PWM noise and supercapacitors

[SeanB]

I didn't like the noise of my fan.  So, I took it out.  I haven't had a fan in my vehicle for years.

That said, this would be a nice little project for a UNO and a couple of FET's.  I have a friend with a custom built car that has asked me to put some LED turn signals on it.  I was thinking about using a UNO but couldn't justify it.  The other day he had a section of 2 inch pipe with a 12V pump in parallel.  The radiator is at the other end of the car and he said if he is at idle for a long time he has to turn on the pump.  I didn't think about it till I read your post.  He needs to turn on the pump and control the speed of the fan automatically.  Now I can justify the UNO.  Thanks!

You either do not drive more than 2km in a trip, or only travel on a freeway and never are in any traffic jams. The only time my fan runs are if I have done a slow trip, or have been in traffic for more than 3 minutes. As the car has a 2 speed thermostat which does the regulation the simple way by using a resistor block on the one fan I rarely hear it. High speed is only going to occur when the engine is approaching 95C instead of 90C regular running temperature. While the ECU can select the high speed with the AC compressor running, it also has a mechanical thermoswitch to act as back up as well.

[Simon]

No, variable speed control is much more efficient. It uses far less energy to run a fan at 50% of its maximum speed, rather than turn if off and on, over a period of time.

This is because drag is not linearly proportional to the speed but goes up to the square of speed. In order to double the speed if a fan, you need to quadruple the power.
http://www.explainthatstuff.com/aerodynamics.html

This is why companies invest money in upgrading their air conditioners to use variable speed drives, rather than just switching the fans and compressor on and off. In the long term, the savings mount up and it pays for itself.

A continuous lower power draw is also easier to manage than a large short ones, batteries last longer on continuous low power than bursts of high power, alternators can feed the required lower continuous power and so the battery can be smaller the alternator smaller. The whole automotive industry is going more efficient and on large electrical systems this has a benefit. Cables won't be expected to carry large sustained currents, voltages drops and losses will be less.

Very well, these points are all valid.

(I think you will find, by the way, that the power of a fan varies with the cube of the speed. From the standard "fan laws" (good engineering rule of thumb), the pressure developed varies with the square of the speed and the flow varies directly with the speed. The required power is proportional to the flow times the pressure rise, so it therefore varies with the cube of the speed.)

Another factor, however, is the aerodynamic design of the fan. It will be optimized for a given speed, and if you run it at a much slower speed it will certainly consume less power, but the power consumed per cubic foot of air moved will go up and it will waste energy for a given amount of cooling.

If a fan is going to be used with variable speed control, the design of the fan blades should be adjusted to suit the normal range of (lower) operating speeds.

Doubling the speed of a fan requires the power to go up by a cube. I was looking for a very fast motor one time for work and spoke to a guy at a company and I couldn't make up my mind if the motor (driving a fan) needed to run at 10Krpm or 11Krpm, he said please go back and check because that little 1Krpm your not sure about could make all of the difference to the motor design and the power required and it can make or break the design.

Lower speeds are not less efficient for airflow at all. Faster fans create more turbolence and turbolence is lost energy and an obstacle to the air itself, if it was such a bad idea why are so many manufacturers designing expensive fans with 15-100% speed ranges ? the world is going brushless not only for the sealed longer life motors but for the ease of control as they are inherently speed controlled motors. I have designed a tiny box that can do what the OP wants because it drives a brushless motor that has all of the power drive electronics built in, my equivalent for brushed fans is vastly larger because of all of the power switching.

[IanB]

Lower speeds are not less efficient for airflow at all. Faster fans create more turbolence and turbolence is lost energy and an obstacle to the air itself, if it was such a bad idea why are so many manufacturers designing expensive fans with 15-100% speed ranges ? the world is going brushless not only for the sealed longer life motors but for the ease of control as they are inherently speed controlled motors. I have designed a tiny box that can do what the OP wants because it drives a brushless motor that has all of the power drive electronics built in, my equivalent for brushed fans is vastly larger because of all of the power switching.

