An actual design question...

First would be to consider if it makes more sense to buy a system off the shelf - 12, 24, 48 VDC in to 120/220 VAC out inverters of that capacity and quite a bit more are an off-the shelf item, if you are not tied to 100-150 VDC as your input. True sine out will result in less heating of the motor than square out (misnomered mod-sine in the industry) will. You'd have to check to be sure it would work with your motor, but there's at least one true sine 300W (continuous) with 600W peak out there for $250 retail - 12V input. Reinventing the wheel can be expensive. Using an appropriate cap to correct the power factor should help, of course, since you have a specific dedicated load.

Second would be to swap the motor out for a 90VDC motor (again, off the shelf item) and simplify your controller problem to PWMing (or otherwise dropping your input voltage) to 90 VDC.

As for the PCB, if there's too much current to keep your traces happy, use big fat wires for those parts. Before that point, use short, wide traces for the power stuff.

Compressors usually present a very high surge load upon start-up. Your design needs to deliver that surge reliably. Also, it may not like "modified sine" which, as Lawrence said, is just glorified marketing speak for a square wave with a pause around the zero transitions. Got to talk to the motor manufacturer. Even if the manufacturer endorses it anything that deviates from sine a lot will cause the compressor motor to run rather noisy.

Big fat traces and nothing less than 2oz copper, preferably more.

Depending again on what the overall design parameters are, a compressor with an unloader (mechanical device that disables compression) can help a LOT with startup - you get the motor and compressor up to speed under no load, then drop the unloader and start compressing, with the rotational inertia of the system (already started and running full speed) helping to get it over the initial hump. More often seen on big compressors, but if it's a product you are building, might be worth including on your small compressor.

Yes, a clutch would help but I've never seen one on a compressor this small.

The Shepherd here next to me has voiced a complaint about item number one ;-)

Item number two has just been consumed, item number three is in the fridge and when my wife isn't here later today I might happen to find a piece that miraculously "fell out of the fridge".

It is tedious to read a question on one group you frequent, then go to another group and find exactly the same question, without your reply there. It's called "multiposting" and it is, for a variety of reasons, irritating and suboptimal to frequent contributors to the related groups.

Please look up "crossposting" on Wikipedia for a counter example. Even though you're posting from Google, I'm sure there's a way to do this correctly.

Not a clutch, usually - rather, a valve that opens the cylinder (thus, "unloads the system" so that rather than compressing, it's just wiffing air in and out. You might find one on a small compressor designed to be run by a gasoline motor, for instance, since automatically stopping and starting the motor is excessively complex in that situation, and of course they are common on large compressors. A clutch would fail a lot faster, and present more of a load to the motor on starting to compress (with the unloader valve, the whole rotating mass of the compressor/motor system is moving.) Without knowing how much of the OPs system is actually open to design, just tossing out options.

Yet another one being to use a 180V DC motor and jut run it at less than full speed, directly (well, fused and switched) connected to the

Like the bypass valve? That's a good idea. It tripped twice on our AC this summer, opens when the thing gets too hot.

Sounds expensive and special order though.

Nope - both 90 and 180 VDC motors are as common as dirt in industrial supply outlets. How expensive varies with how good (or how good a brand used to be, in a few cases of running things in the ground) but they are extremely off-the-shelf. Both MSC and McMaster could have one on your desk tomorrow, if you wanted it that badly. In many cases they'll have it on your desk tomorrow without even making it a rush order or paying extra for shipping, but they usually don't guarantee that it gets there that fast - it just does, anyway. They don't tend to stock the crummy ones, though, nor to they seek to be the lowest price suppliers.

If seeking low price for a one off, electric treadmill motors range from $40 (surplus outlets) down to free (free usually involves prying it out of a treadmill that provides an unpleasant reminder of an exercise program that didn't last) and they are often 90VDC.

Contraiwise, if the OP is designing some product to be manufactured, I'm sure the factory suppliers can find a source for motors priced to suit whatever market it's aimed at (ie, quality or cheap).

Motors, sure, but this would mean he'd have to build his own compressor. Usually small compressors come with an encapsulated motor inside. Even the big 4kW one in our AC is like that, you'd need a cutting torch to get to it.

That's not entirely fair. A proper "modified sine" has a pause (0V out) that runs from exactly 330 dgrees to 30 degrees, and from 150 degrees to 210 degrees. This means that it has no third harmonic content at all and a fairly low fifth harmonic content, which you can filter out, along with the higher harmonics with LC low pass filter. Even without the filter, you've got rid of a lot of eddy current heating.

-- Bill Sloman, Nijmegen

Well, yeah, it'll run. All I can tell you that I ran quite a few smaller motors on modified sine and none of them was too happy about it. Heat wasn't a huge issue but reduced power was.

Do you know why?

Years ago in a power electronics class that I took we did various SPICE designs of motor controllers, and while we just generated "modified" square waves, the inductance of the modeled motors made the output waveforms look

AFAIK most motor controllers use modified square waves?

Haven't looked into that much. For example, in a power outage I cannot run the AC fans of our wood stove at high speeds using modified sine. They just won't go there. With a real sine I can but that inverter is very old, heavy and quite the guzzler. In this case it's ok because we never really run them full bore unless we have a hard freeze. So I prefer the modified sine inverter because it is small and more efficient (battery lasts longer).

On pump motors that's another matter.

Not the ones I designed :-)

But that was usually because the client wanted a quiet operation. On steppers you often slam them in order to ramp up the current as fast as possible (limited by the voltage you have at hand). So those are rarely sine-driven.

Speaking of wood stoves and fans... do you point your fan at the wood stove, or have the fan facing outward from the wood stove? :-)

It's an insert (Quadrafire 3100i) and they are built in. No intake filters . Air gets sucked in on both sides, cirulated and out the top front. A ceiling fan also runs in the living room, direction up, which pushes the air back down at the walls and windows which avoids a draft. Extremely nice compared to central heat.

Ah, ok, thanks. We have a standalone wood stove at present, although it's in the living room around a corner and a good 100' I'd guess from the office, so it's supplemented with the regular electric furnace. (...which blew a fuse and quit working a few weeks ago! They had two 5kW coils plus the blower on one fuse and just a single 5kW coil on another fuse. Both were blown, which worried me for awhile until I realized that in all likelihood the single 5kW coil's fuse had been blown for ages and no one realized the furnace was just operating at 10kW rather than 15kW. I replaced both fuses and so far it's been perfectly happy... unlike me, watching the meter spin, when it's on...)

I think outdoor wood furnaces are rather cool:

---Joel

A Sola ferroresonant transformer might be the ticket.

Cheers

Phil Hobbs

A Sola ferroresonant transformer might be the ticket.

Cheers

Phil Hobbs

Well, thank you for your input. I think I will do it this way:

I will design a modified sine wave version with an unfinished output stage and adjust the parts to optimize according to the impedance of the compressor, and looking at the waveform, I will be able to see the effect relationship and get as close as I can to a good sine. (minimize noise and reflected emf)

Thanks again, guys.