Engine air dehumidifier

[0TreeLemur]

I write to share what I think are results of a pretty cool experiment aimed at preventing engine corrosion in our grounded M20C.

Our beloved C is grounded pending piston pin plug replacement on one cylinder.  Since buying this aircraft in 2017 I've always flown her for at least an hour every two weeks or so, especially during the steamy parts of the year to keep the oil film on the engine parts and help prevent internal corrosion.  Earlier this year increasing amounts of aluminum in the oil filter suggested a possible loose piston pin plug, which was confirmed by endoscopic examination, so we stopped flying in June.  Since it is the steamy part of the year here in the deep south U.S., I immediately became concerned about preventing engine corrosion.

Searching the internet turned up the most common solution of pumping air through a desiccant and feeding that air into the breather tube.   My experience with desiccants in humid environments is that their effectiveness is pretty short lived.   And of course there is Camguard^(TM) and the like, but I've never been too trusting of magic potions.  Call me skeptical.  I also see that there is a commercial electric engine air dehumidifier but I have the parts and expertise to design/test a frostless dehumidifier system and have confidence in how it works.

My engineer brain kicked in, and I prototyped a closed dehumidifier using a Peltier effect device as an aftercooler.  After experimenting with a few configurations, I found a set up that works really well.  It consists of a Peltier cooler in a sealed chamber that also contains a piezoelectric positive displacement aquarium air pump that is ridiculously simple and rated at about 10 l/h (~3 gph) flow rate.  The intake air to this chamber is filtered using a Fram G-2 just like the one used on Brittain PC vacuum systems.   I put an el-cheapo digital temp/r.h. indicator in the cooling chamber just to indicate what's happening in there.   I epoxied a type-K thermocouple to the fins of heat sink on the cool side of the Peltier to monitor its temperature.  The dehumidified air then passes from the air pump through a hose barb bulkhead fitting and flows into a second measurement chamber that contains a research-grade (Vaisala HMP50)  temperature/relative humidity sensor.  I also put another el-cheapo digital temp/r.h. sensor in there as an independent check.  Testing revealed that the el-cheapo R.H. sensors are qualitatively correct, but not particularly accurate.  If they indicate <40% r.h. the dew point is less than 10C at hangar air temps.

The air leaves the measurement chamber through a 3/4" hose and flows into the engine breather tube on the aircraft.  I sealed off the "ice hole" on the breather tube and unscrewed the dipstick.

The slickest part of this system is the programmable controller/data logger.  It cycles the Peltier on/off to keep the thermocouple temperature in a range of 11-13C. This prevents icing.  I wrote an iterative solver to calculate the dew point temperature from the research-grade r.h. sensor data.  If the Peltier temperature exceeds 15C, or if the dew point temperature of the system output air exceeds 15C, the system shuts off the air pump and Peltier, and switches on a red LED on the panel.  This puts the system in a safe condition and prevents a failed component from leading to pumping humid air through the engine.

That system is now pumping dehumidified air through the engine on a continuous basis, waiting for our date with the repair shop to get 'er fixed.  The temperature of the output air is actually a bit warmer than the hangar air because the Peltier doesn't remove much sensible heat, mostly latent heat.  Here is a photo of the system in operation, and a data plot showing:

(1)  horrendous high dew point (humidity) of the ambient air this time of year in Alabama (from ASOS),

(2) ridiculous high mid-day hangar air temperatures,

(3) nightly low ambient air temperature approaching the ambient air dew point, and (from ASOS),

(4) the greatly reduced dew point of the air going into the engine, reduced from 22-24C to below 10C!

I over-instrumented the heck out of this thing.  The total cost of this system if all parts were bought new would be ridiculous, but I used all surplus measurement/control gear and re-purposed a computer case to serve as a chassis.  Oh, and it uses only about 30W of electrical power.  A sniff of the air coming out of the dipstick tube confirms flow through the system.  A rag around the dipstick opening keeps the bugs out.  These things should be common in hyper-humid climates.

Engineering rulez.

 

 

 

[carusoam]

Go engineers!

-a-

[larryb]

Good job! I have had the exact project in mind for a few years. But I never did it because I found I could keep the RH low enough by keeping the engine at 90 degrees F all the time with the pre heater running a PID PWM algorithm. I may still do it though, just because.

[0TreeLemur]

25 minutes ago, larryb said:

Good job! I have had the exact project in mind for a few years. But I never did it because I found I could keep the RH low enough by keeping the engine at 90 degrees F all the time with the pre heater running a PID PWM algorithm. I may still do it though, just because.