You misunderstand. Variable speed is not a bad idea. But fans are designed to operate optimally at a given speed (this is the aerodynamic design of the shape and size of the fan blades). If a fan is designed to be installed in a variable speed application it will be designed to work optimally at the most common speed encountered, probably somewhere in the mid-range. A variable speed fan will (should) therefore have a slightly different design than a constant speed span. If you retrofit a variable speed drive to a constant speed fan it won't perform quite as well as a purpose designed fan in the same job.

(With small fans and low powers like PC case fans this is all a wash and scarcely matters. But with big fans the efficiency and power consumption are important.)

[Simon]

Yes you're probably right I have fan datasheets that show some sort of efficiency line on the graphs but it seems to amount to the highest volume you get which is generally midpoint between maximum pressure and maximum flow which looked kind of obvious to me and being an Italian datasheet I took it at some wankery although speaking to another fan manufacturer last Friday he said their datasheets also have a similar concept but they have an area of best operating conditions on theirs so you will have the pressure versus air flow line and then a section of that determined as being the most efficient and what the fan is designed to work in.

The other thing is you don't want to go over cooling your engine so often some speed control may be necessary in some applications and yield better overall efficiency.

The fan on my car is broken and I generally only gets a problem in the summer if I have to sit in traffic for more than a couple of minutes and then usually the cabin heater is enough to keep the engine stable. I have in fact just placed a cover in front of half my radiator to stop over cooling in the winter and as the fan does not work I have an engine air intake pipe stuffed into the fan's casing so that the engine gets warm air off the radiator which in the winter makes quite a bit of difference to consumption, performance and the time it takes to get the engine to temperature.

[Simon]

recently i have been playing with the excel design aid from texas instrument for a TPS55xxx series PWM controls.

according to the notes, if the PWM control have a large enough inductor, the ripple will be smaller. looking at the 50A draw you said ... it will need a HELL of an inductor. i think after going thru this design tinkering for TPS series, the post capacitor smoothing can only do so much if originally the PWM inductor is already under-sized ... or maybe i assumed wrong, but any experts here think i might be right?

Yes but I believe it you're referring to a switch mode power supply whereas here controller is supposed to be pulsing the fan to convert the design into a very powerful switch mode controller would require very large inductor and capacitor as the frequency is a mere 200 Hz so it's a bit of a tall order. I think the best option is to switch at a much lower speed to get rid of the noise or reduce it to a bit of clicking or speed need increasing but then the MOSFETs might overheat if they are not properly driven on their gates. I suggest the design inside the controller may be a bit cheap.

[Zero999]

recently i have been playing with the excel design aid from texas instrument for a TPS55xxx series PWM controls.

according to the notes, if the PWM control have a large enough inductor, the ripple will be smaller. looking at the 50A draw you said ... it will need a HELL of an inductor. i think after going thru this design tinkering for TPS series, the post capacitor smoothing can only do so much if originally the PWM inductor is already under-sized ... or maybe i assumed wrong, but any experts here think i might be right?

There shouldn't be a post PWM filter, unless it's after a huge inductor, otherwise the current through the capacitor would be huge.

One approach is to make the PWM frequency high so the motor's internal inductance will smooth the waveform, although even this isn't really necessary, so long as it has enough inertia to smooth out the vibrations and it doesn't excite any resonances.

[Simon]

No, variable speed control is much more efficient. It uses far less energy to run a fan at 50% of its maximum speed, rather than turn if off and on, over a period of time.

This is because drag is not linearly proportional to the speed but goes up to the square of speed. In order to double the speed if a fan, you need to quadruple the power.
http://www.explainthatstuff.com/aerodynamics.html

This is why companies invest money in upgrading their air conditioners to use variable speed drives, rather than just switching the fans and compressor on and off. In the long term, the savings mount up and it pays for itself.

