Designing a Quiet Air Compressor

[vederstein]

Now that the Dancer Engine is complete, I'm starting to think about my next big project: I'm going to attempt to build a quiet air compressor. I have some ideas, but first of all, one of the goals is to minimize the air pulses (chuffing) coming out of the pump.

Note that this is pure engineering at this point. Once I have some spreadsheet models to help guide me in the right direction, I'll get into the design process. At this stage I'm only doing an engineering analysis.

I've created three spreadsheets, modelling an air pump with 2, 3, and 4 cylinders. This analysis is just to see the optimal number of cylinders to produce an even airflow without an excessive number of cylinder. I modeled the output with some flow cutoff ±15° to ensure no crossing of ports between the piston inlet and outlet.

Let's take two cylinders, 180° out of phase. The air pulses are very pronounced. The average flow is 0.60, but the range goes from zero up to one.






Now let's look at three cylinders, 120° out of phase. The average flow is 0.92 with a pulse range of .29:





Now let's look at four cylinders with a 90° phase angle. The average flow is 1.22 with a pulse range of .45:





Given this analysis, I'm only looking for the smoothest airflow with the lease number of cylinders. Based on this, my air compressor pump will have three cylinders. I suspected this to be the answer, but now I have the analysis to back it up.

This is very early in the design process. This isn't really even design yet. It's actually engineering. This is what mechanical engineering looks like.

That's enough for now.

...Ved.

 

[johnmcc69]

Nice! Ambitious project!

Like you said, it's all about the engineering.

I'm following...good luck!

john

 

[vederstein]

Next up in the design process is to figure out what I can expect from a motor selection. I want the pump to have the ability for an output pressure of 120psi.

I want to use a commonly available compressor duty motor that runs at a relatively low speed. Also, I don't want to exceed most US single phase circuits. That leaves me with a 1½hp motor @ 1725 rpm.

At that speed and power, the average motor torque is 54.8 in-lb.

Knowing there's going to be frictional and other efficiency losses, I assumed a 15% efficiency loss. Also in every slider crank, the worst case position is when the connecting rod is 90 degrees to the crankshaft:



With these data points, I can play with the bore and calculate the stroke to match up with the tooling I have available and what seems reasonable to me in a design. The design parameters are as follows:

Number of Cylinders: 3
Cylinder Bore: .625"
Cylinder Stroke: 1"

Each cylinder stroke will have a compression ratio of 8.3:1.

Calculating Pump CFM is not straight forward because the hardware store's air compressors specifications are complete bullshit. On a 120V/1ph/60Hz/15amp house circuit you can get a maximum of 1,800 watts which equals 2.4 hp. Go to the hardware store. You'll find ratings that just don't make any sense. How the hell can a compressor running off 120V house power be 4 horsepower!?!? It's total marketing bullshit.

Anyways, using the ideal gas law (Pv=nRT) I calculated that the pump will output .91 cfm @ 120 psi. Another way to think of it is that the pump will compress .332 moles of air from atmospheric pressure to 120 psi per minute.

This calculation results in an air inlet consumption of 8.1cfm.

If you've made it this far. I thank you. I know this post is a bit dry. No fancy renderings or videos, just a bunch of numbers. But again. This is what engineering is.

Attached is the working spreadsheet (OpenOffice .ods format) with all these calculations. If you study it, let me know if I've screwed up somewhere. But the numbers seem reasonable and that's one way to check if the calcs are correct.

Thanks,

...Ved.

 

[ddmckee54]

Ved:

Yup, it's marketing BS alright. I believe that in order to get the most IMPRESSIVE numbers they'll use the motor's current values at startup to calculate the HP. A motor typically pulls 6 times its' normal full load amps during the split second that it's not turning. Now I ask ya, doesn't 4Hp sound a LOT more impressive than the 2/3Hp it actually is?

The big players in air compressors are not above that game either. Several years ago, we installed a 4000 CFM compressor, not gonna name names. The compressor had an 845Hp, 480V 3 phase motor, with a FLA of 900 and a service factor of 1.15. We finally got it running, and that was a struggle in itself. The motor, if started across the line, would pull 5400 amps for about 25 seconds. This would trip the main breakers in the sub-station. We about crapped when we measured the motor's running current at 1050 amps. We thought we had a brand new, bum motor so we called a motor company to come check it out. It was just hunkey-dorree, but when we put it all back together it was still pulling 1050 amps.

I got to looking at the performance curves of the motor/compressor and found out that we were spot on the curves. In order to get the 4000 CFM at our ambient temperature, our motor had to crank out 900Hp or about 108% of its' rated value. This was still within the motor's 1.15 service factor, so the manufacturer was happy. They could SAY that their compressor would crank out 4000 CFM using an 845 Hp motor - that was ALL true. It's just that the motor wasn't loaded to 845Hp when the 4000 CFM was produced. We've rebuilt that motor 2-3 times since we installed it.

