Control options for compressors with single phase & 3 phase AC electric motors.

[tbone1004]

Lubrication is one concern. Cooling the tubes between the cylinders also seems worth considering. With a slower running compressor, the fan is also slower & that makes me wonder if the cooling may become insufficient.

Maybe I might need to poke around with an IR thermometer sometime

You can solve the fan problem much easier than the lube problem though. Removing the mechanical fan and putting an electric fan is easy, but fixing the splash lube is the real issue so if you need to go slow, best to use one that can do it easily.

 

[Gone for diving]

For splash lube, if your concerned,
you could run it abit overfull with oil,
I have an old worn out compressor,
That I fixed up for fun, but parts are near impossible to find.
It needs to be overfull,
if not its noisy.

 

[PBcatfish]

For splash lube, if your concerned,
you could run it abit overfull with oil,
I have an old worn out compressor,
That I fixed up for fun, but parts are near impossible to find.
It needs to be overfull,
if not its noisy.

I have been cautioned not to do that with a breathing air compressor. I was told that the concern is lubricant getting into the compression chamber.

Does anyone know if there is truth to this concern?

 

[PBcatfish]

Oil Pressure switches are easy enough to find. The same is true of delay on break timers. 10 seconds is a commonly available setting. Pressure switches can be found at most auto parts stores. The timers can be found at industrial supply places like Grainger.

A repeat cycle timer would give you the on/off timing you want for the drain. They come in many shapes & sizes. You can get them from Compressor supply guys like August Industries or industrial supply places like Grainger.

Air Pressure switches in the 3,000 to 4,000 psi range are similarly available from August & probably industrial sources as well. Some are adjustable as a process variable.

You could also use a small PLC in place of the timers, but for something this simple, the timers are probably cheaper & easier to set up.

I know how to set up standard industrial controls as described above. I don't have experience setting up remote control from a phone.

For those who like to shop amazon, here are a few searches that turned up some interesting items -
Oil pressure switches - Amazon.com : oil pressure switch
Delay on break timers - Amazon.com : delay on break timer
Repeat cycle timers - Amazon.com : repeat cycle timer relay
High pressure switches - Amazon.com : 5000psi pressure switch

You need to do a little sorting through the things that are listed above, but usable components do seem to be in the mix.

 

[Gone for diving]

As long as the cylinder is higher than the oil when it's not running, (oil hot)
Should be fine, but should not be even close, also make sure the lowest part of the crankshaft doesn't hit the oil to much,
It will throw even more oil around, and rob power, and could cause foaming.
As long as the rings are good, that will not be a problem.
Worse that can happen it blows more oil out the vent,

Most of the time it's better to much oil than no oil, (all within reason)

Catapiller recomend 10% overfill on there dozer transmission and eng,
if working on steep sloped for long periods of time,

 

[Miyaru]

Your blending application seems like a natural fit for a PLC with a couple of analog cards on it. I'm curious what type of processing unit you chose to use. If you were to use a PLC, then you would automatically have a large number of timer functions at your fingertips for no additional charge. A PLC does require some programming though. The general family of programming languages are known as ladder logic.

It's interesting that you chose to use needle valves & motors. My first thought would have been to pulse width modulate solenoids. There is certainly more than one good way to skin a cat.

There are some products on the market that use this approach. Nardi created the TT-Mix Pro which sets you back 5k5, LM.NT created the King Blender, which sets you back 3k23. These boxes can do continuous blending, but that's it.

But technology has advanced rapidly, especially over the last decade. PLCs have been replaced with micro-computers like raspberry-pi, arduino and esp32. These have way more processing power, so why not use it?

Pressure sensors commonly use the current-loop: zero pressure corresponds to 4mA and the maximum pressure of the sensor corresponds to 20mA. Texas Instruments created chips that measure this current, transform it into a digital value and transmit this value over an 2-wire bus straight into the processor. No need for analog cards, since 16 of these chips will fit on a single bus.
Set pressure reached? Shut down the compressor.
Temperature sensors are common and cheap these days as well. Attach a sensor to each stage and the processor knows how hot each stage gets. Too hot? Shut down the compressor.
Shutting down the compressor can be easily achieved by adding another relay into the loop that controls the motor-relay/contactor. That extra relay is controlled by the micro-computer.

