TCS - Train Control System Decoders

Overview

TCS makes a lot of stuff, I won't cover it all.

Quality is OK, but it's not top of the line stuff, it's "upper middle class".

I've found fundamental bugs in their decoder firmware, read below, and the new WOW sound & motor decoder is ok (but buggy), but for $150 it's the wrong price point, it needs to cost $100, or have features of the $190 motor and sound decoders from QSI, Massoth, Zimo, etc. (or have all the bugs fixed and improve the sound)

Note: I am currently persuing updates/upgrades on all my TCS decoders, so the problems listed here MIGHT get resolved.

In response to a thread where again my forum nemesis is questioning my statements on DCC, Stan Ames posted this email from "Dan in TCS Customer Service", posted May 7 2018 on Largescale central forum:  (red highlighting mine)

"TCS is continuously improving our products and services in order to better sever our customers and address their concerns. As a small, family-owned and family-grown business, we have of course climbed some steep hills. However, through hard work and determination we are proud to say we offer the world's best DCC decoders. We are aware of the initial 'bugs' in the WOW501, as it was rushed into production in order to meet the overwhelming demand. We have learned from this experience and responded to the negative feedback by ensuring we never rush another product. Very few WOW501 decoders are in circulation which still have the known bugs still in them. Once we discovered them, we had as many as possible recalled and updated. Decoders manufactured in or after May of 2017 are fixed. The WOW501 uses PWM on both the coil and the fan to provide the proto-smoke feature. We don't use 'fan braking' because it isn't constantly running. This can actually extend the life of your smoke unit. We're going to continue to experiment with this feature so we can implement it on HO decoders and smoke units.  Something else to be aware of is that WOW501 decoders (as well as other V4 decoders) have built-in protections such as over-heat sensors and protections which will turn the sound off if the audio peaks too high/your speaker shorts. I.E. if your volume is too high and you are playing too many sounds at once, the sound will cut out until the fault clears. 

Speed table issues are a hurdle we have done our best to fix, but it is a limitation of V4 and older processor architecture. The processor's breakpoints for the PWM on the motor drives can provide some odd results at specific speed steps. In most cases, you don't notice them, as they are more prevalent at high speeds. If you're running your engine like a Metro car, then sure, you'll notice. A fix was put in place in July of 2016 to address issues with CV5 and CV6. Additional fixes which we consider to have completely addressed problems were implemented in February of 2017. Lockups when using ULST were fixed in January of 2017. We haven't seen these problems on our new development architecture for V5.

We are also aware of the claims of 'erratic behavior', but have never observed this in-house. We have never seen one of our own test decoders lose its programming or stop responding, even after many hours in service with constant programming operations and power cycles. JMRI has been known to randomly do strange things to our decoders, but seems to have been fixed by version 4.7.3. We consider 4.7.3 and newer to be 100% compatible with TCS decoders."

Wow sound decoder for large scale

First Impression

First off, I have to say the Youtube video really set me off in a negative direction, the intro and ending music on the video is, well, over the top. Almost a minute of just the 2001 theme and loud, nothing else. Dear TCS: the WOW sound decoder is nice, has a good price point, but is not even on par with the movie 2001 and Stanley Kubrick. 52 seconds of music with no train! Also "coming in 2013"? I guess the "halleuja" chorus was indicative of TCS' celebration that it was only 3 years late from their prediction?

Sorry guys, I have to poke fun at this. Guys, please edit the 52 seconds at the beginning and the chorus at the end from the video, and take out the "Coming in 2013" it really gives a poor representation of your company!

 Looking at the video, and paying attention:

You hear the dynamo spin up, but the headlight goes on immediately, in a burst. Other decoders go from dim to bright like a real locomotive. The headlight also comes on before the dynamo has spun up, whereas it should come up after.

The next thing I did not like is that the transition from open drain cocks to closed was like a light switch, one chuff they are open and the next closed, no fade and all 4 close at the same time? Yuck. Just sounds artificial.

The first movement with the drain cocks on is not too bad, although you hear there is a sound with chuff and drain cock open, and there is a chuff and drain cock and rod clank, a little weird.

Whistle sounded nice.

The "dynamic chuff" is volume only. Compare this to a QSI or Zimo or Massoth where the sound of the chuff is different under load, heavy load, and coasting.

Also, for example, when you reverse direction, the QSI will give you the sound of the reversing linkage.

So, is it a good sound system, yes. Is it GREAT? No. It does not have all the nuances of sound we are used to having with the top level decoders. Is it worth $160 vs. $200 for a QSI or Zimo or ESU or Massoth? That's your decision. It's not as good, but there is a $40 difference. You need to decide.

