Why Robot Spec Cycle Times Are Almost Never Real

Partner Resource: Epson Robots — Robot Cycle Times: Beyond the Specification Sheet
Content Type: Technical Brief
Best For: Engineers and operations managers evaluating SCARA robots for assembly, pick-and-place, or precision handling applications

Why We Are Sharing This

Every robot buyer has been there. You find a SCARA robot with a cycle time that looks perfect on paper. You quote the application around it. Then during commissioning the real-world number looks nothing like the spec sheet promised. The project timeline slips. The throughput model breaks. The business case takes a hit.

Epson’s brief addresses exactly that problem. Before you read it, though, you should know what you are reading.

This is not a buyer’s guide. Roughly half the document is a two-column table headed “Competitive Claims / How to Respond,” and the rest is a set of “Sales Questions” sorted by application stage. It was written to give an Epson salesperson something to say when a prospect walks in holding a competitor’s faster cycle time. Every caveat in it points outward.

We are sharing it anyway, for two reasons.

First, the technical caveats are correct. The reasons a champion cycle time overstates real throughput do not stop being true because a vendor raised them for competitive reasons. Second — and this is the part worth sitting with — the same caveats apply to Epson’s own numbers, including the ones on the RBTX Marketplace. A document that teaches you to interrogate a competitor’s spec sheet is a document that teaches you to interrogate every spec sheet. Read it that way and it is genuinely useful.

Start with the Footnote

The single most valuable thing on any SCARA spec sheet is the footnote under the cycle time. Here is the one on the Epson T3-B, the model we list on RBTX:

Cycle time based on round-trip arch motion (300 mm horizontal, 25 mm vertical) with 1 kg payload (path coordinates optimized for maximum speed).

Three conditions, all of which flatter the number:

300 mm horizontal, 25 mm vertical. A short, flat, standardized arch. If your part travels 700 mm, or needs a 90 mm lift to clear a fixture, you are not running this motion.

1 kg payload. The T3-B401SS is a 3 kg robot. It is sold as a 3 kg robot. The 0.52-second number is measured at its rated payload of 1 kg — one third of the maximum. Rated and maximum are two different specs and the cycle time only ever attaches to the first one.

Path coordinates optimized for maximum speed. Epson is telling you, in its own footnote, that the path was tuned for the benchmark. That is the entire argument of the partner brief, stated more plainly than the brief states it.

Note also that the test payload is not consistent across the lineup. The LS20-B spec sheet runs the same arch motion at 2 kg. So you cannot compare cycle times between two Epson robots — never mind two different manufacturers — without first checking whether they were measured under the same load.

What the Brief Gets Right

Points, Waypoints, and Path

Your cycle follows the path your part requires, not the path that flatters the benchmark. Extra waypoints, obstacle clearance, an approach vector constrained by fixturing — each one adds time that no spec sheet models. Epson’s point about arm length and righty/lefty arm configuration is a real one: choosing a configuration that keeps your work inside a favorable part of the envelope is a cycle-time decision, not just a layout decision.

Overshoot and Settle

A robot that oscillates before it settles loses time on every single cycle, and settle time is part of cycle time whether or not your throughput model includes it. Epson attributes its settle performance to Quartz MEMS-based motion control, which is a specific mechanism you can ask about rather than a vibe. Ask what settle criterion the quoted cycle time assumes.

Payload derating

The rated-versus-maximum split on the spec sheet is the derating boundary, and it is published. You do not need to guess at a percentage. Find rated payload, find maximum payload, and understand that everything above rated costs you speed. Then ask the vendor for the number at your payload — that curve is not published, and it is a fair thing to request in writing.

Duty Cycle

Some quoted numbers are boost-mode numbers. This is not hypothetical: Epson’s C4-B datasheet states outright that its cycle time is measured at rated payload in boost mode. If you are running two shifts at high cycle rates, ask specifically whether the quoted figure is sustainable indefinitely or is a temporary mode.

