Replacing MTH PS3 decoder

I'd suggest posting this question over on the "MTH DCS...." subforum since I imagine it applies to DCS boards [ the majority ] installed in 3 rail locos as well.

Good luck.

SZ

Decoder?  I presume you're talking about is one of the PS/3 boards?  I can't believe you can't obtain spare parts there!  It seems extreme to replace the whole electronics package for a failure of one of the boards!

Preiten

I'm not sure that the MTH tach reader will work with a normal DCC decoder. I know the engines is PS3 and can be ran by DCC, but I don't think the electronics are compatible. I know MTH has DCC ready engines in HO that you can plug in any DCC decoder and I don't think they have done it O yet.

What I'm trying to say is with out the MTH electronics to accompany the DCC decoder I don't think the tach reader would work on it's own.

The tach reader is just a single component that will give you a logic pulse for each stripe if powered with 5 volts.  There's no supporting circuitry on the MTH tach reader board.

Good point John. It will be interesting to see it work with a DCC decoder.

If the sensor board has the single component as GRJ describes, this would be a so-called reflective optical sensor that needs to be surrounded by additional components to make it compatible with a DCC decoder that accepts a revolution detector.  The component is simply an IR LED and an IR phototransistor.  You must drive/bias the LED/detector at suitable currents.  Doing so may only require a few resistors depending on what your decoder is looking for.  The component is not in and of itself a 5V part or an 18V part or whatever.

Once you get the hardware side working for the "IR rev counter," the number of "pulses" you get per revolution of the drive wheel may not be compatible with your DCC electronics.  That is, the MTH motor flywheel has 48 stripes per revolution.  Then you have the gear ratio between the motor and the drive wheel which may be, say, 20.  So you are getting hundreds of "pulses" per revolution of the drive wheel.  Are you sure your DCC decoder be programmed via a CV or whatever to handle this?

 I only had one large scale ESU decoder and I don't believe it had any ability for a tach input.

here's the one your using?

http://www.esu.eu/en/products/lokpilot/lokpilot-v40/

I still don't see any thing else listed for that. I believe it senses back emf or whatever it's called.

Stan knows about these things and I should bow out.

 I believe you could mail just the board to a proper repair/ service station. Trouble is that what caused the first board to fail may still be present. There would be no way to warranty follow up I believe? Maybe worth it for you in this case though.

Have you found any reason for the failure?

Thank you all very much for your replies and they all help in solving the problem. In case there are other members of the forum with similar problems here is what I have found out so far -

• In their steam engines MTH have divided the electronics into two parts with the main part in the tender and the 2nd part in the engine connected via a 7 wire "wireless draw-bar". I have no idea how the two parts communicate but my guess is that the prime reason for this the limitation of the "wireless draw-bar" 
• I own 5 MTH engines (2 Steam and 3 Electric - beautiful models that run well mechanically). Once I connected the 2 Steamers and the 1st electric engine to DCC, problems with running started to occur (the decoder seems to be overloaded and shuts down). The other two electrics were purchased later and are quite ok). I have been dealing MTH service department since March 2013 and their first conclusion was that new decoder firmware would fix it. I purchased a MTH TIU to enable me to load firmware into any MTH decoder. This worked very well with the 1st electric (Swiss Cargo engine).
• MTH also provided me with new firmware for the Bavarian S3/6 Steamer. I had trouble loading the new firmware from the start and had to jack up the supply voltage to the TIU from 15V to 18V. The problem now is that the engine runs for a short time on a 4 digit address only to then stop because the decoder is gone through a hardware reset - address has been set back to the factory value of 03.
• At this stage MTH asked for the engine to be returned to them and when I enquired at the post office about the cost I nearly freaked out.  I asked MTH to sell me a new decoder board, but there has been silence since (this is about a month ago)! I have heard from others that they have a policy of not selling electronics to "untrained" owners.
• The "Arge Spur 0" club in Germany published an article in their magazine (http://www.hermann-rail.ch/dow...ten-von-mth-damp.pdf) which tackles the problem of "funny running" by removing the wires to the motor and connecting a small (1.1 Amp) Zimo decoder between the pickups and the motor. All the rest of the funtions are still handled by the MTH decoder. I did this mod and the engine runs beautifully but the tender decoder still goes through hardware reset although the engine keeps running.
• This is what prompted me to go to the ESU XL decoder. For this I have so far discovered the following problems -
• • The 7 wire draw bars are too few and too thin for some connection. MTH sell 11 wire "wireless draw-bars" which I am installing (this is a bit of a mechanical nightmare!). I am now past the point of no return since I had to partially destroy the connectors to get the old draw-bar out
  • MTH use LEDs for lights and have connected these to a negative common while ESU have a +ve common. I need to turn the polarity around using Opto-couplers.
  • For the smoke generator heater I have to install a miniature relay because of the current involved.

