The Homemade Heat Pump Manifesto

[MN Renovator]

"Look at it this way. If your condenser needs to be about 5.5x the size of your evaporator to get great efficiency, how good is the efficiency of the units that have the same size either end? My ASHP has a condenser about twice the size of the evaporator. As far as efficiency goes it sucks, but it does the job and is affordable. Would people buy a unit that was three times bigger and vastly more expensive if it was more efficient? Mostly not. It's a design compromise between cost and efficiency. A tradeoff we DIY's don't have to really make. We can afford the extra time, effort and outlay to make the most efficient system possible."

I'd have to argue against that a little bit. The bigger that both coils are, the more BTUs you can transfer for the same size compressor as long as the size of the coils doesn't significantly increase the amount of power that it takes to pump the refrigerant, but that almost isn't a factor for the size these are. I've been looking all over at AHRIdirectory.org data and time and time again I see a higher EER, SEER, and HSPF for a given nominal size when the evaporator coil model number is a size up from the nominal size of the condensor. Some examples are quite extreme if you combine different efficient elements of a system. Take an 18k or 24k condensor, put it into a furnace or air handler that is a variable speed unit, throw the 18k with a 24k coil or the 24k with a 30 or 36k coil and the values can be 12.2 EER and 14.5 SEER when the condensor is a 13 SEER unit. The combination ratings show capacity 101% and power 88%. The combination I used was for Carrier equip, a 80% 313*AAV024045 furnace 44k nat. gas input variable speed blower with a 113A 18k or 24k condenser and an oversized coil. If anyone wants to see what I was playing with.


The big issue with oversizing an evaporator coil is humidity control, you basically need to saturate the coil with condensation before it beads out into the pan and drains out off the evaporator. If the evaporator is huge and your system takes 30 minutes to get enough on the coil to starts removing humidity and the system stops, that soaked coil evaporates all of that moisture back into the house. This is why you MUST not oversize for cooling if you buy an A/C in an area where there is grass outside your house(non-desert), especially with a high SEER system or you run cycles that are too short to be useful. If humidity was zero concern you could probably have a monster indoor coil and probably get an extra SEER point out of it if you had the amount of air flow exact and the system was charged perfectly.

Hope this helps!

[Ko_deZ]

Hi all

Sorry for the wait. I have only had an hour today, so I did not get to make it very nice or anything, but this should work. Tell me what you want in addition to this. If you do understand a little programming, then making changes to the switch/case thing in the loop function at the bottom should be quite easy. The pump relays could of course also be used to control fans for a water to air or air to air system. No de-icing implemented though. I would add an extra state (or case in the main switch/case part) that has a 5 minute delay and turns off everything except an external deicing heater during this period. Easily done if someone wants it.


This is what you need to do for temperature sensors:

For X number of sensors you need:
2 1N914 diodes
X LM35
X 18kOhm resistors

And some whire. Standard twisted pair or a coaxial cable or whatever you like. If the cable is very long ( > 10 meters ), I would add a small capacitor across the sensor and between output signal and ground all the way out by the sensor, as well as one between ground and signal at the arduino, just to remove any noise that is picked up along the way.

LM35 - Precision Centigrade Temperature Sensor

There are loads of differenty types of temperature sensors, but this one is very easy to use. Also, it gives you a quite accurate temperature range of -55C to +150C. Unfortunately the resolution is not very good, but it is more than good enough for our use.

For the output relays, you need to do like this:

http://www.arduino.cc/playground/upl...ing/relays.pdf

You can use any digital output except 1,2 and 13. I use 13 as a status diode thingy. You will see in the code comments. It is on for either 1, 2, 3 or 4 seconds in a 5 second period to indicate which state it is in.
If you wonder why A0 is connected to the two diodes, that is to get an input reference, as this temperature sensor gives an output signal that is 0V compared to ground at 0C, and 10mV up or down is one C. Analog input 0 will be reference voltage for the temperature sensors. You also get these with Farenheit I believe, in that case, some of the numbers must be changed. Get the Celcius, it is the logical way of measuring temperature anyway

The text at the start should explain most of it. I did leave out a few settings, and I don't utilize the indoor and outdoor temperature. I just left the readings in there so that you could add the probes and get a reading on the serial interface.

WARNING: NOT TESTED IN REAL LIFE!
I did do some simple testing here with a breadboard, but that is the complete extent of it.

