Using Hall Effect Sensors as chuff triggers

New or modified article by Eric Timberlake (talk) undergoing review

Upgrading your Chuff Sensor to Hall Effect Sensors for seriously accurate, seriously reliable chuff.

One of the things I like about Garden Scale railroads is that they are mostly not toys, they are solid working models of the “prototype" locomotive, including “realistic” sound.. Many of the current steam locomotives feature built in electronics to simulate the sound of the real thing, including synchronized chuff. I don’t know about you, but that “synchronized” part is extra meaningful to me. The models are big enough that you can clearly see them actuate, and when “auto chuff” is racing while the wheels barely turn, well, that’s a disappointing detail in what otherwise are many very beautiful models.

And then there’s the common case of owning an older model, or, perhaps one that the designers tried, but effectively failed to provide this feature?

In this article, I will discuss my current method to provide chuff input to models that may or may not be "ready for prime time" in this regard..

the premise

Shay Chuff Sensor

A few years ago I was asked to upgrade a mid generation Bachmann 38 Ton Two truck Shay. This locomotive uses an Upright Inline Triple Cylinder arrangement and as a model, it looks fantastic.. However, the built in chuff sensor has proven to be a design failure of epic proportions. (See Photo 1) The trigger switch is comprised of two thin sheets of brass arranged as three simple "fingers" reaching into the motor where the piston-less push rod would poke it and create a signal for the sound card or sound capable decoder. The issue is that the fingers are often out of alignment and corroded/oxidized. This leads to missing chuffs, and to "jitter" chuffs where a single contact results in 10's of chuff triggers all at once due to switch bounce. Working on this friends Shay, I replaced this troublesome assembly with reed switches and tiny magnets.. It was a TEDIOUS job. It took three tries to get it right... Endless frustrating hours of work.. After I got it working I vowed to never again go the route of small expensive reed switches. The root of the issue is the very specific polarization required to operate a reed switch.. A reed switch requires an aligned field from N to S along the length of the switch. This is just plain difficult to achieve in many applications when MOST small neodymium magnets mount with just just the N or S pole facing the target area.

the cure

Fast forward a few years and another Shay lands on my work bench with the same crappy stock chuff trigger. I relate my story to the owner and suggest that I can do better this time. This is the path I took..

First, I ordered and tested a wide range of Hall Effect Sensors from Digikey Electronics. These devices are small transistor like devices with three leads and an array of sensors inside that can detect a magnetic field. When such a field is detected, it turns on an "Open Collector" output which is effectively just a small switch to ground. I ultimately chose an "omni polar" (doesn't care about North or South poles on the magnets) device made by Texas Instruments, the DRV5033 in a modified TO-93 package. That's a form factor with real leads vs. a surface mount micro-dot that is so very common in todays electronics. This device cost $1.67 each at my last purchase. A typical reed switch, especially a small one can top $5 each pretty easily.

the sensor

https://www.digikey.com/en/products/detail/texas-instruments/DRV5033AJQLPGQ1/6571747

The DRV5033 is a “non-electronics savvy” modelers dream device. It is tolerant of up to 38 volts applied in ANY possible way.. You cannot hurt this device in your train, it just may not work if you have it wired up wrong, but it WILL NOT self destruct under normal circumstances.. This makes it ideal for the non-electronics aware modeling nerd - you simply have to know which pins are what. And you could just GUESS if you wanted, though looking at the data sheet is a lot easier than testing all nine permutations of possible attachment combinations..

the magnet

The next step is a little more challenging, but the internet, if you dig hard enough, has everything, right? In this case we are looking for SMALL magnets. These are harder to find than I would have thought, but I DID finally find a supplier that featured VERY SMALL disc shaped Neodymium magnets. I ordered the 1mm around by 1/2mm thick. Actually, I ordered several sizes to support my future experiments and projects..

https://supermagnetman.com/collections/micro-magnets

I picked VERY SMALL magnets for a reason, and that reason is sensitivity. The "stroke" of a Bachmann Shay cylinder is about 5/8 of an inch. If you put a typically available “small”, sat 1/8 x 1/16 of an inch vs. a "micro" magnet, the DRV5033 will trigger when the rod is still 1/4" away. Turn the motor 120 degrees and we find that cylinder one is still close enough to be in trigger mode when cylinder two ALSO triggers, then the same with cylinder two to three another 120 degrees along - the chuff triggers, when viewed as a WHOLE, never de-trigger! To help with this, we use a VERY SMALL MAGNET that won't trigger the DRV5033 until its only a few MM away, giving us time to UN-trigger before the next cylinder reaches the Top Dead Center trigger point.. This is slightly different from mounting magnets to a driver wheel, where you might actually want a stronger magnet to reach the sensor.. Also, placing a magnet in a little pocket in these pot metal wheels seems to weaken the magnetic field a touch in my experience.

