Every variable frequency drive ships with a default acceleration ramp, usually somewhere between a few seconds and twenty. It is a placeholder. The factory has no idea what machine you bolted the drive to, so it picks a round number and lets you sort it out later. Most of the time nobody sorts it out, and the drive spends its life ramping on a guess.
There is a right number for your machine, and the textbook route to it is old and well understood. The problem is that the textbook route asks for one thing the field almost never has: the inertia of the whole drivetrain. This article is about getting to a defensible acceleration time without that number, by letting the drive do the measuring for you. It supports the VFD Commissioning Guide, which is where ramp setting belongs in the lifecycle.
Why the Calculation Stalls
Acceleration time is Newton’s second law for rotation, solved for time. In the customary units of the trade, the time to ramp is the total inertia reflected to the motor shaft, times the speed change, divided by a constant times the average net accelerating torque. The math is settled. It has been in the vendor literature for decades.
The term that stops you is the inertia, the WR² of the entire rotating system reflected back to the motor shaft. That means the rotor, the coupling, the gearing, and the load mass reflected through the square of the speed ratio. On a conveyor that went in three years ago, nobody has that figure, and nobody is going to crawl the drivetrain on a Tuesday afternoon to build it. The number the calculation cannot do without is the number the machine in front of you never comes with.
The other term people get wrong is the torque. It is not the motor’s rated torque. It is the average net accelerating torque, the torque the drive will let the motor make during the ramp minus the torque the load draws at the same speed. The ceiling on that available torque is not a fixed motor property either. It is the current limit you programmed, typically set between 110 and 150 percent of motor nominal current. Getting the load type wrong here is a common sizing miss, which is why the VFD Selection and Sizing Guide spends real time on constant versus variable torque.
The Field Method: Step the Ramp Into the Limiter
Here is the way that needs no inertia figure at all. It needs only the drive’s own display of output current and the relationship between commanded speed and actual speed, both of which live at the keypad or in the monitoring software.
Program a deliberately long, safe ramp. Twice your best guess, or simply generous if you have no guess. Run it and watch two things: the output current against the limit, and whether actual speed tracks the command. If the current stays under the limit the whole way and speed follows the command, you have headroom. Shorten the programmed time and run it again. Keep stepping it down.
At some point the current limit regulator engages, the actual acceleration time stretches past the programmed time, and actual speed begins to lag the command. That crossover is the practical minimum for that exact machine. Set the final value just above it, with margin, because the object is to know where the limit is and to stay clear of it, not to live on it.
The drive’s current limiter is doing the calculus the datasheet could not. It integrates the equation of motion in real time, which is how you arrive at a defensible ramp with no WR² anywhere in sight. When a ramp is set too short and the drive rides that limiter on every start, you get the nuisance overcurrent trips covered in the VFD Troubleshooting Guide. The method here finds the edge so you can back off it deliberately.
Why the Load Type Changes Everything
A conveyor and a fan do not behave the same on the way up, and a single formula quietly lies on one of them. The difference is where the load fights hardest, and it changes how you read the current during the stepping.
A constant torque load, a conveyor or a positive displacement pump or an extruder, pulls roughly the same torque at every speed. The net accelerating torque is one number that holds across the whole ramp, which makes the linear ramp honest and the reading clean. When the limiter engages on a constant torque machine, it engages across the entire ramp at once, because the current is flat from bottom to top.
A variable torque load, a centrifugal fan or pump, is a different animal. The load torque rises with the square of speed and the power with the cube, so at low speed the load asks for almost nothing and at full speed it asks for everything. The net accelerating torque is large at the bottom of the ramp and shrinks toward the top. The limiter does not engage evenly. It engages near the top alone, because that is where the current piles up. On a variable torque machine you will watch the current sit low and comfortable through the bottom and middle, then climb steeply as the drive approaches rated speed.
That single observation, where the current piles up, tells you the load type without any nameplate. Flat across the ramp is constant torque. Climbing to the top is variable torque. And it determines which shortcut is safe to use, because the two loads land their minimum in completely different places.
A Conveyor and a Fan, Side by Side
Take a 7.5 kilowatt, four pole motor on a conveyor, ramping from rest to 1500 rpm at a 150 percent current limit. The mechanism reflects a modest inertia to the motor shaft and the running load is a steady constant torque. Work the physics and the minimum acceleration time comes out near 3.45 seconds. The field reading, stepping the ramp down until the limiter engages, lands at essentially the same place, within rounding. The agreement is direct because the single net torque makes the crossover unambiguous.
