VFD Panel Design Guide: Heat Load First, Enclosure Second

A VFD panel is sized from its heat load, not from its footprint. Add up every watt the panel dissipates, compare that against what the enclosure surface can reject, and only then choose a box. Do it in the other order and you have picked a size before you know the question.

That order matters more than it sounds, because the drive you specified isn’t the same drive once you enclose it. The panel changes its rating.

Sipho learned that from a summer, not a datasheet. A maintenance engineer at a packaging plant outside Durban, he inherited a pump panel that had run for two years without complaint.

In February 2025 it started tripping on overtemperature, always between two and four in the afternoon, never in the morning. Nothing had been changed. The drive had not been replaced and the load had not moved. What had changed was the enclosure around it, which had been repainted and resealed six months earlier to keep out the flour dust.

The seal worked. It also stopped the panel from shedding heat, and the drive’s thermal margin was thinner than anyone had assumed.

This VFD panel design guide covers the calculation, the derating rules, the enclosure standards and the layout decisions that decide whether a panel holds its rating. You’ll get the formulas, a worked example, and the specific numbers that are scattered across manufacturer manuals.

If you are sizing a panel now and want the thermal arithmetic checked before the drawing is issued, send us the drive list and the enclosure and our engineers will work it through with you.

Key Takeaways

• Size the enclosure from the heat load, not from habit. Add the losses of every device in the panel, not just the drive, then compare against what the enclosure surface can reject.
• Enclosing a drive changes its ambient rating. The same drive rated to 50 degrees Celsius as UL Open Type is rated to 40 degrees as UL Type 1, so the enclosure is a thermal decision.
• The derating rate is not a constant. Manufacturers publish anywhere from 1 percent to 2.5 percent output current derating per degree above 40 Celsius. The altitude rule is the consistent one, at about 1 percent per 100 metres above 1,000.
• The effective-surface-area formula and the airflow formula use different constants. Carrying the surface constant into the airflow calculation overstates the fan size by roughly five times.
• UL 508A Supplement SB rates a panel by its weakest link. An unmarked terminal block can hold a panel to 5 kA while every other component is rated far higher.

What a VFD Panel Is, and Why Sizing by Habit Fails

A VFD panel is an assembly, not a box around a drive. It has its own standard, its own thermal budget and its own nameplate, and none of those come from the drive inside it.

Four documents get confused, and the confusion is what produces panels that run hot.

Document	Who produces it	What it fixes
Drive datasheet	Drive manufacturer	What one drive does, and its losses
Heat-load calculation	Panel designer	What the whole assembly dissipates
Panel drawing	Panel builder	Where everything sits
Panel nameplate	Panel builder	The assembly’s short-circuit rating

The enclosure appears on only one of those, and it isn’t the one that decides the size.

A VFD panel design guide that starts with the enclosure is working backwards. The heat load comes first, because it’s the only figure that tells you how much air or surface you need. Everything else on the drawing fits around that.

Habit fails for a simple reason. “We always use 600 by 800 by 300” is a statement about the last panel, not about this one. Two panels with identical drive counts can need different enclosures if one sits in a 25-degree electrical room and the other sits on a mezzanine in Lagos, or if one is sealed against dust and the other is ventilated.

The enclosure size is an output of the calculation. Treating it as an input is the most common way a panel ends up running hot.

VFD Panel Design Guide: Start With the Heat Load

The heat load is the sum of everything inside the enclosure that converts electrical energy into heat. That’s more than the drive.

Add up everything that dissipates

Walk the bill of materials and add a figure for each item. Most manufacturers publish watts loss for their products, and the drive’s own datasheet is the first place to look.

• The drive itself. Typically 2 to 3 percent of its rated power at 97 to 98 percent efficiency. A 22 kW drive therefore dissipates around 550 W.
• Line and output reactors. Copper losses are usually small but not zero.
• EMC and harmonic filters. These carry real losses, especially passive filters.
• A control transformer. Losses rise with load.
• Power supplies, PLCs, relays and contactors. Small individually, and they add up across a busy panel.
• Terminal blocks and busbars. Resistance losses inside a sealed box.

