A better defrost for heatpumps - Page 2

jeff5may · 12-03-17, 10:31 PM

Most all of the defrost algorithms I have seen that actually work reliably are based on Delta T. The generic boards I've employed and pretty much all of the defrost control boards in split systems' outdoor units use 10k NTC thermistors for thermometers. Some run on time and temperature, and others have demand defrost algorithms. They are more predictable and reliable than the stuff I tried to come up with.

The time and temperature defrost control has only one sensor. It reads coil temperature. When coil temperature falls below the setpoint, a timer begins. When and if the timer runs out, defrost begins. These are great for water source systems, but not so great for air source units.

The demand defrost controls have two thermistors to sense temperature. The first sensor in the same spot as the time temperature control: on the heat exchanger. The other is attached to nothing and measures ambient temperature. The controller reads the difference between the two sensors. When the coil gets a little frosty, the delta T isn't affected very much. As the frost grows, the delta T rises only when airflow is restricted. At some point the delta T becomes excessive and starts the defrost timer. If the timer runs out, the controller initiates defrost.

jeff5may · 12-06-17, 10:21 AM

The mini split units also have temperature sensors indoors. Usually there is one sensor that reads indoor ambient temperature and one that reads heat exchanger temperature. In cooling mode, the hx sensor functions as a defrost sensor. In heating mode, it works with the ambient temperature sensor to measure indoors delta T to do other things. I T all depends on how smart the controller is.

geoff · 12-19-17, 01:00 PM

Sorry, I have to spam to hit my fifth post and be able post a link.

geoff · 12-19-17, 01:01 PM

Here is a great write-up using differential pressure to measure outdoor coil frosting:

Heat Pump Defrost Cycle Information

I think he used a 0.00" to 1.00" WC DP sensor.

Acuario, you may have better luck using a single high resolution differential pressure cell. As opposed to using two atmospheric pressure sensors, and taking the difference in software.

MEMPHIS91 · 12-24-17, 10:46 AM

I'm thinking out loud here but maybe we are thinking to hard on this. We are trying to detect ice, not detect the conditions for ice, or even detect the lost of air flow.
I'm thinking place a sensor somewhere past the metering device at a given (x) distance away from the line (we want t detect ice, not frost).
This sensor could be a switch monitoring small voltage drops between two post (y) distance away from each other. That are "tripped" when ice forms between the two post.
Or it could be a heated temp switch that has to be completely iced over before it starts the defrost. If place in the correct location(where the ice is going to build up first) then it should start the defrost in time. Might even have to add a delay.
Just some ideas.

jeff5may · 12-25-17, 10:15 AM

Nono, some frost formation isn't a bad thing. Especially when the outside temperature is below freezing, the mere presence of frost in small to medium amounts is a good indicator that the heat exchanger is putting heat where you want it. The biggest (high SEER or HSPF) heat exchanger designs can pull full capacity up to the point where they look like they are covered in snow. So detecting the presence of frost alone is not a reliable way to control defrost.

On the other hand, detecting excessive pressure drop or delta T is much more predictable and reliable. Like any heat exchanger, there has to exist a pressure and temperature difference to force heat transfer. Deciding how high is too high on either or both parameters is key to efficiency and effective performance. Only after a sufficient amount of restriction exists for a length of time should defrost be initiated.

12-24-17, 10:46 AM	#15
MEMPHIS91 Journeyman EcoRenovator Join Date: Apr 2015 Location: Oxford, MS USA Posts: 496 Thanks: 69 Thanked 87 Times in 61 Posts	I'm thinking out loud here but maybe we are thinking to hard on this. We are trying to detect ice, not detect the conditions for ice, or even detect the lost of air flow. I'm thinking place a sensor somewhere past the metering device at a given (x) distance away from the line (we want t detect ice, not frost). This sensor could be a switch monitoring small voltage drops between two post (y) distance away from each other. That are "tripped" when ice forms between the two post. Or it could be a heated temp switch that has to be completely iced over before it starts the defrost. If place in the correct location(where the ice is going to build up first) then it should start the defrost in time. Might even have to add a delay. Just some ideas. __________________ שְׁמַע יִשְׂרָאֵל ה' אֱלֹהֵינוּ ה' אֶחָד

Thread Tools
Show Printable Version Email this Page
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

12-03-17, 10:31 PM	#11
jeff5may Supreme EcoRenovator Join Date: Jan 2010 Location: elizabethtown, ky, USA Posts: 2,428 Thanks: 431 Thanked 619 Times in 517 Posts	Most all of the defrost algorithms I have seen that actually work reliably are based on Delta T. The generic boards I've employed and pretty much all of the defrost control boards in split systems' outdoor units use 10k NTC thermistors for thermometers. Some run on time and temperature, and others have demand defrost algorithms. They are more predictable and reliable than the stuff I tried to come up with. The time and temperature defrost control has only one sensor. It reads coil temperature. When coil temperature falls below the setpoint, a timer begins. When and if the timer runs out, defrost begins. These are great for water source systems, but not so great for air source units. The demand defrost controls have two thermistors to sense temperature. The first sensor in the same spot as the time temperature control: on the heat exchanger. The other is attached to nothing and measures ambient temperature. The controller reads the difference between the two sensors. When the coil gets a little frosty, the delta T isn't affected very much. As the frost grows, the delta T rises only when airflow is restricted. At some point the delta T becomes excessive and starts the defrost timer. If the timer runs out, the controller initiates defrost.

12-06-17, 10:21 AM	#12
jeff5may Supreme EcoRenovator Join Date: Jan 2010 Location: elizabethtown, ky, USA Posts: 2,428 Thanks: 431 Thanked 619 Times in 517 Posts	The mini split units also have temperature sensors indoors. Usually there is one sensor that reads indoor ambient temperature and one that reads heat exchanger temperature. In cooling mode, the hx sensor functions as a defrost sensor. In heating mode, it works with the ambient temperature sensor to measure indoors delta T to do other things. I T all depends on how smart the controller is.

12-19-17, 01:00 PM	#13
geoff Lurking Renovator Join Date: Dec 2017 Location: Northern BC Posts: 12 Thanks: 1 Thanked 0 Times in 0 Posts	Sorry, I have to spam to hit my fifth post and be able post a link.

12-19-17, 01:01 PM	#14
geoff Lurking Renovator Join Date: Dec 2017 Location: Northern BC Posts: 12 Thanks: 1 Thanked 0 Times in 0 Posts	Here is a great write-up using differential pressure to measure outdoor coil frosting: Heat Pump Defrost Cycle Information I think he used a 0.00" to 1.00" WC DP sensor. Acuario, you may have better luck using a single high resolution differential pressure cell. As opposed to using two atmospheric pressure sensors, and taking the difference in software.

12-25-17, 10:15 AM	#16
jeff5may Supreme EcoRenovator Join Date: Jan 2010 Location: elizabethtown, ky, USA Posts: 2,428 Thanks: 431 Thanked 619 Times in 517 Posts	Nono, some frost formation isn't a bad thing. Especially when the outside temperature is below freezing, the mere presence of frost in small to medium amounts is a good indicator that the heat exchanger is putting heat where you want it. The biggest (high SEER or HSPF) heat exchanger designs can pull full capacity up to the point where they look like they are covered in snow. So detecting the presence of frost alone is not a reliable way to control defrost. On the other hand, detecting excessive pressure drop or delta T is much more predictable and reliable. Like any heat exchanger, there has to exist a pressure and temperature difference to force heat transfer. Deciding how high is too high on either or both parameters is key to efficiency and effective performance. Only after a sufficient amount of restriction exists for a length of time should defrost be initiated.