Centrifugal blower ringing?

[Chris J]

Hi all,
I'm having an issue with an ECM blower in my central air conditioner making a ringing /oscillating noise 130-150Hz.
It's talking to me, I just don't understand what it's saying but I have a feeling someone might.

I installed this system and all associated ductwork 7 years ago. It uses an ECM motor (I believe this is a brushless variable speed motor) with a centrifugal blower and I have a Magnehelic permanently plumbed in so I know the total pressure across the system all of the time. It helps with filter changes and lets me keep an eye on things. The system runs in two speeds, low and high stages, though I cannot tell you what RPM the fan spins at, it's allegedly programmed to compensate for dirty filters and such, so I assume it adjusts it's speed until power consumption is within some spec.

When I first installed it, I found it would howl / oscillate like a tuning fork at a slightly lower airflow setting in high stage. When I increased it by 5% the noise disappeared entirely. I also seem to recall (but may be remembering wrong) that if I let the filter go a little too long, it would start oscillating at times even at the higher setting. The noise comes and goes, kind of like it's almost getting in tune, and then drifting out but the frequency never changes, just the volume. To me, it seems like the noise is RPM dependent and when you're in "The zone" it sings, otherwise it doesn't and it's completely silent.

Recently, it's started doing this both in high stage, as well as low stage which is a much lower airflow. Roughly 1200cfm in high, 800cfm in low and the behavior and frequency seem identical even though the RPM is very different. It comes and goes but the frequency of the sound is never changing. My total static pressure across the system is still normal, around 0.5" in high stage and 0.2" in low stage, it should be silent right now, but it's not. So it seems like something has changed, or is getting worse over time. I considered high humidity being a cause, as the evaporator would be wetter and possibly more restrictive, but I haven't found a correlation of that.

I haven't measured it yet, but if I had to guess I'd say it's around 130-150Hz. Best I can tell this actual noise is coming from the blower it self, not the duct work and I don't think the evaporator, but I cannot guarantee that at this time.

Any and all ideas welcome. Maybe I'm crazy and none of this means anything at all.

 

[Chris J]

Stupid question...
If it is oscillating around 130-150Hz, does that mean the wavelength is roughly 7.5-8.5 feet? And if so does that mean whatever is producing the sound should be around 4 feet long?

 

[Baluncore]

I would start by running my hands over the duct and blower mount structure. If I can feel the vibration, or if my hand can dampen it, then I could work out how to stop it.

Check the blower blades are all still there, and that they have the correct shape.

 

[Flyboy]

Also check the blower for cleanliness. If it’s dependent on rpm, I would see if there’s any dust buildup on the vanes and clean them off.

Otherwise, I’m with Baluncore… ducting related. Could be duct rumble from turbulence, or a particularly specific resonance. I know there’s a well-known example of that in the Avro Vulcan, the Vulcan Howl, which was an odd resonance between the first stage of the engine compressor and the intake ducts, so look at both sides of the ducting to the blower.

 

[Chris J]

Also check the blower for cleanliness. If it’s dependent on rpm, I would see if there’s any dust buildup on the vanes and clean them off.

Otherwise, I’m with Baluncore… ducting related. Could be duct rumble from turbulence, or a particularly specific resonance. I know there’s a well-known example of that in the Avro Vulcan, the Vulcan Howl, which was an odd resonance between the first stage of the engine compressor and the intake ducts, so look at both sides of the ducting to the blower.

So basically, I've created a musical instrument and I need to break it. Last time I checked the blower was spotless, so I doubt anything has changed with that.

Does the frequency tell me anything about what I'm looking for, size wise? I.E. 130-150Hz should be something around 4 to 4.5 foot long?

And, does the fact the frequency doesn't appear to change even though the blower rpm does suggest it's actually not related to the blower and it's squirrel cage at all?

 

[Baluncore]

Does the frequency tell me anything about what I'm looking for, size wise?

Not really. There are too many complex modes available in a practical engineering application.

If you can measure the frequency, and you know the RPM with the number of blower vanes, then you may see a harmonic relationship. But that will not help you identify the resonant structure in the air, or in the duct material.

This sort of problem is usually identified and solved by getting close to the machine. Run your hands over it, give it a hug.

