Geothermal Rescue – Closed Loop Inadequate Flow Damaging Heat Pumps – Edgartown


IMGMar 04 2014 10-41-31We were asked to retrofit a geothermal system because of the failure of three of the system’s five heat pumps.  Compressor overheating and failure was caused by inadequate water pressure in the system, loss of water flow through the heat pumps, and severe short cycling.  Broken pressure gauges that were stuck at a false reading aided in the failure of the heat pumps.

Several of the heat pumps were running in heating when the system was calling for cooling – again this went undetected for several seasons as there were no thermometers on the heat pumps or piping.  Three of the heat pumps were staged together as the first stage which meant that most of the time, they ran for only a few minutes before satisfying the call for cooling.  This short cycling contributed greatly to their early demise.



One of the large circulators sending chilled water from the heat pumps up to the remote air handlers failed due to inadequate water pressure – as these circulators are water lubricated, low water pressure meant air collected in the circulator and it overheated, almost causing a serious fire in the switching relay that powered it.




This project was further complicated by the access to the mechanical room – through a finished hallway and set of stairs.  This meant that extra care had to be taken with our demolition of the failed heat pump installation and with the transport of the new heat pumps through the finished spaces.



To simplify the removal process, we lightened the failed heat pumps as much as we could by recovering the refrigerant, removing the compressors, and removing the heat exchangers.





We were able to save the geothermal field water supply and return pipes and connections and reuse them with the new heat pumps





We replaced the five old ClimateMaster 5 ton R 22 heat pumps with new ClimateMaster R410 5 ton heat pumps.





We had to create a slide system out of Superstrut to support the new heat pumps over the finished wood stairs – we had to also take care with the finished walls of the stairwell.





Each of the five 5 ton heat pumps required extra care to move onto the slide assembly and then to carefully lower down the slide to the waiting cart.






We were relieved to install all five new heat pumps in their support stands with no damage to the finished surfaces or the new heat pumps!







We proceeded to re-pipe all of the new heat pumps using copper and press fittings.  We connected the existing geothermal source supply and return fittings to the new heat pumps with minor adjustments.





We relocated the load side circulators for each heat pump to the ceiling of the mechanical room to improve piping access and simplify service.





We arranged the piping so that all thermometers could be in a row for easy comparison and lower and upper heat pumps would have symmetrical flow direction.





Because the press fittings for copper sometimes led to crooked assemblies during the press process, we pre-assembled our manifolds locked into a strut framework.  This ensured that the manifolds were straight and no fittings were out of alignment.




The press process was so easy and robust that even our STEM interns and apprentices could successfully press their first time!







To ensure that we would have absolutely no air issues or purging problems, we installed air separators on both supply and return lines between the heat pumps and the buffer tank.






We were able to reuse the existing 120 gallon buffer tank because of our improved staging controls.





We installed an abundance of thermometers on all heat pump supply and return pipes, buffer tank supply and return pipes, and on the supply and return to the chilled water system.  This let us know at a glance how well each heat pump was working and let us optimize the heat pump staging.




IMG_4192We installed a new staging control that let us individually stage each of the 5 ton heat pumps, determine their minimum run time, and rotate which heat pump started each cooling cycle.  We also got feedback from the system via the built in data-logger.




We were very gratified to not hear from the homeowner about cooling issues after our commissioning process – the five new heat pumps worked flawlessly with the air conditioning system all summer long.

This is our recent report to a forensic engineer about our findings after careful examination of the geothermal system:

My firm, Nelson Mechanical Design, was retained by the homeowner to evaluate and repair her geothermal installation in her Martha’s Vineyard residence.  The house was finished in 2008 and the geothermal has been somewhat successful at heating and cooling but has had serious reliability issues.  At the moment, three of the six heat pumps are inoperable and there appear to be leaks in both the outside and inside geothermal loops.

The following are some preliminary thoughts and discoveries:


6 Air Handlers:
There are six air handlers with chilled water coils (that are served by 120 gallon buffer tank in basement that can provide chilled water or heated water) and hot water coils (that are served by a propane boiler).

