I received the capacitors yesterday and played around with which one to use. I ended up going with a 47 micro-farad 20 volt cap (Digikey part number 718-1091-1-ND). This cap is 7.3mm x 4.3mm x 2.8mm. in a surface mount package. By milling out a slot under the rear end of the decoder .075" deep and the width of the cap, I was able to get it mounted under the board. Remember I'm putting this in an SD40-2 so there isn't any extra length. If you're mounting it in one of the longer locomotives that this decoder works with you may not have to make the slot. The other caps I tried were a 47 micro-farad 16v and a 22 micro-farad 20v. These two were both a little smaller (6.0mm x 3.2mm x 2.5mm). However since I would have had to cut the slot no matter which I used making it .3mm deeper didn't seem like a big deal. If you use one of these capacitors you should remember that they are polarized- you need to pay attention to which lead is positive and which is negative. On these particular units the positive connection is the end with the silver stripe. The capacitor that Digitrax includes with the decoder has a black stripe on the negative lead. Don't hook them up backwards or bad things can happen. The pictures show the slot cut in the frame. The capacitor mounted, the wires going up go to the cap, the ones going to the right are the speaker. The final picture shows everything mounted and ready for the body to be installed. With the cap installed everything seems to be working. I still need to play with some CV settings and run it a while before I make any decisions on how I like this decoder, but I will say it sounds better with the body on.
Markwr, I would like to see how you did the speaker. John Claudino and I had several conversations last week and decided that there was not enough room to do it on the mid-production chassis. Bob.
This is one of the mid-production chassis and the speaker just barely fits. The bottom of the speaker is recessed less then .010" from the bottom of the fuel tank. The top of the speaker is actually attached to the plastic frame around the motor. Unfortunately I didn't take a lot of pictures while I was working on the speaker whole. Here is a picture of the frame milled out and one with the speaker installed. I still need to seal around the speaker. The way I milled the frame was, I used a half inch forstner bit to drill through the fuel tank from the inside. Using this type bit prevents cracking the plastic. Avoid the spacing tabs at the ends of the inside of the tank. Then I used double sided carpet tape to attach a 2" x 2" piece of lexan to the bottom of the fuel tank. Then drill through the lexan with the same drill bit so you have holes in the tank and lexan that line up. Now put the frame halves back in the fuel tank, place some shims between the halves to hold the correct spacing. Several sheets of paper will work. If you have some way of clamping this assembly go ahead. I held the thing on the bench top with one hand while using the dremel because I was afraid my vice would damage something. Now using a dremel tool with a router base you can start working on the frame. The lexan provides a surface for the router base to rest on and the router base lets you control the depth. Take off a little at a time and then increase the depth. Stay away from the edges of the hole until the depth is sufficent for the cutter to clear the lexan, then you can use the hole in the lexan as a guide by letting the shank of the cutter ride around it. On the cutter bit I used the cutter radius was a little more then the shank diameter. This let the cutter undercut the lexan guide hole slightly, which is how I got a hole the necessary .59" using a .5 inch guide hole. The procedure may sound confusing but it was actually fairly straight forward. If only I had thought to take some pictures. Of course there's always the next time.
Markwr, you need to send your photos to John at Aztec. He is the best at frame mods. As I said before we both looked and could not see a way to install this in a mid-production chassis. I am going to take another look at the installation for my customer. I would rather have a frame done by John than my attempts. I use a Unibit for my frame milling, works very well. Did one in a Kato E-8 B unit that worke out well, cap and all. Bob.
Unfortunately that often tends to be the case with these smaller companies. For example, how long did it take for NCE to get their radio working satisfactorily, and they had so many complaints about the Smart Booster they completely redesigned it. It's even a problem in the bigger companies. MRC is MUCH bigger than Digitrax or NCE, and it took end users to discover that certain address ranges could not be controlled by the Version 1 wireless system.
I received this email from AJ, It might be of some use to some of you dealing with the new N Scale Sound Decoders. I've found on the locomotives I've tested so far that power pickup is probably the biggest issue. If the power strips are bent or have been damaged it is going to cause problems. On a non sound equipped locomotive, the flywheels will carry the locomotive through a loss in power, Now with sound we hear the decoder restarting. Here's AJ's comments: The sound decoder instruction sheets _recommend_ that a hold-over capacitor is installed and wired to the decoder from the pads shown, to ensure operation with intermittent power pickup. This may be caused by dirty track or poor pickup compliance, or both. We include a 330uF/25V capacitor for this purpose. If your track stays clean and the locos run well, then this is not required, since there is some internal decoder power holdup. We did not include this part as SMT on the PCB because; it is quite large and constrains the use of the decoder in _many_ locomotives (particularly N), and a customer can lower this size or increase it, as conditions warrant. A good small choice is a 47uF/16V capacitor, but 25V is safer in case the unit is used on higher voltages by accident. Another good compromise is a 100uF/25V electrolytic, and the SMT version at about 6.3x7.7 mm is quite compact and it is straight forward to add wires to the leads and install. Be sure the capacitor leads and body do not come into contact with track voltages. We used a 32ohm speakers in some designs (mostly HO) to provide optimized power efficiency at the 1W power level. An 8ohm speaker may be used instead, with the master volume (CV58) adjusted lower, so as not to trip the power management limits. The N scale designs use 8ohm speakers standard. Digitrax sound decoders employ a sophisticated and high efficiency power management architecture, and the power hold up at the track voltage levels (12-18V) provides by far the best energy efficiency. Rich
Digitrax provided me with the same information. I eliminated the possibility f the problem being caused by poor electrical contact with the track by connecting the frame halves directly to the bus feed wires with clip leads. This eliminates the connection between the track and the rails and the connection between the trucks and the frame halves. Setting the master volume lower didn't help either, I tried every setting from zero to the default of nine. I think the cause of the problem is related to the comment concerning the speaker impedance. If the circuitry was designed for a 32 ohm speaker, connecting an 8 ohm speaker would result in more current being drawn. Unfortunately the only way to test this theory would be to install a 32 ohm speaker which I don't have. The capacitor I used cost 87 cents, a lot cheaper then replacing the speaker plus it has the added benefit of allowing the engine to cross the occasional bad spot in the track. I'm not displeased with having to install the capacitor. It would have been nice to have one that would fit included with the decoder but it was easy to install and is a good idea even if you're not having a problem.
