Arduino - Digital AC control system for '80 911

[kuehl]

Well, ideally, if you pull a perfect vacuum, at any reasonable temperature, say 70-100F, and toss in, maybe 36-39 oz in your case, you will be pretty close. However, the warmer it is it is easier to pull that perfect vacuum, then again when its warm the vacuum pump suffers too.

A narrow hysteresis might cause your compressor to develop a hysteria.
Since you have an 'electronic' thermostat control you play around with the low temp cut off and drive it down quite low on dry days; we have gone down in the 9F before. The key is to tell when the evap core is icing up. I imagine you could watch the amp load on the evap motor, when it starts to require less juice that will be a tell tale sign, or measure its rpms ; if it winds up you know you don't have mass as well. Might as well plug in a cheap camera and watch the frost action on the core.

[Discseven]

Originally Posted by kuehl

...play around with the low temp cut off and drive it down quite low on dry days; we have gone down in the 9F before.

Curiosity concerning TXV... If you drove evap to 9F, did that include 'playing' with TXV? Or does TXV just keep making evap colder and colder provided refrigerant is provided it?

Originally Posted by kuehl

The key is to tell when the evap core is icing up. I imagine you could watch the amp load on the evap motor, when it starts to require less juice that will be a tell tale sign, or measure its rpms ; if it winds up you know you don't have mass as well. Might as well plug in a cheap camera and watch the frost action on the core.

Camera... excellent idea. And simple (which am not used to.) Will admit to being out of ideas at this stage in AC's evolution. Mind has gone into 'completion mode' ...as in get AC working and screw any and all new AC ideas. Exception in the making...



USB Borescope Endoscope for Android - Bamshoot (New Upgraded) Endoscope Inspection Camera 2.0 Megapixels CMOS HD Waterproof IP67 Snake Camera with USB Adpater and 6 Adjustable Led Light - 16.5FT

Several similar units like this on Amazon. Relatively inexpensive.

FYI---Way Arduino is currently configured, 'evap minimum temp' and 'hysteresis' are not adjustable on-the-fly. Must adjust program on laptop's software, then access Arduino in 911's AC control box and load new program to it via USB connection. Arduino's program does already collect humidity data. Now, for 'automated variable humidity control', would need to know how humidity affects evap's performance parameters. Collecting that info is where we're headed. Then need to know how to program working parameters in Sketch language---that's Dr J's expertise. (Are you reading this John?!) Humidity-based AC control takes 911 AC operation to another level. Insanity here is seductively amusing.

Not to get lost in another lengthy project detour---as am prone to---"variable humidity control" is a curious adventure that's set on idea-shelf for time being. Camera idea is simple... and so immediately doable. Compliments to Charliemon for yet another fine brainstorm.

.

[kuehl]

Curiosity concerning TXV... If you drove evap to 9F, did that include 'playing' with TXV? Or does TXV just keep making evap colder and colder provided refrigerant is provided it?

One of the most common mistakes you can have is playing around with the TEV's 'adjustment screw'.
So many DIY's read something on the internet about someone who did this and claims the major improvements
they got were from making an adjustment.
Any improvement they got usually comes from either having a better vacuum on the system or better charge.
I have plenty of notes from a Parker Engineer I use to chat with frequently and this fact.

Simply, the "adjustment screw" is to adjust the super heat, this is done at the factory
where the valve if manufactured; to match the refrigerant and evaporator's requirements.

Super heat, for a simple explanation,
is a value at the evaporator outlet required to insure all liquid refrigerant
vaporizes rather than having liquid return to the compressor.
Super heat values of expansion valves vary based upon the type of refrigerant,
the tonnage rating of the valve, the size of the evaporator and other factors.

The way we got it down to 9F, other than we had excellent condenser capacity, was to not allow the thermostat to shut off the compressor.

[Dr J]

Originally Posted by Discseven

Collecting that info is where we're headed. Then need to know how to program working parameters in Sketch language---that's Dr J's expertise. (Are you reading this John?!) Humidity-based AC control takes 911 AC operation to another level. Insanity here is seductively amusing.

.

No problemo Karl! If we get a function spitting out evap temp for a given humidity or an x-y table of a few data points it would be very easy to implement.

[Discseven]

Originally Posted by kuehl

...way we got it down to 9F, other than we had excellent condenser capacity, was to not allow the thermostat to shut off the compressor.

...There is confirmation of a crucible---is point of this project's getting around stock thermo control.

Originally Posted by Dr J

No problemo Karl! If we get a function spitting out evap temp for a given humidity or an x-y table of a few data points it would be very easy to implement.

Your relentless partnership in this experimental endeavor is more than admirable John. Camera is ordered to look in on icing. Believe a year's worth of data is necessary to make sense of how humidity affects evap during various seasonal changes. I'll edit 'evap min temp' and 'hysteresis' monthly to push into (what I consider) evap's "ice boundary levels" (per month.) Results will get logged. Make sense of that data at year's end. Then provide you data points. (We're creeping up on complexity of your Ard-driven-EFI. I'd like to see Sketch for that someday.)

FYI---DHT22 temp & humidity sensors are no good for exterior/ambient environments. Have tested a few units "outside." Each unit tested did same---each is unable to collect variable data after experiencing 100% humidity---data displayed for that unit on Nextion after 100% humidity experience... frozen at "99%." Can correct this fail by wiping ity-bity humidity sensor pad off---not appealing.

...Cabin DHT22 has not failed to date. Doubt this unit will ever experience 100% humidity. Since evap is moving cabin-humid air, don't think ambient humidity matters. Say so if you think otherwise.

