12v electric A/C compressor

[prelude]

Hi Lads,

This is a bit of a write up regarding the installation of a 12V A/C compressor so that others may benefit of it, perhaps :)

First, let's answer the "why" question, since our beasts have a proper working A/C to begin with. I decided some time ago that I would like to have the option to run A/C during the night, even if it is only on a ridiculously low setting, just to make it through the worst of conditions. In order to achieve this goal one can go two, maybe three routes:


install a separate system, either 12v or 220v
use one of them portable things
modify the existing system


Mounting an extra system uses up a lot of space and weighs a lot, especially if you go for a domestic split-system since that also requires quite a large and heavy inverter. Those portable things have near useless efficiency and are bulky items difficult to store anywhere and IMHO more of a gimmick. Modifying the existing system brings it's own challenges which I will go into below, it is not for the faint hearted :)

I decided that I did not want to add the extra weight of a separate system and started looking into using the existing climate control system that came from the factory. The easiest way to make that work would be to alter the HEVAC controls so that they can work when stationary (ie. engine not running) and replace the belt driven compressor with an electric unit. After a lot of digging around I found that for 12 volts there are only a few options really; a very ubiquitous 18cc unit and a more rare 20cc variant. Smaller versions also exist like a 14cc and even 12cc but I did not think them adequate for the job. So, I decided to bite the bullet and buy the 18cc version directly from china to cut out the middle man which may or may not have been a good choice... read on :)

In order to make this story about the compressor mostly and not about the HEVAC, which is a story in and of it self, suffice it to say that in the end I will be replacing the current HEVAC controls/computer with an arduino based solution. Effectively I will be building my own climate control that will make use of all the existing sensors, blowers and blend motors. Having said that, to test the compressor and see if it would actually even work I simply plumbed it into the existing system.

Since I will be fabricating a raised air intake for the engine anyway I removed the air box and intake from the engine bay and mounted the MAF sensor directly to the intake plenum and one of those cheap cone shaped "sports" filters directly after it so that I can still drive the car. In this country dust is virtually non existent and sucking in hot under bonnet air is not a problem for the time being. I mounted the electric compressor on the spot where the air box used to live and hooked it up temporarily to see if it would run. Due to a mishap I unfortunately cooked the driver board inside the unit and that is where the first big reason NOT to go electric pops up...

Most, if not all, of the 12v (and higher DC voltages) compressors are pretty much made in the same factory or by the same concept at the very least. The consist of a scroll type compressor driven by a brushless DC motor which in turn is driven by a 3-phase driver. In order to cool the DC motor and the driver board, which is fixed to the aluminium housing with copious amounts of thermal paste, the suction side of the compressor flows through the housing and motor. On top of the driver board sits a controller that knows how the turn the motor in the right direction and provides a soft start by ramping up from 0 to the desired RPM. This controller takes it's input from either a 3 setting lead (4 wires) which allows you to select 2000/3500/5000 rpm or takes a 400h PWM signal where 5% duty cycle is 2000 RPM, 65% duty cycle is 3500 RPM and 95% duty cycle is 5000 RPM. The version you order does not matter since the internal connector is the same and you can simply switch the lead from one to the other. I have not yet tested PWM. In any case, it is the lower driver board that contains a bunch of beefy FET's that failed and I tried to see if I could either repair it, remake another version of get some spares.
Suffice it to say that our chineese friends are not in the habit of producing anything that is even close to durable so spare parts were of course never thought of. You are supposed to consume product and buy more when something breaks. This has put of me off of these units but after some soul searching I had to come to the conclusion that I am already invested in this solution and I had to push on with it. So, I ordered another complete compressor thinking that at least I had a spare one when the mechanicals would wear out.

