Login
Blog
OEE for Supplement Manufacturing: Short Runs, Counters and Clearance
In short: measure OEE on a supplement plant as two loops, the encapsulator or tablet press and the bottle packing line, because they run at different times, in different units.
- Set the ideal rate per SKU. On the bottle line it changes with the count per bottle and the tablet or capsule shape.
- Keep line clearance and cleaning inside planned time, as an availability loss with its own reason code.
- Take stops from the machine signal. Tablet counter jams and capsule machine minor stops are too short and too many for a paper sheet.
- Count sealed, labelled bottles at the end of the line, and count reworked bottles as rejects.
A supplement plant rarely runs one product all day. A contract manufacturer may pack three or four lots on one bottle line in a shift, and clear and clean the line between them.
So the usual OEE picture, one product at one speed, does not fit. This page shows how to set OEE up for blending, encapsulation or tablet pressing, coating and the bottle line.
We sell Fabrico, an OEE and light MES platform, so test what we say here like any other vendor's claims.
The supplement line, machine by machine
The plant has two halves that do not run in step. The bulk side makes capsules or tablets in batches, and the packing line puts them in bottles hours or days later.
Powders skip the encapsulator and go from the blender to a tub filler or a sachet machine. That route has its own page: powder filling line OEE.
Why the encapsulator and the packing line need separate OEE
Bulk bins or drums sit between the two loops. The encapsulator can run flat out while the bottle line is down, and the other way round.
The units differ too. An encapsulator makes capsules per hour, while a bottle line makes bottles per minute, and the count per bottle changes with the SKU.
One blended number hides which half is costing you. Give each loop its own planned time, ideal rate and count.
- Encapsulator or tablet press: count capsules or tablets leaving the machine. Good count is what passes the weight check, the sorter and the metal detector.
- Blender and coater: these are batch machines. Track the batch cycle against a standard per product, and the waiting between batches.
- Packing line: count sealed, labelled bottles at one point, the last machine that all bottles pass.
On the packing line the tablet counter is usually the constraint, so its rate sets the line's ideal rate. On a low count bottle the capper or labeller may be the slow machine instead, so check per SKU.
Where a supplement line loses time, speed and product
| Machine | Time (availability) | Speed (performance) | Product (quality) |
|---|---|---|---|
| Encapsulator | Size part changes, strip-down cleaning, empty capsule feed jams | Capsules that fail to open, sticky powder on the tamping pins, speed turned down to hold weight | Weight rejects, dented or split capsules, unfilled shells |
| Tablet press | Tooling changes, cleaning, waiting for granulate | Slowed for sticking or weight drift | Capped, chipped or out of weight tablets |
| Tablet counter and filler | Channel jams, dust on the sensors, format changes | Many short jams, vibration turned down to stop miscounts | Miscounts caught by the checkweigher, broken tablets |
| Desiccant, capper, induction sealer | Cap chute jams, desiccant reel ends, a cut sachet that forces a clean | Gaps in the bottle flow after each upstream jam | Missing desiccant, cocked caps, failed seals |
| Labeller and cartoner | Reel changes, coding faults, carton jams | Blocked or starved by its neighbours | Skewed labels, unreadable lot codes |
Two of these are easy to miss. A counter jam cleared in 20 seconds never reaches a paper stop sheet, and there can be dozens in a shift.
Our page on micro-stops covers how to count them.
The other is the counter an operator slows down to stop miscounts. The line never stops, so the loss shows only as speed.
A stop also costs product. Bottles left under the induction sealer head during a stop are usually rejected at restart.
Bottles that fail the checkweigher are often opened, and the tablets go back in the hopper. Count them as rejects anyway, because OEE quality is first pass.
A bottle with 61 tablets in place of 60 passes as good. OEE quality does not see that giveaway, so track it as its own loss beside OEE, in money.
Short runs: how to class line clearance and cleaning
On a short-run line this is often the largest block of lost time. How you book it decides whether OEE shows it or hides it.
Clearance means removing every trace of the last lot before the next one starts. Our page on line clearance covers what it includes.
If the line was planned to produce, clearance and cleaning stay inside planned time, as an availability loss with their own reason code.
Booking them as planned stops makes OEE look better and hides the minutes you most want to cut.
Do not book all changeovers under one code either. A count change and a full wet clean are different jobs.
| Type | What happens | How to book it |
|---|---|---|
| Same product, new bottle or count | Change parts, counter recipe, label reel, clearance of old labels | Availability loss, code "format change", its own standard time |
| New product, dry clean | Full clearance, counter stripped, vacuumed and wiped, new bulk brought in | Availability loss, code "product change" |
| New product, wet clean | Contact parts washed, dried, checked and refitted, often for an allergen or a strong colour | Availability loss, code "major clean". If it runs in a slot with no production planned, report it next to OEE |
| Waiting | Line is clean and cleared, but labels, bulk or the sign-off is missing | Availability loss, code "waiting", never inside changeover |
Set a standard time for each type and track the minutes over it. The clean itself follows your procedure and should not be rushed, so look for the gain in the waiting around it.
Some of the time cannot be cut at all. US rules for dietary supplements say that when contact surfaces for low-moisture products are wet-cleaned, "they must be sanitized, when necessary, and thoroughly dried before subsequent use" (21 CFR 111.27(d)(2)).