Nope, not JUST the hardware store guys.

Don

 

[Toymaker]

As you noted in your first post, the more cylinders you have, the smoother the output airflow will be. A secondary benefit of lots of smaller cylinders is lower noise level as it becomes harder to hear individual cylinder pulses. With that in mind, I built a small home-shop air compressor using an automotive Air Conditioner (AC) compressor which uses a wobble plate to drive 10 cylinders. Here's the thread: Shop Air Compressor using Auto AC Compressor
I use a small 1/3 HP motor with a couple a reduction pulleys and the system is very quiet; what noise you hear doesn't sound at all like a typical air compressor.

Thanks to the automotive industry, there are quite a few different AC compressors and designs to choose from. The Denso 10S15C I used is one of the smallest units made, so if you need more air flow, just choose a larger compressor. All automotive AC compressors will typically exceed 200 psi, so no worries getting to your 120psi design pressure.

 

[Richard Hed]

Next up in the design process is to figure out what I can expect from a motor selection. I want the pump to have the ability for an output pressure of 120psi.

I want to use a commonly available compressor duty motor that runs at a relatively low speed. Also, I don't want to exceed most US single phase circuits. That leaves me with a 1½hp motor @ 1725 rpm.

At that speed and power, the average motor torque is 54.8 in-lb.

Knowing there's going to be frictional and other efficiency losses, I assumed a 15% efficiency loss. Also in every slider crank, the worst case position is when the connecting rod is 90 degrees to the crankshaft:

View attachment 141513

With these data points, I can play with the bore and calculate the stroke to match up with the tooling I have available and what seems reasonable to me in a design. The design parameters are as follows:

Number of Cylinders: 3
Cylinder Bore: .625"
Cylinder Stroke: 1"

Each cylinder stroke will have a compression ratio of 8.3:1.

Calculating Pump CFM is not straight forward because the hardware store's air compressors specifications are complete bullshit. On a 120V/1ph/60Hz/15amp house circuit you can get a maximum of 1,800 watts which equals 2.4 hp. Go to the hardware store. You'll find ratings that just don't make any sense. How the hell can a compressor running off 120V house power be 4 horsepower!?!? It's total marketing bullshit.

Anyways, using the ideal gas law (Pv=nRT) I calculated that the pump will output .91 cfm @ 120 psi. Another way to think of it is that the pump will compress .332 moles of air from atmospheric pressure to 120 psi per minute.

This calculation results in an air inlet consumption of 8.1cfm.

If you've made it this far. I thank you. I know this post is a bit dry. No fancy renderings or videos, just a bunch of numbers. But again. This is what engineering is.

Attached is the working spreadsheet (OpenOffice .ods format) with all these calculations. If you study it, let me know if I've screwed up somewhere. But the numbers seem reasonable and that's one way to check if the calcs are correct.

Thanks,

...Ved.

I bought a table saw in the Philippines. The claim on the package was 3 HP. The wattage was 1800watts. LOL. I called the sales people on it, but frankly, they were just too uneducated to know a thing about a horsepower in watts. In Asia, one must look VERHY carefully at all claimjs. Of course, in Rome it was called 'caveat emptor', let the buyer beware. Truthfully it is everywhere, but when advertisers are allowed to openly lie?

 

[Steamchick]

The same with "200Amp" Weld sets! - And "Lifetime guarantees", mean "the lifetime of the battery" - not you or I! And they don't tell you that is only 100 re-charge cycles.
K2

 

[Richard Hed]

The same with "200Amp" Weld sets! - And "Lifetime guarantees", mean "the lifetime of the battery" - not you or I! And they don't tell you that is only 100 re-charge cycles.
K2

What's a 200 amp weld sets?

 

[Toymaker]

What's a 200 amp weld sets?

My TIG-200P, which I purchased thru AliExpress, is rated 20 to 190 Amps; whether that's input or output is unclear, but the power outlets in my house in Thailand, where I live, are 220VAC (single phase) and my breakers are 40 amp. The unit works just fine in my house without blowing the 40 amp breakers, so I've always assumed that the 200 amp rating is for the output and that it likely uses a step-down transformer to lower the voltage and raise the current. This is unit I have: TIG Welder

Maybe K2 can further clarify the meaning of the 200 amp?

 

[Steamchick]

A weld set that can deliver 160 for a continuous weld (say 3 inches long!) Seems nowadays to be sold as a 200A machine, because the initial current can peak at 200A! Also, the service duty is often 5 mins welding per 15 mins cooling.... not good for anything but a few short welds. The thermal cut-out leaves you waiting for it to cool off....
That is advertising.....
K2

 

[Richard Hed]

My TIG-200P, which I purchased thru AliExpress, is rated 20 to 190 Amps; whether that's input or output is unclear, but the power outlets in my house in Thailand, where I live, are 220VAC (single phase) and my breakers are 40 amp. The unit works just fine in my house without blowing the 40 amp breakers, so I've always assumed that the 200 amp rating is for the output and that it likely uses a step-down transformer to lower the voltage and raise the current. This is unit I have: TIG Welder

Maybe K2 can further clarify the meaning of the 200 amp?