Now that I had a working compressor I checked my dodgy test wiring at least 5 times to not burn out another unit and.. it started spinning :) Once I had that all set up I mounted a relais between the original compressor clutch power connector and ground and wired the 12v unit up so that it would switch the high setting. After triple testing that everything worked as it should I drove to my local A/C shop to make an appointment for some custom work. Yesterday I finally got to get the entire system plumbed up and it was a good day of working alongside the experts. Really nice blokes who got my idea and simply allowed me to work along side them in their workshop. I did all the "range rover specific" stuff, or should that be "my handywork" :) and they were there to help with the A/C specific stuff. In the end we removed the flexible hoses that run to the original compressor from the aluminium pipes and replaced them with new ones and some angled fittings that were then routed to the new compressor. All in all not that much material was needed to get that done since the original plumbing already routed through that area, but still half a days work.

Then came the BIG moment! filling up the system and checking if it works. Originally the P38 contains 1250 grams of r134. The guy at the shop figured since we altered the plumbing and removed the original compressor which has a large internal volume that we should reduce that to 900 grams and take it from there. I fired up the engine, turned the HEVAC to "LO" and full blast and waited to see what happens. To my relief the "little compressor that could" was capable of running in a system that was designed for a much bigger capacity compressor! The suction line got nice and cold, the air blowing out the vents was freezing and the cycling (because the evaporator would freeze or any other reason the HEVAC computer would toggle the clutch) worked as well, though not as the belt driven unit would behave.

When we tested the system at around 20c we got 2 bar (130 psi) low side and 9 bar (130 psi) high side which is about right. I took the car for a drive with an eye on the current measurements from my victron shunt and I must say, the A/C worked well. Of course we still need to have a 30C + day which we will have next week so that should be interesting. The total current drawn by the vehicle never got much over 110A (and I know the car uses around 30 when just running) but at idle (in drive at the stop lights) the voltage would drop to around 12.5v so clearly the alternator that I currently run is not able to keep up at such low RPM's and the battery cops a beating for a few seconds.

The one thing I did observe is that the ramp-up of the compressor is quite slow and the HEVAC does not understand this, that is to say it is not designed for that. So when the compressor is disengaged, it does not ramp down slowly, it just stops. But when the A/C request line is brought up high again in stead of the evaporator immediately cooling down again it slowly cools down and since the blowers did not slow down you can feel the humidity from the evaporator entering the cabin just briefly. Suffice it to say that my own climate control algorithm will have to take that into account. Perhaps I will simply run the compressor speed up/down as required to keep the evaporator in equilibrium.

What's next? Well I intend to test the pants of the system whilst my car is still in running condition, ie. before I start work on the other projects. Next week is probably the last week of 30+ temperatures in this neck of the woods so I also plan to play around with the A/C then. I will try the low and medium setting and see what that does and see how the entire thing will cope with a heat soaked car that has stood in the sun for quite some time etc. I reckon that it will be able to handle, just, but that is good enough since my build will remove a lot of windows, replacing them with insulated walls so there will be a larger area to cool, just not one that will have that much light incidence.

To be continued :)

Cheers,
-P

[Old Farang]

You may want to correct the typo:


we got 2 bar (130 psi) low side and 9 bar (130 psi) high side which is about right.

[prelude]

You'r right, that was supposed to be 30 psi at the 2 bar mark. Unfortunately edit grace period has long since passed :(

-P

[prelude]

The "hot" days came and went so it is time for an update to this write up!


The first real test began on a 29C day with around 85% humidity. Not as hot as it can get in tropical/high humidity climates but a nice test nonetheless. Not to get into too much detail: the system can cope. I am not sure if it performs the same as the belt driven unit, surely it cools down a bit slower due to the reduced capacity, but I felt comfortable, the internal air temp dropped to very acceptable levels and the humidity was very clearly decreased. I tested it without outside air to make it harder on the A/C but still it succeeded brilliantly. The compressor was on the highest setting and I observed it cutting out every 5 minutes or so meaning that the evaporator was freezing up ie. there is spare capacity.


