Tsunami CV Changes Won't Save

Before I call Soundtraxx customer support, I thought I would ask here.

I have a loco with factory installed Soundtraxx Tsunami decoder. I have enabled one of the preset speed curve by setting CV 29 and CV 25, but it won’t “save” after the loco is removed from the tracks and then returned. The speed table must be set at the beginning of each op session. Other CV settings, like volume, save between op sesions as do all of my other sound decoders.

Is there something with the speed table CVs that needs to be set in order for the CVs to “stick” between op sessions? BTW, resetting the decoder doesn’t help.

Thanks

Just out of curiosity, are you setting CV29 and 25 with the loco on the programmng track, or on the main?

I’ve had a couple cases where I couldn’t get the speed tables to work until I initially programmed CV29 on a programming track. I don’t know why that is.

Are you setting CV 25 to the speed table you want first, then enabling Bit 4 in CV29 with a value of 1?

The order of how this is done may be the issue. The wording in the tech doc implies, but doesn’t clearly state, that CV25 is set first, then CV 29 enables it. Your wording suggests the opposite.

Also, if you have something non-standard in CV 29, that might be affecting things. Not sure what the default value is, but reverting to that for CV 29 then changing only Bit 4 might clear that up. In fact, restting to default on the whole thing, then trying to set up the speed table might be useful if a little more hassle. That’s where Decoder Pro is handy, as it make it easy to do that and then restore the rest.

I’ve been asked to help a club member who was having trouble with a couple of his decoders (not sure what they were because it was a while ago). He has Digitrax, I have NCE. When I tried to reset the decoder, it didn’t save the new address until I deactivated the verbal acknowledgement (CV 62?) feature. After it was reprogrammed and verified to keep it’s new address (after turning off the power for a few minutes and then turning it back on and selecting the loco’s address), I turned the verbal acknowledgement back on. I have no idea why that feature causes problems, but it’s easy to turn off/on (at least on my NCE…I have no programming experience on Digitrax, I’m sure it’s just as easy).

Thanks for the replies. After further experimentation, its not really the spped table that isn’t saving but simply CV2. I’m setting it to a value of 3, which slows the starting speed to a creep, but shutting off the track or removing the loco returns the speed back to factory default (don’t know the value). This happens regardless of whether or not I use the programming track or the main. The CV can be changed, but won’t seem to save after an interuption of current occurs.

Well, I’m confused. Are you using the speed tables or not? If you are using the speed tables then I don’t believe that CV2 enters into the picture.

OK, couple more ideas.

CV2 will be overridden by the speed table. I think it’s one or the other, at leats that’s how Decoder Pro presents it on the screen. It’e been awhile since I’ve programmed them the old fashioned way, so can’t quite remember how that’s presented like that.

The other possibility is the decoder is locked. If you’ve been just trying to change that one CV and it’s not working, have you tried changing anything else? Does that work? If not, suspect it’s locked and needs to be unlocked to program.

Sorry. Yes, speed table 11 is enabled. What I meant was that the acceleration along the curve seems fine, the starting speed doesn’t save. It reverts back to original speed, even though reading back the CVs shows CV 29 at 18, CV25 at 11, and CV2 at 3; which is how I set it up to creep.

The experiment was to enable speed curve 11 with starting value at 3. A test loco that is faster than this one runs up the back very quickly. Turn off the power, then turning it back on seems to change the starting speed since the faster loco now no longer catches this one. However, reading back the CVs shows the CVs at the proper values of 29/18, 25/11, 2/3, which is the proper settings to get the slow speed I got before.

In order to get the slow creep again, I have to reset the decoder, CV8 to 8, then do the enable speed curve steps with CV2 at 3 again.

I think the decoder is wonky.

I’d contact the manufacturer. I’ve had similar problems with a couple of Tsunamis; they’re fine for a while, then a couple of the CVs change on their own. That’s the main reason I probably won’t be buying any more of them.

I guess I’m not understanding what you are looking at. The speed tables adjust CVs 67 through 94. When you are using the speed tables you set CV 67, which corresponds to speed step 1, to get the loco to start moving. If you are using one of the pre-programmed speed curves, then the speed at step one is set by that curve.

I also don’t understand your CV 29 value of 18, unless you’re using a 2-digit loco address. To arrive at CV 29=18 you are selecting advanced 28/128 speed control, analog conversion off, normal direction of travel, speed table enabled, and 2-digit address. If you are using a 4 digit address, then CV 29 should be 50 for the same conditions. Of course if you really want to use a 2-digit address, then 18 is correct.

The decoder isn’t faulty.You can write whatever numbers to CV2,the decoder will ignore it when the speed table is enabled.It’s in the design.

I don’t clearly understand what you want to achieve.Either you disable the speed table or experiment with other speed curves in CV25.

There are two ways to set speed up in decoders. There is the basic speed curve where you set low (CV2), mid (CV5) and high (CV6) voltages or you can an advanced speed curve (CV’s 67 thru 94)

What you are attempting to do is use both. This is not possible. CV 29 will let you select one or the other but you cant use both at the same time. You will need to change the values for CV’s 67 to 94 if you have it set to use speed tables.

Your value is being saved, however, you have told the decoder to use the advanced speed curves (using CV 29) so it will ignore the values in CV’s 2, 5 and 6.