As i understand it, step-down transformers are not used in the modern TIG welders, it is all done electronically. That means there can be much greater control. I have a transformer type TIG welder in which one has to strike the arc like a match. It was one of the "new" ones made for small businesses back in 1980. I got it for a Lincoln 250 welder. The linconln is very nice, -- infinitely adjustable from 30 to 250 amps. But the TIG welder is COMPLETELY basic. I used it for a bit of practice then put it away as too difficult to use and control unless I absolutely HAD to use it which I never did need.

However, last year I bought a newer model, controlled and run by electronic methods.--that is, the new type of TIG machine. Many years ago, I learned to always check the % duty rating. IMNSHO, one should always try to get 100% duty rating so there need be no down time. If you are working professionally, this is a must. If hobby work, not necessarily a must. Also, this 100% refers to the higherst setting you can weld with. If you are welding at 200 amps, then your machine has a 40% duty cycle. If you are welding at a more likely setting of 40-120 amps, your duty cycle is longer, say something in the order of 60%.. If you are welding at 40 amps, your duty cycle is probably around 100%.

Another thing: since you are in the Land of the Thais, you are in a very hot country--that means the heat will not dissapate as rapidly as in a cooler environment which reduces your duty cycle. Keep a fan on your machine and that will help.

K2--Advertising? Yes, FALSE advertising! The Chinese have been known for centuries to give misrepresentations. However, it really is the same everywhere. In the US, false representation is often a illegal. But a rat can squeeze thru any hole in which it can get it's ratty litltle head thru, so any rat in business will keep chewing till it absolutely breaks the law. Unfortunately for the rats in business, they DO get a reputation, in this case, a "ratputation". -- HA, I just invented a new word, it will now have to be placed in the Oxford Dictionary. Will I be able to patent th word? How about a . . .

 

[Toymaker]

Next up in the design process is to figure out what I can expect from a motor selection. I want the pump to have the ability for an output pressure of 120psi.

I want to use a commonly available compressor duty motor that runs at a relatively low speed. Also, I don't want to exceed most US single phase circuits. That leaves me with a 1½hp motor @ 1725 rpm.

At that speed and power, the average motor torque is 54.8 in-lb.

Knowing there's going to be frictional and other efficiency losses, I assumed a 15% efficiency loss. Also in every slider crank, the worst case position is when the connecting rod is 90 degrees to the crankshaft:

View attachment 141513

With these data points, I can play with the bore and calculate the stroke to match up with the tooling I have available and what seems reasonable to me in a design. The design parameters are as follows:

Number of Cylinders: 3
Cylinder Bore: .625"
Cylinder Stroke: 1"

Each cylinder stroke will have a compression ratio of 8.3:1.

Calculating Pump CFM is not straight forward because the hardware store's air compressors specifications are complete bullshit. On a 120V/1ph/60Hz/15amp house circuit you can get a maximum of 1,800 watts which equals 2.4 hp. Go to the hardware store. You'll find ratings that just don't make any sense. How the hell can a compressor running off 120V house power be 4 horsepower!?!? It's total marketing bullshit.

Anyways, using the ideal gas law (Pv=nRT) I calculated that the pump will output .91 cfm @ 120 psi. Another way to think of it is that the pump will compress .332 moles of air from atmospheric pressure to 120 psi per minute.

This calculation results in an air inlet consumption of 8.1cfm.

If you've made it this far. I thank you. I know this post is a bit dry. No fancy renderings or videos, just a bunch of numbers. But again. This is what engineering is.

Attached is the working spreadsheet (OpenOffice .ods format) with all these calculations. If you study it, let me know if I've screwed up somewhere. But the numbers seem reasonable and that's one way to check if the calcs are correct.

Thanks,

...Ved.

Ved, reading through your two posts, I'm not sure if your primary end goal is simply to have a quiet compressor with smooth airflow, or if part of your goal is the satisfaction of designing and building your own compressor.

If you just want a quiet compressor with smooth output air flow, the easiest choice is likely to use an Automotive Air Conditioner compressor,...(see post #5).

If your goal includes the satisfaction of designing and building your own compressor from scratch, a swashplate compressor is a fairly simple build. I designed and built a 10 cylinder scaled down unit which works very nicely, and because of the double-ended-piston design, you only need to drill and finish 5 holes in order to get 10 cylinders,... the thread for that build is here: Small Wobble Plate Compressor.

