Models
Tranefurnace

XC95m

ComfortLink · modulating · communicating

Trane communicating modulating 95/96% furnace. ComfortLink / AccuLink data pair, not S9V2 24V. The thermostat is on the bus. Capacity is a percent, not W1/W2 / Ht1 / E2.1.

Not S9V2 24V VS. An S9V2 will not become an XC95m by adding a ComfortLink stat. Not S9X2 / S9X1 (24V staged S-series). Not an XR14 / non-communicating outdoor pretending to talk to this IFC.

Heat

Modulating valve + variable inducer (capacity %)

Blower

Communicating VS ECM — CFM follows the bus, not dips

Bus

ComfortLink / AccuLink indoor-outdoor-stat data pair

Vent

96% class condensing · PVC / trap still apply

Sister SKU

American Standard Platinum 95 — badge twin on many jobs

This board

Communicating IFC · ComfortLink

  • Stat diagnostics first. The wall control names the fault in English — believe that when it is alive.
  • Board codes are ComfortLink / AccuLink, not E2.1-as-the-only-story. Do not apply the S9V2 3-character chart here.
  • Data pair polarity. This is not R/W. 24 VAC on the data pins cooks the bus.
  • A 24V stat landed on this IFC will not modulate and may lock comms.

ComfortLink / stat history. Board last-fault only if the stat is dead — write both.

Universal swap

No universal

ComfortLink communicating. No universal.

White-Rodgers

ICM

Honeywell

  • Any 50A55-843 / 50A65-843 / 50M56U-843 / 50M56X-843 / S9200U1000 universal — those are conventional 24V HSI, not this data bus.
  • ICM280 / 2811 / 2810 / 282B — wrong IFC family.
  • S9V2 non-ComfortLink IFC. Spark D67471x boards.

After you pull the dead card

  1. 1OEM ComfortLink IFC. Pair the communicating thermostat.

Field wiring

Written landings for Trane XC95m. Not a factory schematic.

ComfortLink furnace
ComfortLink data pairGas furnace

Trane XC95m / XC80 / XL95 class. Communicating indoor. S9V2 is 24V CoF/Ht1 — different diagram.

Written circuit description only — FieldBench does not redraw maker schematics. Example: R to W closes on a heat call; verify 24V at W1 before you change the IFC.

Open manufacturer PDF

Heat sequence

  1. 1Power. L1 / N → IFC120V
  2. 2Call. Data pair or 24V W → IFCBus or 24V
  3. 3Draft. IFC → inducer → proveTwo-stage PS or a pressure band
  4. 4Ignite. IFC → HSI → valve120V igniter
  5. 5Blow. IFC → serial ECMNo tap, no cap

This indoor

ComfortLink modulating. Pair to XV18 / XV20i.

ComfortLink pair

Communicating stat and outdoor (XV18 / XV20i) on the same bus.

ComfortLink control

Data pair. English events. Not Y.

  • DATA+ComfortLinkData
  • DATA−ComfortLinkData
  • R24V if needed24V
  • CCommon24V

ComfortLink IFC

Do not diagnose this indoor with S9V2 E2.1.

  • DATA+ComfortLinkData
  • DATA−ComfortLinkData

Landing

  • DataStat pairIFC pairoutdoor pair

    Data. Not 24 VAC Y.

IFC harness — communicating two-stage

Call arrives on the data pair (or 24V W). Gas train is still inducer / prove / igniter / valve.

Transformer + fuse

24V still exists for accessories and for a 24V outdoor. Data pair is not the transformer.

  • L1 / N120VLine
  • R / C24V strip24V
  • FUSEOn the IFC24V

24V safety chain

Limit / rollout / door still open the heat path. Codes are on THIS IFC language.

  • DOORInterlock24V
  • LIMITMain limit24V
  • ROLLOUTManual reset24V

Two-speed inducer + dual PS

Low vs high prove. Do not tee the hoses. Do not apply this to a modulating cousin.

  • IND LO/HITwo-speed inducerLine
  • LPSLow prove24V
  • HPSHigh prove24V

HSI + two-stage valve

Two solenoids or a two-stage valve. Staging is the bus or W1/W2, not a tap.

