Small garden tractors, such as ATR5-b0, are not designed to support heavy electrical loads.  Their engines contain integrated alternators to power headlights and maintain the starter SLA, with an output around 40 watts.  This might be sufficient for logic electronics, but actuation and lighting add at least 50 to 100 average watts.  Even optimally efficient CFL or LED headlights will add 35 watts, and the physical actuators (especially steering) will add another 30 to 60W average.  [For instance, pneumatic actuation at 90psi/.06CFM would require a 300W compressor at a 10% duty cycle, or 30W average.]

Automotive alternator

ATR5′s existing 14hp (10.4kW) combustion engine could be retrofitted with an automotive alternator ($100 – $120 with necessary parts) to deliver power in the kilowatt range at the cost of a couple horsepower. Though it is tempting to avoid the complication of a second independent energy source in this way, an external alternator would require extensive chassis modification (positioning, belt tensioning, ground clearance) and an investment that must be customized and dedicated to each new chassis. Moreover, ATR5 would experience power fluctuations in response to suddenly varying drive loads [before the engine throttle could be modulated to compensate]. A modular, decoupled energy source is the general solution.

DieHard 27582

As discussed previously [see Power Sources], there are two feasible options to store and deliver energies in this range: SLA batteries and combustion engines. If average energy requirements are around 100W, a deep cycle SLA ($130 + $40 inverter) can provide the necessary power for 10 hours at 4.4 cents an hour (assuming Peukert constant of 1.4, $0.15/kWh utility rate, 100% efficient charger, 80% charge-to-discharge efficiency, 100% efficient inverter), followed by a charge cycle of at least an hour.  [The efficiency will decrease if the power consumption is concentrated in short bursts. See also DieHard 27582 SLA unverified capacity function.]  If consumption is closer to 1000W, and assuming a minimum 3-hour consecutive runtime (75% runtime duty cycle, assuming 1-hour simultaneous charge), the initial cost would increase to $450 (three $130 SLAs + $60 kW inverter) and the ongoing cost to $0.71 per hour. [Running at 1000W from a single SLA would increase energy costs to $1.10 an hour.]

An efficient 2-stroke gasoline generator, on the other hand, can supply 1000W at an initial cost of only $140 with a 90%+ duty cycle (just an on-site gas refill every four or five hours) and 30% cheaper per output than three SLAs ($0.50/kWh). The 100W performance is much worse — $0.22/kWh, five times the single SLA cost — due to the gas generator’s idle fuel consumption offset of 0.0739 gallons/hr ($0.19/hr).  [See ETQ TG1200 fuel consumption.]

To the left is a graph of the total energy output between recharging/refueling (solid lines) and the cost per delivered energy (dotted lines) for SLAs and gas generators. The single SLA cost/kWh intersection occurs around 410W (not considering initial costs), where both technologies cost $0.77/kWh delivered. For multiple SLAs in use simultaneously, the intersection point moves to the right: 550W for two SLAs, 650W for three.

Thus far, I have assumed that average power and instantaneous power are equal. This is often unrealistic. For instance, a pneumatic air compressor might operate at 500W, but at a 50% duty cycle (triggered by its internal pressure switch with hysteresis) for an average consumption of 250W. Since generator fuel consumption is linearly related to power output, generator efficiency is unaffected by duty cycle. From the generator’s perspective, all that matters is the average power over a complete duty cycle period [and that the instantaneous power output is less than the rated maximum, e.g. 1000W for the TG1200]. SLA efficiency, on the other hand, deteriorates nonlinearly with respect to output power. An SLA operating at a 50/50 duty cycle, for example, will be as inefficient (and cost just as much per kWh output) as an SLA operating at twice the average output power at a 100% duty cycle. The 50/50 duty cycle single-SLA cost/kWh intersection occurs at 310W. Graphed below.

If the duty cycle has a long period — i.e. a 24-hour 50/50 duty cycle (12-hour workday) — common sense suggests that generator $/kWh can be reduced by artificially manipulating the duty cycle to increase the average power: shutting down the generator during the off-cycle and restarting before the on cycle. ATR5 utilizes this trick, with human assistance [as the onboard TG1200 has not been modified for electric start]. This does not work for short duty cycles or if a small amount of output power (e.g. logic electronics) is still required in the off portion of the cycle; ATR5 utilizes low-power battery backups during the off-cycle.

The calculation/graphing spreadsheet is available for download [OpenOffice ODS]. Duty cycle, SLA count, and battery/generator specifications can be modified to test capacity and cost ramifications. For example, upgrading to an AGM SLA with a Peukert constant of 1.25 moves the cost intersection to 830W. The “Lifespan hardware cost” field integrates initial cost into the $/kWh calculations; the relatively high cost of AGM SLAs [even assuming a 1000-cycle lifespan] bumps the intersection back down to 300W. The download is initially configured to reflect the lifespan hardware costs of the 27582 and TG1200.