AI-generated illustration, not a photograph of a specific product
The moment a layout stops being a loop, you need turnouts. In KATO Unitrack the choice is almost always between the #4 and the #6 β one tight and compact, one long and gentle. Two assumptions get made about them constantly, and both are wrong: that the bigger number is simply better, and that the shorter one is cheaper. Here is what actually separates them.
The 30-second comparison
| Item | #4 electric turnout | #6 electric turnout |
|---|---|---|
| Part number, left / right | 20-220 / 20-221 | 20-202 / 20-203 |
| KATO designation | EP481-15 L/R | EP718-15 L/R |
| Body length | 126 mm (about 5 in) | 186 mm (about 7.3 in) |
| Diverging radius | R481 (tight) | R718 (gentle) |
| Nominal exit angle | 15Β° | 15Β° |
| KATO's intended use | Tight spaces, spurs, yards | Main line, higher-speed running |
| Power routing | Switchable to non-power routing | Power routing only |
| Japan list price | Β₯3,300 incl. tax | Β₯2,200 incl. tax |
| KATO USA price | $37.50 | $27.00 |
| In the box | Turnout + auxiliary track set | Turnout only |
Prices checked 2026-08-01 against KATO Japan and the KATO USA online store. Retail prices vary by country and dealer; treat both figures as reference points, not as what you will pay.
If you come from American track, read this first
Under the NMRA convention, a turnout's number describes its frog angle β a #4 spreads one unit sideways for every four units along, which works out to roughly 14Β°, and a #6 to roughly 9.5Β°. So an American modeler reading "#6" reasonably expects a divergence about a third as sharp as a #4.
Unitrack does not work that way. KATO's #4 and #6 are product designations, and the company's own part codes give the game away: EP481-15 and EP718-15. Both diverge at a nominal 15Β°. What changes is the radius used to get there β R481 versus R718 β and therefore how long the turnout has to be to complete the turn.
Both are 15Β°, so why is the #6 longer?
This is the question that confuses most people, and the answer is simply geometry. Turning through the same 15Β° on a larger radius takes more distance. R481 gets there in 126 mm; R718 needs 186 mm. That extra 60 mm buys you a gentler entry into the diverging route, which is exactly what long cars want.
R481 and R718 are not minimum radius ratings. They describe the geometry of the diverging route. They do not license the claim that any locomotive rated for R315 will take a #4 at any speed. Whether your stock runs through cleanly is still something you test.
What is in the box β and why it inverts the price
| Box contents | #4 (20-220 / 20-221) | #6 (20-202 / 20-203) |
|---|---|---|
| Electric turnout | 1 | 1 |
| 64 mm straight track | 2 | β not included |
| Matching curve | R481-15 Γ1 | β no R718-15 |
| Dedicated 60 mm straight, left and right | 1 each | β not included |
| Turnout Control Switch | β not included | β not included |
Read that table and the pricing stops being strange. The #4 is dearer because it arrives as a small kit; the #6 really is a single piece of track in a blister pack. So compare the cost of a finished route, not of the box. If your #6 divergence also needs an R718-15 curve and some straight track, those are separate purchases, and the total may well land above the #4 you thought you were avoiding.
Our take: #6 where trains pass, #4 where trains park
Strip away the detail and there is a rule that is easy to remember: use the #6 where trains run through, and the #4 where trains turn off and stop. That is also how KATO itself frames the pair β the #6 for main lines, the #4 for spurs and yards.
Reach for the #6 when
- It sits on a running line or a station through road
- You own long-wheelbase locomotives, long coaches or Shinkansen sets
- Smooth running matters more than saved inches
- You can afford the longer formation
Reach for the #4 when
- Board depth or yard length is genuinely tight
- It feeds an engine shed, yard, short spur or slow lead
- You need to switch it to non-power routing
- You are willing to test each car over it
Our working preference: if there is room, put #6 turnouts on the main line. Turnouts are where derailments and intermittent contact concentrate, and a gentle divergence removes a whole category of future annoyance. Save the #4 for places a #6 genuinely will not fit, or where trains crawl anyway. That is a preference, not a law β whether your longest car takes a #4 is answered by running it, not by reading. Before anything gets ballasted down, test with your longest, lightest and oddest-riding stock, slowly, in both directions, more than once.
Power routing versus non-power routing
This is one of Unitrack's defining behaviors. Power routing means only the route the points are set for is live; the unselected route goes dead β which hands you a free storage spur, since a train parked on the dead road simply sits there. Non-power routing means both routes stay live no matter where the points are thrown.
- Every Unitrack turnout leaves the factory set to power routing.
- The #4 has a switch on the underside to change it to non-power routing.
- The #6 is power routing only β there is nothing to change.
One misreading to head off: "power routing" is not a mode you select on the #6. It is the name of the behavior β current follows the points.
Both work with DCC
Neither turnout is DCC-incompatible. The only DCC-relevant difference is that the #6 cannot be taken out of power routing. Your actual wiring β feeder spacing, power districts, whether you need extra feeder track or terminal UniJoiners β should follow your command station manufacturer's guidance. Do not let "with DCC the whole layout is live" stand in for a real plan covering short circuits, districts and point-rail contact. If you are still deciding between the two systems, our DC vs DCC comparison covers where the boundary actually falls.
