Problems with TTS J36 sound decoders

[richmarsh]

 Hi

Contacted Hornby Technical Services (a chap called Glen). He states reason for factory CV29 value set at 2 was not for DC/DCC runaway loco's. However, he requested and given (e-mail) factory codes on box. Awaiting response. Seemingly I may have to return these.

[96RAF]

All that value 2 gives you is 128 speed steps which is recommended to ensure TTS timing of things like start delay and chuff transition are correct. CVs 3&4 are set at value 15 for similar reasons. All seen in the manual.

[ColinB]

I am not surprised they want them back, they will probably analyse what is wrong and you have a splendid example. As I said on the previous thread doing what you did shouldn't have made any difference but for some reason it did, which is probably due to something weird happening in the software. At least Hornby are doing all the right things, which will long term, probably mean that they will sort it out. We all make mistakes it is just how quick we at fixing them, that is important.

[SteveM6]

As Rob said CV29 with a value of 2 just gives you the 28/128 speed step function and adding 4 (cv29 = 6) just enables DC running and increases the risk of a DC runaway.

If the loco goes the 'wrong' way add another 1 to the value of cv29, i.e. 3 or 7 in the examples above, and it should then run the other way.

Not a big issue and no need to return it to Hornby.

[96RAF]

For fullness of data for any fault investigation I would be setting CV29 to each possible value and noting effects if any.

There are other CVs that it would be useful to have the read-back values for, such as CV7, CV156, CV157, CV159. Not all of these will have values of interest, as it depends upon the deocder build state as to which ones are populated.

I obviously cannot do this as I don't have hands on with the problematic decoder(s).

 

[ColinB]

RAF96 I would suspect by doing what he did the fault has been fixed. It would be interesting though if you play with those values, what happens. It probably was programmed wrong at the production plant. The thing that is more worrying is that there might be a batch out there with the same fault, but at least there is a post now on how to fix them.

[96RAF]

Treating the symptoms, not curing the disease, which is why the doctor needs more information for the diagnosis.

[ColinB]

Yes, you are right. There is probably something wrong in their code, but you nor I is going to be able to fix it. Usually, with these types of faults it is very difficult to reproduce once it has been fixed. When I have encountered these sorts of faults in my previous profession, it is normally a case of looking at the code and trying to figure out how it could possible get it wrong, so a fix can be put in for the next batch. The purpose of this forum is how to fix it so it works, which has been covered.

[richmarsh]

Hi all

Glad to get folk communicating. I suppose (to keep today's theme Hummm) I got a vaccine for the problem NOT necessarily the cure, because decoder SHOULD function normally on CV29 value 2. That is why awaiting Hornby technical team to investigate and I would presume they will have 'other's' of the same batch. Also I am not going into other CV's of this as this will be investigated by professionals. PATIENCE.

[ColinB]

@richmarsh that is exactly my point. I can see your point having got it to work properly I don't blame you for not wanting to touch anything. I can think of of a few reasons for it doing this, but I will keep them to myself.

[richmarsh]

 HI

Just an update.....been in contact with Hornby Research and development and are invetigating this problem, which appears to be limited to the Uhlenbrock controller. This is due to the fact that, 1) I was informed that ALL decoders are tested on a variety of controllers ( not Uhlenbrock Daisy ll),  2) Hornby have NOT recived any other communicay's of similiar problems with any decoders. PATIENCE.

[96RAF]

It would be unrealistic for Hornby to keep a full inventory of every alien controller under the sun (at each likely revision state where updateable) in order to prove their decoders are 100% compatible. They are designed and usually warranted to NMRA standards and that should be sufficient, but as we all know there are standards and standards.

 

The same goes for checking that Hornby controllers work with all makes of decoder. Take a look at the drop down list in RM to see how many decoders have been validated by Hornby and HRMS. As soon as a decoder or controller problem is identified Hornby R&D looks into it until it resolved as far as possible, even to the extent of liaising with other manufacturers directly.

 

I doubt many if any of the other controller or decoder manufacturers bother going anywhere near that extent of testing and/or investigative support and I am sure the likes of LiaisDCC don’t even test their decoders outside the design clone box.

[ColinB]

Looking at the Uhlenbrock controller (assuming it is the same one I found on the internet) it looks to be derived from the Fleishman twin track controller I have. I may be wrong but the shape etc. look exactly the same except with extra press buttons. So I don't know if they carried over the software or wrote their own, but my Fleishmann works perfectly ok with the TTS decoders. I have had a few TTS decoders that have not worked properly but they have usually failed on both my Fleishmann and my Elite. I have noticed of late that the Elite has had issues writing CVs to certain makes of decoder, so I think with some makes there seems to be a timing issue. I seem to have no issues with Hornby and Zimo decoders. The trouble is with a lot of these decoders you never know how old they are, so a newer batch may be better. I seem to have a lot of issues when the decoder address is set to a different address other that the default value (even a Loksound decoder did it the other day).

[richmarsh]

 The Uhlenbrock DaisyII, I think is similiar to a Marklin controller. The fleishman , U mention maybe similiar to Uhlenbrock 5000 which (to me) looks similiar to a TV remote control. I have had some makes of decoder not operating with my controller. I have not had any problem (so far) with changing address.

[Yelrow]

As a layman, it appears, that had you had a Hornby controller, Elite / Railmaster, most of this would have been avoided.   Could it be that Hornby  cheap TTS, perform best on there own gear. The better quality Licksound ones, fitted in some of their more expensive locos, seem trouble free.. Just putting it out there.