Using solenoids to control the oxygen flow means dealing with an open/close flow. It is not a stable flow that can be fine-tuned with a needle valve. Advances in technology has brought down the prices for stepping-motors and their controllers. Still, it doesn't look like the market has taken advantage of that.

Mixing the gas before it goes into the compressor:
The ideal mixing stick mixes a perfect homogene gas without limiting the flow. Which is impossible. Several solutions like balls or baffles mix gas, but they also cause a pressure drop. The best solution is to use a shape that has been know for over 250 years: a helicoid. It has the smallest surface causing the most turbulence. Ross Cowell used this concept to create the Nitrox Stik and patented it.

Now that 3D printers have become affordable, printing a helicoid is easy. Plus, the patent has expired so anyone can do it. Two engineers in Hungary did this and published their results in this paper.
I used all of that information and designed my own stick:

 

[PBcatfish]

There are some products on the market that use this approach. Nardi created the TT-Mix Pro which sets you back 5k5, LM.NT created the King Blender, which sets you back 3k23. These boxes can do continuous blending, but that's it.

But technology has advanced rapidly, especially over the last decade. PLCs have been replaced with micro-computers like raspberry-pi, arduino and esp32. These have way more processing power, so why not use it?

Pressure sensors commonly use the current-loop: zero pressure corresponds to 4mA and the maximum pressure of the sensor corresponds to 20mA. Texas Instruments created chips that measure this current, transform it into a digital value and transmit this value over an 2-wire bus straight into the processor. No need for analog cards, since 16 of these chips will fit on a single bus.
Set pressure reached? Shut down the compressor.
Temperature sensors are common and cheap these days as well. Attach a sensor to each stage and the processor knows how hot each stage gets. Too hot? Shut down the compressor.
Shutting down the compressor can be easily achieved by adding another relay into the loop that controls the motor-relay/contactor. That extra relay is controlled by the micro-computer.

Using solenoids to control the oxygen flow means dealing with an open/close flow. It is not a stable flow that can be fine-tuned with a needle valve. Advances in technology has brought down the prices for stepping-motors and their controllers. Still, it doesn't look like the market has taken advantage of that.

Mixing the gas before it goes into the compressor:
The ideal mixing stick mixes a perfect homogene gas without limiting the flow. Which is impossible. Several solutions like balls or baffles mix gas, but they also cause a pressure drop. The best solution is to use a shape that has been know for over 250 years: a helicoid. It has the smallest surface causing the most turbulence. Ross Cowell used this concept to create the Nitrox Stik and patented it.

Now that 3D printers have become affordable, printing a helicoid is easy. Plus, the patent has expired so anyone can do it. Two engineers in Hungary did this and published their results in this paper.
I used all of that information and designed my own stick:
View attachment 712466

I am not familiar with the mix pro nor the king blender. I just haven’t looked at the prefab market, so I don’t now what is out there.

I am aware of raspberry-pi, and arduino. I did not know about esp32. The first two have attractive price points & have proven effective in many applications. I should look into them, but I haven’t found the time to do that as of yet.

I’m familiar with common analog pressure sensors that put out 4-20ma, and 0-10vdc signals. I am aware of more complex sensors that put out hexadecimal or binary data directly, some of which can export directly on IEEE, Canbus, Ethernet, Profibus & Devicenet. Some will also put out fiber optic signals. The 0-10vdc are the ones that I find easiest to use, due to ease of troubleshooting. The 4-20ma ones are considered more reliable in some situations. The fiber optic ones are the most reliable in high interference environments. I usually only use things like Ethernet when someone wants me to export data to an excel sheet that shows up in the head office or something like that.

Temperature sensors are basically the same story. Temperature switches, are more simple and less expensive, but potentially also effective for this application.