All of the above has led me to the two decoder solution with two alternatives -

1. I drive the motor from the small decoder in the engine. In which case I have to synchronise the sound using the IR Rev counter in the engine. ESU can use external pulses for this but their manual mentions Hall Devices (Magnets on driving wheels only). I don't know whether the pulses from the MTH counter will work.
2. I could drive the motor from the tender and ESU provide synchronised "chuffing" internally in the decoder. In this case I can also have doubling of the necessary wires for current reasons. However synchronisation from the MTH counter is not possible in this case.

I probably will have to try both alternatives temporarily wired and then decide on one. There are quite a few surface mounted components on the little IR card and I think I have figured out that the grey lead is Gnd (-ve) so I have to try VCC of +5V and later +18V to see whether I can make it work.

Lastly I want to say that MTH Service have been very helpful while we very doing fault-finding in the firmware area. The silence was "deafening" when it came to selling me a new set of boards. I guess they did not want to tell me that. Their models of European prototype are wonderful and I love them. I don't understand why they don't establish a service centre in Australia but this is clearly their marketing decision.

Cheers Chris

Since it sounds like, as you say, you're past the point of no return, I doubt that further MTH help will be useful.

Wow, It sounds like you have been through **** with this.

Dave

 I can only guess in their defense.

  They have not released the PS3 boards yet to the public. You are ahead of their release.

 You have a special case and selling you the boards probably couldn't be answered at the tech level?? Not having a service in your country made the problem worse.

 I'm not qualified to know where the boards would come from? Would they have spares or take them from another engine??? I don't know.

It is extremely unlikely the IR card requires +18V since the PS3 decoders operate in O-gauge "conventional" mode where track voltages can be less than 10V.  I glanced at the ESU online manual and once you get the voltages sorted out you may be in luck with the pulse decoding:

"Using CV 58, you can define how many sensor pulses the decoder has to receive before it generates a steam chuff."

So if the definition of a "pulse" is the sensor signal going low (as I infer from the examples they give of a reed-switch tied to ground and a Hall device with an open-collector-to-ground output), then you will get 24 pulses per flywheel revolution.  Multiply that by the gear ratio and divide by 4 assuming you get 4 chuffs per revolution and stuff that into CV58.

You may need to do some additional fiddling with the smoke.  The manual implies that some smoke generators are simply connected to the track voltage.  The MTH smoke generators will be "destroyed" if directly connected to DCC track voltage.  There are 4 wires to the smoke generator.  2 drive the heater resistors, 2 drive the fan motor.  The voltages going to these two elements is closer to 5V or so and varies with what the engine is doing (more voltage/power when engine is working harder).  So you may need to add more than just a relay.  There was some mention of using a dim-bright control on a lighting output to control a smoke generator so there may be some basis for altering the smoke level with what the engine is doing.

I also saw the comment: 

"Exhaust chuff trigger: the output generates a control pulse for clocked smoke generators..."

This suggests you might be able to "pulse" the smoke fan motor to create puffs synchronized to the sound.  If you can work this out, that would be quite an accomplishment.

Thank you very much for the replies especially Stan2004. I will give +5V a "good going over" and watch for pulses with a CRO. I have counted 24 black/white stripes around the drum on the motor axle and the gearbox translates 4 motor revolutions to a quarter turn of the driving wheels. Since this is a 4 cylinder engine I think I need to set CV58 to 96 - 100 to get 4 "chuffs" per driving wheel revolution.

With the smoke generator things get a little "hairy". The MTH unit has two 18 ohm elements connected in parallel. On 5V this would give 0.5 Amp through the combined elements. By experiment this is insufficient to create enough smoke. On 18 V this gives 2 Amps which, if applied for longer, would probably burn out the elements. Also when the engine was running under address 03 I only measured a 0.4 - 0.5 Amp increase when I applied "smoke" which implies that they chop the smoke unit supply in some way to get sufficient smoke. I resolved to connect the two elements in series (a couple of cuts in the track of the board). On 18 V this gives a current of 0.5 Amps through each element which appears to be sufficient. I still need the relay because the maximum current on the AUX outputs is 0.5 amps and I do want to be able to switch on smoke by command. 