Then subtract what leaves through the heatsink. If the drive is flange-mounted, with its heatsink protruding through the back of the enclosure, a large share of its loss never enters the panel at all. This one decision can change the answer more than any other.

Two corrections apply to the total. Altitude: air cools less effectively as it thins. Ambient temperature: a drive rated at 40 degrees has a smaller margin at 45. Both are covered in the next section.

VFD panel thermal design comes down to two questions, and they have different answers depending on whether the enclosure breathes.

The two formulas, and which one applies

Two equations govern enclosure thermal design, and they’re often quoted together as though they shared a constant. They do not.

For a sealed enclosure, the question is surface area. Heat leaves only through the walls:

Ae = P / [k (Tint – Text)]

where Ae is the required effective surface area in square metres, P is the dissipated power in watts, and Tint and Text are the internal and external temperatures.

For a ventilated enclosure, the question is airflow. Heat leaves with the air:

V = Hloss / (ρ × cp × ΔT)

where V is volumetric flow, ρ is air density at about 1.2 kg/m³, cp is the specific heat of air at 1,005 J/kg·K, and ΔT is the permitted temperature rise.

The constant that differs is k, and this is the trap. In the surface formula, k is the heat transmission coefficient of the enclosure material, about 5.5 W/m²·K for painted sheet steel. Several manual-based versions of the airflow formula also carry a k, but there it is about 1.3. Carrying the 5.5 across into the airflow equation overstates the required fan by a factor of roughly five.

The k-values you need are in the next table, and they vary enough by material to change the enclosure choice.

Enclosure material	k (W/m²·K)
Painted sheet steel	5.5
Stainless steel	~4.5
Aluminium	~12
GRP / polyester	~3.5

Aluminium rejects heat more than twice as effectively as steel per square metre. In contrast, stainless steel, which is what washdown and corrosive environments usually force you into, is the worst of the four.

A worked example

Take a 22 kW (30 HP) drive at 480 V, with roughly 550 W of declared loss at rated load. The design convention worth adopting comes from published municipal standards: a maximum internal temperature of 45 degrees Celsius against an assumed 35 degrees outside, which gives a 10 K permitted rise.

Sealed enclosure. Painted sheet steel. Required surface for a 10 K rise:

Ae = 550 / (5.5 × 10) = 10 m²

A typical floor-standing enclosure of 2,000 by 800 by 600 mm presents about 2.0 m² of effective surface to still air. The enclosure is short by a factor of five. The actual rise would be 550 / (5.5 × 2.0) = 50 K, putting the interior at 80 degrees in a 30-degree room.

Sealed is not an option here, and no amount of internal stirring will change that, because stirring fans move air around inside the box without moving heat out of it.

Ventilated enclosure. Required airflow for the same 10 K rise:

V = 550 / (1.2 × 1005 × 10) × 3600 = 164 m³/h, or about 97 CFM

Cross-check it with the field shorthand, which uses a constant of about 3.1 for air at low altitude:

V ≈ 3.1 × 550 / 10 = 170 m³/h

The two agree within 4 percent, and that agreement is the reason to calculate both ways. A result five times larger is a signal that a constant has been carried across from the wrong formula.

Now change one thing. Flange-mount the same drive so its heatsink sits outside the enclosure, and most of that loss leaves through the back panel instead of into the cabinet. If roughly 15 percent of the total stays inside, the internal load falls from 550 W to about 83 W.

Re-run the surface formula on that figure:

Ae = 83 / (5.5 × 10) = 1.5 m²

The same 2.0 m² enclosure now works, sealed, with margin. The internal rise is 83 / (5.5 × 2.0) = 7.5 K, inside the 10 K allowance. No fan, no filter, no air conditioner, and no dust ingress.

That’s the whole argument for flange mounting in one comparison, and it’s why the mounting decision belongs before the enclosure decision rather than after it.

The practical conclusion for this drive: as an enclosed unit it needs filtered forced ventilation at roughly 170 m³/h. Flange-mounted, it needs neither. Those are different panels with different costs, and both can be correct. What is not defensible is a sealed box sized by habit.

If your drives have a declared watts-loss figure but no enclosure decision yet, we can run the calculation from your bill of materials and tell you where each one lands.

