10-20-12, 12:54 AM | #11 |
Apprentice EcoRenovator
Join Date: May 2011
Location: Tortosa, Spain
Posts: 221
Thanks: 2
Thanked 81 Times in 46 Posts
|
I know it shouldn't affect the operation as it's IR, but have you tested in dark conditions? Sunlight (as you stated) will probably affect it.
Acuario |
10-20-12, 06:23 AM | #12 |
Master EcoRenovator
Join Date: Aug 2012
Location: Toronto
Posts: 958
Thanks: 40
Thanked 158 Times in 150 Posts
|
Often the sensor is placed at the "air on" position as this is where the coldest refrigerant will be and frost starts there and if the two halves of the unit sense the full width of the fin, (from the air on to air off side) and very close to it, it should be able to sense 1mm of frost before it becomes ice and as it is building up over perhaps 100-150mm, I cannot see how it can see through that length of frost as though it was clear.
|
10-20-12, 06:28 AM | #13 |
Master EcoRenovator
Join Date: Aug 2012
Location: Toronto
Posts: 958
Thanks: 40
Thanked 158 Times in 150 Posts
|
The Sanyo is definitely in the open, where my coil is entirely within a cabinet and only subject to reflected light, if at all. I think the tuning can be made to avoid it but it needs to be quite precise which is why 2 parallel systems might work. Temp measuring to determine turn on and IR to determine when it has defrosted, for example.
|
10-20-12, 10:22 AM | #14 |
Lex Parsimoniae
Join Date: Feb 2009
Location: Woburn, MA
Posts: 4,918
Thanks: 114
Thanked 250 Times in 230 Posts
|
tested in dark conditions and under hazy sunlight
Hooked up the voltmeter to the signal output and found there wasn't very
much interference from hazy sunlight directly on the sensor. I believe the sensor lens would have to be aimed directly at the sun, before any big changes in output would be seen. Off-angle sun light and direct light from my 1W LED flashlight had very little effect, when compared to the large changes seen when operating in Bounce mode. (using a white paper reflector). Notice the IR glow of the LED.. My camera can see IR!! The no-reflect voltage of 4.5v is a really good TTL '1'. (Even covered, it never when up to 5V. My power supply might be a tad low). From my white-paper measurements, (simulated frost) it seems like 0.4 to 0.5" from the HX fins would be prefect for detecting frost at TTL levels. It sure looks like no Analog measurement would be needed. But, I would use a Schmitt trigger TTL input chip, to provide noise immunity. 10pcs 74HC14 7414 74HC14N Hex Inverter with Schmitt Trigger DIP 14 | eBay IMHO, water droplets hanging off the lens could cause a false alarm. So, a rain shield would be a good idea. ~~~ If ice forms on the lens pair, it's very likely that IR light will be coupled over to the P-transistor and give you a defrost alarm. Due to the direction of air flow, past the sensor and into the HX fins, Sensor icing could be a problem. The sensor might still be signalling ice on the HX, after a good defrost. Or, maybe the LED will warm up the sensor body and melt off the ice..?. For sure, Real world testing will be needed to tweak the setup.. The sensor might need it's own little defrost heater (100 ohm 1/4w resistor?).
__________________
My hobby is installing & trying to repair mini-splits EPA 608 Type 1 Technician Certification ~ 5 lbs or less.. |
10-20-12, 10:45 AM | #15 | |
Lex Parsimoniae
Join Date: Feb 2009
Location: Woburn, MA
Posts: 4,918
Thanks: 114
Thanked 250 Times in 230 Posts
|
950 nm
Quote:
That Sanyo is in the open, but it's installed next to a wall and there is no way that direct sunlight could get back there and into the sensor. Sanyo #2 is on the north side of the house and it's even darker behind it. From what I can tell about indirect sunlight, so far, the day and night difference in signal voltage levels will be very slight. A fully inclosed HX like yours is the idea solution to solar pollution.. The sensor spec says "• Daylight blocking filter • Emitter wavelength: 950 nm" 950 nm isn't in the power band of the sun.. It looks like visible light is where the peak is.. Edit: Found this doc.. http://www.ro.feri.uni-mb.si/predmet...pplication.pdf
__________________
My hobby is installing & trying to repair mini-splits EPA 608 Type 1 Technician Certification ~ 5 lbs or less.. Last edited by Xringer; 10-20-12 at 11:43 AM.. Reason: pdf link |
|
11-25-12, 04:24 AM | #16 |
Apprentice EcoRenovator
Join Date: Mar 2011
Location: SE Wisconin
Posts: 107
Thanks: 3
Thanked 11 Times in 9 Posts
|
Aquario, any update on the defrost control now that it's gotten a little colder, at least here in Wisconsin? My biggest need right now for my Hot Tub Heat Pump(HTHP) is a reliable defrost setup. I'm currently testing a Dehumidifier control board for my defrost. It's got a thermistor that gets mounted on the vapor suction line just before the compressor and will shut it off when it gets too cold for too long.
