Reverse-cycle (heat pump) heating

[4Springs]

Driving the Brumby can be cold. To get it registered I installed a 12V hairdryer pointing at the windscreen. It does help with demisting but is entirely ineffective at heating the interior. And it is incredibly loud! I installed a heated seat cover last year which is really good. And I have a pair of driving gloves which I use for probably about 5 months of the year.
Driving to work in the winter can be hazardous in foggy weather, as sometimes the windscreen fogs over quickly (one hill in particular is know for this, I think the trees breath heavily right there). The worst is in sub-zero temperatures when the windscreen fogs over and then freezes. Windscreen wipers don't do ice!

There are a few ways of heating in a conversion:
You can install a resistive heater in place of the heater core and heat the air directly.
You can install a small kettle to heat water, and circulate the water through the original heater core.

I, like many people, have been tempted by the idea of using an air-conditioner pump to provide heat in the cabin. In Tasmania this is generally called a heat pump. Heat pumps can be much more efficient than pure resistive heating, in theory you can get 300% efficiency. Great for a range-limited vehicle used often in cold weather.
I, like many people, purchased a Masterflux air-conditioning compressor from EV Works. Masterflux Compressor. This beastie runs off DC voltage, from 120 - 420VDC. Perfect for the Brumby, it can run from the 150V drive battery.
I did originally imagine that I could install a valve to allow cooling as well as heating, but that turned out to be too complicated.

A heat pump needs a condenser (the bit that gets hot) and an evaporator (the bit that gets cold). Ideally I'd put the condenser in the car in the flow of air where the heater core was. This proved impractical because everything is so small and neat. No room to bodge the connections to a condenser even if I could find one small enough and the right shape. So I decided I needed a heat exchanger to heat water, so I could circulate that through the original heater core. I did some reading and found that these things are reasonably common overseas, but no local supplier had ever heard of one. Luckily I found one on eBay from America! Circular and compact, this one was second-hand from a laboratory where it was used to cool something. This is the one I bought: Packless

Here is a picture of the purple Masterflux compressor and the 12V water pump to circulate the water through the heater core. For some reason it is sideways...



For the evaporator, I assumed that I'd use a small radiator like in a normal automotive air-conditioner. I'd run a couple of 12V fans on it. But a local air-conditioning technician guy took an interest in my project, and he convinced me that I could use my motor instead! After all, can you imagine a bigger fan than the air-cooling running through an EV motor? Plus there is the advantage that some of the heat the motor produces would be available to use as heating in the cabin.





So all I had to do was wind copper pipe around the motor. In the photos you can see a bit of the white heatsink compound I used to help with heat transfer. I managed to get the pipe wrapped around with the motor in place, with access from the top and bottom. Once I had enough coils I tightened them up using a couple of gripples (normally used to tighten wire on fences). I then made a clamp out of galvanised sheeting to keep them in place.

[Adverse Effects]

that is cool *pun intended*

[4Springs]

Ok, so how am I going to control this thing?
First off I installed a switch on the dash:



The switch is the one under "R.DEF". That was previously a blanking plate (no rear defroster in a Brumby).
The Masterflux compressor is variable speed. So in theory the speed can be set to exactly what is needed for a particular temperature. Lower speeds use less power. To control the speed of the motor Masterflux provides a controller circuit board.
Here is a photo showing the entire setup:



The large central heatsink is the Masterflux controller. I've made a wooden box for it, since it comes with no housing at all. Directly to the left of it is the top of the 1000A Kelly controller for the drive motor. Note that the Masterflux controller is much larger than the Kelly, although it drives a motor many times smaller!
Behind the Masterflux controller are two silver DC/DC converters, and to the right of them is the round heat exchanger. The compressor is down in the well behind the red button.
In this photo you can also see the clamps I made for the copper tubing around the motor. Above that is the thermal expansion valve and other pipework installed by the air-conditioning people. Note that air-conditioning needs to be installed and serviced by appropriately licenced people...

The Masterflux controller is dumb, it just makes the motor run at the speed that you tell it. It does this via a 0-5V interface. I got quite a long way into building my own controller for this before I realised that you can buy one off the shelf. So I bought myself a Carel controller from RS-Components. A bargain at $182 delivered!
This controller takes two analogue and two digital inputs, and provides a 0-10V output. I used it to control the compressor based on the temperature of the condenser and the evaporator. To do this I bought two RTD temperature sensors and taped them to the pipework. They are visible in the big photo, underneath white tape.
I've programmed the controller to control the compressor speed based on the temperature of the condenser. I set a setpoint and it basically runs flat out until it gets close to that value. Then it runs slower to try and keep the setpoint without overshooting. The temperature of the evaporator is also used, since there is a specification that the evaporator should not get lower than -23 degrees. I've programmed it so that if the evaporator temperature falls below -20 it will disable the compressor.