If you suspect a mechanical, rather than acoustic resonance, you might shut it down, then try tapping the support structure with a rubber mallet, while listening for the resonant note. Use progressively lighter taps as you home in on the resonator. That will only get you to the size of the resonator, you must then use your hands and then fingertips to detect the critical component.

 

[Lnewqban]

What type of associated ductwork was installed 7 years ago?

 

[Chris J]

What type of associated ductwork was installed 7 years ago?

Mostly metal with short 3-5 foot runs of flex duct pulled straight and tight at the ends of each run. The ductwork right at the air handler is rectangular and made from 20 gauge galvanized steel. Everything else is insulated 26 or 28 gauge galvanized round duct. All joints were siliconed on the inside and taped on the outside, I even used rubber gaskets around the damper shafts.

Any place there needed to be a sharp turn by any flex duct has a metal adjustable elbow.

There's dampers on most supplies but they're all wide open. There's two 10" supplies,two 9" and two 8".

The only air filter is at the unit and is a merv 8 20x25 5" thick AirBear pleated filter in their housing that has gaskets and a cover.


There's also the two fittings for the Magnehelic and there's two temperature probes in the duct work, one in the return and one in the supply for each.

 

[Tom.G]

You might try closing/adjusting the inlet dampers (one at a time) to help isolate the noise source.

The same approach can be used on any outlet dampers or registers.

Pay attention to the blower speed as you do so, if it speeds up that just means that it is choked at either inlet or outlet -- that's not a problem in the short term.

Cheers,
Tom

 

[Lnewqban]

The ductwork right at the air handler is rectangular and made from 20 gauge galvanized steel.
...
There's dampers on most supplies but they're all wide open. There's two 10" supplies,two 9" and two 8".
...
There's also the two fittings for the Magnehelic and there's two temperature probes in the duct work, one in the return and one in the supply for each.

That is an impressive ductwork!

It seems that the vibration could be caused by the airflow acting on surfaces that could flutter.

Things that could have influence in 7 years: dust accumulation, corrosion, and deteriorating external insulation.

I would pay close attention to the flat sheet metal surfaces of the supply and return plenums and their insulation, to the butterfly dampers and locking mechanisms, and to the pressure and temperature probes (if slender and reaching deep inro the ducts).

The air sealing of unit-plenums-round ducts could also be failing after 7 years, changing the internal pulses of the flow.
The rigid connections of round ducts to the flat surfaces of the supply plenum could have caused cracks due to fatigue of the former.
An air leak test could be helpful, if everything else fails.

 

[Chris J]

After looking at things a bit there are some possible weak points I forgot about.

I feel most of it is sealed in ways that shouldn't fail for a long time. But there's two "canvas connectors" I made up, one for each side of the air handler which decouple it from the duct work. Since they're seamed it's possible that split open. I'll need to inspect. Also where the return side connects to the air filter I don't think I was able to get in there and tape the bottom joint. It was tight and metal the metal but certainly not perfect.

This is the 20 gauge return plenum before it was installed. Besides the connections you see theres also a 9" on the back towards the open end that goes to the air handler.

Just so there's no confusion this is the suction side of the system. The supply side is much more spread out via two 14" round ducts that go the length of the house

 

[Chris J]

Also check the blower for cleanliness. If it’s dependent on rpm, I would see if there’s any dust buildup on the vanes and clean them off.

Otherwise, I’m with Baluncore… ducting related. Could be duct rumble from turbulence, or a particularly specific resonance. I know there’s a well-known example of that in the Avro Vulcan, the Vulcan Howl, which was an odd resonance between the first stage of the engine compressor and the intake ducts, so look at both sides of the ducting to the blower.

Ok so, as much as I didn't expect this there actually was some dust buildup on the vanes. I assume from the high humidity (blower being after the evaporator) and apparently dust gets through even a merv 8 filter with foam gaskets all around it. Can't be much, because the evaporator was very clean.

Curious, why would a small amount of crud on the vanes cause the howling? Slight imbalance?

Something else I'm not sure if I mentioned previously or not, the static pressure across the return and supply would drop during the noise. It would go back up when the noise goes away. No idea if that's a clue or not, but it was there.

 

[Flyboy]

The disruption in the airflow over the vanes can set up vibrations, and if it hits a resonance it will howl. Did cleaning the dust off help?