The control switch is sold separately
The switch machine is buried in the roadbed on both turnouts β nothing to mount, nothing to wire under the board. But remote operation needs a control switch you buy separately. The current black one is 24-840BK, $12.00 at KATO USA and Β₯770 in Japan; the older blue 24-840 is discontinued and functionally the same. KATO USA lists it as the Turnout Control Switch, which is the same part Japanese documentation calls a Point Switch. Neither turnout includes one, so budget for one per turnout β or a shared switch per pair, if you are throwing a crossover.
Plan the formation before you buy
Two things are worth saying about buying turnouts specifically. First, a turnout is both the most expensive and the most failure-prone piece of track on the layout β it has a moving mechanism and electrical contacts, and it is unforgiving if ballast glue creeps into it. Everything else you can rip up and redo cheaply; a fouled turnout you usually replace.
Second, work out the whole formation before you order: how many turnouts, how many left and how many right, whether you also need curves and straights, and how many control switches. A symmetrical passing loop needs one of each hand β for the #6 that means a 20-202 and a 20-203, not two of one. It is worth laying the plan out in the track planner and reading the piece list off it before anything goes in a cart. If you are mixing brands, note that TOMIX handles route power differently and KATO's logic does not carry over β see the KATO and TOMIX track compatibility guide.
Sources
Checked 2026-08-01. Part numbers, geometry designations, box contents and power routing behavior are from KATO's own product pages; prices are the Japanese list price and the KATO USA store price on the date checked. Reliability comments only restate the manufacturer's own cautions.
- KATO Japan: Unitrack electric turnouts (EP481-15 / EP718-15 designations, #4 auxiliary track list)
- KATO Unitrack: turnouts and crossings (part numbers, Japanese list prices)
- KATO USA: N Electric Left Hand Turnout #4 (SKU 20-220, US price)
- KATO USA: N Electric Left Hand Turnout #6 (SKU 20-202, US price)
Put the turnout into a full plan first
KATO #4 and #6 turnouts: common questions
What is the real difference between the KATO #4 and #6 turnout?
The diverging radius, not the angle. The #4 (20-220 left, 20-221 right) is 126 mm long and diverges on R481; the #6 (20-202 left, 20-203 right) is 186 mm long and diverges on the much gentler R718. KATO's own part designations, EP481-15 and EP718-15, say it plainly: both leave the through route at a nominal 15 degrees, and the radius is what changes. KATO positions the #6 for main lines and the #4 for spurs and yards. There is one more difference that catches people out: the #4 can be switched between power routing and non-power routing, the #6 cannot.
Is a KATO #6 the same as an NMRA #6 turnout?
No, and this trips up modelers used to American track. Under the NMRA convention a frog number describes the frog angle, so a #4 is roughly 14 degrees and a #6 roughly 9.5 degrees. KATO's numbers are product designations: EP481-15 and EP718-15 both diverge at a nominal 15 degrees. If you are laying out a yard ladder or a crossover on paper, plan from the published radius and the 15 degree exit angle, not from frog-number arithmetic.
Why does the shorter KATO #4 cost more than the longer #6?
Because of what is in the box. The #4 ships with a set of matching auxiliary track: two 64 mm straights, one R481-15 curve, and a left and a right 60 mm straight made specifically for the diverging route. The #6 is one turnout and nothing else. That is why the #4 lists at 3,300 yen against the #6 at 2,200 yen in Japan, and 37.50 dollars against 27.00 dollars at KATO USA. Compare the cost of a finished route, not the sticker on the box.
Should a small N scale layout use the KATO #4 or the #6?
If space is genuinely tight and the track is a yard or a short spur, the #4 is a reasonable choice. As soon as the turnout is on a running line, or you own long cars, favor the #6. Be careful with the phrase saves space: what matters is the whole formation including auxiliary track and track centers, not the bare length of the turnout. The 60 mm the #4 saves on paper does not multiply out across a plan the way people assume.
Do KATO #4 turnouts really cause derailments?
Calling it a faulty product is unfair. KATO does note that a tight divergence like the #4 is more sensitive with certain vehicles, and the outcome also depends on wheelbase, car weight, speed, how flat the track is, and how clean the wheels are. Before you ballast and scenic anything, run your longest, lightest and most awkwardly sprung stock over it slowly, in both directions, several times. That test tells you more than any forum thread.
What comes in the box with KATO 20-220 and 20-221?
KATO lists the #4 as one electric turnout plus auxiliary track: 64 mm straight track x2, R481-15 curved track x1, and a dedicated 60 mm straight for the diverging route in left and right, one each. The Turnout Control Switch for remote operation is not included. Those extra pieces are the whole reason the #4 costs more than the #6.
Does the KATO #6 include the R718 curve or a control switch?
No. The #6 is sold as one turnout on its own. The R718-15 curve, any straight track, and the Turnout Control Switch are all separate purchases. Do not read the contents of a V-series track set as the contents of a bare #6 β the sets bundle the turnout with matching track, the single item does not.
Can both the #4 and the #6 be used with DCC?
Both work on a DCC layout. The #6 is not DCC-incompatible; it simply cannot be switched out of power routing. Your feeder placement and district wiring should still follow the advice from your command station manufacturer. Do not let the idea that DCC means the whole layout is live replace a real plan for short circuits, power districts and point-rail contact.
What is power routing and non-power routing?
Power routing means only the route the points are set for is live, and the unselected route is dead β which conveniently gives you a free storage spur to park a train on. Non-power routing means both routes stay live regardless of how the points are thrown. Every Unitrack turnout leaves the factory set to power routing. The #4 has a switch on the underside to change it to non-power routing; the #6 does not.