Allright, here is the code:

/* First some legal mumbo jumbo:

    This program is free software: you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
    the Free Software Foundation, either version 3 of the License, or
    (at your option) any later version.

    This program is distributed in the hope that it will be useful,
    but WITHOUT ANY WARRANTY; without even the implied warranty of
    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
    GNU General Public License for more details.

    You should have received a copy of the GNU General Public License
    along with this program.  If not, see <http://www.gnu.org/licenses/>.


 * This is a crude implementation for a state for controlling a
 * ground souce heat pump machine.
 *
 * Written by Morgan Tørvolt <morgan@torvolt.com>.
 *
 * Analog inputs:
 *
 * T1 - indoor temperature, connects to A1
 * T2 - outdoor temperature, connects to A2
 * T3 - accumulator tank temperature, connects to A3
 * T4 - ground loop temperature out, connects to A4
 *
 * Digital outputs:
 *
 * D3 - Relay for ground loop water pump
 * D4 - Relay for warm side water loop to accumulator tank
 * D5 - Relay for starting the heat pump
 *
 * Defined settings for temperatures and such. Set those below this comment section
 *
 * TEMP_TYPE    : CELCIUS or FARENHEIT
 * TEMP_START   : T3 temperature for when the HP should start
 * TEMP_STOP    : T3 temperature for when the HP should stop
 * TEMP_SHUTOFF : T4 temperature that is critically
 *                cold (freezing) and the HP should stop
 * DELAY_BOOT   : Time to wait when powering up in seconds
 * DELAY_START  : How long to run the water pumps before turning
 *                on the HP in seconds
 *
 * State 0: All off
 *   Wait 2 minutes before doing anything. Blink diode 1 second on and 1 second off
 *   Go to state 1 after two minutes
 * State 1:
 *   If T3 < TEMP_START, go to state 2
 *   Blink diode on 2 sek, off 1 sek
 * State 2:
 *   Turn on D3 and D4 (start water pumps), leave D5 off (HP)
 *   Wait for 2 minutes, then go to state 3
 * State 3:
 *   If T4 > TEMP_SHUTOFF and T3 < TEMP_STOP, leave HP and water pumps running using D3, D4 and D5
 *   If T4 < TEMP_SHUTOFF or T3 > TEMP_STOP, jump to state 1
 *
 */

enum { CELCIUS, FARENHEIT };

#define TEMP_TYPE       CELCIUS
#define TEMP_START      40
#define TEMP_STOP       50
#define TEMP_SHUTOFF    1
#define DELAY_BOOT      120
#define DELAY_START     120

// General program code. Go further down
// if you don't want to know what is going on
// in the background

void blinkDiode( int on, int period )
{
  if( ( ( millis() / 1000 ) % period ) < on )
  {
    digitalWrite( 13, HIGH );
  }
  else
  {
    digitalWrite( 13, LOW );
  }
}

float getTemperature( int input )
{
  int raw = static_cast< int >( analogRead( input ) ) - static_cast< int >( analogRead( A0 ) );
  switch( TEMP_TYPE )
  {
  case CELCIUS:
    // 1024 states across 5V gives 4.8828 mV per step. For celcius, 10 mV = 1C, so we multiply by 0.48828 to get celcius
    return raw * 0.48828;
  case FARENHEIT:
    // °F = (°C × 9/5) + 32
    return ( raw * 0.8789 ) + 32;
  }
  // This should never happen
  return 0;
}

void setOutputs( int output_3, int output_4, int output_5 )
{
    digitalWrite( 3, output_3 );
    digitalWrite( 4, output_4 );
    digitalWrite( 5, output_5 );
}

void setup()
{
  // declare all digital pins as output pins:
  for( int i = 2; i <= 13; ++i )
  {
    pinMode( i, OUTPUT );
    digitalWrite( i, LOW );
  }