VFD Enclosure Sizing: Enclosing a Drive Changes Its Rating

This is the part of VFD panel design that surprises people most, and it’s documented in the manufacturers’ own manuals.

The same drive, two ambient ratings

A drive’s permitted ambient temperature depends on how it’s mounted, because the mounting determines how its heat gets out.

Configuration	Ambient at 100 percent current
IP20 / UL Open Type, heatsink external	-10 to +50 °C
IP20 / UL Type 1	-10 to +40 °C
IP55 / UL Type 12	-10 to +40 °C, derating to 50 °C maximum

That’s one drive, three ratings, and a 10-degree difference between the first two rows. Put the same unit in a Type 1 enclosure and you have taken away a quarter of its usable ambient range.

Some manufacturers expose this directly as a parameter. Yaskawa’s drives, for example, use L8-12 for the ambient temperature setting and L8-35 for the installation method, and the permitted range follows from the pair. That parameter exists because the enclosure changes the answer.

The practical consequence: an enclosure that keeps dust out also keeps heat in, and the manufacturer’s headline ambient figure was measured in a configuration you may not be buying. Check which row of the table applies before you treat 50 degrees as available.

Derating for altitude, and why the rate is not a constant

Altitude is the consistent rule. Drives are rated to 1,000 metres, and above that the output current derates by about 1 percent per 100 metres. That figure is corroborated across INVT, Emerson, CMCO, GE Fuji and others, and CMCO states the valid range explicitly as up to 4,000 metres. Above that, ask the manufacturer.

VFD enclosure derating is really two separate calculations, and they’re often collapsed into one.

Ambient temperature is not a constant, and this is where most published guidance misleads. Drives are rated for full output at 40 degrees Celsius. Above it, the rate varies by manufacturer and frame:

Source	Derating above 40 °C
INVT	1 percent per °C, use not recommended above 50 °C
Emerson PACMotion (enclosed)	2.5 percent per °C
CMCO Impulse (open chassis)	Full load to 50 °C, then 80 percent at 60 °C

As a result, a panel designed on a single printed number will be wrong in one direction or the other. At 50 degrees, the spread between those curves is the difference between a 10 percent derate and a 25 percent derate, which on a 22 kW drive is several kilowatts of usable output.

Three practical points follow.

First, read the curve, not the headline. The headline “derates above 40 degrees” tells you the threshold and nothing about the slope.

Second, derates stack with altitude. A drive at 1,300 metres in a 45 degree room carries both. Calculate them together rather than checking one.

Third, carrier frequency is a third derating axis. Raising the switching frequency above the factory setting reduces output current, and some manufacturers state 10 percent per 1 kHz.

The enclosure derating question is closely related to the faults it produces. Our guide to VFD overheating faults covers what it looks like when this goes wrong in service.

Pick the Rating, Then the Cooling Method

The enclosure rating and the cooling method are linked, and picking them in the wrong order produces a panel that cannot hold its own nameplate.

NEMA 250 and IEC 60529 are not interchangeable

These are separate systems from separate bodies, and they test different things.

The conversion runs one way only. A NEMA type can be mapped to an approximate IP rating. An IP rating cannot be converted back into a NEMA type, because the IP code doesn’t test corrosion resistance, icing, gasket ageing, or resistance to oil and coolant.

NEMA 4 is approximately equivalent to IP66, not identical to it. The tests differ. A NEMA 4 hosedown runs at roughly 65 gallons per minute through a 1 inch nozzle from 10 feet for 5 minutes.

An IP66 jet test runs for about 3 minutes from roughly 3 metres through a 12.5 mm nozzle. The NEMA test is the harsher one, and the two standards committees never coordinated them.

The rule that matters for a panel is this: an assembly is only as good as its weakest component. A NEMA 4X enclosure with an IP54 filter fan in the door is a NEMA 4X enclosure with a hole in it. Every item that penetrates the enclosure, including the cooling device, the cable glands and the door hardware, has to match or exceed the enclosure rating.

Marisol found that out on a food plant in Guadalajara. Specifying a panel for a washdown area in 2023, she selected a stainless enclosure rated NEMA 4X and added a filter fan to manage the thermal load. The fan was a standard industrial unit rated IP54.

