MTH Speed control problem

Are these MTH diesels?

How 's the spacing between the reader and tack tape?   As you know  spacing    should be the thickness of a dime

1/8  inch or more  might explain your problem, however 2 engines having the same problem is confusing.

Anyway check the spacing.

So as I understand it, you get the correct speed below 6-7 sMPH.

It is possible for the speed control to work at slow speeds and become less reliable at higher speeds.  This can be from mis-aligned sensor (e.g., not centered to the flywheel) or distance from sensor to flywheel.  Or from poor black-white stripe distinction when the tape was printed.  When the flywheel spins faster at higher speeds, the reflected optical pulses become smaller/shorter and more difficult to sense.  Then, since the speed controller sees fewer pulses, it thinks the engine is moving slower than it really is and drives the motor even faster making it seem like there's no speed control. 

Yes, I had this exact problem with too wide a gap on the sensor to the flywheel.

That could be a weak sensor or bad spacing.  I agree that two doing it suddenly is somewhat puzzling.

These 2 diesels were in storage a couple of years. I could swear they worked fine when they were put away, so it just didn't happen suddenly. I'm sure the spacing hadn't changed. I wouldn't think the sensors would go bad over time. I'll have to replace one and see what happens. I replaced the speed tape on one with no change, like wise reloaded the file from MTH site.

I can't imagine much else here.  It's a long shot that it's the boards as they're seeing something from the sensors, but that's possible I guess.  I'd put one on the test set and make sure the board is good.

Try  a different sound file? 

Chuck, are these MTH engines?

The diesel rev'ing sound notches up with speed as sensed by the tach circuit.  You may not be able to hear it well over wheel noise with the engine zipping along but if the Protosound rev'ing drops to a lower notch at higher speed this is a good indicator that the sensor is missing pulses for reasons I suggest earlier.  That is, the processor will see fewer pulses and think the engine is going slower and notch down the rev'ing sound. 

If you confirmed 5 sMPH is correct by measuring distance traveled with a stop watch (e.g., 5 sMPH = 18.3 ft in 10 sec), then I think the sound set is probably correct wrt the parameters which set the gear ratio and such.

Marty, yes PS-2. Stan, yes the diesel sound does rev up with speed increases. Gregg, yes I have tried different files, no change. I don't believe I have corrupted files. One example is a Protosound 1 SD-40 upgraded to PS-2 about 5 years ago. It had worked fine. I just tried reloading the files. (3volt) everything works fine but the speed control. Replaced the tach tape with a fresh one. Using the kit supplied motor/tach sensor bracket so spacing is unlikely. I will replace the tach speed sensor tomorrow and see what happens. It has got to be that or the power supply board is my guess. I just have not heard of the sensors going bad 'for no reason' before.

Since you have a way forward, this is just another thought for someone else who stumbles across this thread later.  If you run the engine 10 laps (or whatever) and read out the DCS odometer, if the tach sensor is working you should get exactly the same distance traveled no matter the speed.  This would tell me the tach sensor is working and something else is messing up speed control at higher commanded speeds.

Everyone has hit the likely items.  You can clean sensor also incase it is dirty.  7 pin plug fully seated?  Don't just trust tach holder bracket,  Check gap.  G

I had purchased two consecutive G scale releases and both ran too fast. The funny thing to me was that the first was the last batch of PS2 released and the next was a PS3. They both matched perfectly together at the same speed curve, but much faster than the rest. I was skeptical. I had to grind the truck top mount where the tach reader sits to allow the gap to be closed.

 I could understand the PS3 being off as the tach reader seems much smaller. I had to move that reader in much more. The PS2 reader was swapped with one having a larger  spacer to bring that one in closer. That cure fixed both. It's a fixed spacing so who would have known? Mine ran almost twice as fast as they should. I forget the percentage that I calculated. When set to something like 30 MPH, they would run around 50+MPH?

https://ogrforum.com/t...scale-speed-matching

FWIW, the (1) tach reader in PS3 diesel, both factory & upgrade kits, differs markedly from (2) both the PS2 steam and diesel factory and upgrade sensors, and the PS3 upgrade kit steam sensor.  They are not electronically interchangeable.

