What the Microstrip Impedance Calculator does
This calculator estimates the characteristic impedance of a PCB track - microstrip on an outer layer over one plane, or stripline buried between two - from its width, the dielectric height, the copper thickness and the laminate's permittivity. It also gives the effective permittivity, the propagation delay in picoseconds per millimetre and per inch, and the width needed for a target impedance such as 50 ohms.
The equations used are the IPC-2141 closed forms, which are accurate to a few per cent inside their stated validity window and nothing at all outside it. The result carries that window, and warns when your geometry leaves it.
How to use it
- Choose microstrip for a track on an outer layer above a plane, or stripline for one buried between two planes.
- Enter the track width and the dielectric height - for stripline that is the distance between the two planes, not to one of them.
- Set the copper thickness and the laminate's relative permittivity. FR-4 is roughly 4.2 to 4.6 and falls as frequency rises; your fabricator will give the figure for their material.
- Read the impedance, the effective permittivity and the delay, and check the "in range" indicator.
- Enter a target impedance to get the width that produces it, solved by iterating the same equation so the two figures agree.
Reading the results
Effective permittivity is why a microstrip is faster than a stripline on the same laminate: some of the microstrip's field is in air, so the average permittivity it sees is lower and the signal travels faster.
The delay figure is what matters for length matching. On FR-4 a microstrip is roughly 6 ps per millimetre and a stripline about 7 - so a 10 mm length mismatch is 60 to 70 ps, which is significant on a fast parallel bus.
An impedance figure from any closed-form equation is a starting point. Etching makes tracks trapezoidal, prepreg thickness comes in steps, and Er varies with frequency and glass weave, so a controlled-impedance board is specified by measurement on a test coupon.
Worked example: a 50 ohm microstrip on 0.2 mm FR-4
For a 0.3 mm track on 0.2 mm of Er 4.3 laminate with 35 um copper: 0.8w + t = 0.275 mm, 5.98h / that = 4.349, and its natural log is 1.470. With 87 / sqrt(4.3 + 1.41) = 36.41, the impedance is 53.5 ohms.
Effective permittivity is (4.3 + 1)/2 + (4.3 - 1)/2 / sqrt(1 + 12 x 0.2/0.3) = 3.20, so the delay is sqrt(3.20) / c = 5.97 ps/mm, or 152 ps per inch, and the signal travels at 56% of the speed of light.
To hit 50 ohms exactly the calculator iterates the same equation and returns 0.335 mm - a wider track for a lower impedance, as expected. The fabricator would then adjust for their actual prepreg thickness and etch compensation, which is why their stack-up calculator, not this one, governs the order.
Formulas and scoring rules
- Microstrip impedance (IPC-2141)
Z0 = (87 / sqrt(Er + 1.41)) x ln(5.98 h / (0.8 w + t))Valid for 0.1 <= w/h <= 3.0 and 1 <= Er <= 15.- Effective permittivity (Hammerstad)
Eeff = (Er+1)/2 + ((Er-1)/2) / sqrt(1 + 12 h/w)Between 1 and Er, because part of the field is in air.- Stripline impedance (IPC-2141)
Z0 = (60 / sqrt(Er)) x ln(4b / (0.67 pi (0.8 w + t)))b is the separation between the two planes. Valid for w/b < 0.35 and t/b < 0.25.- Propagation delay
tpd = sqrt(Eeff) / cc = 299,792,458 m/s. A stripline uses Er itself, since it is fully embedded.- Velocity
v = c / sqrt(Eeff)About 56% of c for a typical FR-4 microstrip.
Microstrip against stripline
Microstrip is on an outer layer, so it is easy to probe, easy to rework and faster, because part of its field travels in air. It also radiates, picks up from its surroundings, and its impedance is affected by solder mask and by anything placed near it.
