First confirm which decoder the throttle controls and whether it belongs to a consist. Resolve irregular pickup or mechanical running with the troubleshooting guide before using momentum to disguise it. This article covers a normal commanded stop; automatic braking sections and computer-controlled stopping require their own equipment-specific commissioning.
Separate speed, momentum and the stop action
| Question | What you are evaluating | Keep separate |
|---|---|---|
| How fast does it run once settled? | The steady-speed response to a throttle setting. | Use the speed-matching workflow; this is not a stopping-distance measurement. |
| How does it reach a new speed? | The acceleration or deceleration transition. | A smooth transition does not prove there is enough track left. |
| Which command actually requests the stop? | Normal speed zero, a documented brake function or a system emergency procedure. | Record the exact action; do not combine different actions in one result. |
| Where does the model become stationary? | Travel of a consistently chosen reference point during a controlled trial. | Measure at layout scale in millimetres, not by scaling a prototype braking distance. |
NMRA S-9.2.2 describes CV3/4 in terms of rate, and optional CV23/24 provide adjustments. Those descriptions do not specify where a particular model will stop on your layout. Treat the decoder manual as the implementation guide and the observed route test as the distance evidence.
Named examples: why one CV recipe does not travel
| Source and scope | Verified distinction | What to do with it |
|---|---|---|
| NCE version 3.5 decoder CV list | CV3 and CV4 values extend to 255 in this documented family. | Do not assume another decoder shares this range. |
| Digitrax DN163K1C instruction sheet | CV03/04 use 0–31. Its documented switching feature changes both target speed and momentum response. | Record whether switching mode is active; compare like-for-like trials. |
| TCS non-sound single-stage brake feature | When that brake is applied, CV126 controls the response instead of CV4/CV24; the page distinguishes WOWSound. | Identify the decoder family and brake feature before choosing which setting to investigate. |
| Digitrax KB852, legacy troubleshooting | A large programmed delay can resemble a failure to respond. | Check the saved settings and selected control path before assuming a failed decoder. |
The examples explain a diagnostic problem, not a cross-brand conversion table. “CV4 = 20” is an incomplete test record without the decoder, mode and observed response. Do not copy CV126, a function number or a switching-mode setting into unrelated equipment. Keep the current function map beside the momentum record.
Four gates before using a constrained route
1. Identity gate
GO only when the model, decoder, firmware and selected address are identified. An unknown replacement decoder keeps this gate on HOLD.
2. Mode gate
GO only when normal stop, any brake function and any switching or consist state are understood. A sound effect alone is not evidence of the intended motor response.
3. Distance gate
GO only for conditions actually checked with repeatable stopping room and a retained reserve. A positive calculator result cannot close the other gates.
4. Recovery gate
GO only with accessible documented emergency controls and a saved configuration to restore. End the trial for unexpected motion, loss of control, heat or an electrical fault.
Tool 1: measured stopping-distance budget
Measure from the moment you issue the chosen stop action until the same point on the model is stationary. Measure the available route from that starting position to the chosen no-cross boundary, using the same reference point. If another part of the train projects beyond that point, account for that geometry when choosing the boundary. These defaults are an arithmetic example, not ROKHELM test data.
Balance = available distance − longest measured travel − additional allowance − retained reserve. For the illustrative values above: 800 − 500 − 100 − 100 = 100 mm. A negative balance is a deficit. Zero means the entire entered budget is used; this guide does not treat that boundary case as a pass.
Repeat trials and retain the longest observation, not just the average. The allowance is your explicit provision for remaining uncertainty; it is not a manufacturer-prescribed percentage. The reserve is space you intend to keep unused. Zero allowance or zero reserve is permitted for arithmetic exploration but produces a HOLD result. A positive result with both present still only means the numbers fit.
Do not double-count or omit delay: this measurement already includes travel after issuing the stop action while the system and model respond. It does not include distance covered before you decide to issue that action. If that earlier reaction interval matters, account for it separately and document the method. The tool accepts distances up to 1,000,000 mm as a software input bound, not an equipment specification; decimals are valid, blank fields and non-finite values are not.
Fault matrix: inspect before writing another CV
| Observation | Separate these possibilities | Controlled next action |
|---|---|---|
| Long, smooth run-on after speed zero | Programmed deceleration versus failure to receive the intended command. | End the trial safely, confirm the selected model and compare the saved mode/settings with the manual. |
| CV4 changes but a brake key behaves the same | Normal deceleration versus a separate brake feature. | Find the exact brake-mode documentation; do not cycle through unrelated CV numbers. |
| Stops differently after a function press | A changed control mode versus a random defect. | Record function state and retest the documented normal mode. |
| Two locomotives cruise together but fight during stopping | Steady-speed agreement versus transition mismatch. | Return to separate, uncoupled evaluation before another coupled test. |
| Results vary markedly between repeated trials | An unsteady test setup or an unresolved operating fault. | Suspend the profile; investigate and repeat under controlled conditions instead of hiding variation with an average. |
| Unexpected motion, electrical fault or heat | A control or hardware problem, not an invitation to increase momentum. | Use the documented emergency response, remove power for inspection when appropriate and follow the troubleshooting workflow. |
A reset is not a routine tuning step. Preserve useful settings and use the reset and recommissioning guide only when that recovery decision is justified. The programming-track guide covers isolation and readback; measurements here take place on an operating test route, not by assuming a programming output supports ordinary running.