[richmarsh]

 In response to YELROW. It would appear from Ur post U have not read the beginning of this post (no offense inteaded). ALL dcc controllers and decoders are MEANT to work with one another, although this not always true.  TTS decoders are worth the money, but they gotta work. I already have other TTS decoders and Hornby decoders on my locomotives which work as they should and in my opinion worth the money. I can't give an opinion on their controllers. Also, in this post others have had 'problems' with more expensive decoders, so money don't necessarily means it will or it won't. HOWEVER, love Ur photo, it looks very similiar to my little girl, rescue from Crete !

[Yelrow]

Hi. I make it a habit of reading posts. There  have been similar instances, over some time, with TTS  chips and controllers other than hornby. That was the point i was making. I dont think for a moment that the designers, have all makes of controllers on their bench to test, but they Will have the Hornby ones. Following those tests, they will launch.  Of course chip complies with regulations, etc, but there are, as i remember, batches in the past which had the odd problems. Your particulat controller, is  more  technical than the run of the mill ones, certainly, more expensive.

[96RAF]

Of course this ‘everything works with everything else if built to spec’ is fine if the tolerances are maintained and if each makers’ interpretation of the spec is the same.

 

I wonder how many alien decoder makes Ullenbrock has tested their Daisy II controller with to prove compatibility. Can their controllers firmware even be user updated to incorporate change to accommodate decoders not even made when the controller was launched.

 

Hornby kit even though already NMRA certified has definitely been updated over the years to sort out compatibility with other makes of decoders. Even so you still find folk chanting the urban myths on social media that the Select is rubbish, R8215 decoders are useless, etc. Yes they were then, but they have been improved or replaced since. The problem with the internet is it has no sell by date, after which out of date information should be purged.

 

The same rationale goes for decoders versus controllers. How many decoder manufacturers can say hand on heart they have tested their decoders with every known controller.

[Yelrow]

Yup, twas my point.  I have never seen this controller mentionned on here before. Taint the run of the mill. You really cant, in my view, expect any manufacturer, to buy all other makes of controllers, just to see if their chips work. There is bound to be the odd snag..

[ColinB]

I run a Hornby Elite and the Fleishmann twin track, and I must admit both work ok with all decoders I have used. The issues I do have, especially with the Elite is writing CV values, especially if the decoder is not set at the default address of 3. Hornby TTS and normal decoders work ok as do Zimo ones, but I have had issues with Gaugemaster ones, DC Concept ones and recently a LokSound sound decoder. The Fleishmann does seem to program these although it is hard to compare as the Fleishmann doesn't make writing to CV values as easy as with the Elite. I know from my previous profession that writing values to non voltile memory can cause issues (when you write a CV value), where it can get corrupted (normally caused by removing the power while it is still writing), so I wonder if this may explain a lot of the issues. On my programming setup I have an Elite connected to a metre length of track via twisted pair wires, so unless the unit has got some stray capacitance inside, the unit should be reliable. It could be that something goes wrong when writing the default CV values in the plant, I wonder if blaming it on the controller is a bit of a red herring, I have seen it happen so many times in the car industry when someone cannot explain something. 

[richmarsh]

 WOW. Did'nt expect this kind of diversion to discussions. It is well explained in other areas that i have read, re. problems of compatabilities, when manufacturers imply there should not be.

My experience of this hobby is only 3 years and I dont have a technical background other than my interests. Uhlenbrock Daisy II is not perfect (as is the NORM it appears), but it fits my particular needs for my layout (and budget). Some decoders I have (before purchase of the Uhlenbrock) don't work with the controller! However, a particular type of decoder from one manufacturer works, but the 'other' type does not !!!!

Uhlenbrock do updates, but U need a cable to connect to a computer, and I am not saying in any way one particular product is better than the other. I know, from experience what U/Colin are saying about the car industry, re looking for excuses when they can't locate a fault (expensive), that's another story LOL. 😆

[ColinB]

I don't know if you have said cable, if not what one is it? To update the Fleishmann I had to buy a USB to RS232 cable. Once I had got the computer to recognise the lead, the process was pretty easy. 

[96RAF]

The cable may have to have FTDI chipset (I think the acronym is correct way round), same as required for ESU Lokprogrammer.

[richmarsh]

 Update for DAISY II is by a 63120 LocoNet usb interface. I am not going to bother, concentrating on expanding my loco stock and TTS loco's.

Just for info/fun my layout based on Kyle of Lochalsh, Plockton, and Strathcarron stations/line around 1950's. Researching info, I find difficult, cause I like to loosely base it around reality. Model locomotives difficult to locate, although mainly 'black 5's were used,  HR loco's used (no models, unless kits) but I like variety.

[ColinB]

It doesn't have to have FTDI although generally that is the most common RS232 to USB (or USB to RS232) converter available, it just defines the "chipset" in the cable, a friend of mine wrote his own using a PIC. I have found it is a very good idea to source the cable from a reputable UK source as I found that the cheap Chinese ones don't have PC drivers that work. The other thing that is important is whether the RS232 end is true RS232 with the correct voltages or just a 0 to 5 volt signal, I have been caught out with that in the past. Either way, if the user hasn't got one I am sure the manufacturer will tell which one is the best. I can't remember how I decided on the one I bought, I do remember it was about £15.00 which quite expensive for this type of cable, but it came with a guarantee.