The “nitrox stick” is basically a turbulator. It mixes up what ever you put into it. Quickly modulated spirts of O2 at the inlet should become a steady mix by the time you get to the end. The analysis of the O2 content can also be fed into a PID loop in the software that takes the lumps out of readings if they come in choppy, or if the mix itself is not fully blended. There are a number of ways to handle the potential issues associated with a modulated solenoid. The limiting factor here is the response time and repeatability of the solenoid. It is possible to optimize the carrier frequency of the modulated pulse to account for this. Pulse width modulated solenoids can actually produce some pretty accurate metering, if the system is set up correctly. I’ve done this with powders, granular materials, liquids & gasses in the past. All have given many years of accurate & reliable service. Solenoids do eventually wear out. Usually the electrical end goes bad first, unless you are metering a sticky or corrosive material, which is not the case here.

Steppers offer wonderful precision. What they don’t offer is the reliability of a servo. With a stepper, you send out a pulse train signal & then you hope that the motor gets to where you told it to go. With a servo, there is feedback that tells the amplifier that the motor actually got there, or it didn’t. Corrections are then made by the amp if needed. If the stepper does not have sufficient torque to accelerate or move the load, you can get undocumented lost positioning. The system you propose seems to overcome this potential issue by monitoring the finished PPO2 & using that for a feedback loop. Here too, there is more than one good way to skin a cat. Your choice seems like one of the better ways to approach the situation.

Your information regarding the helicoid is interesting. Thank you for presenting that.

 

[Gone for diving]

I’m familiar with common analog pressure sensors that put out 4-20ma, and 0-10vdc

Can you expand a bit how these work?
I am a bit unclear. For instance...

0-10vdc is that a variable voltage?
Or is it 0v off than 10v for on?
I have some interesting in something like this..C$ 51.70 | pressure transmitter with LCD Display G1/4 12-36V 0-10V 0.5% 0-600bar optional stainless steel pressure transducer sensor

35.39€ 8% OFF|Druck Sender mit LCD Display G1/4 12 36V 0 10V 0.5% 0 600bar Optional Edelstahl edelstahl Druck Wandler Sensor|Drucktransmitter| - AliExpress

Smarter Shopping, Better Living! Aliexpress.com

a.aliexpress.com

 

[Tracy]

Can you expand a bit how these work?
I am a bit unclear. For instance...

0-10vdc is that a variable voltage?
Or is it 0v off than 10v for on?
I have some interesting in something like this..C$ 51.70 | pressure transmitter with LCD Display G1/4 12-36V 0-10V 0.5% 0-600bar optional stainless steel pressure transducer sensor

35.39€ 8% OFF|Druck Sender mit LCD Display G1/4 12 36V 0 10V 0.5% 0 600bar Optional Edelstahl edelstahl Druck Wandler Sensor|Drucktransmitter| - AliExpress

Smarter Shopping, Better Living! Aliexpress.com

a.aliexpress.com

It is a variable voltage

 

[PBcatfish]

Can you expand a bit how these work?
I am a bit unclear. For instance...

0-10vdc is that a variable voltage?
Or is it 0v off than 10v for on?
I have some interesting in something like this..C$ 51.70 | pressure transmitter with LCD Display G1/4 12-36V 0-10V 0.5% 0-600bar optional stainless steel pressure transducer sensor

35.39€ 8% OFF|Druck Sender mit LCD Display G1/4 12 36V 0 10V 0.5% 0 600bar Optional Edelstahl edelstahl Druck Wandler Sensor|Drucktransmitter| - AliExpress

Smarter Shopping, Better Living! Aliexpress.com

a.aliexpress.com

With a pressure switch, it's either on or it's off. With a pressure sensor, also known as a pressure transducer, a pressure transmitter,or a pressure sender, the output is variable as the pressure changes. Most have a linear response. That is if the range is 0-100psi on the input & 0-10vdc on the output, then 1v=10psi, 3v=30psi, 4.5v=45psi, etc. A few have logarithmic response curves or have other types of curves, but those are less common. You typically get your best accuracy when you get a sensor that has a range limit that is near what you expect it to see. If you put a 1,000psi sensor on a 20psi system, you will likely get poor resolution.

That's a cute little unit you found on alie. I would have high hopes for it.