Every packaging line I have ever worked on ends up with at least one indexing station where the motor ramps up, slams to a stop, and repeats that cycle eight or nine times a minute. After the third or fourth rebuild, most maintenance guys stop blaming the motor and start looking at the clutch-brake unit. And honestly, they are usually right. The clutch-brake is the most stressed component on that axis, and it fails in predictable ways if you size it like a afterthought.
1. What the Clutch-Brake Actually Does
A clutch-brake simply routes torque on demand. The clutch half couples the continuously running motor to your mechanism when you need motion; the brake half stops the load when you do not. In a single-shaft unit, the two share one coil and one armature, which keeps the response snappy and the footprint small. I have used both split units and combined units, and for index tables and in-feed conveyors the combined type wins on simplicity every time.
The key number nobody writes down is cycle rate. A motor that runs all day is doing fine. The clutch-brake is the one engaging and disengaging thousands of times per hour, so its duty cycle defines the whole axis, not the motor power.
Rule I follow: size the clutch-brake for the worst-case cycle at maximum load with 30 percent margin on torque, then check thermal capacity before you ever pick a frame size.
Here is a typical duty cycle for a 180 cpm indexing conveyor. The numbers come from a line I commissioned last year, not from a catalogue:
| Segment | Time (ms) | What Happens |
|---|---|---|
| Accelerate | 85 | Clutch engaged, load picks up speed |
| Run | 150 | Constant speed, index moves |
| Decelerate | 70 | Clutch off, brake applies |
| Dwell | 25 | Position held by brake plus detent |
That gives you about 330 ms of work per cycle, which means the unit is being hammered 180 times a minute. If you cannot get the heat out, the armature will glaze in a month.
2. Sizing Torque: Where People Get It Wrong
Most engineers start with the load torque and call it done. That is how you end up with a unit that slips on the third shift. What actually matters is inertia. The torque the clutch must deliver is the sum of the load torque and the torque to accelerate the reflected inertia, and on a lightly loaded, high-cycle index table the inertia term is usually the bigger one by a factor of two or three.
Use the basic form: T = (J_total x omega) / t_acc, where J_total is the reflected inertia of everything the clutch sees, omega is the speed you are trying to reach in rad/s, and t_acc is the acceleration window. You can do this with a spread sheet in ten minutes. I want you to be honest about J_total, because gearboxes and belt reductions do not make inertia disappear. They reflect it through the square of the ratio, which is exactly why a high-ratio gearmotor can still bog down and murder a small clutch.
Add the load torque to the acceleration torque, then multiply by a service factor of 1.3 minimum. I go to 1.5 on food lines because product jams are not a thing you can schedule around. Once you have that number, the catalogue torque table is easy to read. The selection you need to sweat is thermal, not torque.
3. Thermal Capacity Is the Real Limiter
Heat is what kills clutch-brakes. Every start-stop cycle dumps energy into the friction interface, and if the unit cannot shed it, the coil resistance rises, the torque drops, and the armature glazes. The published thermal rating is stated in joules per cycle or watts of continuous dissipation, and you have to compare it against the energy per cycle of your machine, not the average motor power.
Energy per stop is roughly 0.5 x J_total x omega-squared (in joules). Notice the squared speed term: run the axis twenty percent faster and you raise the heat load by 44 percent. I have seen lines fail a clutch-brake simply because someone bumped the speed pot up to push more product through, and never rechecked the thermal budget.
For a 0.05 kg-m-squared system running at 150 rpm with a 0.07 s stop, the math looks like this:
omega = 150 x 2 x pi / 60 = 15.7 rad/s
E = 0.5 x 0.05 x 15.7^2 = 6.2 J per stop
at 180 cycles/min you get 180 stops/min x 6.2 J = 1116 J/min = 18.6 W
Eighteen watts does not sound like much until you remember the unit is only the size of a coffee can and sits inside a closed cabinet. Give it a duty of 60 percent above catalogue ratings and you will be replacing it quarterly.
4. Friction Material and Configured Units
On combined units you will typically meet asbestos-free organic, sintered metallic, or ceramic friction material. Organic is quiet and cheap, and it is fine for low-energy, low-cycle light duty. The moment your energy per stop climbs past a few joules, step up to sintered bronze or ceramic, which hold torque at high sliding speeds and shrug off dust from the friction face.
I keep a mental comparison, because it has saved me a wrong purchase more than once:
| Material | Torque Density | Wear Life | Cost | Best Fit |
|---|---|---|---|---|
| Organic | Low | Short | Low | Light indexing, low cycles |
| Sintered bronze | Medium | Long | Medium | Mid-duty indexing |
| Ceramic | High | Longest at heat | High | Heavy or hot duty |
For pneumatic clutch-brakes, air pressure does the work. The advantage is easy torque adjustment on site by turning a regulator, which is why I reach for them on machines that change product size often. The downside is you need clean, dry air and you lose response time while the valve fills the chamber. On an indexing axis that wants a 40 ms engage, a pneumatic unit will feel lazy unless you size the port and valve generously.