Load Side Inside Geothermal Loop:
The inside geothermal loop (load side) has an air separator, expansion tank, and fast fill connected to the buffer tank.

Heat Pumps 1 thru 5:
The buffer tank is connected to five heat pumps in the main mechanical room.  The heat pumps are controlled by a two stage control  – heat pumps 1, 2,  and 3 are stage one and hp 4 and 5 are stage two.  Each heat pump has a single Grundfos UP26-116 circulator on the load side which is connected to the buffer tank.

Pool Heat Pump 6:
There is a sixth heat pump that is connected via 200 feet of piping to a pool heat exchanger that is the only source of heat for the pool.  There originally was no air separator, expansion tank, or fast fill on the load side (pool heat exchanger) of this heat pump.  We have installed an expansion tank, an air separator, and a fast fill connection on this load side this summer.

Source Side Outside Geothermal Loop:
The source side of all six heat pumps is a horizontal geothermal field connected in parallel to a plate and frame heat exchanger (that is connected to a geothermal source well (open loop) and a geothermal dump well).  The valves connecting the plate and frame heat exchanger to the source side are off (we are not sure why).    The source side of each of six heat pumps is circulated by twin circulator pumping stations (six in all).  These pumping stations use two Grundfos UP26-116 circulators each.


Numerous components of the system have failed or do not operate correctly.  It appears that the air handler and thermostat systems work acceptably. It also appears that the boiler integration with the hot water coils works well too.  It appears that the geothermal system has the following issues (that we are aware of at this moment):

Heat pump 1 (working)
This heat pump is part of the first stage (heat pump 1, 2, 3) and is working now.  It does have extensive evidence of overheating inside the heat pump case.

Heat pump 2  (not working)
This heat pump is part of the first stage (heat pump 1, 2, 3) and was working last summer.  This summer we found it off on Fault 2 which was High pressure in the cooling mode.  We found that the source side pressure was down to 2 psi perhaps due to a leak outside in the geothermal field.  We did not leak test the field at that time due to air conditioning demands.   We repressurized the source side to 25 psi and the fault code was cleared and it cooled successfully.  We returned a few days later and found this heat pump off on the same fault code.  Once again, we found the outside loop had low pressure – this time it was at 10 psi.  We raised it to 25 psi and HP 2 started right up.  Unfortunately, after we had watched it cool successfully for 20 minutes, (by this time we had our ammeter on the compressor leads to document the current draw) it suddenly started to struggle and then the compressor stopped and we had an amp surge of 110 amps.  The internal overload in the compressor tripped and the heat pump cycled on this internal overload up to 100 amps several times before we turned off the power.  After letting this compressor sit for several hours, we tried to restart it but the compressor was locked up.

Heat pump 3  (not working)
This heat pump is part of the first stage (heat pump 1, 2, 3) and has not worked in several years.  Apparently, the previous HVAC service contractor attempted a leak repair unsuccessfully.  We are not sure if there was a component failure before the leak repair – it is possible that a compressor and/or TXV was replaced incorrectly and there was a leak in the heat pump.

Heat Pump 4  (not working)
This heat pump is part of the second stage (heat pump 4 and 5) and is no longer working.  It has a clogged or stuck TXV (extremely high differential in refrigerant pressures) and a leak in the heat exchanger on the source side (we found this by pressurizing the refrigerant circuit with nitrogen and noting that the source outside geo loop pressure went up.)

In the past, when it was working it would go off on fault code 5 which was an fp2 sensor or low load temp.   We found that there was a tripped load pump circuit breaker inside the cabinet and the load circulator was not operating.  We found that the circulator for HP 4 (Grundfos UP26-116) was 50% blocked by corrosion and debris in the volute.  We replaced this circulator with a new Grundfos UP26-116 and HP 4 then worked for several weeks until the TXV issue.

