The Randall Museum in San Francisco hosts a large HO-scale model train layout. Created by the Golden Gate Model Railroad Club starting in 1961, the layout was donated to the Museum in 2015. Since then I have started automatizing trains running on the layout. I am also the layout maintainer. This blog describes various updates on the Randall project and I maintain a separate blog for all my electronics not directly related to Randall.
2016-09-25 - Turnouts ControlCategory Randall
This is where what should be simple becomes overly complicated.
The easy solution is to use NCE Switch-It (1 tortoise) or Switch-8 (8 tortoise) control.
There are a few issues.
This takes power from the track bus or external supply.
- Know issue ⇒ Do NOT take power from same booster powering the same block that is being switched. If there's a short when throwing the turnout, it will remove power from the switch-8 which consequently won't be able to finish switching.
- Power MUST be from either a different booster OR use the direct DC power plug.
- The layout does not use booster power directly. Instead it has 10 separate power districts, each with a short-circuit breaker.
- I do not have a clear map of the booster / circuit breaker / block.
- There is a known power shortage in specific places (mountain). It would not make sense to draw even more power from the bus for these locations.
- The Switch-8 will likely be located next to the control panels which by definition have the turnout power but not necessarily the dcc bus power.
- ⇒ Power via DC jacks, connected to the turnout power supply.
The real tricky part is maintaining compatibility with the current control panels.
The Switch-8 is designed to drive the Tortoise slow-machines directly. That means by default, it would replace the control panel rotary switches, instead of working in parallel with them.
There are 2 uses of turnout control in my automation design:
- In a few select cases, the purpose is to have the automation drive the turnout, e.g. in DCC Jr Eng Day we want the automation to align the Sonora and the T30 turnout based on the trains running.
- In most cases, the purpose of the turnout control is protective. E.g. turnout control is not a primary function of the automation but the idea is to prevent the automation from failing because an operator forgot to realign a switch. For example if somebody forgot to realign the Sonora switch or changed it after the automation had started, the passenger automation would fail. In this case as long as the automation is ongoing, the switch should be protected and the control panel ineffective. When the automation is stopped, the control panel switch should become operational again.
In the first case, parallel control of a turnout could potentially work but it is anyway not desirable. With twin-coils machines, electrical parallel control is possible. This is not the case with the tortoise slow-motion machines.
To make it clear, the ideal solution would be to rebuild the panels using simple push-buttons (one normal and one reverse) and at the same time add proper LED feedback to the panel, which results in clean very easy to use and understand designs. This is however out of the scope of the current project (both in matter of time, budget and skills) and there's a desire to keep the current "vintage" panels even though they are hard to use and very confusing for most operators.
3 possible ways to work around the limitation of working with the existing panels:
- This is the hacky quick-solution that will cost nothing but prove problematic later. Rotary switches are essentially DPDT "inverter" devices. They take the incoming 0-12V 2-wires and either emit them straight or cross their polarity to reverse the tortoise. Instead of powering the rotary switches from the common 12V bus, we can simply power them from the output of a Switch-8, having them in serie. This way the rotary switches basically pass the Switch-8 straight through or they invert it.
- Pros: Use existing stuff as-is. Wiring change exists but is minimal. Quick way to prototype things.
- The Switch-8 would be located near the control panels.
- Cons: Ever more confusing operation. Requires modifying existing rotary switches.
- Major cons: Breaks the whole point of securing the routes. Operator error is not eliminated, at the contrary it is made even more probable. ⇒ Not a viable option besides a quick and dirty test phase. The problem with "quick and dirty" solutions is that from a non-technical viewer they are indistinguishable from properly implemented solutions.
- I will NOT be doing that.
- Equip each tortoise with a relay, selecting from being powered by a Switch-8 or the control panels.
- Pros: This involves no modification to actual panels whatsoever. No cabling changes on panels. It's only where the wires from the panel go that changes.
- Cons: requires one relay per turnout or up to 8PDT (8-poles, dual throw).
- Can't find a general purpose 8pdt at decent price.
