A patch panel is a passive block of ports where the fixed cabling of a building ends and the cords to your switches begin. It has no power, no chip and no configuration.
The cabling standards draw the same line: the fixed run from the patch panel to the wall outlet is the permanent link, capped at 90 meters, and the whole path between two active devices, cords included, is the channel, capped at 100 meters. That boundary is why the patch panel is the thing you certify once and re-patch for years.
- Patch panels are passive: no power, no electronics, no configuration
- The back takes punched down cable, the front presents RJ45 ports
- The permanent link stops at the patch panel, capped at 90 meters
- The channel includes a cord at each end and is capped at 100 meters
- Patch panels are sold in 24 port 1U and 48 port 2U sizes
On this page
- What a patch panel actually is
- The line the patch panel draws: permanent link and channel
- Why you certify the permanent link, not the channel
- Choosing patch panels, in the order the decisions matter
- Does a patch panel degrade the signal?
- Labeling, which is the whole reason the patch panel exists
- Where people go wrong
- Comparison
- FAQ
What it isWhat a patch panel actually is
Ask what is a patch panel and the honest answer is a strip of metal with contacts on it. Take the front cover off the idea and there is almost nothing inside.
Two sides, one piece of metal. The back of a punch down panel is a row of insulation displacement contacts, the same 110 style contacts used on a wall jack.
Each pair of the horizontal cable is pressed into a slot with a punch down tool, which cuts through the insulation and holds the conductor.
The same connections are made on every port, which is why the work is repetitive and why a bad one is easy to miss. The front is a row of RJ45 ports. A vendor listing spells the arrangement out in its own specification fields: side A is 110 IDC, side B is RJ45 female.
No electronics at all. There is no power connector because there is nothing to power. A patch panel does not switch, route, amplify or regenerate anything.
It is a connection point between cables, which is why a patch panel cannot fail in the ways a switch fails, and why a dead port is nearly always a bad termination or a bad cord rather than a broken panel.
The feed-through variant skips the punch down. A feed-through panel has RJ45 on both sides, so a finished cable plugs in from behind instead of being punched down. Feed-through panels are easier and faster to install, they suit a rack full of pre-made network cables, and they add a mated connection where a punch down panel has a single termination.
Patch panels are not switches, and they are not splitters. They are passive cable management for the fixed cables, not network devices, and they allow no Ethernet traffic decisions of any kind.
Every port is one cable going to one place. Plugging a laptop into a patch panel port with nothing patched on the other side connects it to a wall outlet in an empty office, and to nothing else.
Patch panels exist for fiber too. The same idea in fiber is a rack mount enclosure holding adapters and spliced or terminated pigtails. The words change, the role does not: the fixed cables end there and the jumpers to the equipment start there.
Link and channelThe line the patch panel draws: permanent link and channel
This is the part worth understanding, because every number people quote about network cable length comes from it, and because it is what patch panels are for.
The permanent link is the fixed part. Fluke Networks defines it as the portion of a channel that does not change, which excludes the patch cord and equipment cord at the ends.
In a building, it runs from the patch panel in the telecommunications room to the work area outlet or a consolidation point; in a data center it usually runs panel to panel. Its maximum length is 90 meters.
The channel is everything the traffic crosses. It is the connection from one active device to another, so it includes both the patch cords at the switches and the equipment cords at the devices. Its total length should not exceed 100 meters.
The 10 meters is yours to spend. The difference between the two numbers is the cord budget, and Panduit describes it as up to 5 meters of patch cord at each end of the permanent link. A rack with a meter of slack on every cord and a long run in the ceiling is how a compliant installation quietly stops being one.
Four connectors, not more. Panduit also states the connector limit plainly: to qualify as a permanent link, the length must not exceed 90 meters and the number of connectors must be within four. Every extra patch panel between the outlet and the switch spends one.
The panel is the boundary, which is why it is where testing stops. The tester is clipped to the patch panel port because that is the last point in the path that nobody will unplug tomorrow.
What to certifyWhy you certify the permanent link, not the channel
An installer who tests the channel and hands over a clean report has proved less than the report appears to say, and the reason is not intuitive.
The standards give the channel more headroom. Fluke's explanation is direct: combining standards compliant patch cords with a standards compliant permanent link will always result in a passing channel, and because the standards allow the channel more margin, a channel test with high quality cords can pass even when the permanent link underneath it has failed.
The cords are not part of the building. Patch cords are handled, swapped, borrowed and bought from different vendors, whoever was cheapest that quarter. Once the cords used in the test are removed, the test result describes a configuration that no longer exists.
The permanent link is what you own. It is inside the walls, it is expensive to redo, and it is the part a tester can re-measure identically in five years. That is why the specification for a cabling job should say permanent link testing in writing, before the installer starts.
Test adapters matter. Permanent link adapters use a reference grade cord and plug, so that a passing permanent link can be turned into a passing channel simply by adding compliant cords. A channel adapter cannot make that promise in reverse.
Choosing oneChoosing patch panels, in the order the decisions matter
Most of the differences between patch panels in a catalog are irrelevant to the network. Four are not.
Punch down or feed-through. Punch down is the traditional field termination and gives one connection point at the patch panel. Feed-through accepts a finished cable and is quicker in a rack that is being rebuilt often.
Pick one and use it everywhere, because a rack with two different styles is a rack where nobody can predict what is behind a port.
Port count and rack units. Patch panels are normally 24 ports in 1U and 48 ports in 2U, and half height panels exist for a cabinet with no space left. Buy the density that matches the run count, not the one that matches today's switch.
Category. Patch panels have to meet the category of the cables: a Cat6A run terminated on a Cat5e panel is a Cat5e link. Patch panels are sold by category for that reason, and shielded patch panels exist for shielded cable, where the shield has to be bonded rather than left floating.
Fixed ports or keystone jacks. A fixed panel is one piece with its ports built in. A modular panel is a blank frame that accepts keystone jacks you snap in one at a time, so a single panel can carry different cable types and a damaged port is replaced one jack at a time.
Cable management. Patch panels with a rear management bar, and a horizontal cable management panel below them, are the difference between a rack you can work in and a rack where every change risks a neighboring cable. This is the part people cut from the order and regret.
Those four decisions are about maintenance more than performance. Patch panels provide the point where a move or a change is a cord in the front of a rack instead of an electrician in a ceiling, and every choice above either protects that or quietly gives it away.
Signal lossDoes a patch panel degrade the signal?
Anybody asking what is a patch panel eventually asks this one. Slightly, and the standard already knows it.
Every mated connection in a link costs a little insertion loss and creates one more place where the pairs can couple into each other, so a link with a patch panel at each end performs marginally worse than an unbroken run of the same cable.
That cost is accounted for. The connector allowance is why a permanent link is capped at 90 meters rather than 100 and why four connections is the limit: the link budget in the standard assumes the loss of those connections and still leaves the data path inside specification.
So a correctly terminated patch panel is not what makes a link slow. What makes it slow is a bad termination, a pair untwisted too far back, a fifth connection nobody counted, or a run already at the length limit before the connectors were added.
None of that is unique to twisted pair. Fiber patch panels add loss at every mated adapter in the same way, and a fiber link budget is spent on connections and splices before it is spent on distance.
This is also why the certification test matters more than the product. A tester measures insertion loss and crosstalk across the whole link and reports whether the margin is there, which is a statement about the terminations somebody made rather than about the panels they bought.
LabelingLabeling, which is the whole reason the patch panel exists
Patch panels without labels are a wall of identical holes, and the person holding the tone generator at two in the morning is you.
Label the port, not the cord. The port number is permanent and the cord is not. A label on a cord walks out of the rack the next time somebody tidies.
Use the same identifier everywhere. The outlet in the room, the patch panel port, the drawing and the record in the configuration database should all say the same thing. The scheme matters less than using one.
Record what is patched, not just what is punched. Patch panels map to the building, and the patching maps to the network equipment. A record that ends at the patch panel tells you where the cable goes and not what it is connected to.
Photograph the rack after the work. A dated photograph of the front of the patch panels answers more questions than most documentation, and it costs a minute.
PitfallsWhere people go wrong
Expecting the patch panel to do something. They have no power and no logic. If a port is dead, the fault is in the termination, the cord, the run or the switch, and the patch panel is only where the fault becomes visible.
Measuring 100 meters to the outlet. The 100 meter figure is the channel, cords included. The cables in the wall are allowed 90, and a run pulled to 98 meters leaves two meters for two cords.
Stacking patch panels to get from one place to another. Every patch panel in the path is another connection, and the standard allows four connections in a link. A cross-connect added to a link that already had a consolidation point is how a certified installation stops passing.
Terminating the two ends to different schemes. Patch panels are labeled with both wiring schemes, and so is the jack at the far end. Pick one for the site and punch every termination the same way; a pair swapped at one end is the classic fault that still links at a lower speed and fails under load.
Accepting a channel test as the deliverable. It is easier to pass, it hides a failing permanent link, and it stops describing reality the moment the cords move.
Putting the patch panel where the cable is instead of where the rack is. Patch panels belong above or below the switches they feed, with a cable management panel between them. A panel at the top of the rack and a switch at the bottom means every patch cord crosses the cabinet.
Leaving PoE out of the calculation. Patch panels are passive and do not care, but a bundle of network cables carrying power over Ethernet heats up, and heat costs you length and margin on exactly the runs that are already the longest.
ComparisonPunch down against feed-through
| Criterion | Punch down | Feed-through |
|---|---|---|
| Back of the patch panel | IDC contacts | RJ45 ports |
| Needs a punch down tool | Yes | No |
| Accepts a finished cable | No | Yes |
| Connections added to the link | One | Two |
| Speed of a rack rebuild | Slower | Faster |
| Field termination of bulk cable | Yes | Needs a plug on the cable |
| Shielded versions available | Yes | Yes |
| Common densities | 24 in 1U, 48 in 2U | 24 in 1U, 48 in 2U |
The connection count is the row that decides it on a long or marginal run.
FAQFrequently asked questions
What is a patch panel used for?
Patch panels terminate the fixed cables of a building in one place, so that changing which outlet reaches which switch port is a matter of moving a cord rather than re-running cable.
That is the whole infrastructure argument for them. They also provide the point where the permanent link is certified, which is the part of the cabling infrastructure that has to be proved before anyone signs for it.
Is a patch panel a network device?
No. Patch panels have no power supply, no chip and no configuration. They provide a passive termination field, and nothing in them can be managed, updated or misconfigured.
Do I need a patch panel at all?
Not for a handful of drops. Patch panels start to pay as soon as the number of runs makes tracing the cables by hand slow, or as soon as somebody other than you has to work on the rack. The panel is bought with labor, not bandwidth.
What is the difference between a patch panel and a switch?
Switches are the active network devices that forward Ethernet frames between devices. Patch panels are the passive field the switches are patched into. A port on a patch panel provides one connection to one cable at the other end of the building; a port on a switch connects that cable to the data network.
How many ports should the patch panel have?
Enough for the runs plus the growth you actually expect. Patch panels come in 12, 24 and 48 port sizes, and 24 ports in 1U and 48 in 2U are the common ones, and a half height 24 port panel exists for a cabinet with no room left.
What is the difference between a permanent link and a channel?
The permanent link is the fixed cabling between the patch panel and the outlet, up to 90 meters, with no cords. The channel is the whole path between two active devices, cords included, up to 100 meters.
Should the installer test the channel or the permanent link?
The permanent link, and it should be written into the specification. A channel test is easier to pass and can pass over a failing permanent link when good cords are used, and it stops being valid as soon as those cords are unplugged.
What is a feed-through patch panel?
A panel with RJ45 on both sides, so a finished cable plugs into the back instead of being punched down. It is faster to work with and adds one more mated connection to the link.
Do patch panels have to match the cable category?
Yes. The link performs to the lowest rated component in it, so terminating Cat6A cable on a Cat5e panel produces a Cat5e link no matter what the cable cost.
Can a patch panel cause slow speeds?
Indirectly. A poor termination, a pair split at one end or an over-long link degrades the data path, and patch panels are where those faults are made and found. The panel itself adds nothing beyond the loss of one connection.
How many connectors can a link have?
Four, per the TIA standard, which is why every extra cross-connect between the outlet and the switch has to be counted rather than assumed.
Do patch panels work with PoE?
Yes, because they are passive and pass the pairs straight through. What matters for PoE is the cables, the bundle size and the heat, not the patch panels.
What about fiber?
The equivalent in a fiber infrastructure is a rack mount enclosure holding adapters and pigtails. The role is identical: the fixed cables end there and the jumpers to the network equipment start there.
How is patch panel wiring done?
Patch panel wiring means punching each of a cable's eight conductors into the color coded slots on the back of the port, following T568A or T568B to match the wall outlet at the other end. Keep the twists as close to the termination as possible, then label and test every port.
Keep readingRelated concepts
Read next · Cabling and connectivity What UTP Cable Is, and Why the Twist Does the Shielding What runs between the patch panel and the outlet, and why its category caps the link. Open this next9 min- Switching · 10 min Top of Rack Switching, and the Number Nobody Puts on the Datasheet The switch the patch panel is patched into, and why its position in the rack matters.
- Cabling and power · 9 min 802.3af, and Why Your Switch Still Says It What the pairs carry when the run also powers the device at the far end.
- Network design · 10 min The Logical Network Diagram and the Physical One, Drawn for Different Questions The drawing where every patch panel position and uplink gets written down.
- Cabling · 10 min Single Mode vs Multimode Fiber, and Why the Core Size Decides Everything The cabling boundary that records which fiber type and grade a run carries.
- Network fundamentals · 10 min Layer 1 Networking, the Physical Layer, and Why It Is Where Most Faults Start Where the cabling side of Layer 1 is terminated and certified.