Now take a 15 kilowatt, four pole motor on a centrifugal fan at a 120 percent limit. The tempting shortcut is to take the net torque at the top of the ramp, where it is smallest, and pretend it held the whole way. That naive number comes out around 21.6 seconds. The honest answer, accounting for the fact that the load barely fights at the bottom, is closer to 7.6 seconds. The shortcut is not a little off. It is off by a factor of three, and it would push you to program a ramp far longer than the machine needs.
The reason lives in where the ramp time goes. On that fan, nearly half the total ramp time falls in the top third of speed, and the final tenth of speed alone eats more than a fifth of the total. The fan blasts through the bottom and crawls the last stretch. The field signature follows the same shape, current low through the bottom and middle, then climbing hard as the drive nears rated speed, with the limiter engaging at the top alone. That is exactly why the field reading and the honest calculation agree while the naive one overshoots.
Where the Field Method Reaches Its Limits
Naming the limits is the honest course. The method rests on torque tracking current closely, and that holds below base speed. Above base speed, in field weakening, torque falls off and the current limit no longer buys the same acceleration. The method addresses the ordinary ramp to rated speed, not acceleration into the constant power region.
A positive displacement pump can show a breakaway torque at the very first instant that exceeds the running load, so the bottom of even a constant torque ramp deserves a glance. And thermal duty can override the shortest achievable start entirely. A standard motor is rated for a limited number of starts from cold, and a machine that cycles dozens of times an hour has to be sized around the heat of that duty, not the fastest single start it can physically make. The single event minimum and the duty rated time are different numbers, and on a high cycle machine the duty number governs. The VFD Maintenance and Reliability Guide treats start duty as a life question, not only a startup one.
Setting the Final Number
Set the ramp too short and the drive rides its limiter, stresses the machine, and trips. Set it too long and the process waits on a ramp the motor could have finished in a third of the time. The window between those two failures is where the right number lives, and the drive’s current limiter is the instrument that finds it for you.
The physics was never in question. What the stepping method shows is that the field trick and the calculation reach the same minimum, and that you can get there by reading where the current piles up rather than by hunting a drivetrain inertia nobody wrote down. Getting that ramp right is a core part of a clean startup, which is why it sits inside the VFD Commissioning Guide. This is also the kind of hands-on judgment the VFD training program is built around, and the full treatment, including the closed form for the variable torque case and the worked validations, lives in Before the First Fault.
Author: Dr. Carl Lee Tolbert, PhD, CMRP, Wayward Leaders LLC, waywardleaders.com
Frequently Asked Questions
What is the default acceleration time on a VFD?
Most drives ship with a factory default acceleration ramp somewhere between a few seconds and twenty seconds. It is a generic default, not a value matched to your machine, because the factory does not know the inertia or the load you connected. It should be treated as a starting point to tune, not an answer.
How do I set VFD acceleration time without knowing the load inertia?
Program a long, safe ramp, then watch the output current against the limit and whether actual speed tracks the command. Shorten the ramp in steps until the current limiter engages and speed begins to lag, which marks the machine's real minimum. Set your final ramp just above that crossover with a margin, and you never need the inertia figure.
Why does my VFD trip on overcurrent during acceleration?
An overcurrent trip on startup usually means the ramp is too short for the load's inertia, so the current needed to accelerate exceeds the drive's limit. Lengthening the acceleration time spreads the same speed change over more time and lowers the peak current. If a very long ramp still trips, the cause is elsewhere and belongs in a full troubleshooting sequence.
Is acceleration time the same for a conveyor and a fan?
No. A conveyor is a constant torque load that pulls about the same the whole way up, so its ramp reads cleanly. A fan is a variable torque load that fights hardest near the top, and the common shortcut of taking the top torque and holding it can overestimate the ramp by a factor of three.
What happens if I set the acceleration ramp too long?
An overly long ramp does not damage the drive, but it wastes throughput by making the process wait on a start the motor could finish far sooner. On machines that start often, an excessively cautious ramp can also add heat over many cycles. The goal is the shortest ramp that keeps the drive off its current limiter with margin.
Does the current limit method work above base speed?
No. The method relies on motor torque tracking current, which holds below base speed but breaks down above it in the field weakening region. Use the stepping method for the ordinary ramp to rated speed, and treat acceleration into the constant power range as a separate problem.