I picked up a new replacement scroll compressor a couple weeks ago at our local Liebert parts store. I noticed there was a bulliten on the counter regarding alot of digital scrolls out of Lieberts were being brought in for warranty replacement and the compressors were not bad. If i remember correctly Emerson wanted the technician to follow a troubleshooting procedure that they provided before you condemed a comp. I've yet to work on a copeland digital scroll but was told all the new lieberts have them. Has anyone had to replace a bad solenoid or modulation controller?
Sorry for the late response. I used to be a heavy duty HVAC-Talker, haven't lurked here for a while. Was searching for something and this thread came up. I hate to hear that anyone ever has a serivce problem, and wish nothing every broke but it does so we try to do everything we can to support service techs in the field to get our mututal customers up and running as fast as possible.
A couple of notes:
1. The first iteration of our GE MRI chillers was a 25 ton chiller for a 3.0T MRI. We used tandem scrolls; 2 6 tons in tandem per circuit. We then had to design a 15 ton for their new MRI, and at that point tandems were the same price as digitals, and the digitals took up less space, so we used them. There was no ZPD90 (a 7.5 ton) so we went with the ZPD83 (7 ton). When GE designed the bigger MRI, as it was a new model they were SUPER conservative on needing a 25 ton. A 20 ton would have been adequate. So when we designed the "15 ton" we picked the 7 ton compressor over the 8 ton. The 8 ton being the ZPD103. After about 6 months of running we moved to the ZPD103 because of supply issues from Copeland, an increase of units running in the Southwest, and learning that for this new MRI they were no longer over conservative with their heat load estimates. Nothing in the refrigeration system changed, other than the compressor, as the evaporator and condenser were already somewhat oversized, and the expansion valve was OK as well.
In hindsight, this decision was not a good one. We have had an extremely high failure rate of the ZPD103 compressors, whereas with the ZPD83 it has been almost non existent. According to Copeland's teardown reports, the casues of failure were lumped into two major groups; oil dilution, and melting of the discharge bypass tube. When the compressor unloads, the superheat actually shows negative on the screen, because you get an instant rise in suction pressure but a slow rise in suction temperature. This would indicate flooding but it is acutally a consequence of how we measure superheat. Plus we aren't pumping anymore. Best we could figure is that we were getting liquid back to the compressor in very small amounts, over a long period of time causing the compressor failure. Also, our suction piping is very short, so evaporator superheat equals compressr superheat. This is really hard to test for, so we just decided to shove in suction accumulators on all of our machines that have digital scrolls. There is just too much "weird" stuff going on when you unload the compressor that quickly. Before we had the accumulator designed in we changed our suction piping to create more of a "trap" entering the compressor suction.
We are getting far enough down the road in the accumulator design that should start seeing if this has been a beneficial change.
The discharge tube melting issue has not gone away, and we are convinced that this is a Copeland design flaw. We are stuck with having to find workarounds. On the ZPD83 and smaller compressors, the tube to vent the discharge gas to the suction to unload the compressor is through and external tube and valve. On the ZPD103 and up (ZPD182 is the biggest) this tube is internal and the valve body is in the compressor shell (second smaller peckerhead). Copeland claims this failure is due to operating outside of the envelope for extended periods of time. However, we have had this failure happen on units on our test stand that ran for 5 minutes. One acutally failed as soon as we started it after 45 seconds of running. This precipitated a conference call to Copeland. They claimed they can reproduce the tube failure due to excessive discharge temperatures, but it takes 10+ minutes of running it this way to get this failure. The tube melts at 700F (by the way it is plastic). Whenever they do a teardown they say we need to run an LP switch, which we do. OUr workarounds so far is to program a starting pulse and delay to ensure the expansion valve is open before the compressor starts pumping.
The Liebert failures where the compressor turned out to be fine were probably due to another weird problem we experienced. We started seeing compressors that would fail unloaded, some of them on our test stand, or shortly after startup. It would be the same symptom; lack of cooling but the compresor was running. After sending a few of these compressors back (at a huge cost to us) then getting the "there is nothing wrong with" we finally found out what it was. I was doing a startup with our service manager in Dallas for one of our big imaging center customers. Every weird thing we have had happen happens to them. We start the machine (after running it for two days in our shop) and it runs perfect. Anthony and I go get a cup of coffee, and come back later to run it again. There was no load (MRI not online yet) so we had to wait for the pump to heat up the water. 2 of the 3 compressors wouldn't load. I called, pissed, Copealand tech support. The answers I got were wrong, it was obvious the application engineer had never seen a digital scroll, or a DVM for that matter. I called my go to guy at Copeland. I found out that had a bad batch of unloader valves!! Their vendor had an out of tolerance issue. The valve would open (unloading) and stick open. They gave us the suspect date code. We went to our inventory and removed every single valve and they replaced them for us. We have not seen this problem since.
Fans and Head Pressure
The loading and unloading nature of digital scrolls means control is a pain in the ass. Our solution for the fans was to program a "filter" on the head pressure. When the compressor unloads, the head pressure drops like a rock. So the fan speed drops accordingly. Then sometimes shuts off. Then the compressor loads, and the HP rises faster than the fan can accelerate. Then you have an HP trip (hopefully sensor trip so it auto resets). The filter works by reading the HP and comparing the HP every second. The PID loop that runs the fans looks at this filter HP. THe filter does not allow the HP read by the PID function to drop more than 1 psi per second. However, the screen will show the actual HP. The end result is the fan hardly drops in speed when in unloads. The side benefit is that is GREATLY improves EEV function as well. Note that this is a one way filter, if the pressure rises quickly it isn't filtered.
We have several program revs, if you are working with ZPD103 compressors you are on 3 or greater. We are on version 4.5. With version 4 we totally changed the temperature control algorithm to take better advantage of the compressors unloading ability and minimize compressor cycling. We also added the HP filter.
We can do controller swapouts on out of warranty units for a nominal charge. The cost kind of depends on if we have an older controller to swap out (older controller but new program). We have some other new things we are working on to further smooth out EEV control action and improve our alarm handling (try to hide them better unless there is major issue ) .
We also started tech training, we had 14 people for our first one this winter. We should be having another one in the fall. Our goal is to have a smaller class size and have more hands on.
I hope this helps give a little insight to the digital scroll operation and our chillers. Please contact us directly for any futher questions or concerns.
You would have to send it in. We are moving to the new Carel in a few months that can take a USB drive.
We send you a newer one first, so you aren't dead in the water waiting to get the controller back and in case something screws up your aren't totally hosed and you can pop the old one in.
We would need to verify the I/O map to make sure you get the right one. Things to look at are one flow switch or two, and whether or not you can "see" the EEV from the main controller screen. In the main menu bottom item is "EVD Display". You post or PM the jobsite and we can look it up. But it always I like to verify with a real person looking at the system.
Last edited by clydemule; 04-05-2013 at 12:16 AM.