Author Topic: Reactive boost for mosfet inverter  (Read 12510 times)

0 Members and 1 Guest are viewing this topic.

dnix71

  • SuperHero Member
  • ******
  • Posts: 2513
Reactive boost for mosfet inverter
« on: March 28, 2015, 05:24:41 PM »
As some here have suggested, adding a motor to an inverter setup allows starting reactive loads, like an a/c compressor.

I need a practical suggestion for my setup. I have a 20-year-old van and the a/c leaked. It's much cheaper to buy a 120v window unit and an inverter to the van than to pay a shop to fix the van a/c. Plus when I ditch the van, the inverter and a/c can be reused.

The a/c in question is a brand new LG 5k BTU unit that uses 4.1 amps running. The fan would run from the inverter, but the compressor wouldn't start and would eventually trip off the inverter. I have a 8" dual disk bench grinder of from Harbor Freight. The bench grinder will run from the inverter if restarted 7 or 8 times to give it a chance to spin up. Once spun up, the a/c unit will start instantly and run smoothly.

I don't see traveling with a spinning bench grinder, though. The bench grinder would have to be left running with the a/c because the a/c compressor will cycle off and on.

Is there a better way to provide the reactive power? I get nothing from Google except high-tech pdfs on how important it is to have reactive power on grid-tie inverters. From wikipedia, I see the term synchronous condenser for grid balancing, but no references to a consumer grade version.

If I understand the process correctly, the bench grinder isn't actually supplying extra energy, it is simply forcing the voltage and current phases to stay lined up while the compressor starts. It's the slippage between the voltage and current phases that make starting hard.

oztules

  • Hero Member
  • *****
  • Posts: 1477
  • Country: aq
  • Village idiot
Re: Reactive boost for mosfet inverter
« Reply #1 on: March 28, 2015, 09:22:06 PM »
The bench grinder acts as a generator, using the inverter  as the magnetizing current, and the inertia of the spinning disks for power. So thats all that is happening. The grinder in fact will be a PF negative as well as the starting of the AC. so reactive  power was not the problem ..... real power is.... we made the pf worse with the grinder, and got a better result??? improving the PF would help too, but your under done on current to start with from the looks of it.

The AC needs 6-7 time more current to start than to run, The grinder provided this extra current from energy storage in the wheels.

It may be possible to use a larger starting capacitor on the AC, and this will allow start up to be very much faster ( and better PF), but the current will be higher for a shorter time.. if your inverter can do this, then it may work.

Sounds like a HF inverter, and the energy storage in those is feeble, so surge is poor. LF inverters use the battery bank for storage, and can push very heavy loads for a long time before quitting.... but the HF inverter only has energy storage of 1/2E^2C of the hv caps to work with..... so the grinder provides energy storage in the form of inertia, not charge, and this pulls you over the hump.

Once running the AC runs like a transformer, with the secondary at variable ( asynchronous speed)  angle to the magnetising current ( because of the frequency and the speed of rotation), with input current proportional to torque + magnetising current. .. at start up it is a short circuited transformer... hence the huge currents.

A cheap LF inverter like the 3kw powerjack would solve it best, as it has huge start current available.... and it's cheap. W7 powerstar would eat it too, but lossy transformers... or a much bigger HF inverter ( I won't ever touch one again for serious use).

........oztules
Flinders Island Australia

dnix71

  • SuperHero Member
  • ******
  • Posts: 2513
Re: Reactive boost for mosfet inverter
« Reply #2 on: March 28, 2015, 11:03:36 PM »
The inverter is a SunForce Pro Series model 11240-1. I considered buying an ultracap pack made to start cars and attaching the ultracap pack to the back of the inverter.
http://www.ebay.com/itm/V2-Ultracapacitor-Module-KIT-Battery-Eliminator-Car-Audio-Starting-Remote-Solar-/301381327649
Would this work or is the problem inside the inverter itself?

oztules

  • Hero Member
  • *****
  • Posts: 1477
  • Country: aq
  • Village idiot
Re: Reactive boost for mosfet inverter
« Reply #3 on: March 29, 2015, 12:59:23 AM »
Only if your battery pack is a bit feeble or it has small wires.
Your problem is in the pre H bridge stage. It will use a push pull pwm to take the 12v to 160vdc. This is stored in your HV caps. This is all you have to power a surge + the power the little  1000w pwm can contribute at the same time.... so you limited to that supply of energy. The caps on the input will only stabilize the battery voltage giving the pwm a full 13v or whatever your state of charge is.. this may give you a fraction of a boost, but the current limit of the pwm will render any more low impedance oomph unusable...