OK, on to the decoder.

All I can say is WOW! What a mess!

I'm evaluating the Diesel unit and having some issues, and have noticed some disturbing things.

I installed this in an Aristo eggliner, this motor is noted for smooth and low power draw, and it was a convenient test bed.

First, I try the "typical user test", hook up the track and motor and speaker, put it on the rails and select address 3.

WHOA! The dang thing takes off... ok reduce speed to zero, the dang keeps going! Oh well, it stopped when it ran off the ends of the track! At first it seemed that the unit was programmed with an incredible amount of momentum, but reversing the loco did nothing either??!!

So, first impression: FAIL. Read the manual, and notice sound cuts out too, and the manual says use the audio assist menu and calibrate the loco. Hmm, you press F8 4 times to enable the menu, pretty cool... oops, press too slow and you mute the sound, press twice and came up, not four time. Oh, maybe the guys at TCS think F8 is momentary not latching. Anyway since the woman started yakking at 2 presses, I could tell I was in the menu.

OK calibrate the motor, you run it slow and press a key then fast under a load. OK, that surely fixed it. Run the train. Another runaway! What the hell? I noticed a brake function, that seems to do something, but up to 5 presses to get more brakes? How the heck do you remember what "braking level" you are at?

Running the loco I notice the sound cuts out and on usually when you reverse direction. I keep crashing the loco.

OK, read the manual again, notice the part about prototype mode and traditional mode, yeah saw that before... oops! There it is, the prototype mode, where you have to press F7 like a madman, in lower case, it says the DEFAULT mode is the prototype mode. OK, turn that off! OK, now it runs like the rest of the DCC decoders made in the world. (Note to TCS: you can make a decoder work with brakes and also recognize the throttle is turned to zero, EVERYONE else does this).

OK, now I have control, finally I can run on the layout. Good news, the times the sound cuts out completely is less, but NOT gone. Now the thing I notice is that the motor sound kind of runs on it's own, sort of like just running slow, no matter if you are accelerating, cruising, or stopping. It sounds like it is just coming off idle lightly. It makes a bit more noise when accelerating. So, very disappointing. Now come to a stop.... the motor sound will continue JUST the same as if it was running for several seconds.

Really disappointing. I'm sending it back. Really poor and how can someone not notice these things. I'm guessing, adding in my experience with the bugs in the firmware in the Z scale units also, and the big delay to market, that this unit is not fully baked yet. Perhaps in a year or 2 a new firmware update will correct this. The issue I have is that there is no mention of improving this or fixing this. In fact when I brought the speed table issues to light in their forum, all I got was denial and chastisement, but persisting and explaining brought a number of people out in the forums who have identical issues.

So, TCS, the hardware looks nice, the price point is good, the sound is adequate, dual speakers 8 function outputs, now just improve how it works please.

Programming track issues:

I have both Zimo and NCE systems. I rarely have issues reading ANY decoder with either, both do direct, paged and register mode programming,

But I had basically no success with the NCE system, getting erratic results or read fails. I could never read CV7 for example.

With the Zimo system, when you try to identify the decoder, the system will increase the programming voltage if it has difficulty, so it was running at 14 volts, maybe that is the difference. It also retries the commands a few times. I still had maybe 20-30% read failures, but could read often enough to get what seems to be the real values.

The CV's for the firmware date match the audio readout, and I was able to read CV7 most of the time.

So the upshot is that the decoder is not reliable in reading in service mode. One user says, Oh it works, in JMRI go to Preferences->Roster->Programmer and turn off Index CV Caching and Confirm Writes with a Read. These aren't compatible with TCS WOW decoders. No shit, you turn off the readback function. crazy!

Problem with setting "crew announcements" using AirWire

If your loco runs for a while and then puts on the brakes (makes the brake noise and stops), this is most likely the result of a poorly documented "feature".

What happens with this feature on, if you do not issue a command to the decoder for 43 seconds (default setting), then for the next 15 seconds, the system beeps, warning you of imminent "stopping" and after this 15 seconds elapses, the system will set the brakes, you will hear the squeal and the loco will stop.

In a quest for prototype operation, this feature basically emulates the "dead man's switch" in a loco... it makes sense from a prototype view, and from a "normal" DCC viewpoint makes perfect sense. DCC normally spends it's "spare time" sending redundant "speed commands", so this timeout never happens, and if it does, then that usually means something bad has happened, power is there but the DCC signal is corrupt.