A continuous lower power draw is also easier to manage than a large short ones, batteries last longer on continuous low power than bursts of high power, alternators can feed the required lower continuous power and so the battery can be smaller the alternator smaller. The whole automotive industry is going more efficient and on large electrical systems this has a benefit. Cables won't be expected to carry large sustained currents, voltages drops and losses will be less.

Very well, these points are all valid.

(I think you will find, by the way, that the power of a fan varies with the cube of the speed. From the standard "fan laws" (good engineering rule of thumb), the pressure developed varies with the square of the speed and the flow varies directly with the speed. The required power is proportional to the flow times the pressure rise, so it therefore varies with the cube of the speed.)

Another factor, however, is the aerodynamic design of the fan. It will be optimized for a given speed, and if you run it at a much slower speed it will certainly consume less power, but the power consumed per cubic foot of air moved will go up and it will waste energy for a given amount of cooling.

If a fan is going to be used with variable speed control, the design of the fan blades should be adjusted to suit the normal range of (lower) operating speeds.

Doubling the speed of a fan requires the power to go up by a cube. I was looking for a very fast motor one time for work and spoke to a guy at a company and I couldn't make up my mind if the motor (driving a fan) needed to run at 10Krpm or 11Krpm, he said please go back and check because that little 1Krpm your not sure about could make all of the difference to the motor design and the power required and it can make or break the design.

Lower speeds are not less efficient for airflow at all. Faster fans create more turbolence and turbolence is lost energy and an obstacle to the air itself, if it was such a bad idea why are so many manufacturers designing expensive fans with 15-100% speed ranges ? the world is going brushless not only for the sealed longer life motors but for the ease of control as they are inherently speed controlled motors. I have designed a tiny box that can do what the OP wants because it drives a brushless motor that has all of the power drive electronics built in, my equivalent for brushed fans is vastly larger because of all of the power switching.

id like to add that, as air speed increases, the amount of heat it can remove is not linear. it is less efficient ... somewhat like a logarithmic graph

Well at the other thing is having read that page on aerodynamics is that to double the speed of an object through air requires four times the power. The blades of a fan are not much different to the wings of an aeroplane so to double the speed of the fan would require more than twice the power, so I think really all things considered it's about the same either way. The blades may be designed for a certain speed but you won't stop the drag factor. Fans I have controlled will produce quite a decent draft with just 5 amps but to get full speed in fresh air required 20 A.

There is to be considered that generally for more airflow fans get larger. Where I work we supply fans of up to 1.2 m in diameter which go on hydraulic motors in very large vehicles. These fans generally ran at about engine speed and I would guess they are pretty efficient being much larger.

[IanB]

Well at the other thing is having read that page on aerodynamics is that to double the speed of an object through air requires four times the power. The blades of a fan are not much different to the wings of an aeroplane so to double the speed of the fan would require more than twice the power, so I think really all things considered it's about the same either way. The blades may be designed for a certain speed but you won't stop the drag factor. Fans I have controlled will produce quite a decent draft with just 5 amps but to get full speed in fresh air required 20 A.

You've got to be careful here with your details. Drag represents wasted power and is not a good thing. Fan blades, like aeroplane wings, are designed to minimize drag. In an ideal world you would like all of the power from the fan motor to be spent moving air and none of it to be wasted on drag. Real fans cannot be perfect and so they don't have perfect efficiency. However, the area of best operating conditions you mentioned in your other post is the operating region where the efficiency is maximized.

Drag varies with the square of the speed as mentioned in that aerodynamics article, but efficient movement of air through a fan varies with the cube of the speed. Since the primary job of a fan is to move air the cube law wins out.

[Simon]

So what your saying is that if the speed doubles the power requirement is the square but the amount of air moved cubes ? But then I'm told that to double the speed of a fan requires an increase in power of a cube so the power to airflow ratio is linear but the drag is the square anyway ?