I also used one of the digital inputs to the controller. The switch on the dash turns on a relay to provide power to the water pump, and it also goes into the Carel controller. The controller displays "OFF", and doesn't control until the switch is turned on.
On the end of the plastic box are three plugs. The 9-pin D connector has two temperature sensors and the 0-5V motor control signal. The two banana plugs are the 150V supply to the compressor controller.

[Adverse Effects]

that is a nice setup not cheep tho

[jonescg]

Very cool Christopher! I hope to see it in action this November. Always wondered about putting a conventional AC unit in a vehicle...

[Johny]

Fantastic Chris. That engine bay is looking very full.

The system you have put it will help take heat away from the motor as well.

So does it work well?

[reecho]

Awesome...

Can you sketch up a schematic for it?

[4Springs]

Can you sketch up a schematic for it?

Excellent idea, will help explain the next phase, which is testing!

This shows the refrigerant circuit (blue and red) and the water circuit (green). It does not show the electrical circuits (12V or 150V).
I've written in some temperatures to demonstrate how it all works.

The compressor at the bottom takes the refrigerant gas (low pressure) and compresses it to a liquid (high pressure). Compression makes a substance heat up. So if the gas was at ambient temperature then the liquid would be somewhat warmer (say 40 degrees). I've drawn the 'hot' liquid in red.
The liquid flows around to the heat exchanger, where it transfers some of its heat into the water circuit. Note that the water goes through the heat exchanger in the other direction to the hot refrigerant, so the water would leave the heat exchanger at about 40 degrees, and the refrigerant would leave at the same temperature the water entered at (say 20 degrees).
Then we come to a filter dryer. This is a standard component in automotive aircon, it filters particles and water out of the liquid.
The next thing is the thermal expansion (TX) valve. This valve is there to reduce the pressure in the left side of the circuit. Think of the compressor as pressing against this valve - so the right side stays at high pressure and the left is at low pressure. When the pressure is reduced the liquid turns into a gas. It also gets colder, which is part of the magic of air-conditioning.
This cold gas (shown as -20 degrees) can be warmed up by passing air over it. In my diagram I have shown that the fan warms up the gas from -20 to +10 degrees. In my car this is the coil around the motor.

In the water side of the diagram, the warm water comes from the heat exchanger and is passed through the heater core inside the vehicle's cabin. The fan there passes air over the core which cools it to some extent.
I've shown my reservoir there, which is required to allow expansion and contraction of the water. It also removes air bubbles from the system, and gives me a place to add more water if required. I used a plastic juice container, it is just visible in the background of this photo:



The 12V pump then pumps water back up through the heat exchanger.
I left that pump on the other day, I got back to the car after work and heard it humming away. It mustn't use too much power, my little 12V battery was still powering it after 9 hours. Must change it so it turns off with the key...

Next post - results!

[4Springs]

Time for some testing...

On a sunny day of 13 degrees I ran the motor in the garage for some time. First of all I put the controller into Autotune mode, in which it does some testing to see what PID settings are best to control the load it is connected to. This worked fairly well, and I could then set it to various temperatures.

Some pictures tell the story:



Look for the red laser pointer dot to see what is being measured. The first one is the ambient temperature of the car.



The motor is 20° because it has been running for a while.





This bit is the intake to the compressor. Masterflux says in their specifications that this has an absolute minimum of -23°C.



This is the temperature of the refrigerant as it goes into the heat exchanger. 40° was the setpoint.



Nice warm air coming out from the air vent.



And just to show how cold it was in the car itself.



This is the gauge showing how much current I'm drawing. This is the motor running at about 4000 rpm, plus the air-conditioning at the same time.

[4Springs]

Running the motor at speed sitting the garage on a relatively warm day meant that the setpoint of 40°C was achieved fairly readily. The controller would cycle the compressor on and off at the minimum speed to keep that temperature, even with the fan in the cabin blowing at full pelt.
There was a bit of ice on the intake to the compressor, and a little bit of condensation on the motor. A concern has always been that too much condensation on the motor might make things worryingly wet.