OK, so why the problem using AirWire? Because a compromise has been made on some  "wireless DCC" systems. Power is never an issue for normal DCC since you have a 5 to 20 amp supply, but when your DCC command station is in a battery powered handheld, having the transmitter transmitting continuously is an issue. So normally products like AirWire ONLY transmit when there is a change, like you press a function button, or CHANGE the speed.

So what happens is you are merrily rolling along, everything is fine, but if you don't change the speed or push a function button every 1 minute, your train will stop because of the timeout in the "crew announcements" mode, which we have figured out is also a "deadmans switch".

So, the moral here is don't enable this feature in AirWire or any other system that does not work like "normal DCC".

TCS Z scale decoders

TCS makes several mobile decoders, the AZL4 drop in for many AZL locos, the MZA4 drop in for many MT locos, and the Z2 wired decoder for hardwired installations.

As an overview, many people like these decoders because they have 4 lighting outputs (the standard headlights and 2 more) so for extra lighting it's nice to have.

They also make 2 function only decoders, the FL2 and FL4.

My biggest gripe is the custom speed table. A bit of background: very few people use a custom speed table. Usually Vmin, Vmax, and Vmid (if implemented) is sufficient.

Many people complain about different speeds of locos at the same Speed Step, but don't do anything about it, and it's easy with the above settings.

I've used speed matching in G scale for a long time, since I run mainline trains with multiple locos on a train more often than not.

In Z scale, there is a VERY wide variation in speeds, due to different gearing, and something more common in Z scale, identical locos will run at different speeds, due to small differences in friction, mechanicals, etc. It's just a fact of life in this tiny scale where a cat hair on an axle makes a big difference.

So, I use custom speed tables, where you get 28 different values for "voltage" at each speed step. Suffice it to say in 128 SS mode, different manufacturers do different things, some do not support custom speed tables in 128 SS modes (stupid, mostly old designs), but most modern ones do, and interpolate the speed from 28 to the 128 steps.

Suffice it to say here that TCS decoders are crap (yes this adjective/noun is deserved) in this respect, the short answer is that there are 2 symptoms, both I think of a common issue internally.

1. setting a progressive set of entries does NOT give a progressive set of resulting speeds.

2. modifying certain steps in the custom table can produce wild and unpredictable results.

I'm stopping here, to finish later, and fill in the details, and how you can demonstrate this for yourself. I also need to check out the G scale decoder and see if it has the same problem.

My version reported from CV7 is 49.

MZA4

The picture below shows the MZA4. This is a drop in decoder for the MT GP9, GP35, and SD40. There was a "recall" where there was copper very close to where the motor springs go. This copper was connected to the circular "land" that the chassis "pins" connect to, thus creating a short between the motor and the track pickups. Don't know the dates on this. http://www.tcsdcc.com/Customer_Content/Technical_Info/Tech_Info/FAQPage.php?q=64

The solution is to cut/grind away the little chunk of copper in the picture (or you could use kapton tape it seems)

There are 4 functions, counting the two headlights. There are 3 pads to support the 2 extra function outputs. In addition, they indicate the location of the current limiting resistors for the 2 stock leds.

Z2

The Z2 does not have the extra 2 solder pads for the extra 2 functions like the drop in.

TCS Version information:

CV 7 - version number (does not work on my wowsound)

CV 8 - mfg ID (153)

CV 248 - sound set version (wowsound only??)

CV249  - decoder type (wowsound only according to manual) but seems to have the following values:

FL2 is 6
FL4 is 7
MZA4 is 40
Z2 is 41
I got 172 from the WowSound diesel, don't know if this is right or not.

CV250 - month of f/w programming

CV 251 - day

CV 252 - year

(note on the WowSound, the audio readout of this date matched what was in registers 250-252)

BEMF

Many people have performance issues with BEMF enabled.

Years ago, general counsel was to turn off BEMF when consisting, not sure this is still true for all manufacturers

BEMF is enabled by default, It seems that you cannot disable BEMF on WowSound

CV182 has some interesting BEMF features..

bit 0 - automatic bemf / stop on dc enabled / no keep alive

bit 1 - set for special bemf for 3 pole motors

bit 2 - automatic bemf - stop on dc disabled / keep alive enabled,  this is set by default..

bit 3 - disable safety check (need function packet received twice to act on it) (wowsound)

bit 4 - approx 2 second motor delay (wowsound diesel)

TCS complete CV function list:

The "manual" supplied with the decoders, well, it basically "sucks" in my "DCC" opinion. I've seen this where there is a "starter" manual, kept short to avoid confusion, then more manuals for more advanced topics. I am firmly against this, it creates confusion, either by omission of information or the pain in the butt managing multiple documents and where to find what. Make a single manual with the "getting started" in the beginning.

