4 wire fans

Most of the low voltage dc types get speed control via supply voltage variation. The cheap sleeve bearing types are crap, and always fit ball bearing types for long term reliability here. Thay can often be relubricated as well, which extends life.

Other than voltage, speed will vary due to obstructions in the air flow, variations in atmospheric pressure and bearing friction, which on cheap sleeve bearing types, degrades with age. Data sheets should have all the data on speed tolerance.

Stepper motors are not good at high speed, unless designed for that class of service, as the rotors are usually solid, eddy current losses, and will typically get quite hot.

The old synchronous motor driven clocks, 1950's vintage, typically had a mechanism to ensure that they started running in the right direction, not in reverse. Even earlier versions had a shaft that you had to spin to start the motor. Have a 1930's deco instance, Temco, of such design. restored many years ago, that is still running on the wall here, and is accurate within a minute or so, over months, but that shows how accurate the mains frequency is here in the uk, averaged over time.

Most of which you can eliminate by driving them with a better approximation to a sine wave - that is by treating them as the synchronous motors they actually are.

If a square wave is a first approximation to a sine wave, the second approximation to to connect a winding to V+ for 33% of the period, to ground for the next 17%, to V- for the next 33% and to ground again for the last 17%. Don Lancaster listed a whole lot more patterns for his "magic sine waves" which were all pretty horrible, but didn't have much higher harmonic content.

The pulse output from 3 or 4 wire fans is proportional to speed, and has been used for years to close the loop in servers and other kit, to optimise temp control vs fan speed. Some even have the temp sensor in the fan output air flow, to make a self contained unit.

Ime, for identical fans running the same conditions, the speed variation is very small, often in the sub Hz range, with a droning sound as the sync varies. That's with the fan often running at thousands of rpm,, so the variation is almost insignificant.

The motor doesn't care whether it is driven by a sine wave or some switched approximation to one.

The distinction is essentially customer driven.

I think it is because the high-efficiency designs rely on laminar airflow over the blades, which reduces rapidly as the blades become dirty. A 'badly designed' blade will have turbulence to start with - but this won't get much worse when it is dirty.

Big, flat low-temperature-differential cooling plates.

I had a horrible fault on a Volvo that would suddenly produce vast clouds of blue smoke and drop to idling speed when I tried to accelerate (highly unpopular in traffic). After a few moments of roadside idling, it would return to normal.

One day it did it in a place where I could pull off the road safely and look under the bonnet. The corrugated PVC air hose from the air filter to the carburettor was flat but slowly resumed its normal shape after a few seconds. As it was connected to the sump breather, the corrugatios were full of oily condensate, which had been sucked into the engine.

Although the air filter appeared to be perfectly clean, something must have been blocking the pores sufficiently to cause enough suction under heavy acceleration to collapse the PVC hose when it was hot and pliable.

This is what I meant.

You can see that in mechanical timers. Pity the wheels are plastic, or they would be eternal. As it is, they only last a few years.

Makes sense. I'd rely on second-order signals to note that "something is not cooling properly" -- despite fans operating at targeted speed.

One of the advantages of software and 24/7 designs is that you have lots of opportunities for observation: is this behaving the same way it did yesterday? six months ago? So, you can come to more sensible actions instead of simple bang-bang decisions.

Lots of heat to move makes this difficult. Looking into refrigeration, currently -- but that just moves the problem to another location.

But that's its design purpose.

In my vehicle, the hose from the PCV (!) valve vents to the "outside" of the air filter. I.e., vapors have to pass THROUGH the air filter to get to the carburetor. So, you can see the sort of crud that is getting sucked in to the air cleaner via the PCV valve by examining the air filter directly opposite its entry to the assembly.

"Progress" (ever wonder what performance would be like if you just started cutting hoses??)

The suction force of combustion motors is tremendous. It was actually used to power the brakes.

Internal combustion engines are effectively air pumps. It would be interesting to sort out how much air is processed (per mile driven, per engine revolution, etc.)

But, they aren't operated 24/7/365 (as HVAC systems are).

And, the volume of air processed by an HVAC system is pretty impressive (our cooler moves 6000 CFM -- continuously, regardless of whether we are awake, asleep, etc.).

This was a good "in your head" exercise to undertake during my morning walk (so, treat it as you would a slide-rule calculation and hope for one or two significant digits)

Assume 6 engine displacements = 1 cu ft. (5L plant)

One engine displacement = 2 revolutions (VE=1.0) So, 12 revolutions per cu ft.

Assume 1200 RPM (to make the math easier while walking!) means 100 CFM.

As a datapoint, a bathroom exhaust fan is in the 60-100 CFM ballpark.

And, an HVAC system (refrigeration) typically moves 400 CFM/ton WHILE RUNNING (a cooler runs continuously but refrigeration is typically thermostatically controlled and more efficient at cooling -- CFM per comfort unit)

Of course, when you look at the diameter of a tail pipe, the numbers make more sense!

An air conditioner runs 30-60% duty cycle (pretend day+nite) so about 8-16 hours per day -- moving about 16,000 (to 32,000) cu ft (@4T) in that time. Or, roughly half a million (to a full million) cubic feet per month (typical replacement interval)

A car's air filter is roughly an annual replacement interval. How many driving minutes per year? At 12K mi/yr and 40MPH, (city driving) that would be 300 hours or 18000 minutes. At 100 CFM that's 1,800,000 cubic feet

Still seems that we're changing HVAC filters more often!

No, that's a mistake -- about 25% high (I used the size of OUR plant in my calculations -- 5T)

Have we got at cross-purposes here? PVC = polyvinyl chloride, the material the hose was made from.

Ah! I thought you were discussing the Positive Crankcase Ventilation valve which recirculates "fumes" from the crankcase back into the carburetor to be burned off. It's the only hose that enters my air cleaner (the hose being rubber).

For smoothy motion, use PWM microstepping.

I did that to tune the superconducting cavities for the CEBAF electron accelerator at Jlabs.

John Larkin Highland Tech Glen Canyon Design Center Lunatic Fringe Electronics

Actually, the *breather* hose enters the air cleaner; the PCV hose enters the intake manifold directly (so, *could* create black smoke even if the air filter was obstructed).

Which reminds me that its been years since I've cleaned mine. "Sitting" likely causes it to gunk up. <frown>

The motors that drive the turntables of microwave ovens deliberately spin in a random direction. If they hit something and stall, they try turning the other way.

Some newer ones don't have a turntable. They have some sort of RF stirring thing.

John Larkin Highland Tech Glen Canyon Design Center Lunatic Fringe Electronics

That is what Don Lancaster's "magic sine waves" offer.

The mass of the rotor smooths out the rotation pretty effectively anyway

And I got a sub-contractor to do it for IASystems in Cambridge back in