Tool 2: stop-profile record and reopening triggers
Copy this table into your roster or a paper test sheet. A profile is local evidence with limits, not a universal certificate. Use one record per materially different train, direction or braking mode.
| Record field | Evidence to retain | Reopen when |
|---|---|---|
| Identity | Model, decoder model/firmware, controller and active address. | Decoder, firmware, control system or address routing changes. |
| Configuration | Original and changed CVs, active functions and applicable manual revision. | A reset, reprogramming, function remap or mode change occurs. |
| Train and route | Actual cars, direction, test route and chosen reference point. | Train formation, direction or relevant route conditions change. |
| Approach and action | Stable approach setting and precise stop action. | Speed curve, approach speed or stop method changes. |
| Measurements | Each observation, longest travel, allowance, reserve and available distance. | New trials exceed the recorded range or uncertainty remains unresolved. |
| Disposition and rollback | Local limitations, date, operator notes and preserved baseline. | The next operator cannot reproduce the setup or does not know the emergency procedure. |
Illustrative decision: a profile that fits 800 mm in normal mode says nothing about a different train approaching faster or using a brake key. Keep the working profile unchanged while investigating the new case. If the result does not fit, possible planning changes include a lower verified approach speed, an earlier command point or a different documented response setting—each requires another actual trial.
Six steps to a measured stopping profile
1. Identify the active control path
Record the decoder model and firmware, command station, address, consist state and function modes. Read the matching manuals and identify the documented stop and emergency controls before running.
2. Preserve a recoverable baseline
Save readable momentum settings and relevant mode settings. Mark unknown values as unknown. Use the supported isolated programming workflow to make changes only to the intended decoder.
3. Establish a clear test route
Use a clear operating test route with generous stopping room, no other trains and accessible controls. Mark a consistent reference point on the model and begin at low speed with the documented baseline.
4. Change one setting and measure
Change one supported setting at a time. Record the approach speed and exact stop action; measure travel of the chosen reference point from issuing that action until the model is stationary. Repeat under the intended conditions.
5. Compare the worst observation with the route
Enter the longest measured travel, an additional uncertainty allowance and a retained reserve in the distance-budget tool. If the budget does not fit, revise the operating plan and repeat the tests before using the constrained route.
6. Record the approved local profile
Keep the settings, measurements, limitations and rollback path together. Test the actual train and each intended direction or mode separately, and reopen the record after changes that can affect stopping.
Railway culture: braking depends on conditions
Network Rail explains that leaf contamination can reduce adhesion, so drivers may need to begin braking earlier. The modeling lesson is to record the conditions behind a stopping claim. It is not a formula for scaling a full-size train's braking distance or a reason to put prototype rail treatments on model track.
Official and first-party sources
Read 2026-09-03. The NMRA document supplies CV roles; NCE, Digitrax and TCS examples establish product-specific differences. The older Digitrax documents are labeled as such, not presented as universal current specifications. Network Rail provides railway context only.
- NMRA — S-9.2.2, CV3/4 and optional CV23/24 descriptions (2012)
- NCE — version 3.5 decoder CV list
- Digitrax — DN163K1C instruction sheet, momentum and switching mode
- TCS — CV126 single-stage braking, non-sound decoder scope
- Digitrax — KB852, delayed starting or stopping (legacy support article)
- Network Rail — leaves, adhesion and earlier braking
FAQ
What do CV3 and CV4 do in DCC?
CV3 governs programmed acceleration and CV4 governs programmed deceleration. They shape the transition toward a commanded speed; they do not specify a universal stopping distance in millimetres.
What CV4 value should I use for N scale?
There is no single value for every decoder, train and layout. Start from the exact decoder manual and a saved baseline, then use controlled low-speed trials and measured stopping travel to choose a suitable local profile.
Is the momentum range always 0 to 255?
No. NCE's version 3.5 list reaches 255, while the cited Digitrax DN163K1C sheet specifies 0 to 31. These are named product examples, not a range guarantee for other decoder generations.
Does CV4 equal to zero guarantee an instant physical stop?
No. Removing a programmed delay is not a guarantee of zero physical travel. Confirm the active mode and observe the actual model on a clear test route.
Can I convert CV4 directly into a braking distance?
Not reliably from the CV alone. Actual travel must be related to the approach speed and operating conditions. This tool uses measured travel rather than a universal CV-to-distance conversion.
Why did stopping change when I pressed a function key?
A decoder-specific switching or braking mode may change the response. Check the exact function map and mode documentation before rewriting CV4; a function number is not a universal brake assignment.
Does TCS CV126 control braking on every TCS decoder?
No. The cited single-stage CV126 feature is for the documented non-sound decoder system. Its active brake rate replaces the CV4/CV24 response, and the page explicitly distinguishes WOWSound.
Is speed matching the same as momentum matching?
No. Compare steady speeds separately from acceleration and deceleration transitions. A pair that runs together at a settled speed still needs start-and-stop testing before coupled operation.
Should I use normal speed zero or emergency stop for the distance test?
Measure the normal stop action you intend to use in service. Learn and verify the system's separate emergency procedure in controlled conditions; do not mix observations from different stop actions in one profile.
What does a positive distance-budget result mean?
Only that the entered travel, allowance and reserve fit within the entered available distance. It is not a safety certificate, a predicted stop location or approval for untested speeds, trains or modes.
Does the calculator include human reaction time?
The measured travel begins when the stop action is issued, not when a hazard is first noticed. Any earlier reaction travel is outside that measurement and must be accounted for separately when relevant, without double-counting delays already observed.
When should I repeat the stopping test?
Repeat after changing the decoder or firmware, train, speed settings, braking mode, route, direction or other relevant operating conditions. Unexpected variability also reopens the profile rather than being averaged away.
Continue the commissioning workflow
Use speed matching for steady-speed comparison, consisting for group control and block detection for occupancy feedback. Detection alone does not validate a stopping profile.