PrintFarmer ERP
Tell us you’re interested

Organising a print queue across several printers: orders, parts and plates

Run a multi-printer queue by tracking orders, parts and plates as separate records, planning plates from parts still owed, and recording every plate’s result.

Short answer

Track three things separately: the customer order, the parts it needs with a quantity each, and the plates you print, where one plate can carry parts from several orders. Plan plates from the parts still owed, check planned quantities against what is needed before you print, and record each plate’s result as good and failed counts per part, so failures go back into the queue on their own.

Key takeaways

  • One row per order breaks as soon as a plate carries parts from two orders or an order needs parts from two printers.
  • The part is the link: an order owns parts, a plate allocates quantities of parts, and the shortfall per part is what the queue shows.
  • Record results per part, not per plate: 9 good and 1 failed returns exactly 1 part to the queue.
  • The published figures of one person running 20 to 25 printers are Formlabs and Markforged describing their own SLA and composite machines; count your own plate changes instead.
  • Answering ‘where is my order’ is a lookup on the order’s parts, not a walk to the printers.

To run a print queue across several printers, track three things separately: the order (what a customer asked for), the parts it needs with a quantity each, and the plates you actually print, where one plate can carry parts from several orders. Plan plates from the parts still owed, check planned quantities against what is needed before you press print, and record each plate’s result as good and failed counts per part. Everything else in this article, the spreadsheet tabs, the daily routine, answering customers, follows from keeping those three apart.

Why does one row per order stop working?

#

With one printer, a row per order is enough: the order is printing or it is not. It breaks at the second printer, in two ways.

  • One order, several printers. An order for 40 keychains and 2 enclosures is split across machines. The row says “printing”, but one printer finished, one failed and one has not started.
  • One plate, several orders. Three small orders in red PLA fit on one plate. The plate has no row of its own, so when it comes off, nobody can say from the sheet which parts belong to whom.

Operators on the Bambu Lab forum described this gap in 2023: one owner of 13 printers asked to “organize our print history by project, rather than just chronologically”, and another with 30 P1Ps wrote that “Sending print to multiple printer is now become a task” (thread).9

Orders, parts and plates: what is each one?

#
Record What it is Key fields Links to
Order One customer’s request Order number, customer, due date, ship or collect Has many parts
Part One distinct thing to make, for one order Quantity needed, material, colour, nozzle, file Belongs to one order; appears on many plates
Plate One print job on one printer’s bed Plate code, printer, status, file name Carries quantities of many parts

The part is the link. An order owns its parts. A plate does not own anything; it allocates a quantity of each part it carries. A part’s progress is the sum of what its plates produced.

This is the model this article uses, and the structure is not unusual. Printago’s plate merging, for example, offers a scope of “Same order only” (the default) or “Any compatible jobs”, the second letting jobs from different orders share a plate (Printago docs).7

What should the queue spreadsheet look like?

#

If you are building this in a spreadsheet, use two tabs, not one wide sheet.

Parts tab (one row per part):

Part ID Order Part Needed Planned Good Still to plan Material Colour Nozzle Due
P-201 ORD-0101 Keychain “ASHA” 40 0 0 40 PLA Red 0.4 28 Nov
P-202 ORD-0103 Keychain “RAVI” 25 0 0 25 PLA Red 0.4 2 Dec
P-203 ORD-0102 Cable clip 12 0 0 12 PETG Black 0.4 30 Nov
P-204 ORD-0102 Enclosure base 2 0 0 2 PETG Black 0.4 30 Nov

Plate lines tab (one row per part per plate): plate code, printer, status, part ID, planned, good, failed.

“Planned” on the parts tab is the sum of planned quantities on plates that have not come off yet. “Good” is the sum of good counts from plates that have. “Still to plan” is Needed minus Good minus Planned. When a plate comes off, its quantities leave Planned and its good counts join Good, so anything that failed reappears in Still to plan without anyone retyping it. Filter the parts tab to “Still to plan greater than zero” and sort by due date: that filtered view is your queue.

How do I batch parts onto a plate?

#

Group the queue by what has to match on one bed, in this order:

  1. Material and colour. One spool, or one set of AMS slots, per plate.
  2. Nozzle and layer settings. A 0.6 mm nozzle job does not share a plate with a 0.4 mm one.
  3. Deadline. Fill the plate with the earliest-due parts first, then top it up with later ones in the same material.
  4. Sorting cost. Mixing orders saves a plate but adds sorting. Mix freely when parts are visibly different (named keychains); be careful with identical parts for two orders unless you count them out straight away.