Next year, for sake of eliminating exterior sensor's data-inability, replacing DHT22 with SHT10 would be of interest. https://www.adafruit.com/product/1298 If this is asking too much, say so.

.

[Discseven]

Iyzer endoscope. 2MP camera. Rechargeable battery---by way of USB cable. Camera's supposedly water proof to IP67 standard (submersion up to 3' for 30 minutes.) Image taking options = still & vid. Adjustable resolution: 320x240, 640x480, 1280x720. Semi-rigid cable = 11.5' length. Amazon purchase... price range for exact same product: $19.99---$69.99. Product received (including box and printed instruction sheet) displays no maker's name, website, or ANY contact info. Zero product support anticipated.




Accessories: mirror, hook, magnet. Hook and magnet don't make much sense as these screw on around camera's head. (Maybe shafts bend?) Mirror works well.




LEDs for power ON and wifi CONNECTED. ...no info provided for battery status. Supposedly battery's good for 40 minutes.








Optimal focal range as I tested it = 2.5” > 4” (product's specs say differently.) Not a clear image outside noted range. LED lights around camera---can adjust brightness with scroll-dial on wifi module---a gotta-have control.




Looking up to bottom of evap. Would prefer to see more evap---wider angle---but cannot get camera any further away from bottom of evap than it is here (unless running it up through plenum's drain---not happening. Cable has been passed through resistor/intake---camera stays in plenum. For $20, is interesting tool to have. $70... not---probably get better quality than this for $70.

.

[Discseven]

Temp and humidity sensors: SHT10 / DHT 21 (top… left / right.) DHT is good, inexpensive sensor for relatively dry spaces---in-cabin for instance. Outside of cabin it’s useless---humidity sensing fails after unit experiences 100% humidity. Temp continues to work. 3 units were tested with same result. DHT in driver’s side door jamb is being replaced with what’s hoped to be a more functional sensor… the SHT...



Azzumed this was going to be plug & play swap. Nope. DHT is 3-wire. SHT is 4. Arduino's program needs to be modified and as does wiring, junctions, and distribution board. Planned attending this well down stream but “DR. J” / John did a light-speed move attending Arduino’s reprogramming. Once John green-lighted SHT... I started dealing with restructuring hardware (according to John’s directions.) Thank you for your blinding speed here John! Much better to see displayed data correct (than otherwise.) Also... very impressed with your record keeping on system---couldn’t ask for a better collaborator here.




Pins 6 & 7 (on Arduino UNO) that were vacant are now used for SHT. Old DHT wires stand idle. 3-wire cable run through car for DHT… that also stands idle. New 4-wire cable is made & run.




Terminal is added on distribution board for SHT. “Data In” comes from Arduino’s Pin 7. “Data Out” goes to sensor along with: ground, power, and clock (yellow wire from Pin 6.) 10k Ohm resistor is added between data and power circuits. I believe John has benched this… I have yet to proof config as one junction pends.




SHT is mounted where DHT was. Velcro is fastener. Cable's fed through stock grommet hole in door jamb into frunk. As I post this... am looking at a few hours to attend a variety of AC refinements today---pending junction included.

.

[Dr J]

Yep, I benched it. It works!

[Discseven]

... Figured you had John. Testing continues... looking at changes in displayed numbers to see/feel if numbers make sense given car's environment. What's sought is if & how sensors are working. (AC is not yet charged.)


Sunny. Mid afternoon. City driving with windows open. Rained earlier in day.


Early morning. Car's been parked in garage overnight. Rain in distance---offshore.


Afternoon. Car's been parked in full sun but is slightly overcast. Just started driving. One window open.

.

[Jonny H]

It’s quite difficult to place ambient air sensors for accurate readings. You need air flow over the sensor and to keep it away from panels that radiate heat and direct sunlight obviously. Any metal bracket will also conduct heat to the sensor giving a false high reading.

I position mine under the fender in the inlet air to the condenser but not too close to it. The mount is a plastic tube, a few inches long to get the sensor away from the panels.

[Discseven]

Originally Posted by Jonny H

It’s quite difficult to place ambient air sensors for accurate readings. You need air flow over the sensor and to keep it away from panels that radiate heat and direct sunlight obviously. Any metal bracket will also conduct heat to the sensor giving a false high reading.

I position mine under the fender in the inlet air to the condenser but not too close to it. The mount is a plastic tube, a few inches long to get the sensor away from the panels.

Agree with you Jonny. That's fine location you thought of. Am tempted to reposition there. Curious to know what model sensor do you use?

[Jonny H]

Originally Posted by Discseven

Agree with you Jonny. That's fine location you thought of. Am tempted to reposition there. Curious to know what model sensor do you use?

10k NTC thermistor. Simple and reliable but output does need to be linearised. Also, the current through the NTC can cause it to self heat, so it can’t just be left ‘on’. We use a dedicated thermistor IC to manage this. Temperature only, no humidity

[Discseven]

Originally Posted by Jonny H

10k NTC thermistor. Simple and reliable but output does need to be linearised. Also, the current through the NTC can cause it to self heat, so it can’t just be left ‘on’. We use a dedicated thermistor IC to manage this. Temperature only, no humidity

Going to look into that model. Thanks for notes on it Jonny. Thought about putting ambient in "air stream" and have second thoughts. Am thinking SHT10---way it's designed---could suffer a clogging of the mesh that encapsulates the sensors if placed in an air stream. Going to leave it in door jamb.

[Discseven]

Charged yesterday. Am traveling next 2 weeks. Will post results soon after returning.

[Discseven]

Done. Results are here: AC pressure & temp results from Arduino Digital Control and Air-Bucket builds