  • HSI120V igniterLine
  • MV / HV24V low / high solenoids24V
  • FLAMEµA DC24V

Serial variable-speed ECM

IFC talks to the motor on the serial harness. No F-taps. No dual run cap on this blower.

  • SERIALIFC-to-motor harness. Unplugged motor = CFM / comm fault, not a tap problem24V
  • CFMSet in the IFC menu / model plug / CoolCloud — not a dip switch on most of these24V

Landing

  • LineCallIFCinducerprove

    Low then high

  • LineIFCHSIvalve

    120V igniter, 24V or serial valve

  • 24VFlame rodIFC

    µA DC

  • DataIFCserial blower

    No F-taps. No PSC cap.

Prove it

  • ComfortLink event list. Indoor and outdoor both on the bus.

Do not

  • Do not use the S9V2 / S9X2 24V diagram on XC95m.
  • Do not land W on a random terminal to force heat.

DATA+/DATA− means the ComfortLink pair on the silk.

Sequence

  1. 1

    Call

    Stat requests capacity %. Valve and inducer modulate. Blower CFM follows the request — there is no W1/W2 / Ht1 display language.

  2. 2

    Prove

    Variable inducer must hit the pressure band for that firing rate. This is not a discrete two-switch S9V2 prove.

  3. 3

    Ignition

    Still HSI + flame rod. A valve that will not come off minimum looks like a no-light, not an S9V2 E2.1.

  4. 4

    Outdoor

    On a ComfortLink pair (XV20i class) the outdoor is on the same bus. A dead outdoor can take the indoor down with a comm fault.

Faults on this IFC

Step 1, then 2, then 3. Expected reading and the fail branch sit under each check.

Comm

Data bus

lockout

Indoor / stat / outdoor dropped off ComfortLink. Not an S9V2 E-decimal.

  1. 1 Data pair, not 24 V

    On Trane XC95m indoor/outdoor comm is a data pair (AB, S1/S2, U, ComfortBridge, ComfortLink — whatever THIS plate uses). It is not a 24 V Y. Check polarity, continuity, and that nobody landed it on R/C. Data pair landing and polarity. One communicating control.

    Expect: Pair continuous, not shorted to chassis, polarity per the diagram.

  2. 2 Power both ends, then the outdoor status

    Indoor transformer and outdoor control both alive. Read the outdoor LED / 7-seg. Water in the outdoor control is the usual dead-end after a storm. No 24V on the data pins. No extra ‘common’ spliced onto the pair.

    Expect: Both controls powered. Outdoor showing a comm or address state, not a blank board.

  3. 3 Address / bias, then the control

    Communicating plates need a matching indoor address and sometimes a terminating bias. Swap is last: known-good pair + power both ends + correct address. Isolate devices: stat only, then indoor, then outdoor. A failed XV20i drive can look like a dead furnace IFC.

    Expect: Indoor sees the outdoor. No comm code after a power cycle.

Pressure band

Inducer / vent at that rate

lockout

Modulating prove failed at the requested capacity. Not a single discrete PS swap.

  1. 1 Watch one full try — do not reset

    Write the Trane XC95m code. Watch inducer, HSI glow, valve click, flame, µA. The step that dies is the repair. Resetting only burns another try into a harder lockout. Vent, trap, collector, inducer. 96% water still kills draft at every rate.

    Expect: Call → Prove → Ignition → Outdoor

  2. 2 Gas and the valve

    Cock open, inlet pressure, then manifold on the stage that is failing vs the rating plate. 24 VAC at the valve during trial? Coil ohms? A valve that never opens looks like a ‘bad igniter’ if you did not watch the sequence. Pressure transducer / hose on modulating IFCs — there may not be two click-style switches like S9X2.

    Expect: Inlet and manifold on the plate. 24 VAC only during trial.

  3. 3 Flame prove, then the lockout timer

    Glow + valve + no prove → rod, ground, or IFC sense. Prove then drop → µA / PS / limit. Soft lockouts on this family often auto-reset on a long clock — do not sit on resets. Do not pull the S9V2 E3.1/E3.3 unplug playbook unless THIS IFC actually has those discrete switches.