Worked example: three plates, one partial failure

#

Assume the slicer fits 30 keychains on a plate.

  • PL-031: 30 of P-201. P-201 now has 30 planned, 10 still to plan.
  • PL-032: 10 of P-201 and 20 of P-202. P-201 is fully planned (30 + 10 = 40). P-202 has 20 planned, 5 still to plan.
  • PL-033: 12 of P-203 and 2 of P-204, black PETG, for ORD-0102.

That leaves 5 of P-202 waiting for the next red plate.

How do I avoid printing too many or too few?

#

Before a plate is released, compare each part’s good and planned quantities together with what it needs.

  • Over-planning: someone puts 30 of P-202 on PL-032 instead of 20. P-202 would then have 30 planned against 25 needed, 5 extra. Sometimes that is deliberate (spares for a part that often fails), so allow it, but write down why.
  • Under-planning: a part with a quantity still to plan and an early due date that is on no plate. The filtered parts tab catches it; a plate list never will.

In a spreadsheet, a conditional format on “Good plus Planned greater than Needed” does the first check. The second is the queue view itself.

Which job goes on which printer?

#

A short rule list covers most decisions:

  1. The printer has the right material and colour loaded, or it is quicker to change than to wait.
  2. The nozzle and bed size fit the plate.
  3. Long plates go on at the end of the day, so they finish when someone is there to unload.
  4. Earliest due date goes on the first free printer.
  5. A printer that failed its last plate gets something cheap and short until you know why.

Some tools do this matching for you. SimplyPrint matches queue items to printers by build size, temperature, nozzle, custom tags, material and colour, and its AutoPrint starts the next matching item once a printer is free and its bed is clear (SimplyPrint help).4, 5 Printago’s queue matches jobs on material, colour, tags and availability, where available includes being marked “clear & ready for next print” (Printago docs).6 Bambu Farm Manager, a free Windows application for Bambu Lab printers that runs on the local network, lists “Smart queuing: Assign jobs based on printer availability” (Bambu Lab, July 2025).8 These descriptions are from each vendor’s documentation as read in November 2026.

So the matching itself can be automated. What remains: unless you fit an automatic bed-clearing system, a person clears the bed or confirms it is clear, and a tool only weighs a customer’s deadline or a printer’s bad week if someone told it.5, 6 Rules 3 to 5 still need a person.

How do I label what comes off the bed?

#

Every plate gets a code (PL-032), and every plate gets a printed slip listing its lines: part, order, planned quantity. When it comes off:

  1. Count good and failed per part against the slip.
  2. Bag or tray the good parts by part, with the plate code and order number on the bag.
  3. Put failed parts in a separate bin; do not throw them away until the count is recorded.

With named keychains, sorting is obvious. With identical parts for two orders, they are interchangeable until bagged, so count them out by order at the bench rather than trying to track individual pieces.

How should I record a print result?

#

Record per part, not per plate. A plate has three possible outcomes, and each one does something different to the queue:

Result What you record What happens next
Good Good = planned for every part Parts move toward packing
Partial Some good, some failed, per part Good parts move on; failed quantity returns to “still to plan”
Failed Failed = planned Every part on it returns to the queue; check the printer first

In the worked example, PL-031 comes off with all 30 good. PL-032 comes off with P-201 at 9 good and 1 failed, and P-202 at 20 good. P-201 now has 30 + 9 = 39 good against 40 needed, so exactly 1 keychain reappears in the queue and goes onto the next red plate with the 5 remaining of P-202. Nobody had to remember it.

What a failure costs is a separate question, covered in costing failed prints. The filament it used is covered in tracking filament by weight.

What does the daily routine look like?

#

At the desk, start of day

  1. Enter new orders and their parts.
  2. Filter the queue: still to plan greater than zero, sorted by due date.
  3. Build today’s plates, check planned against needed, print the slips.

At the printers, each time one finishes

  1. Unload, count against the slip, bag by part.
  2. Record good and failed counts straight away.
  3. Load the next plate for that printer and mark it printing.

At the desk, end of day

  1. Replan anything that failed.
  2. Put long plates on for overnight.
  3. Check which orders are now fully good and ready to pack.

How many printers can one person run?

#

The figures you will find are vendor accounts. Formlabs writes that “one person can efficiently manage a group of 20+ printers” in a section about a fleet of its SLA printers (Formlabs).1 Markforged described its own 25-printer farm of composite FDM machines, where one person removes parts and sets printers up for the next job (Markforged, 2018).2 SimplyPrint’s guide, based on a survey of 22 farm operators, says the constraint is “operator attention”, and its cost table lists a second operator at 25 printers (SimplyPrint).3 None of these is a hobby-class FDM shop doing custom orders.