I've driven to work a few times now, a trip of 15 km. The temperature in the morning has been something like 6° in the garage and 3° air temperature once I get going. The system is not coping with these temperatures.
There is a second temperature sensor on the intake to the compressor. This is connected such that if that position gets to -20°C it cuts out the control signal to the compressor. On the way to work this is what is happening - it gets too cold and cuts out, then warms up and cuts in again. The setpoint is 30°C but it is not reaching that. The other night I drove home at highway speeds with an ambient temperature of 1°C, and the heater core was only heating to 16°C.
This performance is good enough for demisting, but does not heat the cabin much. The Brumby cabin is not exactly insulated well, so it takes a fair heater to keep up with the heat loss from the windows, thin floor coverings and various holes where the breeze comes through.

So how can I get this system working better?
The first thing to try is to insulate all the exposed pipes that carry warm refrigerant and water. No use losing what was hard-gained! The air-conditioning place used a self-adhesive black rubbery insulator. I haven't been able to find any online yet, I'll keep on looking.

My guess is though that I need to get the cold line heated up better. Going from -20° to 6° should be possible if I can get enough airflow. I might be able to fit a little radiator in the line somewhere with fans on it. Or perhaps I can heatsink the copper pipes better. Currently they are covered by a flat piece of galvanised sheeting. It would be great if I could make a heatsink to clamp on instead, and put fans on that heatsink. A nice aluminium heatsink would make it look pretty fancy too!

[jonescg]

It's so interesting reading these thermal shenanigans... It got to 19'C today and Perth folks were lighting the fire

Definitely wrap closed cell foam around the pipes - that stuff can make a world of difference. Would it be easier to wrap a heating element around the pipe just before the heat exchanger? I know its a resistive load and kind of defeats the purpose of the RC-AC, but it might be a fairly simple approach. You would need to be able to control it so it only came on when ambient was 10'C or cooler or something.

But I'm a bike guy... AC is kind of foreign

[antiscab]

When you drive with a cold motor, what rpm range do you keep it in? With my impulse 9, until it reaches around 40 deg c, I find the most efficient system efficiency means keeping the motor between 2000 and 3000rpm. Basically shifting losses from gearbox and bearings to the motor.

I wonder if you just need to run your motor a bit harder to generate enough heat, rather than relying upon sucking it from the air.

[4Springs]

Definitely wrap closed cell foam around the pipes - that stuff can make a world of difference. Would it be easier to wrap a heating element around the pipe just before the heat exchanger? I know its a resistive load and kind of defeats the purpose of the RC-AC, but it might be a fairly simple approach. You would need to be able to control it so it only came on when ambient was 10'C or cooler or something.

Yes, something like this might be good to add a bit of heat on the cold mornings. A heater could be on any of these components:
1. The motor - i.e. put an electric blanket on it. Could be run from 240VAC, so it could add a lot of heat to the motor while plugged into the wall. This has the disadvantage of heating components that I don't want heated (gearbox etc.), and a lot of the heat would be wasted once the motor was started. I'd want to take it off during summer.
2. The pipe into the compressor. Much like jonescg suggested, wrap a resistive heater around it to heat the gas inside. Would need to be on a thermostat.
3. The refrigerant pipe into the heat exchanger.
4. The water pipes.
5. The water itself, with a kettle such as this one from EV Works: Hot Start.
6. The air itself, with a resistive heater inside the cabin.

I've ordered these from what I assume is least efficient to most efficient.
My guess is that an element wrapped around a pipe would not be able to deliver enough heat to be effective. I'm thinking 500 - 1000W might be required. And it would need to be quite well insulated to make sure the heat went into the pipe and not out into the air.
I like the idea of a kettle, since the element would heat the water directly. That particular one from EV Works though looks like it would probably add a lot of volume to the system. My estimate is that the system currently holds about 300ml of water, and that kettle looks like it would probably double that or more. Looking at ebay there are other promising-looking options, including an under-sink jobbie that claims 3.5kW for $89. Might almost be worth getting that just to see if I could make it work on 150VDC...

A resistive heater inside the cabin is difficult in the Brumby as space is limited. But I assume this would be the most efficient way to spend my Wattage. I do currently heat the air inside the cabin with a 240V heater (i.e. can only run when the car is plugged into the wall). This is installed with a remote control, so when I get up in the morning I can pre-heat the cabin with the touch of a button.

When you drive with a cold motor, what rpm range do you keep it in? With my impulse 9, until it reaches around 40 deg c, I find the most efficient system efficiency means keeping the motor between 2000 and 3000rpm. Basically shifting losses from gearbox and bearings to the motor.

I wonder if you just need to run your motor a bit harder to generate enough heat, rather than relying upon sucking it from the air.