The way you are doing it is pretty darn reliable.  A heater did not occur to me.  I guess probably because our hangar gets so stinkin' hot during the high-humidity time of year.   Looking at the temperature data above, our hangar stays warm all night too, so the risk of condensation is pretty low, unless the air in the crankcase has a dewpoint above the nightly low engine temperature, which I don't know.  I like this solution.   I'm keeping an eye on it to assess reliability.

[carusoam]

The challenge with sump heaters.... 

There can be a part of the engine that isn’t getting heated.... one end is warm and dry... the other end is experiencing cold and condensation...

So... lots of instrumentation to know it isn’t raining at the far ends...
and possibly, some insulation...

[larryb]

For me it was an accidental discovery. I instrumented for RH and found I could control it with temperature. And that makes sense since the warmer air holds the moisture in suspension. And if the moisture is suspended in the air it is not available to cause corrosion on the surface of the metal.

But I live in a dry area. I would expect in a more humid environment one would need a more active solution.

[squeaky.stow]

Looks like a great idea! I had been looking at building a desiccant system but was not keen on the idea of having to bake the beads periodically. Is this design something you would be willing to share for the DIY-but-not-engineer types amongst us? I am thinking that if you have already proven the concept, then the instrumentation side of it would not be as necessary. Just the operational parts.

[jetdriven]

Engine preheated warms the engine, and the RH is lower, but the total moisture holding capacity of the air is higher when it’s warmer.  More water, more corrosion.  
blackmax.  500$.  Plug and chug and it’s dry all the time.  No I am not conpensated for saying this. 

[0TreeLemur]

On 7/18/2020 at 9:11 PM, squeaky.stow said:

Looks like a great idea! I had been looking at building a desiccant system but was not keen on the idea of having to bake the beads periodically. Is this design something you would be willing to share for the DIY-but-not-engineer types amongst us? I am thinking that if you have already proven the concept, then the instrumentation side of it would not be as necessary. Just the operational parts.

Sorry @squeaky.stowI hadn't checked this thread in a while.

I'd be happy to share, but I used a Campbell Scientific data logger for the measurement and control system.  I had that on hand, which would cost you about $1500 to buy outright.   It could be done cheaper using a Raspberry Pi or maybe even an Arduino with cheaper sensors.  The blackmax system mentioned by @jetdriven is definitely cheaper than my solution. 

My solution uses a Peltier CPU cooler in a plenum with an piezoelectric fish tank air pump rated at 10 liters/h.   A type K thermocouple epoxied on the fin of the cool-side heat sink measures that temperature.  The fish tank air pump pumps air from the cooling plenum into another plenum that contains a research grade ($400) temperature/r.h. sensor as a qa/qc check on the cooler.   The logic on the controller is the part I'm really pleased with.   Some calibration was required to keep it from icing up.  The Peltier is switched on any time the thermocouple temperature exceeds 15C, and it switches it off when the thermocouple hits 13C.  It cycles off/on about once every 2-3 minutes using a digital I/O driven relay board.    I have some logic in there for start-up because it takes a 3-4 minutes for the calculated Td in the second plenum to get below 15C.  After that startup period, if Td>15 C, it shuts off the system because something is wrong.   That prevents blowing humid air through the engine.

Data in the plot above show that this thing consistently discharges air with Td<10C, which is pretty good for where I live.   Let me know if you have any questions.

BTW- the shop removed the bad cylinder where the nitride coating was coming off and "cheese-grating" one of the piston pin plugs.   Photos below show the good one and the bad one.   Produces a lot of aluminum that shows up in the oil filter.

-Fred

The good & the bad.

 

[carusoam]

Great piston pin pics, Fred!

Thanks for sharing them.

Best regards,

-a-

[DXB]

On 7/18/2020 at 10:25 PM, jetdriven said:

Engine preheated warms the engine, and the RH is lower, but the total moisture holding capacity of the air is higher when it’s warmer.  More water, more corrosion.  
blackmax.  500$.  Plug and chug and it’s dry all the time.  No I am not conpensated for saying this. 

Not sure that's exactly true - it's not the total moisture content of the air that determines corrosion risk but rather the RH, which determines risk of condensation on the metal.  Thus heating the engine uniformly well above ambient temp should be effective.  There is some concern with the oil film coming off more quickly if the engine is hot all the time, offsetting benefit from reduced RH. Using a dehumidifier doesn't come with that  potential issue.