Inertia

The brief is right that inertia is the most frequently overlooked factor, and right that a gripper carrying a long or offset part loads the joints in a way payload alone does not capture. It is the most important thing in the document.

Where We’d Push Back

Inertia is not unknowable

The common summary of this brief says inertia can’t be determined from a spec sheet and therefore requires an application study. That overstates it, and it conveniently routes every reader to a sales conversation.

The allowable limit is on the spec sheet. The T3-B publishes an allowable moment of inertia at Joint #4 of 0.003 kg·m² rated and 0.010 kg·m² maximum. What isn’t on the sheet is your tooling’s inertia — and that comes out of your CAD package’s mass-properties tool in about ten minutes, once the EOAT is modelled.

Epson also documents the handling explicitly. The LS6 spec sheet notes that the published figures assume the payload center of gravity (CoG) is aligned with Joint #4, and that if it isn’t, you set parameters using the RC+ INERTIA command. The C4-B datasheet points to the manual for the parameter calculation method. So the workflow is:

Pull moment of inertia and CoG offset from your EOAT model.
Compare against the robot’s rated allowable inertia.
If you’re over rated, expect a speed reduction, and ask for the magnitude.
If the CoG is offset from Joint #4, plan on setting INERTIA parameters — and get the cycle time quoted with those parameters applied.

An application study is the right call when the tooling doesn’t exist yet or the geometry is unusual. It is not the only path, and a competent integrator gets most of the way there before anyone books a call.

“Three tiers of precision” doesn’t work the way it sounds

The brief’s tier framing implies a straightforward trade: more precision, less speed. Within a single robot that’s true — a tighter settle criterion at a point costs milliseconds at that point, and specifying camera-grade precision where a loose nest would do is wasted time.

Across the lineup, it inverts. Compare two 400 mm, 3 kg-max, 1 kg-rated SCARAs from Epson’s own catalogue:


Arm length	400 mm	400 mm
Repeatability (J1+J2)	+/- 0.020 mm	+/- 0.010 mm
Payload (rated / max)	1 kg / 3 kg	1 kg / 3 kg
Standard cycle time	0.52 s	0.42 s

The more precise robot is also the faster one, by about 19%. Precision tier isn’t a speed trade-off — it’s a price trade-off. What you give up by buying down the range is both accuracy and throughput. Frame it that way when you build the business case, because “we don’t need the precision” is not by itself a reason to accept the slower robot.

Worth noting too that the three-tier structure the brief describes is dated. Epson’s current SCARA catalogue filters by five series — LA, T, LS, GX and RS. If the brief’s tier language is what you’re navigating by, you’re navigating by an old map.

The Five Conditions to Get in Writing

The brief tells you to request an application study. Before you get that far, this is the short list to send back with any quoted cycle time. It costs a vendor nothing to answer and it will tell you a great deal about who you’re dealing with.

At what payload was this measured — and what is the robot’s rated payload, not just its maximum?
What path geometry — horizontal distance, vertical lift, number of waypoints — and was the path optimized for the benchmark?
What precision or settle criterion was applied at each point, and what is the published repeatability?
What moment of inertia was assumed, what is the rated allowable, and was the payload CoG aligned with the wrist axis?
Sustained or boost — can this figure run continuously across a full shift, and what is the sustained figure if not?

The RBTXpert Take

The most useful frame this brief provides is the distinction between a well-defined application and one still in development. Each stage demands different questions. A well-defined application can support a full application study before equipment selection. A development-stage application benefits from surfacing the key variables early, specifically precision requirements, EOAT geometry, and inertia, before those variables become commissioning surprises.

Read the brief for the mechanisms, not the conclusion. The mechanisms — path, precision, overshoot, payload derating, inertia, duty cycle — are real and they apply to every SCARA on the market, Epson’s included. The conclusion the document is engineered to produce is that you should call Epson. That may well be the right answer for your application. Make it your answer, not the document’s.

Read the full Epson brief Here .
Check out Epson on RBTX Here.