The assembly delivered IP54, and the first washdown put water where the fan’s filter was. The fix was a heat exchanger matched to the enclosure rating, which cost more than the fan and should have been the first choice in a washdown area.

The cooling ladder

Work down this list and stop at the first method that fits. Each step costs more and adds more to maintain.

Method	Use when
Passive convection	The heat load is small against the enclosure surface area
Flange-mounted heatsink	The drive’s own heatsink can sit outside the panel, and the environment allows it
Filtered forced ventilation	The surrounding air is clean enough to bring inside
Air-to-air heat exchanger	The enclosure must stay sealed but ambient air is not too hot
Air conditioner	The enclosure must stay sealed and ambient is hotter than the internal target
Liquid cooling	Very high power density, or an environment where air cooling is not viable

However, filtered ventilation is the cheapest and the most abused. It brings the outside air, and everything in it, into the panel. In a cement plant or a flour mill that’s a maintenance liability. The rule that governs the choice is whether the air around the panel is clean enough to put inside it.

Two points about the top of the ladder. Flange mounting is the cheapest way to remove heat, because it uses the drive’s own cooling system and needs no additional device. Where the environment allows it, check it before anything else.

And air conditioning is expensive in two ways, because a sealed enclosure cooled by an air conditioner has to reject the heat load plus the heat the conditioner itself generates. Locating the panel somewhere cooler is often cheaper than cooling it where it stands.

VFD Panel Design Guide: Layout and Clearances

Clearances

Drives need free air above and below them, and the figure is model-specific. Check the manual for the frame you are using, and treat these as the range to expect rather than a rule:

Guideline	Figure
Beneath floor-mounted enclosures, for cable termination	At least 18 in (460 mm)
Beneath wall-mounted enclosures	At least 12 in (300 mm)
Between stacked drives	Manufacturer’s clearance plus the wireway width
In front of the panel	At least 1 m

VFD panel layout starts with the airflow path, and the airflow path is the first thing a crowded drawing destroys. Duct, terminal strips and cable bundles all compete for the same space the manufacturer left open.

Wire duct is not clearance. A drive with 50 mm of duct directly above its heatsink has no clearance, whatever the drawing measures. The most common real-world cause of an overtemperature trip is a panel that satisfied the clearance figure on paper and obstructed the airflow in practice.

Cutout sizing should allow for the next drive, not just this one. Published municipal specifications call for the mounting cutout to exceed the manufacturer’s minimum by roughly 2.75 in wider and 3.5 in higher on each side for a 40 HP drive, scaling with frame. Replacing a failed drive with a different model is normal; re-cutting a panel door is not.

Cable segregation

The rules here are consistent across manufacturers and they aren’t optional.

• Keep power and control cables apart. Around 20 cm (about 8 in) of separation is the common requirement inside a panel.
• Do not run them in parallel. Separate them into different areas of the panel, or different wireways.
• Where they must cross, cross at 90 degrees. A right-angle crossing couples far less energy than a parallel run.
• Terminate shields with a 360-degree loop using a cable clamp or gland, not a pigtail. A short lead to a ground terminal is a poor high-frequency path and is a common cause of intermittent faults.
• Segregate noisy and sensitive devices, and keep drives away from analogue signal wiring and encoder cables.

Control wiring practice on site is covered in more depth in our VFD installation best practices guide. This section is about where things sit on the drawing.

Two filters often belong in the panel, not next to it. Input reactors and filters reduce harmonic distortion at the supply, which is covered in our comparison of the line reactor and the harmonic filter. Output filters manage the voltage stress on motor insulation and are covered in the dv/dt filter versus sine wave filter comparison. Both take space in the enclosure and both dissipate heat, so they belong in the heat-load calculation on the first pass, not as an afterthought.

The Standards That Govern a VFD Panel Design Guide

The UL 508A VFD Panel Rating and Supplement SB

UL 508A, Standard for Industrial Control Panels, is the North American standard for the assembly. The current edition is Edition 3, published in 2018 and last revised on 26 June 2025. It covers panels operating at 1,000 V or less, in ordinary locations, with ambient conditions assumed to be between 5 and 40 degrees Celsius unless the panel is marked with an ambient temperature rating.