Stripline is buried between two planes, which makes it shielded, quieter and more predictable - and slower, because the whole field is in the laminate. It cannot be probed, and it needs at least four layers. High-speed designs typically route critical signals as stripline and accept the extra layers.
When you need controlled impedance at all
A track behaves as a transmission line when its propagation delay is a significant fraction of the signal's rise time. The common rule of thumb is that it matters when the round-trip delay exceeds the rise time - at 6 ps/mm, a 1 ns edge starts to care at about 80 mm, and a 100 ps edge at about 8 mm.
Below that, a track is just a wire and its impedance is irrelevant. Above it, mismatches cause reflections that show as overshoot, ringing and, in the worst case, double-clocking. Note that the rise time, not the clock frequency, is what decides: a 10 MHz clock from a modern driver can have a 300 ps edge and behave like a far faster signal.
Limitations: what the result does not prove
- These are closed-form approximations valid only inside their stated windows: 0.1 <= w/h <= 3.0 for microstrip, w/b < 0.35 for stripline. Outside them the result is not reliable and the calculator says so.
- Er is not a constant. It falls with frequency, varies between laminates and varies locally with the glass weave, which is why fibre-weave skew exists on fast differential pairs.
- Etching produces trapezoidal tracks and solder mask adds a thin higher-Er layer over microstrip; neither is modelled here. Both lower the impedance slightly.
- Differential pairs are not covered. Their impedance depends on the coupling between the two tracks as well as on each track's own geometry, and needs a two-conductor solver.
Privacy: where your data goes
Everything you paste, type or drop is processed in this browser tab. It is not uploaded, logged, stored or sent to analytics. Session recording and tag-manager scripts are switched off on this page.
Standards and sources
- IPC-2221 - Generic standard on printed board design - checked 19 Sep 2026
- IPC-2141 - Controlled impedance circuit boards and high-speed logic design
- Texas Instruments - High-speed layout guidelines
Frequently asked questions
What track width gives 50 ohms on FR-4?
It depends entirely on the dielectric height. On 0.2 mm of Er 4.3 laminate a microstrip needs about 0.335 mm; on 0.1 mm it is roughly half that. The height to the nearest plane is the dominant variable, which is why the stack-up has to be settled before the layout.
What is the difference between microstrip and stripline?
Microstrip runs on an outer layer above one reference plane; stripline is buried between two. Microstrip is faster because part of its field is in air, and it radiates; stripline is slower, shielded and more predictable, and needs at least four layers.
How fast does a signal travel on a PCB?
About 6 ps per millimetre for a microstrip on FR-4 - roughly 56% of the speed of light - and about 7 ps per millimetre for a stripline, which sees the full permittivity of the laminate. In imperial terms, around 150 and 180 ps per inch.
What is effective dielectric constant?
The average permittivity the field actually experiences. A microstrip's field is partly in the laminate and partly in air, so its effective Er sits between 1 and the laminate's value - typically about 3.2 for FR-4. It is what sets the propagation delay.
How accurate are these formulas?
The IPC-2141 closed forms are within a few per cent inside their validity windows, which is enough to choose a starting width. They are not a substitute for a field solver or for the fabricator's stack-up calculator, and a controlled-impedance order is verified by measuring a test coupon.
Does solder mask change the impedance?
Yes, slightly. Mask over a microstrip adds a thin layer of higher-permittivity material above the track, which lowers the impedance by a couple of ohms. Fabricators account for it in their stack-up models; this equation does not.
When do I actually need controlled impedance?
When the track's delay is a significant fraction of the signal's rise time - the usual rule is when the round-trip delay exceeds the rise time. Rise time, not clock frequency, is what decides, and modern drivers have fast edges even on slow clocks.
Can I use this for differential pairs?
No. A differential pair's impedance depends on the coupling between the two tracks as well as on each track's geometry, and it needs a two-conductor solver. Use your fabricator's differential calculator or a field solver for those.
Last reviewed by the A2Z.Tools team against the sources listed above.