I'm creating a single list here, because some are not listed, and many are listed in a confusing way, and some are on separate documents.  Here's my cheat sheet and comments.

Things in bold are stuff that is "different" or "interesting".

CV 1 - short DCC address (default 3)

CV 2 - start speed, (default 0)

CV 3 - acceleration momentum (default 1)

CV 4 - deceleration momentum (default 1)

CV 5 - top speed (default 0)

CV 6 - mid speed (nice) nice (default is 0, straight line between CV 2 and CV 5)

CV7 - decoder version

CV 8 is decoder reset, set to 2 for reset, resets most stuff. Read back for manufacturer id, should be 153 for TCS

Factory reset can be done on the main track or on the programming track. To reset CV's to their factory setting write a value of 2 into CV8. The reset occurs the moment the decoder gets the reset command. (Earlier decoder versions had to go through a power on/off before the reset occurred.) CV's that will reset to factory default: 1, 2, 3, 4, 5, 6, 17, 18, 19, 29, 30, 49, 50, 53, 54, 55, 56, 57, 61 . The user-loadable speed table is not affected by the reset

CV  10 sets BEMF cutout speed

defaults to 0, i.e. BEMF always on... range is almost exactly mapped to speed steps, so if you want it to cut out at ss 10, set to 10. This might be helpful in consists.

CV 13 - analog power control "turn off black or red powered functions"

analog power conversion (DC mode) must be enabled in CV29... also refers to "brake on DC". Apparently if the loco crosses from DCC into a DC "block", the loco will slow to a stop (subject to CV4 setting), "convert" to analog power, and then accelerate to  a speed appropriate to the DC voltage level. Weird, but I think some of this is legacy stuff for when DCC was young. (Thanks for the explanation: Garth Hamilton) default is 255, don't know what different values do.

CV 15 is decoder lock

to unlock set to zero or equal to CV16, setting CV 15 to 7 locks all decoders that are at same address, same address decoders need different CV16 subaddress.  (I guess this implies POM)

CV 16 works with CV 15

convention is to set to a number for groups, can be set to anything apparently, see the manual for a confusing explanation, but =1 for mobile, =2 for sound, =3 lights only as defaults, you can set it to anything if you want a unique lock.

CV 17 & 18 - standard long address registers (default 0)

CV 19 - standard consist address register (default 0)

CV 21 - consist lighting control - affects function output wires, green, purple, brown wires in a consist. (only green and purple on Z2)

green is 1, purple is 2, brown is 4. add them in if you want them active in consist.

CV 22 - consist lighting control - headlight/white wire = 1, rear headlight/yellow wire = 2, set these for active in consist as in cv 21

so lead loco wants cv22 =1, rear loco wants cv22 = 2

Note: there is something erratic in this programming, had to program locos several times for it to work right, even though reading CV22 gave the right result. Also seems that changing any other lighting configuration wipes this setting out.

CV 23 & 24 - standard acceleration and deceleration in consist (default 0)

CV 29 - follows standard (default 6) Note that 28 speed steps and DC mode are the default, the manual is really backwards, implies 6 means 14 steps and disable DC mode

bit - value - meaning when "on"/set

0 - 1 - reverse direction (default forward)

1 - 2 - enable 28 speed steps (default)

2 - 4 - enable DC mode (default)

3 - 8 - enable railcom

4 - 16 - activate loadable speed table (default use standard speed table)

5 - 32 - enable long address (default short address enabled)

CV 49-52 control the lighting effects for the outputs for white, yellow, green, purple respectively. (white and yellow are front and rear lights) (note Z2 only has white and yellow)

set to 0 through 13 for front light only, add 16 for reverse light only , add 32 for both - see their table http://www.tcsdcc.com/Customer_Content/Literature/Decoders/Z_Scale/MZA4/MZA4_WEB.pdf

CV 56, 57, "dither" frequency and voltage - when enabled, it is enabled all the time. (defaults 3 and 10 respectively)(enable with cv 61) (My opinion is the TCS BEMF is fine, even in consists)

Like "kick start" in NCE, only works if BEMF is off. Both need to be greater than 0 to be working.

56 is frequency, default is 3, 1 is highest freq, 255 is lowest

57 is voltage, default it 10, lowest is 1, highest 255

TCS recommends using dither of 15 as a starting point, and increase by 5 at a time until you get proper low speed starting. Then play with frequency by 1 at a time to see when it improves.