I tested it today on the line just after the cap tube where it frosts up first and it did indeed shut down the compressor after a few minutes, but than it stayed off for a half hour until I hit the switch that determined if the bucket was in place to simulate taking it out and putting it back in. I think it's getting cold enough to shut the compressor off, but in 33* weather, maybe not getting warm enough again to bring it back on. After all this was designed for a basement in the 50 or 60* range. I'm going to try it with the sensor closer to the compressor like intended and see if that makes a difference. I have a hunch if I insulate the line from the compressor to past the thermistor, the heat of the compressor will warm the line enough during the defrost to let it come back on. My A/C doesn't have a reversing valve btw so it's just compressor off, fan on, defrost. What do you all think? Will it work or should I pursue a new direction? I'm thinking if I can't get the thermistor to bring it on reliably I may build a simple timer that essentialls simulates the bucket being removed and put back, bringing the compressor back on, after a set amount of time, maybe 5-10 minutes. The other option is to use a FischerTechnik interface, thermistors, and relay, design my own demand defrost system, and make a program for it in RoboPro. That would be a fun challenge. Adam Adam |
11-25-12, 05:03 AM | #17 |
Apprentice EcoRenovator
Join Date: Mar 2011
Location: SE Wisconin
Posts: 107
Thanks: 3
Thanked 11 Times in 9 Posts
|
Some Googling found me this: Patent EP1510768A1 - Dehumidifier with a defrost control system - Google Patents
It's a patent file for a dehumidifier demand defrost. I think I could implement this with the Fischertechnik parts. It would solve my issue of the set temperature value system I believe my current Dehu Defrost has. I think it just senses coil is below 32*F, waits a set amount of time, and defrosts for a minimum set time and until the sensor is above 32*F. The obvious flaw being I can't run my HTHP in <32*F which is a good portion of the winter, when I want it most. Adam |
11-25-12, 06:17 AM | #18 |
Master EcoRenovator
Join Date: Aug 2012
Location: Toronto
Posts: 958
Thanks: 40
Thanked 158 Times in 150 Posts
|
Adam,
What you need is a reversing valve or easier still, a hot gas bypass defrost system. Basically, all it does is take the hot gasses straight out of the condenser and puts it directly into evap bypassing whatever regulating device you have (TXV or Cap tube). It only needs a solenoid and a sensing system with a timer. Last edited by Mikesolar; 11-25-12 at 06:25 AM.. |
11-25-12, 10:01 AM | #19 |
Apprentice EcoRenovator
Join Date: May 2011
Location: Tortosa, Spain
Posts: 221
Thanks: 2
Thanked 81 Times in 46 Posts
|
No it has not been needed yet - we still haven't got down to those temperatures yet where it freezes up -about the lowest we've had so far this autumn has been 7 degrees.
This coming week the temperature is due to drop to zero so I'll know more then. |
11-25-12, 10:45 AM | #20 |
Supreme EcoRenovator
|
Adam,
If you are using this unit without a reversing valve and at temps approaching freezing, you will have to provide the evaporator coil with some form of heat to thaw it out when it freezes. It sounds like your dehumidifier control is doing its job, but without any added heat the thermistor never gets warm enough to reset the control. There are a few users experiencing this bug right now, namely hv23t and ecomodded, on their hacked heaters. If you are indeed using a dehumidifier, the evap fan runs in defrost mode while the cxr does not. In this case, you can add your choice of resistive heat element and a 120V relay with the coil wired between the fan and the cxr hot wires to sense when your unit is in defrost. The relay contacts will operate the heater element and warm the evap coil. Then when the frost has melted, the thermistor will sense the warm coil and terminate defrost. This seems like the simplest, cheapest way to accomplish defrost in your situation given the info you have provided. |
|
|