Hamburger

If you want a sound decoder without motor drive and use ESU you should have a look at the LokSOUND FX (function) decoder which is a hybrid of sound decoder and function decoder. The cost for LokSound Fx plus a LokPilot is not much higher than for a normal Loksound 5. (ESU 58210 for 8-pin, 58219 for 21mtc where AUX3/4 can be set by CV to either logic level or amplified output). https://www.esu.eu/en/products/loksound/loksound-5-fx/

On old Fleischmann motors often one pole is connected directly to the chassis and therefore to track voltage. This causes severe problems and can even destroy a decoder. This has to be checked and fixed. Maybe you have to cut the respective connection directly at the motor shield. Secondly, old Fleischmann motors often have built up a lot of dirt/abrasion in the gaps between the commutator plates so this has to be removed with a toothpick and the commutator has to be cleaned with alcohol or similar.

Because CVs don't usually change by themselves. What should a reset result in then? The forum is full of comments that a reset didn't help or just made things worse.

You need 2 capacitors: one for the main power supply (16V or 25V depending on the track voltage) and a second for the internal electronics (6V and about 1/4 of the capacity of the main capacitor). The connection pads are different depending on the board version. Instructions can be found on the Internet and in the ESU manual for the LokPilot V3. To be on the safe side, the voltage at these points should always be checked again. LokPilot manual

The one thing I almost never do is a reset. First of all, I thoroughly clean the rails and pickups, especially if the problem occurs with more than one loco.

If you compare all DCC standards, there is an address range that is not clearly defined: the range between 100 and 127. Some components expect a short address there (CV1), some a long one (CVs 17&18). Therefore, it is better to skip this area altogether to avoid problems later.

Yes, exactly. That's when I got to know this system, and it convinced me on all counts, except that new features are only introduced when they've been on the market for 1 or 2 years (maybe that's exactly the reason why everything is working). The system has the most expensive wireless controllers, but they are also the best and always work. No wifi, normal radio, no interference from smartphones. But I think that's not what the questioner wants.

The best is clearly Massoth. But I wouldn't buy it because of the price. US controllers are mostly based on the NMRA, which differs from the NEM in some points and therefore some features of European decoders cannot be used optimally or it even leads to problems. Hornby is not quite up to date in development and also somewhat limited. Maerklin has too many problems with DCC (this is intentional). All other European controllers can be recommended. Again, you get what you pay for.

Could be a problem with the flash memory, but in this case you should lose all settings and the model would not respond at all. However, you can read a lot on the ESU forums of people having similar problems, and in most cases the problem was that they didn't follow the upload procedure correctly to set what the default values should be for that decoder. When writing decoder data, a window pops up asking whether the old default values should be overwritten. If the project works correctly, the checkbox for it should then be clicked before writing.

You can set CV58 at medium speed, too, but then it becomes an iterative process. (Recommendation: Set CVs3&4 to a low value (3), use a rolling road or a test circle, set your programming device to PoM and 28 speed steps and program directly when running the loco). Try to set CV58 at both speed steps 8/28 and 12/28. This will affect CV57, so, if CV58 fits more or less (exact value not necessary at this very first step), then try again CV57 at low speed. This might affect CV58 so try this again at the medium speed steps. Then you have to try again with CV57. Probably you have to repeat this some times. You will come to a point where you will have to alter the values of CV57/58 by just 1. Then you almost got it. After this you can set CV249 which will reduce the chuff rate at the very high speed a bit in order to make it sound better (you won't be able to "see" it, as Paul mentioned, it is just for good acoustic impression.) Usual CV249 range is between 20 and 60. The higher the value, the lower the cuff rate. This is similar to a "dynamic limiter".

It's not CVs 3&4 that affect the chuff rate. It's especially CVs 2&5. Why? This is easy: for example CV57 sets the chuff rate at slow speed. If you change CV2 you change the slow speed and therefore the CV57 can no longer fit because it was adjusted to a different speed. If CV2 is changed, logically CV57 must also be adjusted. The same applies to CV5 and CV58.

The shunting mode should be available with all "good" decoders. For example, on ESU you can individually and seamlessly set both top speed and acceleration/deceleration for shunting mode.

Maybe this helps (CV7): JMRI decoder selection identification

That is correct. This is even different with some controllers: they have a dI/dt cut-off, which means that if the current rises by more than a certain value within a very short time, this is defined as a short circuit. A controller with 5A can even switch off when the current increases extremely quickly from 0.5 amps to 2.9 amps. In this case, an additional circuit breaker with 3.0 amp will not respond.

Oh, but some already give the source. And also mention the sound designer. I suspect that if this information is missing, the sounds are either not authentic and/or the sound designer is not that well known.

DCC address 0: compare it to a walking person: 1. Pure DC (e.g. battery) is like riding on a treadmill (like those found in some airports). 2. DC rectified from AC is like normal walking. 3. AC on a DC motor is like stepping forward and then stepping back. 4. AC on an AC motor is like stepping forward, turning 180°, stepping backward, turning again, etc. 5. DCC directly on a DC motor is like doing a very small or small hop forward, followed by a very small or small hop backward. 6. DCC on address 0 is like doing a very, very big hop forward, followed by a very small or small hop backward. One can imagine that when a human moves like 6., it will take a lot of energy and only strong people will be able to keep it up longer. Less robust or very old people will work up a sweat very quickly and may even injure themselves. Accordingly, a motor will heat up considerably and it depends on the type of motor how long it can withstand this. At 3 . or 5 . even only heat is generated, a movement does not take place or hardly takes place.

So is there another manufacturer trying to implement another proprietary standard?

In the ESU forum there is an offer to help you ...

Correct. The only straightforward solution is to switch it off. But there is a 'quick'n dirty' workaround by setting CV2=1 and by staying on speed step 1 for a while when driving off.

Nevertheless, no matter if it is LS3.5, LS4, LS4 select, LS5 or LS5DCC, there are ways to find out the correct version and to perfectly match them, so if both are CV8=151, no trouble.

As the Accurascale 37 has a Loksound 5 and the Bachmann 37025 probably also has a Loksound decoder there should be no problems. To be on the safe side you should read CV 08 on the Bachmann loco, if it is value 151 then it is an ESU decoder. Nevertheless, depending on how the sound project is set up you might have to alter a lot of CVs - do you have LokProgrammer or JMRI?

CV 19 values above 127 just change direction of travel. Real consist address then is value in CV 19 minus 128.

Maybe it's the address? There is a known problem with addresses between 100 and 127: the standards do not clearly define whether these are to be considered short or long addresses. This applies to both decoders and controllers, so some combinations may work and others may not. My suggestion is to use an address between 1 and 99 for a CV19 consist. And in general I would personally avoid all addresses between 100 and 127, which has turned out to be the best way in practice.

This is a known issue. Phasing occurs when both sounds are the same, but are played very, very slightly differently. I can only think of a solution for ESU V4 and V5, because you can easily pitch the engine sound there using CVs. 3% up or down on one of the two locomotives is usually enough to achieve an attractive result. I suspect Legomanbiffo does the same for double loco projects.

When you start out with sound, be sure to note that the function keys can all be reprogrammed separately from momentary to latching. Good sound projects have features that should only be active as long as you press the button, for example a playable whistle, or a brake button. With some sound projects, you can get different sounds or reactions by pressing the button briefly or long. You cannot use all these possibilities if you can only simply switch the functions on or off. Your retailer can certainly explain or demonstrate this to you, if not, find another retailer.