Heat pump 5 (working)
This heat pump is part of the second stage (heat pump 4 and 5) and was found several weeks ago not operating due to a fault code 5 which was an fp2 sensor or low load temp.  We found that the circulator had failed (broken shaft) and was not operating.

We replaced this circulator with a new Grundfos UP26-116 and HP 5 now works.

Pool heat pump 6  (working intermittently)
During the summer, this heat pump went off on a high pressure fault.  We investigated and found that the load side had very low water pressure and had no air separator, expansion tank, or fill station.  We installed those components and repressurized the load side.  We also found that the load side circulator (Grundfos UPS 40-80/2F) had failed.  We replaced the circulator and we did get the heat pump running.  Over the next few days, it tripped several times on a high pressure fault 2.  We noted that the existing control board had no operable LED lights so we suspected damage due to the poor quality of power on Martha’s Vineyard.  We replaced the CXM control board.  The heat pump continues to occassionally trip on fault 2.   During operation we got the following measurements:

9.3 F rise on loop going to pool heat exchanger (out at 98.3 F and in at 88.8 F)

4.6 F drop on source going thru pool heat pump (in at 59.7 F and out at 55.1 F)

11 psi rise across the circulator on the pool heat exchanger load side

The pool heat exchanger has never been cleaned as far as we can tell.  Plans show 1-1/4″ piping to and from the pool heat exchanger (200′ total developed length) and it appears to be piped with 1″ piping.

Geothermal inside loop load side

There appears to be a leak in the inside load building loop.  During the investigation of the fault codes, we determined that the pressure gauge had failed and was indeed stuck at 15 psi – though after replacement of the pressure gauge we found the system pressure was at 2 psi.  There is an air separator, expansion tank, and fast fill station on the inside loop – we did note that fill valve was off.  We pressurized the system to 25 psi.  The pressure did drop several pounds over several days.  We now have the fast fill on to maintain system pressure but there is no water meter on the inside loop.  Perhaps the air separator is not working correctly (we have seen this occur due to corrosion of the fine screen element inside separators which reduce the ability to collect the smallest air bubbles in solution).  It is difficult to tell if there are leaks in mechanical room – when the air conditioning is off the building humidity is high (coastal environment on Martha’s Vineyard) and extensive condensation occurs on the outside of all piping, valves, and fittings leading to puddles on the floor which may mask possible system leaks.

Geothermal outside loop source side

There is a leak in the outside loop as well.  According to Sean Fennessy, engineer of record, during construction of the home there was a leak discovered after the horizontal loop was covered up and a sealant was used to fix the leak.  When we were brought into this project, there was no expansion tank, fill station, or air separator on the outside loop.  During the investigation of the fault codes, we also found that the outside loop pressure gauge had failed.  When we replaced it, we found that the pressure in the outside loop was 2 lbs.  We also noted that the twin circulator stations connected to each heat pump were cavitating seriously.  We repressurized the loop and did note a pressure drop over several days.

We installed a fill station and water meter on the outside field to quantify the leak.  From 8/9/16 to 8/16/16 the loop outside lost 0.11 gallons based on my reading of the meter.  So over the course of a year this would be about 5 gallons.

Over the course of 8 years of operation the system has lost 40 gallons – enough to wreak havoc on the heat pumps due to no fill station or expansion tank but not a significant leak by any standard.  Of course this is assuming that the leak rate is linear and has not increased over time.



Control system short cycling
 It is worth investigating the role that the controls may have played in shortening the life of the heat pumps.  Because the five heat pumps are all 5 ton units and three heat pumps are ganged together as stage one (heat pump 1, 2, and 3) and two heat pumps staged together as stage two (heat pumps 4 and 5), it is a distinct possibility that heat pump 2, 3, and 4  failed due to too many short on/off cycles.  We noted that the buffer tank chilled water differential was 2 F for the first stage set point.  This meant that 15 tons of heat pump cooling would be applied every time the buffer tank temp rose 2 F above the set point.   As soon as heat pump 3 failed, there was a greater burden on heat pump 1 and 2.   It is highly likely that heat pump 1 will fail early due to the excessive duty cycle.

