A kilowatt-hour number on its own doesn’t tell a plant manager much. “This machine used 340 kWh last week” isn’t a decision anyone can act on. “This machine’s energy cost 6 cents per part, up from 4 cents last month” is. The gap between those two statements is one division, attributed kWh over good parts made, and it turns a utility abstraction into a number that sits next to labor cost and material cost on the same page.
What the Energy page actually prices
Spall’s Energy page reads a metered branch, one or more machines behind a shared power meter or clamp-on sensor, and reports three things for the window you’re looking at: total kWh and dollars for the branch, that same total split down to each machine, and a ranked list of machines burning power while stopped. None of those three numbers is a cost per part. The page prices electricity, not output. Getting to cost per part means pulling one more number from a different screen, good parts produced in the same window, and doing the division yourself.
That’s not a gap to apologize for. Cost per part depends on what you’re comparing power against, cycle-weighted output, shift output, a single job’s run, and no single dashboard tile can guess which one you mean. The Energy page gives you the numerator. Performance or the shift handoff gives you the denominator.
Getting the numerator right: attribution, not a flat split
When a branch covers more than one machine, the page doesn’t split the meter’s total evenly by default. It weights by run state, a machine that ran 70% of the window gets roughly 70% of the branch’s attributed kWh, with idle machines still carrying a small standby share rather than a hard zero, because a machine sitting pressurized and powered isn’t drawing nothing. If a machine has no run or stop signal configured, the split falls back to even, and the page says so in plain text next to the number instead of presenting a weighted-looking figure that was actually a guess. Check that label before you trust an attributed number for a specific machine. An even split on a three-machine branch means each machine’s kWh is off by however far its real duty cycle sits from a third.
A machine with its own dedicated submeter skips attribution altogether, its share comes from its own meter, not a formula, and that’s the most accurate path when the hardware exists. Rated-kW weighting is a third option, useful when nameplate capacity is known but run-state telemetry isn’t, though it assumes every machine draws its full rated load whenever it’s counted as on, which is rarely exact.
Getting the denominator right: which parts count
Good parts, not gross parts, is the number to divide against. A machine that ran a full shift but scrapped a third of what it made didn’t get a third-cheaper energy cost, it spent the same power making parts that never shipped. Divide attributed dollars by good output only, the same first-pass yield discipline that already governs every other per-part cost figure, and a rising scrap rate will show up as a rising energy cost per part even if the machine’s power draw never changed. That’s not a bug in the math, it’s the real cost of the scrap, paid twice, once in material and once in the power it took to make the reject.
Pull the good-part count for the same window and site as the Energy page’s attribution, from Performance or the per-shift production view. Matching windows matters more than it sounds, a week of energy data divided by a single shift’s part count produces a number with no relationship to reality.
Why idle energy is the part everyone underestimates
The idle-power finder ranks machines by dollars burned while stopped, and that number has a direct, sometimes uncomfortable relationship to cost per part. A machine that sits powered and pressurized between jobs is spending money whether or not another part ever comes off it. Spread that fixed idle cost over fewer parts, because the shift ran short or a changeover ran long, and cost per part climbs even though nothing about the machine’s actual cutting or forming energy changed. Spread it over more parts, a longer run with fewer changeovers, and the same idle draw gets diluted into a smaller per-part number.
This is the practical reason idle power belongs in the same conversation as cost per part, not off in its own sustainability report. A machine that never gets shut down between jobs is paying a fixed energy tax on every single part it makes, and the size of that tax depends entirely on how many parts share it.
A worked example
Take a machine attributed 40 kWh for an eight-hour shift, at a set rate of 12 cents per kWh, for $4.80 in energy that shift. It made 240 good parts. Energy cost per part: $4.80 divided by 240, two cents. Now take the same machine on a shift where a tooling problem cut output to 80 good parts, same 40 kWh because it still sat powered through the delay. Energy cost per part: six cents, three times higher, without the machine’s power draw changing at all. The energy number didn’t get worse. The output number did, and cost per part is what makes that visible in dollar terms instead of burying it in a percentage.
Where the rate itself comes from
Every dollar figure on the Energy page cites where its rate came from, set directly on that machine, inherited from a parent node, or the app’s default when nothing’s been configured. A cost-per-part figure built on an unconfigured default rate is still useful for spotting trend and relative comparison between machines, but the absolute number is only as real as the rate behind it. Set a real rate in Rates and Costing for any machine whose energy cost per part is going in front of a plant manager as a dollar figure, not a placeholder.
Quick recap
- Cost per part isn’t a single tile. Divide the Energy page’s attributed dollars by good parts from the same window and site.
- Attribution weights by run state by default, with idle machines carrying a standby share, not zero, check the label when it falls back to an even split.
- Use good parts, not gross output, in the denominator. Scrap pays for its own energy twice.
- Idle dollars are a fixed cost spread across however many parts actually got made in the window, fewer parts means a higher per-part tax.
- A cost-per-part figure is only as real as the rate behind it. Set one in Rates and Costing before treating the number as a dollar fact.