    Expect: Stable µA after light-off. No new PS/limit code.

Limit / airflow

Overtemp / CFM

warn

Communicating heat CFM is a bus request, not a tap.

  1. 1 Airflow before you touch the limit

    On Trane XC95m a limit is airflow or overfire until proven otherwise. Do not replace the limit first. Filter, blower wheel, A-coil, and the heat-speed setting for THIS motor (tap, model plug, or communicating CFM — not a sister SKU chart). Stat / IFC airflow configuration vs coil and filter.

    Expect: Rise in the rating-plate window. Static vs this cabinet’s blower chart.

  2. 2 Measure static and temperature rise

    Total external static vs the Trane XC95m blower chart for this cabinet size. Supply minus return rise after 8–10 minutes. A new 5" media cabinet or closed supplies will trip this all winter. TES. Media cabinets on Platinum jobs are the usual.

    Expect: Rise typically 35–65 °F — rating plate wins.

    If fail: Fix the restriction or the heat-speed setting. The limit is doing its job.

    If pass: Airflow in range → step 3 (overfire / limit itself).

  3. 3 Overfire, then the limit circuit

    Clock the meter / manifold on the stage that is tripping vs the rating plate. Then ohm the limit after it cools — a limit that never recloses is failed. Rollout in the same circuit is a different call (see rollout on this door card). Rollout is still a hazard — inspect before reset.

    Expect: Manifold on plate. Limit closed when cool.

Ignition / flame

Prove / valve / HSI

lockout

Trial without flame, or unexpected flame. Confirm the ComfortLink text — do not map it to S9V2 E2.1 / E5.x by habit.

  1. 1 Data pair, not 24 V

    On Trane XC95m indoor/outdoor comm is a data pair (AB, S1/S2, U, ComfortBridge, ComfortLink — whatever THIS plate uses). It is not a 24 V Y. Check polarity, continuity, and that nobody landed it on R/C. HSI, gas, rod, ground.

    Expect: Pair continuous, not shorted to chassis, polarity per the diagram.

  2. 2 Power both ends, then the outdoor status

    Indoor transformer and outdoor control both alive. Read the outdoor LED / 7-seg. Water in the outdoor control is the usual dead-end after a storm. Modulating valve offset — a valve stuck at minimum never looks like a healthy light-off.

    Expect: Both controls powered. Outdoor showing a comm or address state, not a blank board.

  3. 3 Address / bias, then the control

    Communicating plates need a matching indoor address and sometimes a terminating bias. Swap is last: known-good pair + power both ends + correct address. Shut the cock on unexpected flame.

    Expect: Indoor sees the outdoor. No comm code after a power cycle.

Workflows

Identify XC95m vs S9V2

First minute on an S-series-looking cabinet.

  1. 1 Prove the call and the live code

    Communicating stat + data pair + modulating valve = XC95m family.

    Expect: Write whatever Communicating IFC · ComfortLink is showing. ComfortLink / stat history. Board last-fault only if the stat is dead — write both.

    If fail: No call at the IFC → stat / harness. Do not change gas parts.

  2. 2 Isolate the first failed step in the sequence

    24V W1/W2 and CoF / Ht1 / E2.1 = S9V2. Use that chart. Do not stay here.

    Expect: Prove: Variable inducer must hit the pressure band for that firing rate. This is not a discrete two-switch S9V2 prove.

  3. 3 Confirm with a number, then the next part

    An S9V2 with a ComfortLink-looking stat still speaks 24V. The bus does not magically appear.

    Expect: Bus: ComfortLink pair, not 24 VAC. Polarity matters. · Stat events: English fault name + timestamp. Write it before you pull power.

    If fail: Stay on this plate. The out-of-range reading is the part.

    If pass: Reading in range → the next step in this workflow is the repair.

ComfortLink dead, furnace idle

Stat blank or ‘communication’ and no heat.

  1. 1 Prove the call and the live code

    24 VAC still feeds the stat/IFC — ComfortLink still needs C. Then the data pair.

    Expect: Write whatever Communicating IFC · ComfortLink is showing. ComfortLink / stat history. Board last-fault only if the stat is dead — write both.

    If fail: No call at the IFC → stat / harness. Do not change gas parts.