To make my prototype device I used a small slice of readily available protoboard with solder holes on 0.1 inch centers.. I sliced and sanded till it matched the outline of the original switch assembly, then started mounting components, which in this case are FEW. Three of the DRV5033 Hall Effect sensors, three capacitors and.. Oh. That's it.. The data sheet of the DRV5033 says you should place some "decoupling" .01uf capacitors near the device, so I did. Most of the work was simply soldering a bunch of wire around to make the connections to all the devices.. A little testing, then some black spray paint and you get a new solid state three way chuff sensor that fits right where the original did.

A little black spray paint and the entire assembly becomes as invisible as the original.. Photo 2 shows the results of my rather mashed together prototype.

how to attach the sensor

As far as wiring it up, this was relatively simple. Pin one, V+ went to the battery plus. Pin two, ground, went to battery minus. Pin three, the "output switch" was routed to the Chuff Trigger Pin, in this case a revolution 57002S decoder with a dual "trigger input” jack on the decoder and included wire harness.. These inputs aren't described in the manual at all, which is disappointing.. However, a little testing showed me that "trigger one" set off the whistle, and "trigger two" set off a chuff.. A trigger on the chuff input seems to disable auto chuff, so no additional configuration is required. The last thing required is a COMMON GROUND - this is critical in the electronics world. The Revolution 57002S in this case appears to feature a 100% isolated 3 volt power supply, so ground on the decoder and battery minus need to be connected together so that the Hall Effect sensor output and the decoder share a common ground reference. Some brands of decoders and sound cards actually have +5 volts and system ground available to power the sensors so ground is ground, so to speak. Check your decoder manual for more details on the trigger input and power configuration that might be available..

works everywhere you can put the sensor

While I have focused on the Bachmann Shay, the same concept applies to any locomotive. Typically I will configure "Single Chuff", which is NOT prototypical, but for some reason seems to simply sound better on a model. Most people I know who have modern Bachmann locomotives with a choice leave the "2-4 switch" in the 2 position. This is equivalent to SINGLE CHUFF at the recessed end of the stroke. To achieve this is pretty simple MOST of the time. I use a carbide bit and dig a small hole in the back of a convenient wheel, then epoxy a small magnet into the divot. Next you glue the DRV5033 to the chassis so the magnet passes over it. Bingo. Chuff. ACCURATE chuff that works every single time.. If you want real prototype chuff, you just have to arrange four magnets around the wheel spaced 90 degrees apart, or two on one wheel and another two on a different wheel. Be sure to DOUBLE CHECK stroke alignment before you start digging holes in your wheels!

The Shay cannot actually present double chuff - there’s just no place to mount another round of sensors at the bottom of the stroke. So “single chuff” it is..

pick your wheel carefully

A short note on wheel selection here - to get accurate chuff, you have to use one of the driver wheels that are directly connected to the pistons. You can’t (or shouldn’t) use a small wheel or a wheel in the tenter, for instance. Using a non-driver wheel technically will make a chuff, there’s just no relation to the driver wheels and piston position. Many stock loco’s use this sort of arrangement - I just took apart an Aristocrat 4-6-0 that had a hall effect sensor on the tender wheels, which in no way could create accurate chuff at any speed, much less slow speeds, which is where we really care the most..

If you choose to use two sensors, you can just tie the output pins (pin 3) together without fear. Remember, the "open collector" output is just an electrical switch to ground - you can stack as many outputs together as you want. The active output on any one device will short the chuff trigger to ground and create a chuff sound without affect other devices on the line.

variations

Replacement Shay Chuff Sensor

I was motivated by the apparent success of the second Shay project. But assembling the prototype board was, just like the original Reed Switches, “tedious”.. Confirming that the concept works and see how the owner seems happy with the end result, I sat down and designed and ordered a few circuit boards (See photos 3 and 4) that would make the assembly process significantly easier, solder down eight components and three wires and that's it - new solid state chuff sensor.. I ordered six each of these boards and the components to populate them. If you are a Shay triple cylinder owner and wish your chuff would work right, contact me at shayhalleffect@mtrcycllvr.org and perhaps I can make a guinea pig out of you too, er, I mean, help you out.. You need a sound capable decoder with a hard wired trigger feature installed. I suspect this would work with the stock DCC sound decoder without issue, but I haven't had a chance to attempt it.. There are only so many Shay's out there, right? So the six boards I made will likely last a long time..

Here's a link to a way too long video of the Shay being put together using a prototype Hall Effect Sensor assembly.

related video

https://youtu.be/UT7dVgtUHes

38 Ton Two Truck Shay provided by fellow RCGRS Member Phil Yokers. Thanks, Phil! A privilege working on your model!

--Eric Timberlake (talk)