[Simon]

this graph explains well
http://www.nmbtc.com/fans/white-papers/Fan_Efficiency_Curve.jpg

like a capacitor ESR dipping lol

Yes that is the sort of graph i have seen, but then efficiency compared to what ? 25% of what ?, You would think that the best operating point would be 100% so the graft is a little confusing to me and seems to be simply a representation of where you get the most air moved. For example in an air conditioning system we have at work the customer has put a silly amount of ducting on the outlets. So my quest has not been to look at so-called efficiency graphs but to actually find the fan that will develop the most pressure because if we don't develop pressure nothing will move in the first place. What needs to be remembered when you are working with fans is that most of the time you are trying to push the air someplace, either through ducting for air conditioning or through a radiator for engine cooling so you will need to overcome the pressure drop of what ever there is after the fan or if the obstacle is behind the fan it still has to draw air across that pressure drop.

[Zero999]

this graph explains well
http://www.nmbtc.com/fans/white-papers/Fan_Efficiency_Curve.jpg

like a capacitor ESR dipping lol

Is that with the motor run at full speed?

It seems to be a plot of the efficiency vs load, which makes sense. When there's no pressure increase (uncovered), no work is being done, so the efficiency is zero and when there's the fan is covered the pressure is maximum and no work is being done  as the air flow is zero, so the efficiency is zero. In the middle there's a sweet spot with  some pressure increase and reasonable airflow.

[IanB]

So what your saying is that if the speed doubles the power requirement is the square but the amount of air moved cubes ? But then I'm told that to double the speed of a fan requires an increase in power of a cube so the power to airflow ratio is linear but the drag is the square anyway ?

No, what I am saying is this:

The "Fan Laws"

1. The volume of air moved (cubic feet per minute, cubic metres per hour) varies directly with the speed
2. The pressure rise (psi, bar, kilopascals) varies with the square of the speed
3. The motor power (hp, kW) varies with the cube of the speed

Law 3 arises from laws 1 and 2, since the power required depends on the air flow times the pressure rise.

[Simon]

So what your saying is that if the speed doubles the power requirement is the square but the amount of air moved cubes ? But then I'm told that to double the speed of a fan requires an increase in power of a cube so the power to airflow ratio is linear but the drag is the square anyway ?

No, what I am saying is this:

The "Fan Laws"

1. The volume of air moved (cubic feet per minute, cubic metres per hour) varies directly with the speed
2. The pressure rise (psi, bar, kilopascals) varies with the square of the speed
3. The motor power (hp, kW) varies with the cube of the speed

Law 3 arises from laws 1 and 2, since the power required depends on the air flow times the pressure rise.

2. Right so is that 2x speed = 4x pressure or 4x speed = 2x pressure

I'm assuming you need 4x the speed to double the pressure.

[IanB]

So what your saying is that if the speed doubles the power requirement is the square but the amount of air moved cubes ? But then I'm told that to double the speed of a fan requires an increase in power of a cube so the power to airflow ratio is linear but the drag is the square anyway ?

No, what I am saying is this:

The "Fan Laws"

1. The volume of air moved (cubic feet per minute, cubic metres per hour) varies directly with the speed
2. The pressure rise (psi, bar, kilopascals) varies with the square of the speed
3. The motor power (hp, kW) varies with the cube of the speed

Law 3 arises from laws 1 and 2, since the power required depends on the air flow times the pressure rise.

2. Right so is that 2x speed = 4x pressure or 4x speed = 2x pressure

I'm assuming you need 4x the speed to double the pressure.

The pressure varies with the square of the speed, so 2x speed = 4x pressure (= 2²), 3x speed = 9x pressure (= 3²), and so on.

[T3sl4co1l]

TBH ... does anyone think the implementation i inferred from the simulations (a few post ago) are practical? or the "bridge snubber" is rubbish lol ... cos there is no way i have any chance to test in a real car

I can't see what it is you're simulating, but it doesn't look representative.  But that's okay, you'd need an inductor in the 10-100mH range to handle 200Hz.  The obvious answer for any other situation is: by throwing caps at it, you're just going to make things hotter and (electrially) noisier.

