No announcement yet.

Idle back inverter genset project

  • Filter
  • Time
  • Show
Clear All
new posts

  • Idle back inverter genset project

    I recently posted a pic of a genset I put together using an old 2 kilowatt Winco alternator and a lawnmower engine. It worked fine but wasn't very fun to use. The Winco alternator weighs 65 lbs alone and the entire unit was over 100 lbs. That's a bit much for me to cart around in casual use.

    The other day the local Canadian Tire store had 1000 watt inverters on sale for $100. I bought one and decided to rebuild the genset as an inverter system with full idle back electronic servo controlled throttle. The inverter is modified sine wave and is rated at 1000 watts continuous, 1200 watts for five minutes and 2000 watts for 15 seconds for starting up to 1/2 hp motors.

    To ensure that the inverter would be able to pull sufficient surge current I included a small lead acid utility battery in the design. This also allows for short term use without starting the motor. It does add some weight but not that much. The entire system is still much lighter than it was with the Winco unit.

    I machined up a flywheel for the lawnmower engine to smooth out the operation and improve starting. I happen to have a reliable source of used aluminum flywheels, already spin tested, that started life as diamond grinding wheels. In this engine the crank is very slightly bent so I rebalanced it using the flywheel by drilling out some weight on the rim. The runout is only about 20 thou so not too bad and it runs smoothly now. When I have more time later I will take it apart and straighten the crank.

    I salvaged the 12 volt alternator from our old Toyota Corolla that happens to be lying back in the trees on my property. It's a 55 amp unit which makes it a good match to the inverter and the 3.5 hp engine. The frame is welded up from thin wall tubing so doesn't weigh much. A few more parts from the junk box and I built a control panel. It has a volt meter to monitor the battery condition and an amp meter to keep an eye on the inverter load.

    {next slide please}

    Mounted below the inverter is the electronic control for the throttle servo. There is also a handy cigarette lighter socket on the panel for aux power. I will also be including a couple of posts to connect jumper cables for vehicle boosting/charging. One switch on the panel turns on the charging system and the other is a kill switch for the motor. The kill switch is a three position toggle and has stop, run and start positions. In the start position it commands the servo controller to hold the throttle open the right amount for starting.

    The electronic servo control is a simple analog comparator using a LM358 op amp. It has a zener diode reference voltage on the non inverting input and a potentiometer resistance divider on the inverting input. That pot acts as a setpoint adjustment for rpm. The op amp gain is controlled by another pot that gives variable negative feedback. The output of the amp drives a TIP120 power darlington transistor that actuates the servo motor on the throttle.

    {slide four}

    A key part in this is the small servo used to operate the throttle. All it is is an old 24 vdc motor that was originally an ordinary gear motor unit until it broke a pinion. I took off the gearbox and soft silver soldered an actuator arm to the stub shaft. It draws no more than 300 ma on 12 volts when stalled so it can easily withstand 100% duty cycle operation in the stalled condition. It actuates the throttle against the small return spring on the throttle plate shaft.

    Most people don't know that small dc motors can be used as partial rotation servos with very good response time. I have done a lot of work in the past with analog controls and first studied analog computers back in the 60s. Many of the earlier Xerox machines I worked on had extensive analog controls before the switch to all digital. This is a good example that supports the old bumper sticker that used to be seen in Silicon Valley that said "One Op Amp is worth a Thousand Gates".

    Designing a control such as this is not a straight forward exercise. Even though the circuit is very simple the mechanical system is an integral part of the feedback loop. The parameters of the feedback system must be carefully tuned to prevent oscillations or runaway conditions. It doesn't matter if the controller is analog or digital since the same characteristics must be emulated in a digital system. The controller must be matched to the entire feedback loop which includes the response time of the engine to acceleration, load changes and deceleration.

    In this system the engine is regulated to provide as constant a voltage as possible from the alternator. Unlike an ordinary governor that tries to regulate the rpm of the engine in this system the rpm is allowed to change to whatever is needed to supply the required output. The controller simply measures the system voltage and commands the engine to accelerate or allows it to decelerate as needed to maintain the battery in a charging state of 12.5 volts or higher.

    The unit will supply up to about 200 watts of power with the engine at idle. With a 500 watt load it still doesn't run at much over 1000 rpm and doesn't really start working until the load nears 800 watts. It keeps up just nicely at 1000 to 1100 watts.

    The big advantage is that if all you need is a few lights the genset runs at or near idle and consumes much less fuel than an old style system that must run at full rpm all the time in order to maintain output voltage and frequency regardless of the actual load. This system operates in demand mode all the time and when the load is light it automatically throttles back to only what power level is needed to meet that demand. Such a system is much more efficient, quieter and longer lasting than a constant rpm system. It uses far less fuel when used to power loads that vary a lot.

    I made a short video that shows the response of the control system to load changes. I used the light stand to place varying loads on the genset. In the video it starts with idling on a 100 watt load. I then switch on a 500 watt lamp and the controller immediately commands the throttle to increase power. Since the battery is providing surge capability to the inverter no complex systems are required to ensure against an under voltage condition during rapid increases in load. I then switch on another lamp for a total load of 1100 watts. Once again the control commands the throttle, this time to full throttle. Then I begin removing the loads back down to 100 watts. I again switch on a 500 watt load for only a couple of seconds and off again to demonstrate the controller's response to a sudden short load increase which it handles with virtually no oscillation when the load is removed.

    The video is here in Windows Media Player format. It's less than one megabyte.

    Also here in mp4 format:

    Note that the crummy tinny microphone on the camera makes the engine sound like it is running with a bad knock since it can't pick up any of the bass frequencies. It does a good job of picking up only the mechanical noise of the engine. That is partly due to the muffler I constructed that makes the engine quiet enough that you can hold a conversation in ordinary tones right beside it.
    Free software for calculating bolt circles and similar: Click Here

  • #2
    Very interesting Evan. Nice work on that.


    • #3
      I forgot to mention that my total out of pocket cost to build the unit was around $150. That doesn't include the worth of things such as the gauges that have been lying around for a couple of decades...
      Free software for calculating bolt circles and similar: Click Here


      • #4
        Very ingenious design Evan, the packaging of the various subsystems is very efficient as well.

        There is a lot of demand for a unit such as the one you have built. With the large number of complaint's for users of the typical 3600 rpm, 2 pole generators being mostly related to high noise, high fuel consumption, and short life, I would think you would have a marketable product there. With a little refinement I'm sure you would have no problem selling such a unit.

        Good project Evan, I like that a lot, and thanks for sharing the details before it goes into production.
        Home, down in the valley behind the Red Angus
        Bad Decisions Make Good Stories​


        • #5
          Just curious, did you try the setup with the engines normal governor?


          • #6
            The normal governor won't do the job. It is trying to maintain the same rpm which is not what is needed. This is especially the case in a Tecumseh engine as it uses a flyweight mechanical governor. It would be fine for an old style genset where it needs to maintain 3600 rpm.
            Free software for calculating bolt circles and similar: Click Here


            • #7
              That's some amazing work Evan. Must be big fun to have the electronics skills needed to create such an elegant solution to sneakily complex problem...from stuff you have lying around. Me, I can't even find my good junk anymore, much less find a use for it.

              "Accuracy is the sum total of your compensating mistakes."

              "The thing I hate about an argument is that it always interrupts a discussion." G. K. Chesterton