CV 61 - BEMF control and dimmed lighting (default 1, BEMF on, no dimming)

bit - value - meaning when "on"/set

0 - 1 - enable BEMF (always active on WowSound)

1 - 2 - enable function button control of BEMF (won't come on unless bit 0 is 1) (default is function key 6) (always active on WowSound)

2 - 4 - maybe enhanced button control, auto control - f2 on/off, direction for reversing

3 - 8 - button control of braking (trolley)

4 - 16 - enables headlight dim when stopped (rule 17 enabled)

5 - 32 - enables the "opposite" light dimmed (rear light dimmed in forward)

6 - 64 - maybe used in button control of motor output (f2 forward, f3 reverse)

? - 68 - ??? WTF ?? Automatic Button Control of motor (looks like this function is really bit 6 and bit 2)

7 - 128 - button control of brake lights (trolley)

CV 64 - sets dimmed brightness

2-6 for LEDs, 12-18 for bulbs. Default is 15 - need to experiment

CV 67-94 (custom speed table)

note: seems you have to do this in service mode, not pom

CV 105, 106 - user registers

CV 125-132, supplemental acceleration curves "variable momentum", that "enhance" CV3 and 4. (not sure these are implemented)

For acceleration, cv 129 - 132 take over from CV3, cv 129 is the rate 2 starting point, and cv 130 is the rate 2 acceleration, cv 131 is rate 3 starting point, and cv132 is the rate 3 acceleration.

Similarly, cv125 and 126 are the rate 2 "pairs" for rate 2 deceleration, and cv127 and 128 are the rate 3 pairs.

CV 136 - sets which function button can turn BEMF on and off

Default is 6, but this is off by default in CV 61 (bits 0-7 correspond to functions 5-12, 1=5, 2=6, 4=7, 8=8, 16=9, 32=10, 64=11, 128=12 (very weird)

CV 248 - sound set version (wowsound only??)

CV249  - decoder type (wowsound only according to manual) but seems to have the following values:

FL2 is 6
FL4 is 7
MZA4 is 40
Z2 is 41
I got 172 from the WowSound diesel, don't know if this is right or not.

CV250 - month of f/w programming

CV 251 - day

CV 252 - year

(note on the WowSound, the audio readout of this date matched what was in registers 250-252)

Tips / Problems:

Speed control:

I noticed that if you set CV 5 or CV 6 to anything other than zero, then the speed of the loco does NOT increase with increasing speed steps, but has a "glitch" approximately ever 4 speed steps.

If you set cv 5 nonzero, and then step up from SS 0  by increments of 1 SS, (use 128 ss, it's easier to notice) you will hear/see the loco increase speed with every step, except at the "glitch points", where increasing the speed by one step actually results in a lower speed than the previous step.

The glitches I noticed were at SS 5, 9, 14, 18, 23, 27, 32, 36, 41, 45, 50 (did not bother going higher)

I found this out because I was working on a different problem, in a Marklin Mikado, increasing the speed near max, results in a runaway at full speed, uncontrollable. It appears the motor noise masks the ability of the decoder to receive commands.

By setting CV5 to about 70, this tendency was almost completely eliminated, and when the "runaway" occurred, usually a second speed command brought it back.

I surmise the first problem above is caused by bad coding, a math error in the firmware. Thus I reasoned a custom speed table would solve the "speed step glitch". It did.

The problem has been noticed by many people, although the "loyal" followers will deny it, often stating they have installed "dozens" and never seen the problem.

Here's a link to a discussion and description of the issue:

https://www.therailwire.net/forum/index.php?topic=40182.0

Apparently a TCS rep acknowledged the problem in October 2016, so maybe we will have some resolution.

 The stuff below is messed up, will fix as I finish the CV table and incorporate it.

advanced consisting and cv21 and 22

make sure all lights are off before doing any programming

 TCS UWT-100 Throttle

New in 2021, a small, universal wireless throttle with an LCD screen.

• Uses Wi-Fi
• Has 4 speed up down buttons and a roller wheel just linke NCE
• uses 2 AA batteries
• comes with wrist strap, a nice feature, the connection loop can have a lanyard attached (my preference)
• entering letters is assisted with the thumbwheel, a nice idea
• the throttle interfaces to:
  • Digitrax LNW1
  • MRC
  • WiFi Traxx
  • JMRI
  • LCC??

Pro's

• slim design convenient
• interfaces to different protocols
• firmware updated via Wi-Fi and Internet
• If using JMRI, you can scroll through roster on screen
• backlit
• function labels (4 top buttons) updated by JMRI
• built in flashlight

Con's

• expensive, $300
• other than speed controls, no labelled buttons
• no clock on display
• throttle only displays 128 ss mode, even if you are in 28 or 14 SS mode
• loco recall only 2 locos

Has