The static pressure drop is odd, though. Not sure what to make of it.

 

[Baluncore]

Curious, why would a small amount of crud on the vanes cause the howling? Slight imbalance?

Yes.

Take a look at the dust pattern on the blades.
The dust might build up on the leading edge, at the stagnation point. That converts the leading edge from a round 'C', to a sharp '<', which can chop airflow above and below, inducing a vibration of the blade, and the fan.

 

[Tom.G]

The static pressure drop is odd, though. Not sure what to make of it.

As @Baluncore surmised, if the build up is on the fan blades, the airflow could be seriously affected / reduced. That would reduce the differential pressure.

In any case, the energy creating the noise has to come from somewhere.

Cheers,
Tom

 

[Chris J]

The disruption in the airflow over the vanes can set up vibrations, and if it hits a resonance it will howl. Did cleaning the dust off help?

The static pressure drop is odd, though. Not sure what to make of it.

I started working on everything and then got sidetracked, so I still need to finish cleaning and put everything back together. It's still cool out, so I haven't been in a huge rush.

After thinking about it, I suspect the blower wheel is as dirty as it is just because of the high humidity it's exposed to (essentially 100% RH), and this system spends most of it's life running in low stage. Am I correct that high humidity combined with low rpm would result in dust basically getting glued to it?

Here's a picture showing some of the accumulation. It's not filthy, just a lot worse than I expected. Maybe it's not even enough to matter?

 

[Flyboy]

Eeeyup. Although the dust accumulation will happen regardless of humidity, higher humidity will absolutely accelerate the issue.

 

[Lnewqban]

... I assume from the high humidity (blower being after the evaporator) and apparently dust gets through even a merv 8 filter with foam gaskets all around it. Can't be much, because the evaporator was very clean.

It is possible that the dust is leaking in though the shaft opening, or any other path into the central low-pressure zone within the blower case.

Please, see:
https://blog.airprofan.com/gas-tight-fans/

..., so I still need to finish cleaning and put everything back together.

I would suggest doing a dynamic balance at 130-150Hz (or at least static balance) of the impeller between finishing cleaning the vanes and putting everything back together.

(Edited to add link).

 

[Chris J]

It is possible that the dust is leaking in though the shaft opening, or any other path into the central low-pressure zone within the blower case.

Please, see:
https://blog.airprofan.com/gas-tight-fans/


I would suggest doing a dynamic balance at 130-150Hz (or at least static balance) of the impeller between finishing cleaning the vanes and putting everything back together.

(Edited to add link).

No, the motor fits into one side of the squirrel cage and is located in the unit. I didn't see any places I thought air could get in other than the less than perfect filter housing.

Are you saying there's a chance balancing issues could make the squirrel cage it self sing at specific speeds and always the same frequency noise?

 

[Lnewqban]

Are you saying there's a chance balancing issues could make the squirrel cage it self sing at specific speeds and always the same frequency noise?

Yes.
Please, see:
https://en.wikipedia.org/wiki/Resonance

 

[Chris J]

Between the weather, and me coming down with what I think was the flu (it wasn't fun, regardless) I finally got things cleaned up and back together but barely got to test anything.

The only thing peculiar that caught my attention was the blower wheel wasn't centered in the housing, remember it pulls from both sides. I don't know if it would ever matter, but it had almost no gap on one side, and a good 5/16"+ of a gap on the other. I'm sure it wasn't rubbing, but I don't know if the large gap on one side could do something funky as far as oscillating etc.

 

[Lnewqban]

Between the weather, and me coming down with what I think was the flu (it wasn't fun, regardless) I finally got things cleaned up and back together but barely got to test anything.

Could you please post a diagram clarifying your explanation?

 

[Chris J]

Could you please post a diagram clarifying your explanation?

I don't have a diagram handy, but here's a few pictures that show the design of the blower assembly. These are not of the actual one I'm dealing with, but it's the same style. The motor is mounted inside of one end of the wheel but with a decent gap around it so it can pull air through. The blower pulls air in both sides of the wheel. In my case, there was a fairly large gap between the wheel and housing on one side, and almost none on the other (motor side). I can't imagine the tolerances are very critical on something like this, it's just something that stood out.