I also tested with the compressor at the medium setting. The car by then was already properly cooled and no longer heat soaked but on medium the compressor would simply no longer cut out. The internal temperature and humidity felt fine though and I reckon that the medium setting is perfectly serviceable though it would take forever to cool down a heat soaked car, if at all.


The second test I did was on a 30C+ day with lower humidity (around 55%) with the car stationary. I hooked up one of my lithium batteries over the car battery and put a 750Watt charger on top of that. I did a lot of tests with the car facing into the sun running all sorts of different combinations of compressor speed, condensor fan speed and blower speed.


My first observation is that one should not underestimate the power of the VC fan! Even at idle it will pull quite a bit of air through the radiators and condensor. There is a topic on here somewhere regarding replacing the VC fan with an electronic unit (in a p38). I tried to run the compressor without the condensor fans running and the compressor would see very high currents, even in the low setting, up to around 70A+ and would shut itself off. I am not sure if that is because of an internal overpressure switch or an overcurrent protection but it is nice that it protects itself in any case. Note: I had the compressor wired up directly so the climate control had no control over it. That means that even on the lowest setting, producing much less pressure/flow than the original compressor, the relatively huge condensor (compared to the compressor) could not dissipate enough heat to be of any use. Mind you, the engine was turned off for about half an hour or so, so there would be a lot of residual heat coming from the radiator still and the sun was shining on the condensor.


Switching the condensor fans on, even on the low settings would allow me to run the compressor at any speed and the system would cope, sort of. The suction line would not get as cold anymore in the end (of course) and the superheat would not be within spec any more I reckon but the compressor would not shut off. Maximum current I noted was around 130A which is technically outside of spec for this unit. It is rated at 97A for the motor though the documentation also gives 1.3KW as it's rated consumption at full speed. These two do not match, unless you start playing with the voltage. The controller board is rated at 1.74KW and as mentioned has over current protection. 1300watts /12v = 108Amps and 1740watts/12v = 145Amps so I am not sure what is what.


Later, as I tested it during driving I noticed a pattern: the compressor currents would be high when idling for a prolonged period of time at say a stoplight, which makes sense when the voltage drops since it is a BLDC driven motor that simply keeps it's RPM steady and just starts using more current to maintain the same power levels. Another big difference however would be that the most likely the condensor would be hotter and thus the pressure would be higher in the condensor. It was clear that once I pulled away from a light and hit speeds over 40-50kph for any amount of time (longer than say a minute) the current through the compressor would reduce. I did not notice this behaviour during stationary tests since the maximum airflow over the condensor was limited to the two fans on high. Clearly, that is enough to help the A/C out in the original setup with the belt driven compressor probably pushing the pressure and thus the temperature way up and the VC fan always on doing part of the work but not just running the condensor fans without help from the VC fan. On the electric setup I did not notice any problems or cutouts but I did not want to try stop and go traffic or a traffic jam even at this point since the voltage would drop significantly and my battery was taking quite the beating. In my new alternator setup this will be solved but alas I won't be able to test that until I hit warmer places again.


The third test was in 30C+ just driving the car around doing some moving, shopping, the normal stuff. For the first part I wired the compressor up for high speed operation. Everything operated just fine and the temperature in the cabin, driving face into the sun was perfectly doable. With the climate control set to 20c I measured around 11C coming out of the air vents (which is very near the compressor cut off point, probably due to the temperature between the fins of the evaporator reaching 4c which is when the ECU turns the compressor off) and the blowers on I would guess at most 50% once everything settled down. The compressor would not shut off unless I drove underneath the trees for a while and the light sensor would ramp down the blowers causing the evaporator sensor to cut the compressor. Of course the air blowing over the evaporator would very quickly re-engage the compressor, within 30 seconds or so, but since the ramp up time from 0 to 5000rpm for the compressor is hard coded to be about 60 seconds you could feel a distinct difference in the cabin for that minute. Not uncomfortable but noticeable.