If your battery is  sagging on start up, then short leads, or cap pack may help.... but you need to add more caps to the HV pack inside, this will give the H bridge a bigger reserve to bang into the load, doubling it may get you started... may not. If it had some chance before, it will have a show now... if it had no chance before, it probably needs more than twice... but you also run the risk at some point of over current on the Hbridge fets..... but i suspect they can take a pretty fair o/load for short periods.

Physically getting them into the thing may be a problem, but external mounting is possible with thick wires... never tried. This is where LF inverters shine, as they drive  the Hbridge directly from the battery bank, so have basically unlimited surge....... well...limited only by how much the fets can take.... mine is possibly 150X6X50 or 40kw or more for a limited time. The transformers then become the limiting factor ( mine are 90lbs or more now without the kilos of copper windings).

So thats your problem, surge storage inside the unit for the 160v or whatever the USA units run ... ( I assume 110x1.414... ish)

The length of tome of the surge will be determined by the inertia the rotor has to overcome versus the current developed in the rotor bars, working against the stator fields. The start cap will effect how this magnetic dance plays out, as it is the thing that helps drag the a phase around so there is a rotating field. If the start winding is wound with thinner wire, then this will also change the phasal relationship of the windings... so it may not even have a capacitor to keep it simple, but most refrigeration stuff I have worked on do have caps, and a weird replacement for the centrifugal switch that you would normally have in a higher torque start motor.

So try a larger cap in that position as well, it will not run after it has started as it uses a ntc in series with the start fields, so as it gets hot it's resistance goes up very high, and effectively you have disconnected the cap by this mechanism.

If you get an inverter type AC, you will probably have no start issues, as it is a three phase brushless DC motor, and it will have it's own caps there to do the job for you.


.................oztules
« Last Edit: March 29, 2015, 01:11:39 AM by oztules »
Flinders Island Australia

OperaHouse

  • Hero Member
  • *****
  • Posts: 1308
  • Country: us
Re: Reactive boost for mosfet inverter
« Reply #4 on: March 29, 2015, 06:24:49 AM »
One more thing that might get you through startup.    Get an on delay timer for the fan and set it to one minute or more. That way you won't be starting two motors at the same time.

dnix71

  • SuperHero Member
  • ******
  • Posts: 2513
Re: Reactive boost for mosfet inverter
« Reply #5 on: March 29, 2015, 02:25:49 PM »
The inverter voltage sags and the battery in the vehicle is being pulled down by the surge and this is with the engine running and a fairly new battery. I suspect the inverter simply can't supply the surge. Once running, it doesn't seem to use much power at all.

I went to eBay and looked up LF inverters and see that the smallest is rated 3kw but at split phase 220v. tapping off one side gets a lot less power.

My larger inverter also complains that the battery voltage is too high with the engine running. I have to play tricks to get it to stay on long enough to even test. The smaller inverter I have doesn't complain about supply voltage but it won't run the a/c even with the grinder boost. The smaller inverter will start the grinder, though, by brute force. It just keeps at it until the grinder is up to speed.

Cheap US a/c and fridge/freezers have obscene locked rotor amps. I had an upright box freezer that drew 140 watts running but would pull well over 800 watts to start. It used a Klixon instead of a capacitor. I added a large cap and the start amps dropped to 400 watts + but that was still too much for the so-called 1000 watt inverter, even being bolted within inches of the battery bank.

MOSFET amps for automobile ghetto blasters usually need a super capacitor with it to supply the current surges demanded by bass notes. That internal wiring arrangement would be what inverters should use to handle high start loads.


dnix71

  • SuperHero Member
  • ******
  • Posts: 2513
Re: Reactive boost for mosfet inverter
« Reply #6 on: March 29, 2015, 04:05:33 PM »
I opened the a/c unit and the compressor is a Quigan rotary. YZG-A045Y2D2T1
It's supposed to have a 30mfd/250v capacitor but the cheap scumbags that made it for Home Depot substituted a klixon.
A rotary compressor should not be that hard to start. The terminals are clearly marked. I need to order a cap from eBay and attach it and see what happens.

oztules

  • Hero Member
  • *****
  • Posts: 1477
  • Country: aq
  • Village idiot
Re: Reactive boost for mosfet inverter
« Reply #7 on: March 29, 2015, 05:00:09 PM »
Yes, a capacitor should help the torque, and this may get it started in time..

They are relying on higher wire resistance in the start windings only for the phase change from the looks... the cap should make this more efficient.... try different sizes too.

