Howling noise and lower temperature from Independence unit

I have a 2nd Independence unit that I keep around as a spare. I don't usually run it. However, I did run it recently as a test comparison and noticed some strange behavior. It first manifested itself in a 'Did not reach operating temperature' error in cooler conditions. This means the unit started normally and progressed through the warmup stage, but never reached the target operating temperature.

In investigating it, I noticed that shortly after ignition, the unit would start making a 'howling' noise. If I didn't do anything, it would continue to make this noise and in many cases, never reach operating temperature. I could stop this noise if I covered the plume tube exhaust with my hand, or open the top net cover. But as soon as I remove my hand or close the cover fully, it would resume. Here is a log of the thermistor temperatures of both a normal unit (MM1), and the abnormal unit (MM2):



The Independence has a 2 stage A/D measurement and the two units were started at the same time. You can see the orange line (MM2) progressed a bit faster until after they switched to the lower A/D range. At that point, it started to trail the red line (normal MM1). At 14:01:54, I opened the cover, the noise stopped, and then I closed it loosely, whereupon the sound did not return. After that point, you can see that the MM2 combustion changed and the unit started to run much hotter, at a normal temperature. I can repeat this multiple times, and even if I start the unit with the cover loose, it will still make the noise until I open it fully, and then close it loosely, after the sound begins.

Clearly there is some flow issue that is affecting the combustion, resulting in the howling noise. If I shut off the gas while it is making the sound, the sound immediately stops, so it is combustion related. I just can't figure out yet what is causing this difference in combustion. I will note that while it is making the noise, I can sometimes smell a rotten egg smell, so it is some sort of abnormal combustion. I can't figure out why changing the airflow slightly by covering the plume tube, or opening the cover is changing the process. The air flow from the air intake trumpet to the net to the fan is all unobstructed, as is the flow to the carburetor. Leaving the cover loose can only affect the air suction pressure only slightly, given the size of the input trumpet.

I know this is not a new problem, as before I modified these units, the MM2 unit used to howl sometimes as well, despite appearing to run normally. But because I didn't have visibility into the temperatures, I didn't know it was running low. I can't believe it is THAT sensitive to slight changes in airflow...

I'm still investigating, but any other experiences would be appreciated.

As I mentioned in another thread, this sort of thing happened to my Defender in 2019. See the blog entries. I changed the regulator and nozzle, which fixed the problem, so far.



Edited 1 time(s). Last edit at 07/11/2021 04:13PM by dev.

The nozzle and regulator were the first things I suspected as well, but I had actually swapped the whole regulator/solenoid/nozzle assembly into the working unit, and it behaved exactly the same as the working unit, so that eliminated that. The one thing I didn't do was to do the swap it the other way and see if there is any difference, which I may do at some later point.

The reason why I think it's to do with the airflow was due to an abnormally I noticed in the howling unit:



If you look at the left deflector circled above, you can see that it is misplaced from the proper deflector circled on the right. The deflectors mostly channel air coming from the heatsink out the sides of the unit. It doesn't seem like they should affect the air going into the carburetor air intake (black plastic item). I'm not sure how the left deflector got bent like that, but I bent it back into shape like the right side but, but the unit still does the same thing.

At this point, I may swap the whole heatsink/catalyst tube/plume assembly from the working unit into this housing and see if does the same thing. It would be weird is some slight differences in the shape of the housing alters the airflow enough to cause this problem.

I was just fighting the howling on my Independence last week, never had it before. I had just replaced the exhaust gasket and once everything was back together the unit howled when it first started, but then stopped until later that night (wife was not happy)

I took it apart a couple times, and found the same thing that you circled on those deflectors, but they didn't seem to make a difference.

When it was running and howling I flipped the head unit over and it sounded like the plume tube was where the noise was coming from. It kind of reminded me of the sound of blowing across the top of a Coke bottle. Not exactly the same but some type of airflow resonance. The plume tube seemed like it was sitting tighter in the suction funnel, so I twisted and turned it until it loosened up some and the howl went away and so far hasn't returned.

I'm not sure what caused it or fixed it. The closest I have is that the new exhaust gasket is thicker than the temporary paper one I was using and it got the plume tube out of alignment.

sevans01 wrote:

I'm not sure what caused it or fixed it. The closest I have is that the new exhaust gasket is thicker than the temporary paper one I was using and it got the plume tube out of alignment.

Woo hoo, another Independence owner! Welcome!

I have also changed gaskets before but they don't seem to have made a difference. It's a strange issue as once you can get the noise to stop (by altering airflow via messing with the cover or playing with the plume), it then performs normally until you restart it. I presume you are running an unmodified unit, and am running it 24/7. That's how I used to run mine to and that's why the howling was only occasionally heard on startup, but once stopped, it would run with no issue after that. However, now that I am running my own controller and am implementing a two cycle a day program, the howling is manifesting itself every startup and I can't be going out to play with the cover each time.

At this point, although it can be made to run properly 24/7, I want to get to the bottom of it. I think I will swap the whole active unit (solenoid, nozzle, carburetor, cat converter, heatsink, and upper plume tube) from one case to the other just to see if there is anything in the airflow of each case that makes a difference.

Yes, I'm running the stock board 24/7. When it was howling I could get it to stop, but it would come back right away. I should probably restart the unit to see if it howls on startup again.