  Serial.begin( 9600 );
}

int           state      = 0;   // We start in state 0
unsigned long waitUntil  = 0;
bool          diodeState = 0;

 void loop()
{
  delay( 50 ); // No need to rush things, and also make sure we don't overflow the serial interface
  Serial.print( "State: " );
  Serial.print( state );
  Serial.print( " Temps: " );
  Serial.print( getTemperature( A1 ) );
  Serial.print( " " );
  Serial.print( getTemperature( A2 ) );
  Serial.print( " " );
  Serial.print( getTemperature( A3 ) );
  Serial.print( " " );
  Serial.println( getTemperature( A4 ) );
  switch( state )
  {
  case 0:
    setOutputs( LOW, LOW, LOW );
    if( millis() > DELAY_BOOT*1000 )
    {
      state = 1;
    }
    blinkDiode( 1, 5 ); // on 1 of 5 seconds
    break;
  case 1:
    setOutputs( LOW, LOW, LOW );
    if( getTemperature( A3 ) < TEMP_START )
    {
      waitUntil = millis() + DELAY_START * 1000;
      state = 2;
    }
    blinkDiode( 2, 5 ); // on 2 of 5 seconds
    break;
  case 2:
    setOutputs( HIGH, HIGH, LOW );
    if( millis() > waitUntil && millis() - waitUntil < 20000 ) // Just making sure that things does not get screwed up when the unsigned long overflows.
    {
      state = 3;
    }
    blinkDiode( 3, 5 ); // on 3 of 5 seconds
    break;
  case 3:
    setOutputs( HIGH, HIGH, HIGH );
    if( getTemperature( A3 ) > TEMP_STOP || getTemperature( A4 ) < TEMP_SHUTOFF )
    {
      state = 1;
    }
    blinkDiode( 4, 5 ); // on 4 of 5 seconds
    break;
  }
}

[BradC]

Originally Posted by MN Renovator

I'd have to argue against that a little bit. The bigger that both coils are, the more BTUs you can transfer for the same size compressor as long as the size of the coils doesn't significantly increase the amount of power that it takes to pump the refrigerant, but that almost isn't a factor for the size these are.

Ahh, but we are talking about two different things. I'm talking about the extra efficiency that can be gained by having a much larger condenser than evaporator. The larger the condenser, the lower the condensing temperature and therefore the lower the compression ratio and power input required.

The other thing is I'm talking about water not DX.

Now, when you have a reversing valve having grossly mismatched HX presents a problem of refrigerant management, therefore it's going to be far more efficient to swap the water connections rather than the refrigerant with a reversing valve.

[Vlad]

Simply by oversizing especially largely oversizing your condenser you will shift your suction pressure and evaporation temperature to lower area. It will lead to freezing coil and ice build up(air coil) or freezing water cooled evaporator.

When I was in refrigeration school we did this many times by opening wide water regulating valve on water cooled units. Our instructor knew what happened just by looking at ice at suction at compressor. He was giving us s... because of converting walk-in cooler into walk-in freezer.

Yes your amps going down, but you don't remove more BTUs. You just shift temperature range. Usually water regulating valve is just cracked open it means that you use only small part of condenser.

All parts of refrigeration system MUST MATCH each other. It doesn't matter how huge your evaporator or condenser, your system capacity will not be bigger then compressor capacity.

In theory everything is possible, but in our DIY projects we use standard mostly scrap parts so we have to follow their parameters.

[Vlad]

Originally Posted by BradC

....the extra efficiency that can be gained by having a much larger condenser than evaporator. The larger the condenser, the lower the condensing temperature and therefore the lower the compression ratio and power input required.

So, by removing MD you have compression ratio = 0, then your power input will be 0 and efficiency will be infinite.

[Vlad]

Originally Posted by Xringer

I've been using SSRs on my Sanyo ASHPs and they don't seem to heat up
their heat sinks much at all. Slightly warm to the touch maybe..

Xringer, how about LRA which is usually much higher then SSR rating? Can you just size SSR based on compressor FLA and just ignore LRA because it is so short period of time. Also can you put 2 SSRs one for each phase? Sorry about dumb questions but I have never used them.

[BradC]

Originally Posted by Vlad

Yes your amps going down, but you don't remove more BTUs. You just shift temperature range. Usually water regulating valve is just cracked open it means that you use only small part of condenser.

Not true at all. Your suction temperature drops because your compressor is not working as hard and therefore can flow more mass of refrigerant. Therefore you *DO* remove more BTU's. Refrigeration in BTU or Watts or whatever is precisely linked to the refrigerant mass flow. It is after all the phase change of the fluid that provides the effect. More fluid, more heat moved.

Originally Posted by Vlad

When I was in refrigeration school we did this many times by opening wide water regulating valve on water cooled units. Our instructor knew what happened just by looking at ice at suction at compressor. He was giving us s... because of converting walk-in cooler into walk-in freezer.