I have allowed for two AUX outputs to drive the smoke unit - one to drive the elements and the other to drive the fan. I need to get some information about how to use the clocked smoke facility - I think this would be a good feature.

I am certainly not "sore" on MTH, technical people often have to follow company orders even if they know better. MTH is only one of three suppliers for O gauge central European rolling stock at very reasonable prices. I hope they keep this line going! 

I have someone in Australia who might be able to help you with MTH, he has had to deal with MTH as a dealer with warranty work.

Strapco Australia

 Tom Crozier

PH 03 97964454

MTH will never catch on in Australia till it thinks more latraly about small markets and how to grow them, as my friend Tom has tolled me If he didn't have another business he'd starve selling MTH. The other dealers I know will only order MTH through Walther's. I love my MTH and like running DCS but its problematic in a small market like Australia, but give Tom a try he's a really nice guy, and the only person Id order directly from MTH, If I could even find any one else.       Stephen    (cTr...Choose the Right) 

Actually, MTH has two 16 ohm elements and uses a pulsed current to drive them.  They are effectively driven by 6 volts, and at that level they can produce plenty of smoke, as is evidenced by the many MTH locomotives I have here!   Trust me, 5 watts through the two elements gives you plenty of smoke!  Lionel runs around 9 volts through a 27 ohm resistor in their TMCC stuff, it still gets good smoke, however not up to MTH standards.

It also appears there are CV's to program the brightness of the outputs which I assume is done by varying the duty-cycle:

Brightness: Defines the brightness of each individual function output in 32 steps (0 to 31).

While I appreciate the reasoning of hacking on the smoke unit elements to convert from parallel to series, you might consider putting this duty-cycle control to work for you.   As GRJ says, 5 Watts puts out a healthy level of smoke.  So if 16V (after rectification) is applied to the 2 parallel 16 ohm heaters (8 ohms effective), that's a heating power of 32 Watts which would destroy or melt something.  Suppose you drive the heating resistors with a FET (about $1).  If the FET is driven by an AUX output with a duty-cycle set to "5" (range is 0..31 steps), that would apply about 5 Watts to the heating elements.  Each step would alter the power by 1 Watt which ought to give you suitable resolution for experimenting.  This would be in place of a relay.

Another thing I saw in the manual is a firebox function on an AUX output.  Apparently it automatically increases brightness (duty-cycle) to a firebox light proportional to what the engine is doing.  Perhaps this could be applied to the smoke heater to increase the smoke when the engine is working harder.

Separately, you previously made a comment about using opto-couplers to "invert" the polarity of the MTH-installed LEDs.  Are you sure you need opto-couplers?  I'd think you could just use a 5V logic gate inverter IC chip which should be much cheaper...maybe 25-50cents USD for an IC with 6 or 8 inverters with current sinking (active low) capability to drive LEDs.

Thank you guys for your interest. I will try and contact Tom Crozier here in Hallam in Victoria.

For the smoke generator MTH must be chopping up the supply ie. varying the duty cycle and be able to switch 2 or 3 Amps of current, while doing it. The max current allowed on the AUX outputs is 0.5 Amps, higher than this and the protection circuit operates. MTH must have designed their heater output circuit to handle this situation. I don't think I could use a relay to chop up the output of "dimmed" AUX port. Putting a transistor switch there is also problematical.

I agree with all your power calculations. The two elements in series on 18V were actually developing 9 Watts and this produced plenty of smoke. I can probably use an AUX port to drop this to 5 W if this is sufficient. 5V on the original parallel arrangement produced 2.7 Watt (ExE/R), which I found definitely to be insufficient.

You are probably right about the opto-couplers. I just happen to have a bunch of LTV-844 Quad Opto-couplers with nothing to do so I thought of using them.

Don't get me started on MTH's inability to understand support overseas. I am in Australia and have never been given any help by MTH. I gave up and got an agency to install Electric RR product. So, whenever an MTH loco goes, I just install TMCC. Sad situation, they are so blinkered, but produce trains for the overseas market and are unwilling to help with tech support. Crazy stuff.

Hi Dave,

I have read your thread of last year and I was going to get in touch with you to see whether you had any more success later. Obviously you don't and I am sorry about this. I talked with some US friends about 18 months ago mainly about marrying up conventional DC with DCC and casually mentioned MTH. The response from all was the same: MTH is high quality, looks good, runs well but the electronics is RS. I think it has a lot to do with that Mr. MTH took DCC, "bent" it a little to make DCS out of it. This was probably done to lock people into MTH. They have now seen the error of their ways and, since PS3, made their decoder switchable between DCC and DCS. I think this is where a lot of the problems lie. 