This staging scheme would result in very short run times for the heat pumps.  A crude calculation shows that for the 120 gallon buffer tank, Stage 1 would have a 4 minute run time with 15 tons of heat pump (hp 1, 2, 3 staged together).
Perhaps this control should be replaced with a true staging control that can intelligently turn on and off and rotate lead of the 5 heat pumps based on buffer tank temp and PID of tank temp (how fast it is dropping,  how big of a drop in temp, etc.).

Each heat pump ideally needs to have a minimum run time of at least 10 minutes (?)  – in our experience short run time is a big part of heat pump failure over time.

Failed or absent check valves on heat pumps on the outside loop side
We noted that heat pump 4 (when it was operational) and heat pump 5 experienced extremely high water temperatures on the source side and would fault on high refrigeration pressure  – an inability to reject heat to the outside loop.  We confirmed that we were indeed getting rising refrigerant pressure – it climbed up above 300 psi before shutting off on both heat pumps.  We confirmed that we had loop pressure in the outside loop (it was at 20 psi so there was no cavitation in the outside loop pumps).  We noted that the twin pump station circulators were getting quite hot so we thought perhaps stuck impeller, debris, etc.  We disassembled both twin pump stations (removing push/pull circulators from both heat pump 4 and 5 stations) and found volutes clean and in very good condition and impellers spinning freely.

We put our pressure gauges on the operating heat pump 1 and got a 7 psi pressure drop across the outside loop heat exchanger (source side).  We measured 7 psi drop across both of these heat exchangers in heat pump 4 and heat pump 5.

We measured the inlet and outlet water temp of the outside geo water into and out of heat pump 1 and were getting 73 F inlet with a 3 degree rise.

We measured the inlet and outlet water temp of outside geo water into and out of heat pump 4 and 5 and found that both heat pumps had rapidly rising water temps that climbed up to 115 F before tripping the high pressure fault.

At that point we knew we were getting backwards flow – we closed the supply ball valve from the outside geo loop into the failed heat pump 2 and failed heat pump 3.

Instantly, our water temps plummeted and we were getting inlet temps of 73 F into both heat pumps 4 and 5.  Fascinating interaction of backwards flow through the inactive heat pumps to defeat the heat pumps 4 and 5 through high pressure lock out.

Perhaps in alternating years, when the first stage was reversed, and heat pumps 4 and 5 started first, were there issues with backwards flow?  I don’t believe there are check valves on the circulators in the twin pump stations or in the piping or perhaps they were asked for and not installed – hard to tell due to the insulation on the piping.

Heat pump failure due to reduced heat transfer from a sealant
Pressure/Temp ports are not present on each heat pump which makes determination of heat transfer rate for each heat pump more difficult  – it is a distinct possibility that the sealant used on the outside geothermal source loop to “fix” the leak during construction coated the heat exchangers in the six heat pumps thereby reducing the heat transfer from the hot refrigerant to the source loop.  This would lead to longer run times, higher temperatures, and higher pressures.

Leaks and Corrosion in outside and inside loops and geothermal system
Loop pressurization issues have caused problems with this system.  Our experience has taught us that using glycol fill stations and expansion tanks on outside loops help keep systems working well.  In order to not cover up leaks, it is important to install a water meter and watch it.

We found a lot of corrosion in the load side circulators of heat pumps 4 and 5.    We would suggest testing the source loop water, the load loop water, and the well water for pH and any minerals present.   Perhaps oxygenated water was added to the system during repairs in the past and the air separator did not adequately remove the entrained air bubbles leading to corrosion in the ferrous parts of the load side.

Other Issues Most Likely Present that we did not uncover….
As our investigation of this system coincided with a need to have it working to meet the air conditioning load, it was impossible to locate, quanitify, and resolve all issues with this system.


As we proceed with this project, we add pictures and descriptions.

Our goal is to be Martha’s Vineyard’s premier plumbing and HVAC shop – we appreciate the opportunity to show you some of our work.