My point from my first reply still remains... if it does everything the way you want, fine.  Don't complain about it.  It is exactly what it is, no more and no less.  Attempting to address undesirable features (like beating the shit out of the motor at 200Hz) is like trying to train an elephant in ballet.  You need a fundamental redesign, a device which isn't made for crap.

Tim

[Leadfootin]

First I would try some fairly high current flyback diodes, say 10 or so amps and check the noise difference. 50V minimum, try paralleling a few smaller ones to test out. Often an old UPS or inverter will have some Schottky's suitable for the job.

[Simon]

So what your saying is that if the speed doubles the power requirement is the square but the amount of air moved cubes ? But then I'm told that to double the speed of a fan requires an increase in power of a cube so the power to airflow ratio is linear but the drag is the square anyway ?

No, what I am saying is this:

The "Fan Laws"

1. The volume of air moved (cubic feet per minute, cubic metres per hour) varies directly with the speed
2. The pressure rise (psi, bar, kilopascals) varies with the square of the speed
3. The motor power (hp, kW) varies with the cube of the speed

Law 3 arises from laws 1 and 2, since the power required depends on the air flow times the pressure rise.

2. Right so is that 2x speed = 4x pressure or 4x speed = 2x pressure

I'm assuming you need 4x the speed to double the pressure.

The pressure varies with the square of the speed, so 2x speed = 4x pressure (= 2²), 3x speed = 9x pressure (= 3²), and so on.

That makes sense as in my work application where we need pressure we are going for a fan that is the same size but it runs a bit faster. However pressure does not always equate to more air flow as if the pressure drop is low putting more pressure in with no ability for airflow will still not get you far hence I suppose those efficiency graphs although it's the characteristics of the whole system that will tell you if a speed change will make much difference.

A lot of cars seem to have put large fans in, far more powerful than needed, Bosch developed a brush-less 12V 48A fan for Audi quite a few years ago now, and I've seen audi's that have a mechanical fan in that runs all of the time anyway so the electric fan is either well overpowered or is an extra safeguard so running full speed and drawing all of that power is a really bad idea.

[SeanB]

Audi fan will be a viscous coupled unit, the extra fan is there for when the AC is running and developing high head pressure from being stuck in traffic, and typically will have a 2 speed switch and a resistor. In traffic on a hot day with AC on you will hear it cycling on in a 20 second cycle on 30 second off or so. It will run as well when the radiator reaches 87C.

[Simon]

The fans I refer to are called QBA fans and they are speed controlled with a 5V signal straight from the ECU, the last thing you want is one of those going full blast, no one has yet matched them in the automotive market that I'm aware of. If they loose a signal they will default full speed.

[max_torque]

The reason you have speed controlled fans on a passenger car these days is not directly to do with any "efficiency" of the fan, motor, or controller, and EVERYTHING to do with the fact the fan peak power requirement is massively intermittent!

During testing and sign off, the cooling system will be loaded under a worse case scenario (usually a 45degC ambient hill climb in a low gear at max vehicle train weight).  As modern cars are powerful (at peak output) the cooling system needs to be able to reject a LOT of heat (around 1hp per flywheel hp), so under this condition it takes a powerful fan to produce sufficient Air flow to remove that heat at a sensible DeltaT.  However, for the vast majority of it's operation, your car is under  a much lower engine load, and ram airflow is sufficient to provide the heat transfer.

As a result, older cars just switched a big powerful fan in when needed, but as cars became quieter and NVH expectations grew, OEMs started fitting dual speed fans, and then fully speed controlled ones to limit the impact of suddenly turning on up to 750W of fan all of a sudden!

Currently, pretty much all cars use speed controlled brushed motors, with some manufacturers moving to brushless motors for reliability reasons (no brushes to wear or carbon build up etc)