That's good to0 know RJR....Thanks

Back to Chuck's problem.... I think I might  try a new tape. Maybe one with less strips just to see what happens. It may slow it down a bit.

If the speed is proper at slow speed, I don't see less stripes curing the issue, it'll just make the scale MPH calibration way off and move the problem to a slightly higher speed.  There is a problem with either the sensor, wiring, or the boards.  I'd check the wiring first, then replace the sensor.  Of course, I'd probably connect the board to my test set and see if the board works with a totally different sensor and tape that is known to work properly as the first step.

Just a note on mine, I cleaned the flywheels, checked the wiring (and replaced anyways), and swapped out the sensors. For some reason these two new engines needed to have the tach closer than usual.

 Could that be that the boards were weak? (something in the circuit wasn't up to spec?) I have had upgrades where the tach just had to be within spec. These two had to be tight. Both of these engines have run great ever since (four now out of a hundred total, needed adjustment).

 I still have one that runs slower than normal.

Engineer-Joe posted:

...

 Could that be that the boards were weak? (something in the circuit wasn't up to spec?) I have had upgrades where the tach just had to be within spec. These two had to be tight. Both of these engines have run great ever since (four now out of a hundred total, needed adjustment).

 I still have one that runs slower than normal.

Right.  Frequently discussed are a) incorrect spacing from sensor chip to flywheel, b) dirty or mis-aligned sensor, c) dirty black-white striping, d) incorrect soundset which calibrates the stripes-to-sMPH conversion.  But to your point the circuit components themselves introduce variation.  The tach sensor chip consists of an LED emitter that generates the Infrared beam aimed at the flywheel and a phototransistor detector that senses the variations in reflected energy from the black/white stripes.

The datasheet for the Omron EE-SY124 doesn't have that much info except for this table:

For example, for a given "Light current" which would be when the sensor is reflecting off a white-stripe on the flywheel, you can have a 6-to-1 variation in how much current the phototransistor generates to the circuit.  The LED too will have variations though the way they specify it is a harder to explain in one sentence. In any event, these are worst-case specs which might never be encountered.  While it's the circuit designers job to account for component variations, it's an engineering tradeoff between complexity, cost, etc.

As for your one slow engine.  Was this a "factory" flywheel-soundset combination or one of those upgrades where you printed your own black-white sticker with 46, 48, 50, etc. stripes?  And do you have some actual numbers or percentages on exactly how slow it is at different speeds?

One other useful bit of info from the datasheet is the sensing distance.  Greater than 1mm the response falls off pretty sharply.  The "sweet spot" seems to be from about .75mm to 1mm for the spacing to the tach strip.

I finally found the problem. Apparently spacing was the problem. The heat from the motor has apparently warped the kit supplied motor/tach bracket just enough to take it out of its 'sweet spot' as John says. It warped just enough not to be noticeable from eyesight, but to pull it away from the sensor tape to give the sensor a false reading.

Another notation to put in my service notes.

Thank you EVERYONE for helping me with this problem. Good to know that there are so many helpful people on this forum!

Chuck 

Very good to know.  Also, <0.75 falls off drastically.  Unfortunately, sensor holders lack a fine adjustment screw.

"Don't just trust tach holder bracket,  Check gap."

So this was it.  

While there is a finite peak, if the threshold is at 40ua to trigger a count, the gap is less restrictive isn't it?

I find that over time, some sensors are more susceptible to gap, than others.  G

Sorry Stan, just got back from the dealer for my truck.

 Anyways the slow engine is all factory as best as I can tell. I did buy it used. Although it appeared to have very little use if at all. It is the MTH gauge 1 F series red and silver Santa Fe "A" engine. The number escapes me right now. I'd guess somewhere around '07 man date? I believe it was more than 20 % slow? I could not use it with my others as it would get shoved off the tracks. However it still runs fine years later on it's own. I had a bunch of issues all crop up at once and this engine got overlooked for repair. It does run after all!