Since I only needed my compressor to develop about 10 psi, the only math I did for the design was to calculate expected output volume at various RPMs; at only 10 psi max, the compression load on the motor is smaller than the frictional loads.
Since all the cylinders and pistons are contained within two halve of a cylindrical solid of metal, the process of drilling, reaming, and honing all 10 cylinders was in practice reduced to making only 5 cylinders as the machining is done while both halves are bolted together.

I hope this gives you a few more ideas and choices.

 

[vederstein]

I'm not sure if your primary end goal is simply to have a quiet compressor with smooth airflow, or if part of your goal is the satisfaction of designing and building your own compressor

I'd say my end goal is another project.

I already have an air compressor from California Air Tools and it's one of the quietest on the market. My first compressor (circa 1998) was a "5hp" Craftsman direct drive. It put out loads of air, but damn was it LOUD! I'd guess around 90 dBa. Several years ago I replaced it with my current unit from California Air Tools which runs around 65-70 dBa. It's noticeable when it turns on but at least you don't need to leave the garage when it turn on.

I've had "quiet air compressor ideas" in my head for several years now. Now It's time to see if those ideas can bear fruit.

Perhaps I do want the satisfaction of designing my own compressor or at least (in)validating my ideas.

...Ved.

 

[vederstein]

For a week or so, I've been working in the CAD system to flesh out the next, and what I think is the largest source of air compressor noise: Check Valves.

A cheap, loud, single cylinder air compressor has two check valves. One for and inlet on the cylinder's "suck" stroke and one for air exit on the "push" stroke. Those valves can be balls, reeds, or whatever but they're slapping open and closed dozens a times a second.

The threshold of hearing is 20Hz at the low end and 20,000Hz at the high end. The opening and closing "clicks" from the check valves are easily within the range of human hearing.

So, I want to eliminate check valves. Better if I can eliminate valves altogether.

Many moons ago, I designed and built a 5 cylinder rotary steam engine. This is one of the quietest engines I own. I think that I can use this concept as a basis for my air compressor. The valving is accomplished through the porting between the crankshaft and the engine block, not unlike a wobbler steam steam engine with the porting on the radius of the crankshaft.

There's two ways to look at this. I can rotate the crankshaft and port through the block or I can rotate the block and port through the crankshaft. If I rotate the crankshaft, I'll have an offset weight reciprocating at motor speed, 1725 rpm. or ~29 Hz which is in the range of human hearing. If I rotate the block, the mass is higher, but the noise from the crankshaft throw will be eliminated.

Therefore the block will be rotated and the porting is to be through the crankshaft.

I have a bit more CAD work to figure out, but I show that concept in a follow up post.

...Ved.

 

[vederstein]

This short video is showing where the design for the air compressor is going....



P.S.: I realized that my inlet air porting is faulty in the CAD rendering. I'll need to think a bit more about it....

 

[Toymaker]

For a week or so, I've been working in the CAD system to flesh out the next, and what I think is the largest source of air compressor noise: Check Valves.

...Ved.

I agree that some valve types are fairly noisy, but the reed valves used in auto A/C compressors are so quiet that I cannot hear them. Now, at 72 years old, my hearing isn't as good as it once was, and hearing the compressor run for yourself will be far better than my written description, so with that in mind, here's a video with sound: Shop Air Compressor. During the video, I'm speaking at a low conversational level and you can easily hear my voice over the compressor noise.

 

[vederstein]

here's a video

Yup, very quiet indeed. It sounds like most remaining noise is from the motor.

Yet, the more I think about, I want the satisfaction of designing and building my own design. This is more for validation of my ideas that anything else.

Thanks,

...Ved.

 

[Toymaker]

Yup, very quiet indeed. It sounds like most remaining noise is from the motor.

Yet, the more I think about, I want the satisfaction of designing and building my own design. This is more for validation of my ideas that anything else.

Thanks,

...Ved.

I'm sure that's a decision everyone on HMEM understands and supports, and looking over your past projects on YouTube, I'm sure that whatever you build, it will be awesome

 

[davidyat]

Ved and the others, I have to uncross my eyes after reading your engineering reports. Most of the time I just don't get it. Ved wanted me to learn some 3D CAD software and I bailed out when pretty much I couldn't even draw a straight line!!! I admire those who understand it. For me, just show me the plans so I can make some metal chips in my shop.
Grasshopper

 

[Asm109]

There is a major difference between a compressor valve timing and an IC engine.
Engine timing needs don't change.
Compressor timing varies with the amount of pressure in the tank.
When the tank is at zero pressure, the valve is open for the entire stroke and all of the air goes into the tank.
When there is pressure in the tank. The air above the piston cannot go inside the tank until the pressure exceeds the tank pressure.
The reed check valves do a seamless job of meeting this requirement.