The same rules say the plant's batch production record must include the "date and time of the maintenance, cleaning, and sanitizing of the equipment and processing lines used in producing the batch", or a cross-reference to the logs that hold it (21 CFR 111.260(c)). That record is the plant's own duty, and OEE stop data is no substitute for it.
The "waiting" row matters most to contract manufacturers. Customer-supplied labels or a late release can hold a clean line for an hour, and that is not a changeover problem.
How to set the ideal rate and the planned time
The ideal rate belongs to the SKU, not to the line. On the bottle line it depends on the count per bottle, the tablet or capsule shape and the bottle size.
A counter that fills 60 count bottles at 40 a minute will fill 120 count bottles far slower. Use the best sustained rate you have proven for that SKU, not the nameplate speed.
On the encapsulator the rate depends on the capsule size and how the powder flows. Sticky botanical blends often run slower than a free-flowing mineral blend.
Contract work brings new products with no history. Start with the rate of the nearest product, correct it after the first run, and look into any run that shows performance over 100%.
Planned time is the time the line is scheduled to run an order. A line with no order is a capacity question, not an OEE loss, so show it beside OEE.
For the count, older counters and semi-automatic lines often have no usable PLC signal. A sensor after the capper works, and so does a camera that counts parts, which also suits hand packing of cartons and kits.
A worked OEE example for one packing shift
This is example data. One bottle line runs a 480 minute shift with three short runs of 60 count bottles.
Two changeovers with clearance and cleaning took 45 minutes each, 90 in total. Other recorded stops, mostly cap chute jams and a label reel fault, took 30 minutes.
The many short counter jams are not in that 30 minutes. They show up below as lost speed.
The ideal rate is 40 bottles a minute, or 1.5 seconds a bottle. The line counted 12,240 bottles, and 12,000 were good.
- Availability: 360 running minutes of 480 planned is 75%.
- Performance: 360 minutes at 40 a minute is 14,400 bottles. The line made 12,240, which is 85%.
- Quality: 12,000 good of 12,240 is 98.0%.
OEE is 0.75 x 0.85 x 0.980, or 62.5%. Check it the short way: 12,000 good bottles x 1.5 seconds is 300 minutes, and 300 of 480 is 62.5%.
The full method is in our guide to OEE calculation.
Clearance and cleaning cost three times the other stops, so the first project here is the changeover, not the counter. The 54 minutes of lost speed comes second, and nobody wrote it on a stop sheet.
What to check in OEE software for a supplement line
Run these with any vendor, including us. Bring one real shift with at least two product changes as the test case.
- Two loops: can the encapsulator and the packing line each have their own OEE, in capsules and in bottles?
- Rate per SKU: does the ideal rate change when the product changes? Ask who keeps the rates, and what happens when a new SKU has none.
- Changeover types: can clearance, dry clean, wet clean and waiting have separate reason codes and standard times? Ask to see the minutes over standard.
- Short stops: does a 20 second counter jam show up? Ask for the count and total minutes of short stops per machine.
- First machine to stop: when the capper jams and the counter backs up, does the report separate the capper's own stop from the blocked counter?
- OEE per run: with three lots in a shift, can you see OEE for each lot and customer? Ask how the system knows the run changed.
- Rejects: where do reject counts come from, and what happens to bottles that are opened and reworked?
For a wider view of the market, see our review of OEE software.
Where Fabrico fits
Fabrico is a cloud-based OEE and light MES platform, on the web, iOS and Android.
- Stops: they come from the machine signal, through the PLC or a Fabrico IoT device. The operator adds the reason on a tablet or phone.
- Micro-stops: detected from the machine signal, which is how short counter and capsule machine stops get counted.
- OEE and downtime: availability, performance and quality by line and by shift. Losses can be shown in money.
- Counting: the AI camera is an add-on that counts parts. You can open the video of a stop, and in the standard setup video is streamed, not stored.
- Your data: REST API and webhooks, and any list exports to Excel or CSV.
- Maintenance: a stop reason can be set up to open a maintenance task.
- Not in Fabrico today: production planning and scheduling. Your run sequence stays in your planning tool.
OEE software is not a batch record or a cleaning log. Those stay in the systems you use for them today.
What people ask next
Should line clearance count against OEE?
Yes, when the line was planned to produce. Keep it inside planned time as an availability loss with its own reason code, and set a standard time for it.
Should we measure the packing line in bottles or in tablets?
In bottles, because the ideal rate in bottles per minute already carries the count per bottle. Use capsules or tablets on the encapsulator or press.
What is a good OEE for a supplement packing line?
We know of no sourced benchmark for supplement lines, so we give none. A short-run line will sit below a line that runs one product all week, so compare each line with its own past weeks.
Can we measure OEE on an old tablet counter with no PLC count?
Usually, yes: stops can come from a simple machine signal, and the count from a sensor after the capper or a camera that counts parts. Ask each vendor to show it on your counter.
How do we compare runs for different customers?
Compare each run with its own ideal rate and its own changeover standard, not with other products. A 120 count job will always make fewer bottles a minute than a 30 count job.
Want to run these checks on us? Book a demo or Contact us.
Related articles
Real Fabrico tiles, sample data.