5. Gap Setting and the Maintenance Trap
Every unit has a working air gap, typically 0.25 to 0.5 mm on small electromagnetic units. You set it with a feeler gauge on the armature. If you set it too tight, the armature drags, the coil runs hot, and the friction face wears unevenly. Too loose, and you sacrifice torque and gain a harsh slam as the armature travels. The slide-over-rotation adjustment is simple, and the feeler gauge should be part of the PM kit for that machine.
Here is the trap that got me early in my career. The torque seems to fade a little after a few months, so a well-meaning technician cranks up the air pressure or the current “to compensate.” That masks a worn friction face and accelerates glaze. The right move is to measure the gap, and when it exceeds the wear limit in the manual, change the friction disc, not the current.
Track torque fade and gap growth together. When gap grows past spec, the fix is a disc, not a pressure knob.
6. Wiring, Timing and the Controller Side
The clutch and brake coils on a combined unit are usually two separate windings in one body, and the controller switches them alternately. A dead band between release and engage is what you want, so the coil does not see simultaneous power and cook itself. Check the datasheet for the overlap guard, and set the dwell so the motion comes out crisp without hammering the frame.
Switch DC electromagnetic units with a solid-state relay or a proper clutch-brake controller, not a plain contactor. I watched a contactor-wired index table weld its own contacts after a year because every cycle broke the coil current at the contact gap. A cheap SSR fixed it and the process stayed quiet. Also add a flyback diode across the coil if your controller does not clamp the inductive kick. Without it, the back EMF will eat the driver board and the tech will blame the PLC.
Timing matters more than people admit. The engage delay on a solenoid-driven unit shows up as repeatability error in the index position. If your pick-and-place needs +/-0.05 mm at the nest, a tired, slow brake shows up as drift you will chase for days. Put a prox or a light barrier on the final dwell and let the controller trim the release point by a few milliseconds. On one line I trimmed the engage lead 6 ms and the registration error dropped from 0.4 mm to 0.08 mm.
7. Duty Cycle Table for a Typical Line
Here is the operating budget I hand to the maintenance crew, filled in for the 180 cpm line from section 1. It makes the thermal check concrete.
| Item | Value / Note |
|---|---|
| Cycles per minute | 180 |
| Energy per stop (J) | 6.2 |
| Heat dissipated (W) | 18.6 |
| Serviced air gap (mm) | 0.30 |
| Gap wear limit (mm) | 0.55 |
| Friction disc change interval | 6 months at 3 shifts |
That service interval assumes the gap is checked monthly. Skip the monthly feeler-gauge pass and you will be replacing the whole unit instead of a 40 dollar disc.
8. Conclusion and a Hard-Earned Checklist
Honestly, the clutch-brake on an index axis deserves more engineering attention than the motor. Motor oversizing is free insurance, but clutch-brake oversizing costs you response time and cycle speed. So size torque with reflected inertia, check thermal energy per cycle, pick sintered or ceramic friction material for real duty, set the gap with a gauge, and let an SSR do the switching. Do that and the axis will run crisp for years.
Before I send the drawing to the shop, I run this short list:
- Torque with 1.3 to 1.5 service factor, based on reflected inertia, not load only.
- Thermal energy per cycle under the catalogue rating at the fastest speed.
- Friction material matched to energy and cycle count.
- Air gap at 0.25-0.5 mm, checked monthly.
- SSR (or proper controller) switching, with flyback protection.
You will thank yourself the first time the machine runs a full shift without a glazed disc. I have been on both sides of that fence, and the extra ten minutes of sizing math is worth every one of the rebuilds it prevents.
One last thing about stock and spares. Keep a spare friction disc and a spare set of springs on the shelf for high-cycle units, because lead time on a made-to-order disc will take your machine down for a week. As a habit I spec the same clutch-brake frame across several machines on the line, so one spare kit covers all of them. Standardizing the frame also means the maintenance team learns one gap-setting procedure instead of three, and that consistency is worth more than a few euros of catalogue savings.
Also keep an eye on the mount. A clutch-brake bolted to a flimsy sheet-metal bracket will flex under each cycle and eat its own bearing over time. Mount it to a machined face or a stiff bracket, align the bore square to the shaft, and use a proper taper-lock or keyed hub rather than a set screw that walks loose. I have opened machines where the only reason the unit was still working was that the bracket had worn itself into a perfect alignment, which is luck, not design.
If you are wiring this into a safety circuit, remember the brake is your holding element and it must be fail-safe. On the pneumatic side that means spring-apply, air-release so a loss of air pressure stops the machine instead of letting it drift. On the electromagnetic side, the coil current must be monitored or the brake must default to applied when power is lost. Every safety story I have been called into on an indexing line traces back to a holding element that released when it should have held.