I generally try to keep the motor running fairly fast, probably 3-5000 rpm. This is mainly to get the airflow as high as possible.
It's mostly downhill on the way to work, at least to start with, so I'm conscious that the motor isn't working very hard. I was hoping that running harder would translate to more heat, but it didn't seem to eventuate the other day on the highway. Running at 200-300 amps for 20 minutes in the summer would normally produce a motor 70°C or higher. But the other night with 1° air (and the heatpump) cooling it down it was barely 20° when I got home. So I think I'd have to run it really inefficiently to get it hot enough. Use 30kW of electricity to get 1kW of waste heat!

[antiscab]

there is another option:

pipe in an evaporator radiator into the refrigerant line before the motor.

so ambient air heats the refrigerant to slightly before ambient, before the refrigerant goes to the motor to get heated above ambient.

probably expensive to redo, but should improve heat output without using more wh

[T1 Terry]

(1) Send the -20*C gas through a heat exchanger (condenser) between the fan and the motor so you get to use the cold for motor cooling before it goes through the case cooling tubes.
(2) As the cooled water leaves the heater core in the drivers cabin, send it through the controller heatsink plate to cool the controller and preheat the water before the heat exchanger.
(3) Last idea is the most energy inefficient but a proven concept, a few Peltier effect modules on a heat exchanger where the water exits from the controller heatsink plate before the AC heat exchanger using the ambient temp over a heatsink on the cool side for the module to extract what ever heat it can and pass that to a heat exchanger heatsink on the water side so the water is further preheated before entering the AC heat exchanger.

If the water flow could be diverted to a radiator rather than through the cab heater for summer driving the Peltier effect modules either turned off or reverse connection to move the heat out of the water, then the same cooling effect could be gained via the condenser core between the fan and the motor as well as the motor case cooling.

T1 Terry

[4Springs]

pipe in an evaporator radiator into the refrigerant line before the motor.
....
probably expensive to redo

Yes, I even have an evaporator radiator which I originally purchased before wrapping up the motor. Ideally this would utilise the airflow coming from the motor, since that is slightly warmer than ambient. But space is at a premium around there, so it would have to go at the front where the original radiator used to be. Would be prone to icing up, but once iced up it would be only as bad as it is right now...

(1) Send the -20*C gas through a heat exchanger (condenser) between the fan and the motor so you get to use the cold for motor cooling before it goes through the case cooling tubes.

I'm not sure what you mean here. Possibly what antiscab suggested above?

(2) As the cooled water leaves the heater core in the drivers cabin, send it through the controller heatsink plate to cool the controller and preheat the water before the heat exchanger.

In this weather the controller does not get as warm as the "cooled" water. So doing this would warm up the controller and cool down the water! It would be better to do this with the refrigerant, except that I'm not confident the Kelly Controller is able to operate at -20°C. In all I don't think my controller is a good source of heat, my measurements have always seen it cooler than the motor, and at about 120 times less mass there is a lot less heat there.

(3) Last idea is the most energy inefficient but a proven concept, a few Peltier effect modules on a heat exchanger where the water exits from the controller heatsink plate before the AC heat exchanger using the ambient temp over a heatsink on the cool side for the module to extract what ever heat it can and pass that to a heat exchanger heatsink on the water side so the water is further preheated before entering the AC heat exchanger.

Not sure I'm following all the ins and outs of this suggestion, but Peltier effect is pretty inefficient isn't it? So it would be more efficient to just put a resistive heater element intto the water or refrigerant. At least then it is about 100% efficient.

The hottest I've ever seen the motor in summer is about 80°C. The hottest I've seen the controller is about 60°C. Typical winter temperatures are about 40 and 30. The motor and controller are both air cooled, there is no provision for water cooling on either of them. And as you can see I have no requirements to cool them, I think at 80°C I still have a fair way to go before I'd be worried about the motor being too hot.

[reecho]

Chris you may of overcharged the system perhaps???

Could explain the freezing up on the suction line...

[4Springs]

Cold again, so I've been working on this heatpump, trying to get it more efficient.
I've decided to put a heatsink on the evaporator coils around the motor. I had various ideas, but eventually settled on something fairly cheap and simple. Took quite a lot of work to make though.