Stan, I think I remember that the slow engine was off more, as the speeds went up. I believe it was the exact opposite as the two mentioned above that ran fast? (guess ex.: So maybe it would run at 20 MPH when set to 24 and 30 when set to like 36?? (Maybe higher speed differences?)). Last summer I ran it alone and haven't run it this year at all.

I should write this stuff down. My old computer crashed and I lost tons of stuff.

GGG posted:

While there is a finite peak, if the threshold is at 40ua to trigger a count, the gap is less restrictive isn't it?

As I see it, the relevant information from the plot is the "focusing" behavior of the component.  That is, the component has two windows, one with emitter, other with receiver.  As you get further away from the flywheel the strength (uA current) of the reflected optical beam decreases in a non-linear (exponential) fashion as one would expect since the light is spreading in multiple dimensions.  But as you go closer to the flywheel you have a different effect - the separation or geometry of the windows is such that the reflection off the flywheel can't even make it to the other window!  That's why the current decreases as you get closer and goes to zero if the sensor is touching the flywheel.

Note that the plot is for a specific emitter current and a "shiny" metal surface to reflect off of.  If the LED current is different, then the vertical axis (uA reading) will change though the shape of the curve will stay the same.  Likewise, if the target is a painted white stripe then the vertical axis will also change though the shape of the curve will stay the same.

OK, so to your question which I'll presumptively modify to change "40uA" to "something lower" since we don't know what the vertical scale is for the tach.  I'm not sure there's enough information.  The design problem in the flywheel application is looking for contrast between black and white.  Even with the black stripe there will be some reflection.  So setting the threshold lower may indeed allow you to detect the white stripe over a wider range of gap...but that lower threshold would also increase the "false-positives" where a black-stripe reflects enough to masquerade as a white-stripe.

Engineer-Joe posted:

Stan, I think I remember that the slow engine was off more, as the speeds went up. I believe it was the exact opposite as the two mentioned above that ran fast? (guess ex.: So maybe it would run at 20 MPH when set to 24 and 30 when set to like 36?? (Maybe higher speed differences?)). Last summer I ran it alone and haven't run it this year at all.

I should write this stuff down. My old computer crashed and I lost tons of stuff.

Well, if you want to re-open the cold case file we can perform some experiments.  For example, if the speed is a fixed percentage lower at ALL speeds that would be a good clue.   Of course first off I'd want to confirm the correctly scaled soundset was available and loaded.  I'm not so familiar with One Gauge products but IIRC there are some Standard gauge engines that aren't all calibrated to same scale speed.

Thanks Stan. I try and get back with info. I believe it has the correct G scale soundset installed? It was so long ago that I forgot.

Update: I just put it on the tracks with a brand new set of PRR F3s #70-2007-1.

http://mthtrains.com/70-2007-1

The SFe now runs about 1 to 2 MPH faster than those do? Maybe it finally broke in? It used to be a slower engine???

stan2004 posted:

So setting the threshold lower may indeed allow you to detect the white stripe over a wider range of gap...but that lower threshold would also increase the "false-positives" where a black-stripe reflects enough to masquerade as a white-stripe.

This is exactly what I ran into when doing my Chuff Generator project.  It took several hours with the 'scope and a handful of resistor values to come up with a good drive level for the 1mm spacing for the sensor.  At first, I was driving the LED too strong, and move reflective tach tapes like the MTH ones that came with earlier upgrade kits, would not trigger properly.  I finally settled on a drive current that seems to balance the requirements at my chosen 1mm sensing distance.

Engineer-Joe posted:

Thanks Stan. I try and get back with info. I believe it has the correct G scale soundset installed? It was so long ago that I forgot.

Update: I just put it on the tracks with a brand new set of PRR F3s #70-2007-1.

http://mthtrains.com/70-2007-1

The SFe now runs about 1 to 2 MPH faster than those do? Maybe it finally broke in? It used to be a slower engine???