I experimented with your esp8266 based board (which is awesome by the way) because I found a bunch of corrosion on my board and cleaned it up enough to run, but don't know how long it will last. My board turned out to be a little different it says it is a 2010 rev B, same size and out-board connections so I had the PCBs made and just ordered all of the components to build it.
I worked though enough of the issues to get it to run pretty reliably on my unit except for a couple that could be my unit (or my board construction) but 1 showstopper for now - I run cordless and my TEGs don't generate enough power to keep it running. I think I was about 250mA over budget

Thank you for posting your work on the Independence board!

sevans01 wrote:

I worked though enough of the issues to get it to run pretty reliably on my unit except for a couple that could be my unit (or my board construction) but 1 showstopper for now - I run cordless and my TEGs don't generate enough power to keep it running. I think I was about 250mA over budget

I will respond to this in the Independence thread..

How about a short posting containing a sound recording of howling vs not howling? The forum accepts mp3,mp3,m4a, and ogg.

Here's a recording I made when it just about starts. It is pulsing a bit here, but it gets more constant and louder as time goes on. In the middle of the recording, I covered the plume exhaust and it stopped the sound for a bit.

howling1.m4a

What I'm finding is that if I don't do anything, it will howl till it reaches an steady temperature which is less than normal. During this time, I can try all kinds of things such as opening the cover, removing the net, covering the exhaust, and while I can sometimes stop the howling, when I close the cover, it starts up again. But once it reaches a relatively high, but not normal temperature, I have a better chance of opening the cover, stopping the howling, and then I can close the cover and the howling doesn't resume. At that point, it starts to get hotter and it will then reach normal operating temperature. But every time I restart it, I have to go through this gyration....



Edited 1 time(s). Last edit at 07/19/2021 02:26PM by Independence.

I listened to your recording, and was struck by how regular the sound was, both in pitch, and amplitude modulation. It is far different from the irregular, turbulent "wind" sound produced by the Defender. Your sound reminds me of a rolling 'R' trill. The Bandmate app displays two pitches around concert D (but off tune).

I hear two separate frequencies, one for the high tone, and the other for the modulation. The only thing that I can think of that might produce the pure tone high sound is the fan, with the low pulsing modulation caused by some sort of wobble in the fan bearings. How about lowering the fan speed and seeing (hearing) if the high tone decreases in pitch? If it is the fan, these things generally can have the dust cap removed, and re-lubricated with motor oil. Clean the bearings if accessible. I have done this on several very irritating noisy notebook computer fans with good success.

If it is not the fan, then this is a really interesting phenomenon. Another thing to consider is the nozzle or valve, but I believe you have ruled that out. I would be surprised if it were the regulator because I would expect the length of hose smooth the flow and attenuate the sound.

Uploading the A/V file types mp3 mp4 m4a ogg should be working now.

I did suspect the fan as well, but then I had swapped the whole combustion assembly into another quiet housing with a different fan, and it did exactly the same thing, so that would seem to rule it out. I also measured the fan rpm and it was within spec. I even went to the trouble of ordering a replacement fan, and it measured out exactly the same. I also did vary the fan rpm while the sound was present and while it sometimes varied the pitch a bit, it wasn't able to stop it.

At this point, I believe it's a combination of the combustion process and fan airflow through the system that is causing some kind of resonance. The second I shut off the gas solenoid, the sound stops. I had cleaned the inside of the catalytic chamber and it didn't make any difference. Opening the cover or covering the exhaust plume also varies the input air pressure to the carburetor, probably affecting the combustion process as well.

I haven't done any more experimentation as I currently have the unit running cordlessly 24/7 in it's quiet working state at the edge of my property. We've been having very wet weather this summer in the North East, and it is catching THOUSANDS of mosquitoes. I haven't checked it in two days but I suspect the bag will be full when I get a chance to check it.

...catching THOUSANDS of mosquitoes...

I am happy to hear that this whine or howl is not affecting your catch rates. Come to think of it, your 299 Hz D 4 tone might act as a super mating call attractant. At least they are not being repelled. This frequency, however, is lower than that of a typical mosquito's wing beat. Perhaps the trap's horn is 'tuned' to this frequency by design, although this is not mentioned in the patents.

For reference, this site reports:

Mankin reported that the wingbeat frequency produced by the flight tone of mosquitoes varied on the basis of genus and gender. Wingbeat frequencies of Ae. aegypti and Ae. albopictus fall within the range of 400-500 Hz in females and 700-900 Hz in males.

Another site shows 350–664 Hz, closer. A site with a mosquito recording sounds like G# 4 about 406 Hz. There are other recordings at other frequencies.

Dev wrote:

I am happy to hear that this whine or howl is not affecting your catch rates.

Ah, no, I don't let it run with the howl. It DOES eventually reach operating temps, especially in warmer weather, but it's not as hot as the normal unit. I play around with leaving the cover open at lesser angles, with no howl, as it warms up, until I can get the temperature above 165C, at which point I can fully close the cover and it doesn't howl. At that point, it is running like a normal quiet unit. This is fine for a cordless unit that runs continuously, but would be annoying for my AC unit that powers up twice a day.

I braved the back woods and retrieved the unit today. Check out it's 5 day haul!: