Sparkles and dropouts over twisted pair

Likely causes

• Copper-clad aluminium or stranded cable in the channel

  ~25%

  Cut end shows a silver-grey core instead of bright copper, or the conductors are stranded. Sparkles are constant, never clean, and get worse the longer the run.

• Channel too long, or too many connection points, for the format being pushed

  ~20%

  Sparkles scale with content: a static desktop is clean, high-motion 4K is a snowstorm. Drop to 1080p and it goes away entirely.

• Alien crosstalk from other HDBaseT runs in the same bundle

  ~15%

  Clean until the neighbouring extenders are powered up. Unplug the bundle-mates and the picture clears. This one hides completely during a single-room test.

• Marginal termination: excess pair untwist at the punch, split pair, kinked or over-tensioned cable

  ~15%

  Cert tester fails NEXT or return loss and locates the fault within a metre or two of a connector. Visually, the pairs are opened up further at the punch than the category allows.

• EMI coupling: run parallel with mains, ballasts, VFDs, elevator gear

  ~12%

  Dropouts time with something else in the building: lights switching, a motor starting, a dimmer moving. Watch the picture while someone cycles the load.

• Screened channel bonded at one end only, or bonded nowhere

  ~8%

  Sparkles or dropouts on the video coincide with audible hum in the audio on the same rack: a digital link shows shield-current problems as errors, not as bars. Changing the screen connection at one end visibly changes the artefact, which identifies the screen path as the coupling route, even though the fix is bonding it properly rather than lifting it.

• Field-terminated plug or unshielded keystone on a 4K run

  ~5%

  Fault appears only through that one outlet; patch straight past it and the picture is clean.

What to bring

• certification tester with distance-to-fault
• spare solid Cat 6A on a drum
• known-good HDMI cords
• source resolution control
• phone light and loupe
• punch tool and spare jacks
• clamp meter

Steps

1. 1

   Step 1: Bracket it with resolution

   Force 1080p60, then 4K30, then 4K60 4:4:4. Note which rate is clean. A format-dependent artefact is a channel margin or extender ceiling problem; sparkles at every resolution point to a hard cable or termination defect.

   If that doesn’t do it

   Sparkles at 1080p too: expect to find a real physical fault, not a margin issue.

2. 2

   Step 2: Lay a temporary cable on the floor

   Run a drum of solid Cat 6A between the two extenders. Read the safety note before you route it. A clean picture on the temp cable indicts the installed link without anyone re-terminating anything. This is the highest-value test in the whole list and it costs an hour of walking.

   If that doesn’t do it

   Sparkles persist on the temp cable: the fault is the extender pair, the HDMI jumpers or the source.

3. 3

   Step 3: Read the jacket print and look at a cut end

   Confirm solid copper, correct category, and plenum or riser rating for where it actually runs. Count every connection point in the channel: outlet, patch panel, cross-connect, coupler. Couplers in an HDBaseT channel are a red flag.

   If that doesn’t do it

   Cable and channel check out: go to crosstalk and EMI.

4. 4

   Step 4: Power down every other extender in the bundle

   Then bring them back one at a time. The sparkles returning as a specific neighbour powers up is alien crosstalk, and the fix is separation or screened cable, not a re-termination.

   If that doesn’t do it

   No change with all neighbours dark: continue.

5. 5

   Step 5: Certify the link and read the diagnostics, not just pass or fail

   NEXT, return loss, length and distance-to-fault. Run your tester's HDBaseT or extended-frequency video limit if it has one. A channel can PASS the TIA data limit and still fail video on return loss or alien crosstalk margin, so read the margin figures, not the verdict. A failure clustered at one end is a termination; failures distributed along the run point at the cable or its routing.

   If that doesn’t do it

   Passes with healthy margin and still sparkles: strongly suspect alien crosstalk or the extender's video ceiling.

6. 6

   Step 6: Re-terminate both ends, then re-certify

   Only now. Keep untwist at the punch to the minimum the category allows, maintain the jacket right up to the termination, respect bend radius, and use the screened connector the cable calls for. Re-test before you close the ceiling.

   If that doesn’t do it

   Re-terminated and still failing: the cable or its route is the problem. Plan a replacement run away from the noise source.

7. 7

   Step 7: Sort the screen bonding: both ends, not one

   A screened channel is bonded to the grounding system at both ends, through screened jacks, screened panels and equipment chassis, with the telecom bonding infrastructure itself installed per ANSI/TIA-607 so both ends sit on the same common bonding network. A screen bonded at one end only is an antenna at HDBaseT's frequencies and gives you no shielding where you needed it; bonded nowhere is worse. If bonding both ends produces noise, the two ends are sitting at different potential. That is an electrical bonding defect to correct, not a reason to lift the screen.

   If that doesn’t do it

   Still artefacting with the screen correctly bonded at both ends: replace the run away from the noise source, or move to fibre-based extension for the distance.

More Audiovisual guides

• No picture through HDBaseT extender
• Camera or mic missing in the soft codec
• HDCP error on one display
• Hum or buzz in the sound system
• Projector dark, dim, or shutting down
• One zone dead, or amplifier in protect

All 54 guides