Work it out from your own plate changes instead. Assume 3-hour plates, a 9-hour working day, and 6 minutes per plate change including counting and recording. Each printer turns 3 plates during the day plus 1 overnight that is unloaded the next morning: 4 changes per printer per day.

Printers Changes a day Minutes at 6 each Hours
10 40 240 4
15 60 360 6
20 80 480 8

At 20 printers, plate handling fills 8 of the 9 hours, with nothing left for planning, packing or customer messages. Longer plates, faster changes or a second person move that line.

How do I answer “where is my order?”

#

From the sheet, not the printer room. Filter the parts tab by the order number and read needed, good and planned, then look up the plates still carrying its parts. “38 of your 40 are done, the last 2 are on a plate that finishes at 4 pm” is two lookups. If you sell over chat, this guide to Instagram and WhatsApp orders covers the rest of that thread.

Where PrintFarmer fits

#

PrintFarmer keeps orders, parts and plates as separate records in the shape described above. The production queue lists the parts still outstanding, and selected rows carry into the plate builder. Planning more of a part than it needs is flagged, with an override that needs a reason. A part can need several filaments. People on the floor get a Ready plates view laid out for a phone, and a plate’s result is recorded as good and failed counts per part; a plate with any failure goes to reprint planning, and the failed quantities come back onto the production queue. The printers screen shows status, active plate, queue depth and queued hours, all from what people record rather than from the printers.

It does not talk to printers. As the plate builder says, “Assignment is manual; nothing is scheduled for you.” There is no printer telemetry, no sending jobs to printers and no auto-nesting, so it sits alongside a farm manager rather than replacing one. Comparing farm tools and ERP versus farm manager cover that split. Tell us you’re interested. It is not launched yet.

Frequently asked questions

How do I keep track of which printed parts belong to which customer order?

Give each order a list of parts with a quantity needed, and give each plate a code and a list of which parts and how many it carries. When a plate comes off, bag its output by part with the plate code on the bag, then sort into orders by count at the bench.

What columns should a 3D print job tracking spreadsheet have?

Use separate tabs. Parts: part ID, order number, part name, quantity needed, quantity planned, quantity good, quantity still to plan, material, colour, nozzle, due date. Plates: plate code, printer, status, and one row per part on the plate with planned, good and failed counts.

Can parts from different customer orders go on the same build plate?

Yes, if they share material, colour, nozzle and print settings and fit the plate. The cost is sorting afterwards, so only mix orders when the parts are easy to tell apart or the plate record says exactly how many of each it carries.

What should happen in the queue when a print fails?

Record good and failed counts for every part on the plate. The failed quantity is no longer covered by any plate, so it shows up again as outstanding and is planned onto a new plate like any other part still owed.

How many 3D printers can one person run?

It depends on how often plates need changing, not on the printer count. Formlabs says one person can manage 20 or more of its SLA printers, and Markforged described one person handling part removal and setup for its own 25-printer farm in 2018, but both are vendor accounts. Count your own plate changes per day and multiply by the minutes each takes.

Do print farm tools assign jobs to printers automatically?

Some do. SimplyPrint and Printago match queued jobs to printers by material, colour, tags and availability, and Bambu Farm Manager lists queuing by printer availability. Unless an automatic ejection setup clears the bed, a person still has to clear it or mark it ready, and the tool only knows a customer’s deadline if someone entered it.

Sources

All sources accessed .

  1. Managing 3D Printer Fleets (Formlabs) formlabs.com.
  2. 3D Print Farm: Behind the Scenes (Markforged, published 2018) markforged.com.
  3. How to start and run a 3D print farm (SimplyPrint, June 2026) simplyprint.io.
  4. Auto-matching queue items to printers (SimplyPrint help) help.simplyprint.io.
  5. Print farm workflows: the print queue, AutoPrint and automation (SimplyPrint help) help.simplyprint.io.
  6. Print queue (Printago docs) docs.printago.io.
  7. Plate merging (Printago docs) docs.printago.io.
  8. Bambu Lab introduces local fleet control with Bambu Farm Manager (Bambu Lab blog, July 2025) blog.bambulab.com.
  9. We need better tools for managing our print farm (Bambu Lab forum thread) forum.bambulab.com.

Run one real order through it

PrintFarmer is an ERP for 3D print shops. The enquiry, the quote, the plates, the parcel, the invoice and the payment sit on one record, so the answer is on the screen and not in a chat thread.