That ambient clause is the one to notice. UL’s default assumption is the same 40-degree ceiling the drives are rated to, and a panel installed where the air is hotter than that needs both a marked ambient rating and a design that supports it.

Supplement SB is where panels actually fail. It defines how a panel’s short-circuit current rating (SCCR) is determined, and the method is a weakest-link analysis. The panel’s marked SCCR must be equal to or greater than the available fault current at the point of installation, and the panel must not be installed where the available fault current exceeds it.

When a component’s SCCR is not marked, the standard supplies a default, and the defaults are low:

Component	Default SCCR
Motor controller, 0 to 50 hp	5 kA
Motor controller, 15 to 200 hp	10 kA
Switch units	5 kA
Power distribution blocks	10 kA
Terminal blocks	10 kA
Bus bars	10 kA

This is why a panel full of 65 kA-rated gear can still be marked 5 kA. One unmarked terminal block drags the whole assembly down to its default, and the rating that reaches the nameplate is the lowest one in the chain.

The rating can be raised, in one of two ways. Current-limiting devices upstream can be given credit, but only if they’re a recognised current-limiting class such as a Class J, RK1 or CC fuse, or a listed circuit breaker. Or the specific combination of devices can be tested together as a unit and rated on that test.

The consequence for the designer: a high SCCR is only valid with the exact protective device it was tested with. Substituting an equivalent breaker voids it. Specify the device, not just the rating.

Hassan met that on a submittal review in 2024, on a municipal pump station in Alberta. Every major component in the panel had been selected for a high fault rating, and the panel drawing carried the number. The reviewer asked for the component list.

One terminal block had no marked SCCR, which put the assembly’s rating at that component’s default. The panel was rebuilt around current-limiting fuses, which was the right answer, but the schedule slipped by three weeks.

A note on what this does and doesn’t mean. A drive that carries UL Recognition is not defective or non-compliant, and specifying one is normal practice.

The distinction is between a component being recognised and an assembly being Listed. A panel builder’s file covers the assembly, and the assembly is what carries the rating. The obligation sits with the panel design, not with the drive.

IEC 61439 and design verification

Outside North America the governing standard is the IEC 61439 series. IEC 61439-1:2020, Edition 3, covers general rules and carries two corrigenda, one from 2021 and one from 2023. IEC 61439-2:2020, Edition 3 is the product standard for power switchgear and controlgear assemblies, and it is read together with Part 1.

Part 1 is not a product standard on its own. Nothing can be declared compliant to Part 1 alone.

The structural idea worth understanding is the split between design verification and routine verification. Design verification is performed once, by the original manufacturer, on the design or a prototype. Routine verification is performed on every completed assembly before it leaves the works.

Among the characteristics subject to design verification are temperature rise, short-circuit withstand, degree of protection, clearances and creepage distances, and electromagnetic compatibility. Three routes are permitted: testing, comparison with a tested reference design, or calculation.

Temperature rise is the part that connects to everything above. The standard sets limits on accessible surfaces, at 70 degrees Celsius for metal and 80 degrees for non-metal, and it provides a calculation route through IEC/TR 60890, a technical report on verifying temperature rise by calculation. In other words, the thermal arithmetic in this guide is not a rule of thumb. It’s the calculation method the standard points to.

The division of responsibility matters commercially. The panel builder is the one who has to demonstrate the temperature rise, which means the panel builder needs the watts-loss data from every device supplier. Asking a drive supplier for the declared loss figure at the design stage is not a favour. It’s an input the standard requires.

The enclosure and cooling decisions also need to survive the assembly document that describes the panel. Our guide to the VFD specification template covers the submittals that prove it.

VFD Panel Design Mistakes to Avoid

Six mistakes account for most panels that run hot.

1. Sizing the enclosure before calculating the heat load. The size is an output. Any panel dimension chosen before the arithmetic is an assumption waiting to be tested in July.

2. Counting only the drive. Reactors, filters, transformers, power supplies and control devices all dissipate. In a small panel with an 18-pulse arrangement, the filter can be a meaningful share of the total.