When I would step into a heat soaked car and let the climate control decide how to handle it the air recirculation would not engage, I can't recall ever see that happen in the range rover anyway. This makes life for the A/C a bit harder but since the blowers are hardly ever (if at all) switched to full by the ECU the airflow is somewhat limited. When given time for the compressor to ramp up etc. the air out of the vents would be around 14c and within a few minutes the blowers would calm down a bit. I reckon that it is around 25% slower than usual to do so which is perfectly acceptable.


The second half of the day I hard wired the compressor to use the medium setting with the same usage of the car. The air out of the vents would settle around 15-16c with the blowers in the higher setting and it would take a long time for them to settle down. Still, it was comfortable in the car albeit it a bit more noisy. After a 30 minute highway drive the blowers did settle down and the medium setting was perfectly capable of maintaining the set temperature with the outside temps still hovering around 30c. Suffice it to say the compressor never cycled during this test.


I did not even attempt this at the lowest setting until the late afternoon. With outside temps still 28-29c but the sun lowering I switched to low on the compressor on an already cooled car. The air temp from the vents now climbed up to 17c but that is still lower than the requested temperature and without much external heat entering the cabin (low sun) I drove all the way home without discomfort but one could notice a higher humidity level.


Finally I hooked the entire system up to my mates A/C service unit to see what the pressure were and how it all handled after all my "abuse". At 28c with the system off (for over an hour) I got 6.8bar which is about bang on. With the compressor on the low setting the suction side is around 3.5bar, which is high-ish but still within spec and the high side is around 12. Increasing the compressor speed did not result in significant higher pressure on the high side but the low side would creep down to slightly below 3 bar. This is without the engine running and only the condensor fans on high. Seeing as my initial test during the fill-up I did get around 2bar with the engine running, the VC fan does cool the entire thing down better I am concluding.


All in all it was quite the journey and although I am not done playing around with the system :) I think for now we can make a few observations:
in high European temperatures, say below 40c this compressor will work just fine on the P38 system albeit slower
the current draw on the highest setting means that you need a beefy electrical system
first off a battery type that can live underneath the bonnet and is also capable of delivering high currents for a few minutes without complaining
Your alternator needs upgrading and preferably needs to put out 100A+ at idle
The wiring in the car (ground lead, positive lead from the alternator) will most likely be insufficient and need upgrading
Using an existing climate control or A/C system may lead to different experiences


So what's next?


Well, with all the testing done so far I need a better condensor cooling solution that much is clear. The fans that landrover fitted in the front are actually too big and the sides push air around the condensor in stead of through. As mentioned; these were fitted to help the A/C function in stop and go traffic when the VC fan would simply not be able to push enough air and to help the engine cool in extreme circumstances due to the same reason. This will not be an easy task since the radiator/condensor in the P38 is practically square so the only solution is a very big thermo-fan or 4 smaller ones. The two that are currently mounted have proven to be insufficient although more modern and better fitting fans would no doubt help, I doubt 2 fans would suffice in the end. More on that in the thermo-fan topic ;)


The ramp up time of the electric compressor is also something that needs looking into. It is not changeable AFAIK. Preferably the compressor ramp up to the highest needed setting and down when possible. Since the climate control unit can not be used when stationary (engine off) it was on the nomination to be replaced anyway so the new system will try to manage the compressor speed as best it can.


Finally I already have a design for a new alternator setup that will provide up to 380A of power, more than enough to run the A/C at max, replace the VC fan with the beefiest thermo-fan if I decide to go that route, run the car AND have power left to charge my batteries. This setup should also provide ample power at idle so that the voltage over the starter battery will not drop significantly. Still, I am looking to replace the old worn-out one with another and I am still considering my options. A smaller and lighter unit would on the one hand be preferable, possibly an AGM or else an EFB. Suffice it to say it won't be a DCS[bigwhistle]


That's all I've got for today!


Cheers,
-P