..........oztules
Flinders Island Australia

OperaHouse

  • Hero Member
  • *****
  • Posts: 1308
  • Country: us
Re: Reactive boost for mosfet inverter
« Reply #8 on: March 29, 2015, 05:12:01 PM »
My fridge just had just a PTC on the start winding.  I replaced it with a 200uF cap I think when I was trying to get it to start with a 1K inverter.  Never actually tested it both ways, the cap is still there and that is where I got the 120A reading.  Try disconnecting the fan and see if it will start.

oztules

  • Hero Member
  • *****
  • Posts: 1477
  • Country: aq
  • Village idiot
Re: Reactive boost for mosfet inverter
« Reply #9 on: March 31, 2015, 04:23:43 PM »
Thanks Opera, I meant PTC not NTC in above post.

Nix, you may need a much higher capacitance than 30 uf too. Because of the time constraints, Opera's 200uf  may be closer to what you will need to get a solid phase change, and the best chance of getting the thing rotating before the surge power is spent. Once the rotation speed is near 2/3rds-3/4 of operation speed, the current should ease of... just need to get it up there.

............oztules
Flinders Island Australia

dnix71

  • SuperHero Member
  • ******
  • Posts: 2513
Re: Reactive boost for mosfet inverter
« Reply #10 on: March 31, 2015, 05:10:31 PM »
I spoke tto soon about it not having a capacitor. Here is the wiring diagram.

I have a 30uF oil-filled run cap on order. It looks like I would just piggyback it across the R and S compressor terminals to get twice the boost. Is that correct? There is a PTC going to the common compressor terminal.

joestue

  • Hero Member
  • *****
  • Posts: 1759
  • Country: 00
Re: Reactive boost for mosfet inverter
« Reply #11 on: March 31, 2015, 09:28:47 PM »
what you see on top of the compressor is the OLP component in the schematic.. overload protection.
the run capacitor provides sufficient starting torque for these motors. it is in the schematic, with terminals marked C H F.
the fan also has a run capacitor, surprisingly.

I do not recommend significantly changing the run cap on the compressor, you might increase it slightly, but if you increase it too much you can burn it out the winding, but before you get to that point the efficiency is decreaed.
if you want to hit unity power factor you'll need to install another correctly sized run cap across the line.

the capacitor may be buried inside the control cabinet, unless you've already found the ptc or klickson, whatever you called it.
My wife says I'm not just a different colored rubik's cube, i am a rubik's knot in a cage.

oztules

  • Hero Member
  • *****
  • Posts: 1477
  • Country: aq
  • Village idiot
Re: Reactive boost for mosfet inverter
« Reply #12 on: April 01, 2015, 04:03:53 AM »
I fail to see any PTC or any other device to help starting, it is using the power factor run cap for the pf and starting as well. I normally only see this on low torque start applications ( bench grinder, centrifugal pumps etc.)

A ptc and a 100ufor more cap would probably shorten the start time, but a big cap  must not be hard wired to the start field. ..or it will cook. It must be removed ( centrifugal switch or PTC etc) from the circuit in less than .5sec or so and then the run cap can take over and pull the pf back to nearer unity.

Is there a PTC there I am not seeing?


..............oztules
Flinders Island Australia

bart

  • Full Member
  • ***
  • Posts: 200
  • Country: us
Re: Reactive boost for mosfet inverter
« Reply #13 on: April 01, 2015, 06:33:49 AM »
   Google search for a "110v hard start kit". Includes a Potential Relay and start cap.

dnix71

  • SuperHero Member
  • ******
  • Posts: 2513
Re: Reactive boost for mosfet inverter
« Reply #14 on: April 01, 2015, 01:23:32 PM »
I was calling the Klixon a PTC. The black disk on top of the compressor with 2 wires. It's the same thing used in light aircraft. There is a metal resisitive wire inside and a bimetal disk. Too much current heats up the disk and it pops. It will self reset after a few minutes of cooling down. It protects the unit from short-cycling in hot weather. There is also some kind of protection built into the wall end of the plug, too. There is at least a ground fault circuit in it, because it has a test and reset buttons.

I saw warnings about hard start kits. It the relay fails you will fry the compressor.

According to my Kill-A-Watt meter the PF running is 0.98. I don't know if I believe that. That's an insanely perfect PF. The running current the KWM shows does not quite match the nameplate, either. Maybe they are overcompensating the fan PF to balance the compressor. It takes at least 10 seconds for the fan to spin up, with or without the compressor. The fan itself is a squirrel cage instead of a common flat disk blade. A rotary compressor shouldn't be hard to start. We replaced our 120 gallon shop air compressor last year and went rotary. You can be standing next to it running and still carry on a normal conversation with someone. Rotaries are sweet.