My main reason for searching these forums like OGR, ESU, Digitrax etc. is to find someone who may have actually done a full conversion of an MTH engine to someone elses DCC. Sadly I have found no-one. However I am getting a lot of helpful hints particularly on this Forum and I am sure I will get there!

If I was running American O Gauge models there is plenty of competition to MTH eg. Williams, Lionel, Atlas, Roco and even the old faithful Rivarossi/AHM. However I am running German, Swiss, Italian trains and there is only MTH with the biggest range, Lenz and Brawa as well as Kits from Schnellenkamp etc.

I used to get all my MTH supplies from Trainworld in NY but about 12 months ago they suddenly advised me that they would no longer stock the European 2 Rail scale wheel stuff. Checking with other suppliers in the US revealed the same: Will deliver to the USA only. I think this is pressure from the European dealers who charge much higher prices than MTH. Probably service is also good over there.

Anyway "by hook or by crook" we will get these MTH engines working on DCC.

Cheers Chris

Yes, the MTH smoke heater electronics definitely chops the voltage.  In part this is because it operates over a relatively wide range of track voltage (i.e., conventional mode) and maintains a relatively constant heater power by varying the duty-cycle.  I understand your point that the peak currents when chopping exceed your decoder's limits.  So speaking of optocouplers, that would be a place to use one.  That is, if you add a bridge rectifier up in the engine to supply DC heater voltage, you could use an optocoupler to isolate the decoder's current-limited output to drive the heater with no high pulse currents flowing through the tether.

As for using the LTV-844, I ask, "Do you feel lucky?"  Opto-couplers are notorious for widely varying CTRs.  The datasheet says 20% to 300%.  While these are worst-case, that's still a 15:1 ratio.  Obviously you also "waste" the input LED current.  So if the MTH LEDs draw, say, 10 mA, then with a nominal 100% CTR you must "waste" 10 mA on the opto's input side to get that 10 mA on the output side.  OTOH, a logic gate has essentially zero current draw on the input side.  I realize the optocouplers are "free" to you, but I'm fascinated by your dogged determination to get create what amounts to a "DCC-ready" O-gauge engine starting from a production MTH PS3 chassis.  I figure whatever you come up with might be duplicated by others (who don't have LTV-844's lying around!).

It also seems there's got to be some way to sell or trade your PS3 electronics with someone in the US.  While not quite apples-to-apples, I'm thinking of the PS2 boards where guys here pay over $100 to replace boards.  That ought to recover some of the cost off modifying the PS3 engine to be DCC-ready.  Comparing MTH's PS3 DCC to the ESU DCC is beyond my pay grade but it seems like the ESU decoder has some neat features not offered by the MTH DCC implementation.

I had a similar experience with MTH a few years ago even here in the states. They would not sell me replacements for a damaged PS-2 board set and then stopped bothering to even respond to emails. The dealer I normally buy MTH stuff from was even baffled and got nowhere.  I finally gave up and rebuilt the engine with ERR/Lionel boards - it's worked perfectly ever since.  MTH no longer has to worry about selling me anything with electronics. You have to support what you sell - which in my experience, Lionel & ERR are doing very well.

Hi Guys,

I have been doing some more experimenting today by modifying my other MTH steam engine the French 2-3-1 Chapelon as per Christian Schmid's article using a ESU LokPilot decoder. I had no trouble getting the engine going but although the stationary sounds were coming out of the speaker, the running engine would not "Chuff". However I was able to do some measurements even with a CRO and can now report -

Purple = Ground (-ve) = 0V

Red     = Output (+ve going) = 4.5V approx

Grey    = Vcc   =  +5V

So it looks like I will have to go for the full modification to an ESU Loksound XL decoder, but may be the IR rev counter will work. I have also found today an overwhelming reason that I need to use Opto-couplers for the voltage inversion for the LEDs. I found that the LokPilot chip does not bring out Gnd and I have to use + 18V, +5V and Gnd supplied from the LokSound decoder mounted in the tender. In this way full galvanic separation can be provided between the two decoders. Of course currents will have to be limited by series resistors.

Just a little bit more reasoning -

Why MTH? - Provides a unique range of European rolling stock.

Why ESU? - They provide Sound files for the particular european engines

Why 2 decoders? - To fit the system into 11 wire connection between engine and tender.

I will let you know how I am getting on.

Thank you for your help!

Chris