I bought some aluminium bar from the local salvage shop. 20x25mm, they charged $5 per kilo. I cut it into short bars like this:

IMG_20180811_134926497.jpg (180.74 KiB) Viewed 2236 times

It was interesting making these. Dad has an old power hacksaw that he purchased at a clearing sale. He knew that it "worked" (i.e., went back and forth when plugged into a power point), but had never used it. So my first few weeks were taken up with getting the antique machine going properly.
Cutting the bars to length was pretty straightforward once the hacksaw was working. But then I wanted to cut the slots in the top. I found that the hacksaw would take more than one blade. I ended up putting four blades on - two sets of two, to cut two slots at once. The blades left a thin sliver of aluminium between them, so I drilled a hole below each slot. Then put it back in the hacksaw to finish it off.
The holes in the bottom of the bar are to take the wire rope that holds them in place.
Painted black to radiate heat better. As per this interesting video: https://www.youtube.com/watch?v=f1QL9ve ... Vs12Aw7Btb

Here is a photo from the underside of the motor, with the heatsinks in place:

IMG_20180812_095043169.jpg (277.52 KiB) Viewed 2236 times

The space in the middle is for an engine mount, not yet fitted. Same with the flat heatsinks - they need to be thinner here to fit under the bracket. You might be able to see the wire rope running through.
Here is a photo from the top:

IMG_20180812_115146525.jpg (283.26 KiB) Viewed 2236 times

I'd hoped for a nice arrangement of heatsinks all the way round, but the refrigerant pipework gets in the way. I decided to use this gap for the threaded rod, which is required to put tension on the wire, and thus keep the heatsinks in place.

The new heatsinks are on, and I ran a few tests. They definitely get cold, but it is very hard to tell if there is any difference in a static test. I'll drive to work over the next week and see if I think it is any better.
I did get out a fan at one point, and found that it makes a large, and very quick, difference. The ambient air was about 5 degrees at the time, and the heatsinks were down to -7. Directing a 12V fan over the heatsinks brought the temperature up to about zero within a minute or two. I'm not going to race out and put fans on them just yet though - I imagine that when running there should be lots of air movement around the motor.

Next step could be to fill in those gaps around the copper pipe with a heat conductive compound. I think with this arrangement I'd be able to force a goo from a tube in under each heatsink in turn.

[Nevilleh]

I admire your tenacity! Hope you get it working satisfactorily.
My old BMW E30 had air con and I got my brother, who is an air con specialist, to help me retrofit it. We drove the compressor off a v-belt pulley mounted on the front end of the drive motor and used the electromagnetic clutch to operate it. We re-routed the pipework so as to cool the existing radiator and heat the original cooling evaporator. He did some magic with the pipework so the expansion valve was round the other way. The heating was amazing as that little evaporator was trying to suck the heat out of the original large condenser (radiator) and it warmed the whole car within several minutes of driving. Of course, it only worked while driving!
I sold the car a few years ago and never took any pictures, but the conversion was not hard.You do need to convert a car that had air con in the first place though.Helps to have an air con man on tap too, as I had no idea how much refrigerant the thing would need- or even how to get it in there!

[4Springs]

Well, I bought some (thermally) conductive goo, as discussed over in this thread: viewtopic.php?f=51&t=5555
I ended up buying the stuff that Richo suggested, part number 8497400 from Element 14. This is a white paste with the consistency of putty. I squeezed it in around the coils as much as I could to make contact between the coils and the motor. This was not easy in the close confines of the engine bay, so I probably did a half-decent job over about one third of the motor. I also put the stuff under each of the heatsinks that I could reach, probably about two thirds of them. The goo dries out to a hard material that will break up with a bit of effort, so I should be able to get it off again if I ever need to. I did notice that it really does stick well to paint. On an experimental piece it took the paint off one of the heatsinks when I prised the dried adhesive off.

The goo is white, so doesn't look good on the otherwise black motor. I seem to have lost the picture I had of it unpainted, but here is the final, painted job:

IMG_20180915_154623905.jpg (174.79 KiB) Viewed 1891 times

This stuff has made quite a difference. The cabin warms up much more quickly now. The compressor is cycling off due to the outlet reaching the setpoint, rather than the inlet getting too cold.

[T1 Terry]

Great to see you have had a win with this project. I very recently became the new proud owner of PHEV Prius and stumbled across a problem regarding the heater. It runs the ICE motor to heat the water for the heater so just wastes fuel, so I'm looking at the possibility of either reworking the air con to get it to reverse cycle, or a heat exchanger from the inverter and HV motor cooling with circulation pumps to get the water to flow through the heater core. No idea if there is enough heat generated fast enough for this to be viable, where the reverse cycle air con would work every time and maybe employ a heat exchanger in the HV cooling circuit to heat the refrigerant.

T1 Terry