low pressure at start up in certain situations). This means in the event of refrigerant loss the compressor will run for 3 minutes without refrigerant on each cooling signal.


That is not flaw. That is OK and compressor could work for 3 minutes without any damage.

Flaw is that it doesn't count number of such events in measure of time and then decide when is problem persistent or intermittent




I wondered if anyone had come across these problems, and also if anyone had done any mods to improve.

Cheers

If you want, you could send low pressure trip signal to BMS and then program BMS to count low pressure trips in measure of time and decide when it is fault in refrigerant system.

For HP you could also send HP trip signal to BMS and program to prevent further running of unit until inspection and manual reset by human decision thru BMS

[Refrigerologist]

Hi all,

I've recently installed a few Liebert Himod M47 DX units to a data centre. Instead of the standard Liebert controller these have Trend fitted. On studying the units and drawings etc there appears to be a few flaws in their design.

1. The LP switch is in the compressor contactor coil circuit with no relay and thus does not create an alarm in the event of a trip. This is made even worse as there is a timer which shorts out the LP switch for 3 minutes each time the compressor starts (to allow it to run on low pressure at start up in certain situations). This means in the event of refrigerant loss the compressor will run for 3 minutes without refrigerant on each cooling signal.

2. The HP switch, upon tripping does energise the alarm relay which is nice. This obviously is then flagged by the Trend and in my case is logged on the BMS.
Thing is, (and I know it doesn't really matter) when the HP switch goes open the HP relay drops out. The alarm relay then energises via a NC contact on the HP relay. The alarm relay gets it's supply from the cooling signal from the Trend, therefore if there is no cooling demand there appears to be no alarm.

I wondered if anyone had come across these problems, and also if anyone had done any mods to improve.

Cheers

With regard to your problem, when a normal Liebert controller is installed these units have optional alarm report boards that can be installed. These can provide up to 9 different alarms. Otherwise you are stuck with a general alarm signal, which does not distinguish between critical and non-critical alarms.

I have just modified 4No.M66UA Himod systems, out of the 9 alarms quoted only 5 were useful.

How this could be done with a Trend controller I have no idea, but would suggest you contact Emerson Network Power A/C technical department for an answer.

As for the low pressure switch 3 minute delay, this applies only to the start up of the compressor, and is mainly to prevent unnecessary tripping in the event that the refrigerant has migrated to the remote condenser during cold ambients when combined with long periods in standby mode! As Nike says, this is not a flaw! The low pressure alarm is just delayed for a few minutes after compressor start, there is no delay after the initial 3 minutes has elapsed.

[Allrounder]

If the units are new definately go back to the manufacturer.

[Tiesse]

With regard to your problem, when a normal Liebert controller is installed these units have optional alarm report boards that can be installed. These can provide up to 9 different alarms. Otherwise you are stuck with a general alarm signal, which does not distinguish between critical and non-critical alarms.

I have just modified 4No.M66UA Himod systems, out of the 9 alarms quoted only 5 were useful.

How this could be done with a Trend controller I have no idea, but would suggest you contact Emerson Network Power A/C technical department for an answer.

As for the low pressure switch 3 minute delay, this applies only to the start up of the compressor, and is mainly to prevent unnecessary tripping in the event that the refrigerant has migrated to the remote condenser during cold ambients when combined with long periods in standby mode! As Nike says, this is not a flaw! The low pressure alarm is just delayed for a few minutes after compressor start, there is no delay after the initial 3 minutes has elapsed.

I don't believe the LP switch is wired into the Trend controller as an input, as such, the Trend controller cannot report what it can't see, whereas the HP reports as a fault as it is wired as an input.

The delay on start up is to account for potential migration issues on start up following long periods of inactivity. This is quite normal (Isovel amongst others used to employ this method, mostly caused by the compressor being indoords & the condenser & receiver being outside - particuolarly on systemns with a long pipe run).

In regard to the limitations of using a Trend controller, although no use to you in this case, another option would have been to use the standard controller (now iCOM) & connect it to a Hirolink to give a modbus output & then an SXNC that can sit on the client's LAN. This then offers all of the inputs/outputs/alarms that the standard comtroller sees.