The above quote proves my point. A walk-in is a balance between the heat you are removing, vs the heat infiltration and product load. If you drop the temperature in the room, you *must* be moving more heat unless you have radically re-insulated.

To counter this you can do one of two things.
- Increase heat load on evaporator resulting in more refrigerant mass flow and therefore a higher suction pressure. In my case I can do this by speeding up the fan.
- Reduce compressor speed to reduce the mass flow. My in-build unit will do this with the VFD.

The closer your evaporation and condensing temperatures, the more efficient your unit.

As your compression ratio drops, your compressor can move more refrigerant mass per revolution. This is to do with dead space at the top of the chamber, the efficiency of the valves and other factors. The effect is less on scrolls, but on Recips you can lose up to 30% of your mass flow by pumping to a much higher pressure. Plus of course your current draw increases.

If you have controls to allow you to tweak the other variables in the refrigeration cycle then you can easily take advantage of the increased efficiency. In the case of the walk-in you describe, you can't and therefore you cause the refrigeration system to find equilibrium outside the parameters it was designed for.

This is precisely how VRV systems utilize the best available parameters to get as efficient as possible.

On VRV systems, my comment about radically oversized condensers vs evaporators causing an issue with refrigerant management? I saw one today that has a 4 circuit condenser. When the unit hits the reversing valve, it only uses 2 of the circuits as evaporators, so half the available liquid capacity. Clever!

[Vlad]

Originally Posted by BradC

Not true at all. Your suction temperature drops because your compressor is not working as hard and therefore can flow more mass of refrigerant. Therefore you *DO* remove more BTU's. Refrigeration in BTU or Watts or whatever is precisely linked to the refrigerant mass flow. It is after all the phase change of the fluid that provides the effect. More fluid, more heat moved.

You are going wrong direction.

1. You lowered discharge pressure by oversizing condenser.

2. Because of #1 you feed less refrigerant through MD (and mass flow as well)
#2 is true because it is harder to push liquid refrigerant with less pressure.

3. Because flow rate is lower but compressor keeps running and sucking refrigerant it lowers suction pressure.

4. Lower suction pressure=lower temperature.

5. Lower suction pressure lowers discharge pressure.

You only move more mass when you have it, it means the lower suction pressure the less to move. So amps go down because less work is done. Take all refrigerant out and check amps, they will be very very low.

For this reasons refrigeration compressors are differ from AC. Refrigeration compressors have to deal with less dense vapor(low temp=low pressure=low density), but you still want them to pump enough mass, so you put bigger pistons or vanes or scrolls or... They simply have bigger displacement.

It is all cycle. One affects the other.

[BradC]

Originally Posted by Vlad

You are going wrong direction.

1. You lowered discharge pressure by oversizing condenser.

2. Because of #1 you feed less refrigerant through MD (and mass flow as well)
#2 is true because it is harder to push liquid refrigerant with less pressure.

In the case of a fixed expansion device (orifice or cap tube) I agree with you completely.

A TXV or properly controlled EEV will open up further to maintain superheat thus increasing the mass flow in the system in line with the reduced condensing temperature. You move more heat from the evaporator. If you don't have the heat to move, then your SST drops until the system reaches an equilibrium.

Additionally, as the liquid entering the MD is cooler it has greater enthalpy and therefore you increase the evaporator capacity on a mass flow basis.

I'm not saying reducing condensing temperature is a good thing to do in isolation, but in my case where I'm designing the system cycle from scratch and have complete control over the metering device, it's a wickedly easy way to gain both capacity and efficiency.

My point is as a DIY'er, if you are aiming for super COP's it's something you should really consider. Now if you are heating water for a hydronic application, then oversizing the condenser is going to buy you nothing, but if you are trying to cool a space (like I am) it's like printing free money.

Based on the power savings converting from air to water condensed and going with a bigger HX, the water-side setup (pump, pipes, filter, HX) will pay for itself in less than 24 months. What's not to like?

[Pegasus]

Originally Posted by BradC

In my "Brads Gas and Gear" thread, I posted an Openoffice spreadsheet that will allow you to calculate the surface area required to achieve certain HX goals.

I use that in combination with "Coolpack" to define my requirements.

Thanks a lot. I will do my homework and report back !!!