3. Carrying the wrong constant between formulas. The surface-area formula uses a heat transmission coefficient around 5.5 W/m²·K. The airflow formula uses roughly 1.3, or the field constant 3.1.

Mixing them overstates fan size by about five times, and a fan that’s five times too large is noisy, expensive and still not the right answer.

4. Assuming the datasheet ambient figure applies in the enclosure. It may have been measured in a configuration you are not building. The Type 1 rating is 10 degrees lower than the open-type rating on the same drive.

5. Fitting a cooling device that doesn’t match the enclosure rating. A filter fan in a washdown enclosure reduces the whole assembly to the fan’s rating. Match the component to the enclosure, or change the cooling method.

6. Leaving the drive with no clearance above and below. Wire duct, terminal strips and cable bundles all occupy the space the manufacturer reserved for airflow. The clearance in the manual is free air, not free space in a cabinet that happens to be that tall.

And one habit worth breaking: designing for the drive you specified rather than the drive that will replace it. Cutouts sized for a different frame, and doors without enough room to swing, turn a two-hour replacement into a panel rebuild.

Frequently Asked Questions

How do I calculate the heat load in a VFD panel?

Add the watts loss of every device inside the enclosure, not just the drive. The drive typically dissipates 2 to 3 percent of its rated power at 97 to 98 percent efficiency. Then subtract whatever leaves through a flange-mounted heatsink, because that heat never enters the panel. Compare the total against what the enclosure surface can reject using Ae = P / [k (Tint – Text)].

What clearance does a VFD need in an enclosure?

It’s model-specific, so check the manual for your frame. As a range, expect roughly 75 to 150 mm of free air above and below the drive, more between stacked units, at least 12 to 18 inches beneath the enclosure for cable termination, and at least 1 metre in front of the panel. Wire duct and cable bundles don’t count as clearance.

What is the difference between NEMA and IP ratings for a VFD enclosure?

They’re different systems from different bodies. A NEMA type can be mapped to an approximate IP rating, but an IP rating cannot be converted back, because the IP code doesn’t test corrosion, icing or oil resistance. NEMA 4 is approximate to IP66 and is tested more harshly. An assembly’s rating also equals its weakest component.

Does a VFD need derating inside an enclosure?

Yes, and the amount depends on two things. Ambient temperature: drives are rated at 40 degrees Celsius, and derating above that varies by manufacturer from about 1 percent to 2.5 percent of output current per degree. Altitude: about 1 percent per 100 metres above 1,000 metres.

The two stacks. Read the manufacturer’s curve rather than applying a single number.

What is the SCCR of a VFD panel?

It’s determined under UL 508A Supplement SB by a weakest-link method: the lowest-rated component in the power circuit sets the panel rating, unless current-limiting devices or a tested combination raise it. Unmarked components fall back to defaults as low as 5 kA for switch units and small motor controllers. The panel’s marked SCCR must equal or exceed the available fault current where it’s installed.

Conclusion

Four decisions determine whether a VFD panel holds its rating.

Size the enclosure from the heat load. Add every device, subtract what leaves through a flange-mounted heatsink, and compare the result against the enclosure surface using the right constant.

Check which ambient rating applies to your mounting. The enclosure is a thermal decision, and on the same drive the Type 1 rating can be 10 degrees below the open-type figure.

Read the derating curve rather than a headline number. The rate varies by manufacturer, and it stacks with altitude.

Match the cooling device to the enclosure rating. A mismatched fan reduces the whole assembly to its own rating, which is a defect written into the drawing.

If you want the thermal calculation and the SCCR worked through before the panel is built, our engineers do that as part of selection support. Send us the drive list, the enclosure and the environment and we will tell you what the numbers say.

If you are building this into a repeatable panel design, keep the calculation with the drawing. A VFD panel design guide is only useful if the arithmetic travels with the panel, so the next engineer can see why the enclosure is the size it is.

The standards will keep moving. UL 508A was revised in 2025 and IEC 61439 carries two corrigenda, so check the edition before you cite one. A panel drawing is a dated document, and the date is part of its content.