Grams, metres and why the two never quite agree
Slicers report filament usage in both weight and length, and the conversion between them runs through density and diameter. One metre of 1.75 mm filament is a cylinder 0.175 cm across and 100 cm long, which is 2.405 cubic centimetres. Multiply by PLA's nominal 1.24 g/cm3 and you get 2.98 grams per metre, which is why a 1 kg spool of 1.75 mm PLA holds roughly 335 metres. The same spool in 2.85 mm holds about 126 metres, because the cross-section is 2.65 times larger.
The density is the loose part of that chain. Manufacturers publish figures for the base polymer, but pigment loading, filler and blend changes shift it. Natural PLA sits near 1.24, heavily pigmented and glitter-loaded PLA runs higher, and foaming or lightweight PLA can drop well under 1.0 once it expands. If your slicer reports weight, none of this affects the money. If it reports length, a five percent density error is a five percent cost error.
| Material | Nominal density (g/cm3) | Common range | Grams per metre at 1.75 mm |
|---|---|---|---|
| PLA | 1.24 | 1.17 - 1.30 | 2.98 |
| PETG | 1.27 | 1.23 - 1.38 | 3.05 |
| ABS | 1.04 | 1.02 - 1.08 | 2.50 |
| ASA | 1.07 | 1.05 - 1.10 | 2.57 |
| TPU 95A | 1.21 | 1.14 - 1.24 | 2.91 |
| Nylon PA6 | 1.14 | 1.01 - 1.15 | 2.74 |
| Polycarbonate | 1.20 | 1.19 - 1.22 | 2.89 |
| Carbon-filled PLA / PETG | 1.15 | 1.05 - 1.30 | 2.77 |
Those are the common nominal values, not a specification. Check the spool you are actually feeding. Nylon in particular covers a family: PA12 sits near 1.01 and PA6 near 1.14, and vendors are often vague about which one is in the box.
The slicer estimate is a floor
Slicers compute filament usage from the extrusion paths they generated, which means it is an exact answer to a question slightly different from the one you asked. Real usage tends to land a few percent above it: purge and prime lines, the blob wiped off before the first layer, ooze during travel, retraction that does not perfectly reverse, and filament that is 1.73 mm where the slicer assumed 1.75. Three percent is a reasonable default allowance and the field is there so you can change it. If you weigh finished prints against the estimate on your own machine, use your own number and set the allowance to whatever you measured.
Power draw is the number people get most wrong
The figure that belongs in the wattage field is the average draw across the whole print, not the rating on the power supply. A 350 W supply spends most of its life delivering a fraction of that, because the heated bed is on a duty cycle once it reaches temperature and the hotend draws almost nothing by comparison. A common open-frame machine printing PLA at a 60C bed averages somewhere around 80 to 130 W. Push the bed to 100C for ABS and add an enclosure that has to stay warm and the average climbs a long way, into the hundreds. If you want the real number, a $15 plug-in energy meter left on for one full print settles the argument in a way no estimate will.
Even when the wattage is right, electricity is usually the smallest line here. A 14 hour print at 100 W is 1.4 kWh, which is about 24 cents at 17 cents per kWh. That is worth knowing and rarely worth optimising.
Depreciation and failures are assumptions, not measurements
Two lines on this page are modelling choices and you should treat them as such. Machine depreciation divides the printer's price by an assumed number of print hours before it is worn out or replaced. Three thousand hours is a plausible working figure for a hobby machine, but it is a guess about the future, not a property of the printer. Set it to whatever horizon you actually plan around, or leave the machine cost blank and price the print without it.
The failure allowance divides the machine-side cost by one minus the failure rate. At an 8 percent failure rate that is a division by 0.92, an uplift of about 8.7 percent. The assumption buried in that is that a failure consumes the entire print, all of the filament and all of the time, which is the conservative end of reality. Most failures announce themselves in the first layer or the first hour, and a print that fails at 4 percent completion costs you 4 percent of the plastic plus the annoyance. If your failures are mostly early, the true uplift is well under what this shows. If they are late-stage warping and layer shifts at hour eleven, this is about right. Your time is deliberately not multiplied by the failure factor, because the failed attempt did not need sanding.
What this is not
This is the cost of producing a print, which is a different question from what to charge for one. Pricing has to cover the prints nobody buys, the design time, the failed customer revisions, packaging, shipping, the platform's cut and the fact that a maker who charges cost plus a small margin will run out of enthusiasm before they run out of filament. Cost is the floor under a price. Where you put the price above that floor is a business decision this page has no opinion about.
Questions people ask
How many metres are in a 1 kg spool of filament?
For 1.75 mm PLA at 1.24 g/cm3, about 335 metres. For 1.75 mm PETG, about 328 metres, because PETG is denser. For 1.75 mm ABS, about 400 metres, because ABS is lighter. In 2.85 mm the numbers drop by a factor of 2.65: roughly 126 metres for PLA, 124 for PETG and 151 for ABS. Set the mode to length in this calculator and it will show you the metre figure alongside the grams, using whichever density and diameter you selected. Note that a spool sold as 1 kg means 1 kg of plastic, not 1 kg on the scale; the plastic or cardboard core adds another 150 to 250 grams.
Does the density really matter for cost?
Only if your slicer reports length rather than weight. If you enter grams, density is used purely to show you the equivalent metre figure and it never touches the money. If you enter metres, density is directly proportional to the answer, and a spool of heavily pigmented PLA that is genuinely 1.30 rather than 1.24 will cost you five percent more than this page says. The cheap fix is to enter grams whenever your slicer offers them, which every mainstream slicer does.
Should I include machine depreciation for a hobby printer?
If you are printing for yourself, arguably not, because the printer was bought as a thing you wanted rather than a cost of production, and charging your own prints for it does not change any decision you are going to make. If you are selling prints or quoting for someone else, absolutely include it, because the machine wears out and the replacement has to come from somewhere. The middle case is deciding between printing something and buying it, where depreciation belongs in the comparison. Leave the printer cost field blank to drop the line entirely.
Why is my print more expensive than the filament calculator on the spool vendor site?
Because most of those calculate filament and nothing else. Filament is typically half to two thirds of the true cost of a print once you count power, wear and the attempts that failed. A 145 gram print on a $22 spool is $3.19 of plastic; add 14 hours of power, a share of a $400 machine over 3000 hours, a nozzle budget and an 8 percent failure rate and you are past $6 before anyone has touched it with sandpaper. Neither number is wrong. They answer different questions.
What failure rate should I use?
Count your own for a month rather than guessing. Rates vary enormously with what you print: simple flat parts on a well-tuned machine can run under 2 percent, while tall thin models, bridging-heavy geometry, ABS without an enclosure or a first layer you did not watch can push well into the double digits. Eight percent is a placeholder here, not a recommendation. The other thing worth tracking is when failures happen, because an early failure costs almost nothing and a late one costs everything, and this calculator treats them all as late.