I hope this helps.

Thanks. I will try the other prescribed ways you suggest, and seek the manual for further guidance on this. The thing is, I’ve had other factory Tsunamis that saved this particular performance from session to session, which is why I presented this as a problem with this particular loco and not how I set up the speed curve.

The real goal is not the shape of the speed curve itself, but a creepingly low start speed. So if you or anybody have a series of CV entries that slow the start speed on a Tsunami without even messing with speed curves, I’m all ears.

I’ve tried messing with CV2 by itself, and with 5 and 6, but the start speed doesn’t change…unless I enable a speed curve.

If you set CV2 to something in the range of 20-30, the loco should leap into motion like a jackrabbit. If the issue is that the loco starts like a jackrabbit even when CV2 is a small number, like 1-3, then you’re looking down the wrong rabbit hole (keeping the whole rabbit thing going here…). You need to be looking at the BEMF settings and adjusting those.

If the loco doesn’t start moving until you crank the throttle way up, and putting fairly large values in CV2 (check the manual for the max, it’s not always 255) does not change this, it still could be the BEMF settings.

Also make sure you aren;t accidently setting CV3, forward momentum. A large value here will make it seem like the loco doesn;t start moving until you crank up the throttle, but in reality it is just the momentum delaying the throttle change.

Always start after a reset, that way you have a baseline to work from.

–Randy

I don’t know what the problem is, but there is something some of you don’t understand about his configuration. In addition to a straight line curve and custom speed table, the Tsunami also has several preset curves you can select from. Doughless is not using a custom table, but using one of the preset curves. To use a preset curve, you program CV29 as if you are using a 28 step table and you select either a preset curve or the 28 step table with CV25. CV2 is still supposed to work when using a preset curve.

Hmmm, not so sure of that myself. That’s definitely not how JMRI Decoder Pro presents it, but as I said it’s been awhile since I messed with things via the old school interface via my NCE.

BTW, while CV 5 and 6 are presented in some programming menus with the Tsunami, they aren’t actually used or settable in the Tsunami. So you can set them all day long and shouldn’t make any difference. Effectively, CV 5 and 6 are the “poor man’s speed table” but if you want that with the Tsunami, you just have to use the user uploadable speed table.

It wasn’t specified whether this is a steam or diesel loco. If steam, you’re not really doing what works best to get good low speed performance. I have some notes culled from various sources that will unlock that for you. I can post them again here, just ask. I don’t have any diesel Tsunamis but presume similar “CV menus” are availabel to do them same for them. If you want to experiment on your own, download the Tsunami users guide for steam or diesel and turn to the section on setting up the Hyperdrive where it discusses CVs 10, 209, 210, and 212.

Thats correct Robert. I believe the 67 to 94 CVs are for making a custom speed curve that is different from one of the preset curve choices Soundtraxx provides. I like speed curve 11. Its a preset parabolic curve bent to the slow side in the lower speed steps, which is what I want.

In practice, setting cv29 and cv25 and cv2 to the values I described earlier achieves the slow startups that I want. In fact, setting CV2 to 4 instead of 3 makes start ups noticeably faster, so changing cv2 during the preset speed curve enabling process does seem to matter.

Again, I’m getting the performance that I want with the cv changes, it just doesn’t seem to save in between sessions even thought the cv readback shows those three cvs to be exactly how I had set them. That’s what really is the stumper. If the CVs went back to default or changed after the track being shut off, it would make a little more sense.

I need to run through everything tonight again. Maybe I’ll mess with BEMF a bit.

BTW, its a diesel. Brand new from one of the on-line dealers.

I’d give Durango a call. They can set you straight on exactly what’s going on. I’m kind of curious what they say about CV2, too.

My NCE decoders seem to have their speed tables affected if CV2, 5, and 6 are set at significant variance, because I’ve resolved wonky running by setting them to more appropriate values. Never noticed the Tsunami’s doing this, though, which doesn’t mean it doesn’t happen, just that I’m human[*-)]

Yes, CVs 67 to 94 are what get adjusted to make a custom speed curve. However, I believe that all you are doing when you select one of the preset curves is populating those same CVs with whatever values the preset curve wants to put in them.

I know this because if your were to use JMRI to program the speed curves and selected one of the presets, you would see on a graph the settings for each speed step.

Concerning the interaction between the CV2 value and the speed curves, I believe that I have read that there is some effect on the curve from CV2 ***. However, I don’t believe that the value in CV2 replaces the value in CV67 if that is what you are trying to accomplish (Loksound decoders act differently).

The easiest thing to do to be able to see exactly what the speed curves are doing is to use JMRI.

***(Edit) If you look at page 2 of the technical manual, http://www.soundtraxx.com/manuals/tsunami_diesel_technical_reference_0213.pdf, it shows that whatever value you put in CV2 will offset all the points on a given speed curve by the same amount after speed step one. So to me this means that all the values you are trying to put in CV2 are having no effect other than messing up whatever your preset speed curve wants to give you for speed step one. If you don’t want to have this occur, you need to set CV 2 to a value of zero.

Good catch. Now I recall seeing this somewhere before. I’m fairly certain that’s what’s going on with my experience with something similar with NCE as I described earlier.