Capacity Planning

8 test ports or 1? Size the capacity before you buy

8 min read · Xie Yang · 2026-09-07 · For production supervisors, QC managers and buyers sizing test capacity on a charger or adapter line

Quick answer. Size charger test capacity from station time, not daily output. Per-station capacity = available seconds ÷ test time per unit (H × 3600 ÷ T); stations needed = daily output ÷ that number. Example: 3,000 units, 45-second tests, 7-hour shift → 560 per station → 6 stations. Multi-channel usually wins on annual labour cost, not purchase price.

The question I hear most often is: we run 5,000 units a day, what machine do we need?

Sounds reasonable. It is also how lines end up with four idle testers, or one machine with a queue in front of it. Daily output is a result. What decides how much tester you need is how long one unit occupies the station and how many hours the station actually runs. Two lines with identical output can need completely different setups, because one tests in 30 seconds and the other in four minutes.

Three formulas. Put your own numbers in and you will know, before talking to any supplier, whether you need one machine or six. More useful still, they tell you whether the bottleneck you are trying to fix is the tester at all.

Size by takt, not by daily output

Three numbers, from your own line:

Then:

per-station capacity = H × 3600 ÷ T
stations needed = Q ÷ (H × 3600 ÷ T)

Worked example, round numbers purely to show the method:

T = 45 s · H = 7 h (25,200 s) · Q = 3,000 units

per-station capacity = 25,200 ÷ 45 = 560 units/shift
stations needed = 3,000 ÷ 560 = 5.4 → 6 stations
Two things that catch people out

Two things fall out of this in practice. First, test time dominates: moving T from 45s to 30s takes you from six stations to four with nothing else changed. Second, effective hours are always lower than people assume. A nominal 8-hour shift with changeover and breaks is often 6.5 to 7 real hours — and using 8 in the formula is how lines under-buy.


The expensive part is the operator

This is where single-station and multi-channel actually diverge, and it is rarely in the hardware price.

A single-station tester generally needs someone in attendance. Load the unit, start, read the result, sort pass/fail, unload. One person, one station. Six stations means six people on that stretch of line.

A multi-channel box with real plug-and-test behaviour changes the ratio. The operator loads a batch, the machine runs the protocol sequence and disposition on its own, and the operator comes back to unload. One person covers several ports, as long as the load/unload cycle is shorter than the test cycle.

Run it on annual labour, with your own operator cost:

SetupOperators on testAnnual labour
6 single stations66 × your annual operator cost
1 × 8-port multi-channel1, if load/unload fits the cycle1 × your annual operator cost

I am deliberately not putting a salary in that table. Operator cost varies enormously by region, and a number I made up would be worse than no number. Fill in your own.

Where the money is

In most manufacturing regions the annual labour difference dwarfs the hardware delta. That is why multi-channel is usually an operating-cost argument, not a purchase-price argument.

Check this before you count the saving One operator only covers multiple ports if (time to load + unload one port) × number of ports is comfortably shorter than test time T. If your cycle is short and the part is fiddly to fixture, the operator becomes the bottleneck and the saving disappears. Time your own load/unload first.

When multi-channel is the wrong answer

Three situations where single stations, or fewer ports, are the better call:

Hold off on multi-channel when

  1. You run high-mix, low-volume with frequent changeover. Change model every hour and the setup cost is multiplied across every port sitting idle. Flexibility beats parallelism here.
  2. T is already short. Under roughly 20 seconds a unit, the tester is probably not your constraint. Look upstream — assembly or packing is more likely holding the line.
  3. Capacity is not saturated. Running two shifts a week and buying for peak means buying idle hardware. Utilisation first, capacity second.

Five questions before you request a quote

Pre-quote checklist

  1. What is my measured T? Not the vendor's figure. Your product, insertion to result.
  2. What is my effective H? Nominal shift minus changeover, breaks, and the downtime you actually get.
  3. Can one operator really cover N ports? Time the load/unload, compare against T.
  4. How often do I change model? Frequent changes push you toward flexibility over port count.
  5. Is the tester actually my constraint? Compare T against assembly takt. If test is already faster, more ports buy nothing.

What the 8-port unit covers

The GTI6031H-8×2 is the multi-channel machine in our range. 8 test ports, specified as plug-and-test with no dedicated operator required, at ±0.1% typical voltage and current accuracy, covering 20+ fast-charge protocols — QC2.0/3.0, PD3.0/3.1, PPS, PE and the major brand protocols. It is aimed at multi-port fast-charge heads and adapter lines.

Two things I would rather say up front, because this is exactly where equipment sales go wrong. First, "no dedicated operator" only holds under the load/unload condition above. If your part is awkward to fixture, verify it on your part, not on a datasheet line. Second, I have not quoted an electrical range here. That is in the datasheet and varies by configuration. Ask for it for the specific model you are considering.

Sources and scope The sizing formulas are standard takt arithmetic — no vendor data, verifiable on a calculator. The worked example (45 s, 7 h, 3,000 units) is a hypothetical illustration of the method, not a client's line. Operator cost is left as a variable on purpose, because it is region-specific. The GTI6031H-8×2 port count, plug-and-test / no-dedicated-operator specification, ±0.1% typical accuracy and protocol list are from the product datasheet.