  2. 2 Isolate the first failed step in the sequence

    Isolate the outdoor. An XV20i with a cooked drive or open data pair will take this indoor down.

    Expect: Prove: Variable inducer must hit the pressure band for that firing rate. This is not a discrete two-switch S9V2 prove.

  3. 3 Confirm with a number, then the next part

    Do not ‘convert it to 24V’ by landing W on a random terminal and walking away — you lose modulation and may lock the IFC. Board codes vs stat events — believe the wall control text when it is alive.

    Expect: Bus: ComfortLink pair, not 24 VAC. Polarity matters. · Stat events: English fault name + timestamp. Write it before you pull power.

    If fail: Stay on this plate. The out-of-range reading is the part.

    If pass: Reading in range → the next step in this workflow is the repair.

Modulating prove / no heat at a rate

Runs at low %, dies when the stat asks for more, or never lights.

  1. 1 Prove the call and the live code

    This is a pressure band, not S9V2 E3.1 shorted-PS. Trap and vent first.

    Expect: Write whatever Communicating IFC · ComfortLink is showing. ComfortLink / stat history. Board last-fault only if the stat is dead — write both.

    If fail: No call at the IFC → stat / harness. Do not change gas parts.

  2. 2 Isolate the first failed step in the sequence

    Then inducer actually ramping. A seized modulator that only has one speed will fail the band.

    Expect: Prove: Variable inducer must hit the pressure band for that firing rate. This is not a discrete two-switch S9V2 prove.

  3. 3 Confirm with a number, then the next part

    Only then the transducer / IFC. Do not start at the S9V2 switch pair.

    Expect: Bus: ComfortLink pair, not 24 VAC. Polarity matters. · Stat events: English fault name + timestamp. Write it before you pull power.

    If fail: Stay on this plate. The out-of-range reading is the part.

    If pass: Reading in range → the next step in this workflow is the repair.

Test points

Bus

ComfortLink pair, not 24 VAC. Polarity matters.

Stat events

English fault name + timestamp. Write it before you pull power.

Valve

Modulating, not W1/W2 solenoids. Position follows % request.

Gotchas

  • AS Platinum 95 is the badge twin on many jobs.
  • S9V2-C in a parts list is not an invitation to use the S9V2 E2.1 chart.
  • Pair outdoor is often an XV20i / 4TWV0 class communicating unit — outdoor drive faults are a different chart.
  • Do not apply S9V2 Molex-tap or dip-switch stories here.

Related plates

Install this plate

Install XC95m as ComfortLink modulating communicating. Not the 24V S9V2 IFC.

Cat IV PVC · modulating · ComfortLink bus · communicating VS · XV20i typical outdoor

Bus

  • ComfortLink communicating stat and outdoor. This is not CoF / Ht1 / E2.1.
  • Capacity is modulating. Pair the outdoor (XV20i common) on the same bus.

First fire

  1. 1Pair ComfortLink first. Min-fire, then ramp. Outdoor must be visible on the bus.

Do not on Trane XC95m

  • Do not apply S9V2 24V menu setup as the primary control.
  • Do not convert this to a 24V furnace as a ‘fix.’

Maker literature

FieldBench does not host manufacturer files. Each link opens the maker’s own literature page in a new tab.

  • XC95m product pageProduct pageOpen manufacturer PDF
  • Trane owner’s guidesOwner guidePublic owner manuals. Install / service literature lives in Trane e-Library (dealer).Open manufacturer PDF
  • Trane e-Library (public search)Literature searchSearch the plate (S9V2, S9X2, XV20i). Full I/O often needs a dealer login.Open manufacturer PDF
  • Trane e-Library searchLiterature searchSearch Trane XC95m on the maker’s literature page.Open manufacturer PDF

Warranty and bulletins

Affected models, serial range, symptom, remedy, date. Maker files open on the manufacturer site.

All bulletins
  • Residential limited warranty

    warranty sheet

    Trane furnace warranty hub

    Registered residential furnace warranty · Limited warranty certificate

  • Dealer portals

    dealer

    Where remaining service letters live

    Current-year service letters not on a public literature page · Maker dealer portal