- Specs: 8PDT or 2x 4PDT, Form C, Non-latching? (explanation)
- 2x General purpose 12V relay 4-poles dual throw [jameco MY4-DC12 part 2167461 $4 each or same at mouser $7 + support socket jameco PYF14A-E-R part 2167470 $2-4 each -- relay spec here, socket here]
- The relay could be connected to the Switch-8 as a turnout itself, thus allowing the automation to trigger the relay. That would reduce the capacity of each Switch-8 to 7 turnouts + 1 relay. ⇒ Must double check that the non-latching current draw is compatible with the Switch-8.
- The relays and the Switch-8 would be located near the control panels.
- Cost is reasonable at $16 per Switch-8 + wiring.
- This is an alternative to solution A using a relay. In this case however the Switch-8 does not control the turnouts. Once in automated mode, the turnouts are fixed in the desired mode. When automation is disabled, the turnouts are operated using the rotary switches.
- Cons: Involves as much work as solution A (need to have relays and double-wire each turnout near its command point) but does not add possibility to control via NCE.
Solution D (retained solution):
- Use the NCE Button Board add-on for each Switch-8 (part 05240152, PDF).
- Each power is powered/connected to a Switch-8 and has 8 Normal/Reverse inputs. Simple 3-wire wiring to the Switch-8.
- Each rotary DPDT switch would be rewired to act as an input on the Button Board.
- Pros: Buttons can be locked out via CV programming on each board.
- Cons: Requires modifying each rotary switch.
- If the existing rotary switches are SPDT or DPDT, "convert" them to SPDT connected to the button board. Must use the GND.
- If the rotary switches are 3PDT or more, can use a free input.
- Either way, the existing cabling of the rotary switch becomes irrelevant.
- Note: The PowerPro Command Station remembers the last ACCY state (so we won't have that status via USB, maybe Serial?). On a Switch-It, the Command Station does not see ACCY toggled using the wired inputs on the board. This limitation seems fairly academic since here we'll be using the USB which has no such feedback.
- Note: Not explicitly stated in the Button Board page, but are input triggered all the time or on low-edge? On the Q-Snap it's clearly low-edge (floating to grounded) and on the Mini Panel it also clearly states it's triggered on low-edge to ground. We shall assume the same here.
- Note: The CV 556 to lock the buttons should achieve the desire of locking out panels ⇒ Controller can still be used to throw the turnouts. Resetting the system will unlock.
- The Switch-8 and the button-board would be located near the control panels.
- Cost is at $30 per Switch-8 + wiring. Plan is to have 2x Switch-8 in phase 2 then 5x in phase 3 so that would be a projected cost of 5x$30=$150 max.
- Use the NCE Mini Panel [NCE part 5240230].
- This has 30 input, all grounded SPST contacts.
- The mini panel acts as an NCE cab, directly on the bus. It sends commands to the Switch-8 using the accessory switch commands, as the USB or ProCab would do. It is thus completely isolated from the Switch-8 and does not need to be located nearby.
- The mini panel inputs cannot be queried from the bus. Their sole purpose is to generate commands, in this case 2 inputs would generate Normal or Reverse for one turnout.
- Each rotary DPDT would be rewired to act has 2 SPST inputs on the board, thus 15 rotary switches per mini panel. Each one is thus equivalent to about 2 Button Boards.
- There's a lot of programming involved. Every single input needs to be mapped to the corresponding turnouts. Done once and can be clearly documented.
- The fancy macro programming is not useful in this case for mainline.
- There is no way to trigger an input from USB.
- Pro: The fancy macro programming would be good for yard ladders *except* there is no way to trigger it from USB and I have no intention of DCCing the yards turnouts.
- Cost is $50 per mini panel. Plan is to have:
- 1x in Passenger-2, so $50 budget (10 turnouts out of 15, +1 in JED-2 crit).
- 1x in JED-1 (1 turnout out of 15, +6 in JED-2 crit).
- 1x in JED-2 (+5+3) mountain in non-critical phase.
- 1x in JED-2 (+10?) valley in non-crit phase.
- Budget is start with 1 ($50), expand to 2 (JED, another $50), these can cover the critical turnouts. If doing the later non-crit, might need 2 more ($100) for a grand total of $200 but not getting them upfront.
- Cons: The overall cost will thus be a bit more than with Button Panels and it will be a lot more work to program. Wiring is about the same.