LOL.  The engine must have ears and heard us talking about it and quit mis-behaving!   Well, if we're now talking about 1-2 sMPH differences the question becomes 1-2 sMPH at 10 sMPH command or 100 sMPH commanded.  Obviously if the former you still have a problem!  The good news is if it's "just" a 1-2 sMPH error you probably have the correct calibrated soundset loaded so it becomes more likely a tach/sensor issue if you want to proceed.  If so, I'd want confirmation that the new F3s are indeed operating at the correct scale speed (stopwatch and measuring tape) if that's the speed reference.

As an observation, it's improbable that the soundset changed itself while sitting in the box.  So if the speed characteristics changed from slow to fast while sitting in a box it's more probable something changed in the tach sensor system like fading toner or pigment on the black stripes or whatever.   As a general statement when the sensor has trouble reading the stripes, this usually means it senses FEWER stripes than actual, so it drives the motor harder which then makes the engine run faster than it should.  So since your engine went from too SLOW to too FAST, I'm playing the odds and suggesting it's probably something in the tach sensor (as opposed to an incorrect soundset).

Than there is the practical aspect.  I have seen tach gapped at a variety of different distances and at the edge of thin flywheels and work fine.  My point was despite the graph the gap opens as threshold lowers.  So it is not as super sensitive as being portrayed.  Actual performance trumps theoretical what ifs. G

True George, I suspect that some sensors are just a bit weaker than others and need to be more precisely adjusted to function.

Well I'd like to test further but now it quit? It won't respond to anything. So I need to tear into it and see why. It was running +1or 1.5MPH faster today at 30 MPH. I had to add it to my new phone WIFI thing and add the new set of F's as well. I could see if the battery was low that the new address didn't stick. However it won't run in conventional either. Maybe it didn't like being in the limelight?

 The problem really is that I have not checked the new set of F's against anything to see if they are slow? They probably need a break-in? I was excited to get them out of their box for testing this year and hoped they'd run in consist with the SantaFe. I pulled the SFE off the tracks to run the PRR's alone. I will test them against the rest of my stable and see what's wrong with the problem child this year. Urrrggg.

OK, it seems like an easy one. I should have known as it's the last thing I did when it quit. There's power going to the polarity switch. There's no power going to the board. So the switch isn't passing power. The wiring is good. I don't have this slide switch in stock, I just have mini toggles that I use in upgrades. I'll try and bypass it and do more speed tests tomorrow.

 As a note, there's smoke oil everywhere but in the smoke unit. Maybe that led to the switch failure? This isn't the first one for me that's failed though (third or fourth?). I prefer the mini toggles as they seem more robust.

Engineer-Joe posted:

It was running +1or 1.5MPH faster today at 30 MPH.

If it's then +2 to +3 sMPH fast at 60 sMPH commanded that would be good to know.  So if there's one "bad" dis-colored or damaged stripe of the 48 which blends into the adjacent stripe, you effectively get 46 stripes.  The fewer the stripes, the faster the engine goes. 

If you only have 46 stripes you will go 4% faster.  At 30 sMPH commanded that would be 1.25 sMPH faster which is suspiciously coincidental with your observation that it runs +1 or 1.5MPH faster.

Thanks Stan for the help! I would have tested more. I decided to run the brand new engines for a few laps to break them in. I shouldn't have touched that switch until later!

I just decided to install my toggles as they seem to last much longer. I'm grinding the weight to allow it to fit. I should have known the switch was bad, as the program could not find the engine. I reversed the engine and threw the polarity switch and it found it. I threw the switch again to run on a different loop and that was that!

 I forget the number of stripes used on MTH G scale diesel engines? I remember the flywheel is painted from the factory. I've always feared that the grease would cause problems when it gets real hot outside in the sun. Whenever I tear into an engine, the flywheels have been almost perfect. They ride in an enclosed area with shields on the shaft to seal them from grease that's right next door on the gearing. I have MTH diesels that have run flawlessly for years! Recently, some engines popped up for sale that may have questionable backgrounds. I'll still buy them as overall they are built very well. There are a few weak links, like these polarity switches that seem to fail sometimes. Yet I do have some that have never failed? I don't know!!!

That switch is a problem area, I've removed it for several people.

Stan, I just counted the stripes on a spare motor and I believe there's 24?