I adjusted a trim pot on the inverter to allow for higher input voltages. That inverter has always complained about anything above 14v, even thought the book says 15v. I have a supercap battery on the way from eBay. I watched someone on youtube start a V8 engine with that particular battery. I'm thinking a supercap should be standard on all inverters that are vehicle mounted. The best place for the inverter is next to the battery, but that means under the hood with heat and vibration. It also exposes the alternator to the surges as the car battery ages.

The next hurdle will be finding a way to mount the a/c so it can't be vandalized when the vehicle is parked. In the hood I live, if it stuck out from the side someone would bend the fins or poke the tube and huff the freon.

oztules

  • Hero Member
  • *****
  • Posts: 1477
  • Country: aq
  • Village idiot
Re: Reactive boost for mosfet inverter
« Reply #15 on: April 01, 2015, 02:26:50 PM »
I still think that Joe is correct and that is the o/load. If it was a PTC, the unit would never run at all, as it is in the common leg, not the start leg... it is an over temp/overload thinggy.

A PTC in a fridge will go high resistance very very quickly, and would certainly not allow the fridge to run if placed there if it was the normal refrigeration ptc.. usually a ceramic looking  disc in a plastic/bakelite like box.

The thing you describe is a o/load using the bimetal disk and heater.

Perhaps isolate the fan on start up, and this may get you over the line.

............oztules
« Last Edit: April 01, 2015, 02:32:08 PM by oztules »
Flinders Island Australia

oztules

  • Hero Member
  • *****
  • Posts: 1477
  • Country: aq
  • Village idiot
Re: Reactive boost for mosfet inverter
« Reply #16 on: April 01, 2015, 02:47:28 PM »
Something like this

Flinders Island Australia

bart

  • Full Member
  • ***
  • Posts: 200
  • Country: us
Re: Reactive boost for mosfet inverter
« Reply #17 on: April 01, 2015, 05:42:17 PM »
   The OLP being a klixon is for high temp. shut down of the compressor. Have used lots of hard start kits and have rarely had any problems with them, but those have been on split phase, residential type condensers. These units are typicality fairly robust and will handle starting components going bad. The window units are probably import and who knows how well they are made. Me, I'd put the hard start on.

dnix71

  • SuperHero Member
  • ******
  • Posts: 2513
Re: Reactive boost for mosfet inverter
« Reply #18 on: April 01, 2015, 07:24:18 PM »
In case you were wondering, the switch in between the thermostat and the capacitor with 2 terminals is the "energy saver" switch. It causes the fans to run all the time or forces them to cycle with the compressor, so I can use it to spin up the fan fully before turning the temp dial switch to start the compressor. There is a main power off/on toggle switch also. This being an economy unit, does not have a remote control. My inverter does have provision for a remote, but I have never used it.

I just ordered a hard start kit/relay and cap http://www.ebay.com/itm/201307414615
« Last Edit: April 01, 2015, 07:30:50 PM by dnix71 »

TDC

  • Jr. Member
  • **
  • Posts: 75
  • Country: us
Re: Reactive boost for mosfet inverter
« Reply #19 on: April 04, 2015, 03:15:22 PM »
I'm confused, is the vans alternator capable of powering the inverter load?

dnix71

  • SuperHero Member
  • ******
  • Posts: 2513
Re: Reactive boost for mosfet inverter
« Reply #20 on: April 04, 2015, 06:30:07 PM »


The van is just barely capable or the inverter' max load. The van has factory air, hence a better alternator than usual, but it was not installed properly and it requires a lift to access the idle pulley. I won't mention the name of the licensed shop that botched that job, but I no longer do business with them. The alternator should have either had a smaller pulley or be wound to produce power at idle rpm. Even without the factory a/c load, at night it sometimes stops working because it isn't spinning fast enough. I have dash gauges and know to pay attention. Either a gentle foot on the gas or put it in neutral and rev it a bit will get the charging system goin again.

I am considering going with LED headlights to reduce the vehicle's electrical load.
« Last Edit: April 07, 2015, 12:22:25 PM by DamonHD »

OperaHouse

  • Hero Member
  • *****
  • Posts: 1308
  • Country: us
Re: Reactive boost for mosfet inverter
« Reply #21 on: April 07, 2015, 05:41:57 PM »
So it really worked.  I didn't give that $10 capacitor booster much of a chance.  Is it just two wires in parallel with the cap or three wire?

dnix71

  • SuperHero Member
  • ******
  • Posts: 2513
Re: Reactive boost for mosfet inverter
« Reply #22 on: April 07, 2015, 08:54:59 PM »
OperaHouse it's just 2 wires and it's exactly like Oztules showed, jumpered across the R and S terminals. With that kit the inverter chirps twice but the compressor has already spun up. The fans still take a long time but high speed is 40 watts and low is 30 watts. Might as well leave it on high.