We've been discussing the Athearn F45 with the Soundtraxx Tsunami decoder on another thread. Right out of the case, probably with dirty, oily wheels and new motor brushes that may not have set in optimally, this engine never stutters, stalls, restarts. They have done something right with their choice of components, ie caps, speakers. I can't imagine the frustration of a modeller who is not electronically gifted going through the installation of the Digitrax decoder, only to have their prized engine unable to travel a few inches without stopping and restarting constantly, then to be told by the manufacturer that their track is dirty and that is the cause of their problems. I was eagerly awaiting this product from Digitrax as an alternative to MRC but clearly, this is a fairly complex mechanical and electrical install, with a separate capacitor not installed but recommended. This is a poor first impression for Digitrax and N scale sound. It will be hard for them to overcome this, unless they own up to their role in the problem and provide a fix, rather than confusing them with false explanations. Like the early impression of the Apple I Phone 4. Eventually Apple acknowledged their role and provided a fix ( I don't have one, so I don't know if it works), probably restoring their credibility. BTW, exactly how is the SMT cap attached to the decoder?
I've been an electronics technician for over thirty-five years. I wouldn't claim this makes me "electronically gifted" but I do know which end of the soldering gets hot. Hopefully none of my postings made me seem frustrated because I'm really not. I've viewed the entire process as a challenge to overcome. The main reason installation problems with a new product are a problem is you never know if you installed the unit wrong, the unit you purchased could be defective or if there could be a design problem. Experienced people make mistakes once in a while and can screw up an installation. Occasionally you get a DOA component from even the best manufacturers. Design problems occur occasionally with new products and it's not easy for a manufacturer to test a product for every possible situation. This is kind of a long winded way of saying the installation problems wouldn't prevent me from using this decoder again, I now know I need to install the cap and when I ordered them I got five. As far as mounting the capacitor - in a previous post I showed three pictures. The second picture shows the cap with wires attached to it and run through the gap between the frame halves. The third picture shows the wires attached to the decoder. The points were the cap attaches to the decoder are shown in the decoder manual. Just be careful with the polarity of the cap as I mentioned in that post. Now for the things that might prevent me from using this decoder in the future: 1) It doesn't sound as good as the Athearn FP45 Soundtraxx units I have or the Kato F7A/F7B units I put LokSound decoders in. The bell doesn't ring, it kind of makes a "tink" sound. The rest of the sounds are better but overall the other decoders seem to have better frequency response. 2)I can't get this decoder to accept programming on the main (POM). It programs on the programming track fine. This wouldn't be an issue but I find it's easier to optimize motor settings with the engine running. At this point I'm not sure if it's this decoder or a compatibility problem with my command station. Since this is the only decoder I have that doesn't work with POM I'm leaning towards it being the decoder. Even with these two issues I might still use this decoder again for the simple reason it fits. Issue 1) isn't that bad and if it really bothers me I'll avoid the bell. 2) is more of a problem, but once I have the basic settings figured out I shouldn't have to make changes very often. Also I can purchase this decoder for about half what the others cost ($55.99), plus $0.87 for a cap. I would like to hear from someone else who has installed one of these to find out their impressions.
Richard, the caps shouldn't take any noticable time to charge up. Adding more capacitance should help. From the picture you attached it looks like you should be able to add two more caps in the space available. Just add them in parallel to the existing cap and make sure to keep the polarity correct. The other option would be to get one cap with a higher capacitance. You have a lot more room then I did with the SD40-2. Also, this decoder seems more sensitive to dirty track then the other sound decoders I have. I agree with your observation about the speaker not sounding good without an enclosure. I used the fuel tank and motor cavity as the enclosure and just putting the shell on made a noticeable difference in the sound. As far as the problem with the decoder not programming on the main i'm glad to here yours works. I had taken the locomotive to a local train store and asked them to test the programming on their track. They couldn't program it on the main using an NCE system. Unfortunately their Digitrax system wasn't working at the time. At least with yours working I know it should be capable of programming on the main. I may try getting a warranty replacement from Digitrax.
I did a SD-40-2 mid production with this decoder and a 10mmX18mm speaker with the help of John Claudino at Aztec. I used the small cap the Markwr reccomended. The frame was milled to allow the speaker to be under the board. John took just enough off the back of the frame to make this work. The cap is at the rear of the frame. The sound is great, needed to be set much lower than factory defaults. No problems with resetting, now some 5-6 hrs run time. The only problem is you can not purchase just frames, only complete mechs. This makes it much harder for John to do swaps. I sent him 2 locos so I now have a frame ready, no speaker available (Digitrax supplied one for this test) and waiting for the nest shoe to drop. I and my customer are very satisfied with the results. Bob.
That's very good. I too have been in touch with John C. and with AJ Ireland at Digitrax to try and get those 10 x 18 mm speakers, but AJ won't give them to me (yet). I know ESU Loksound also advertises them but they also don't have them yet. They will be standard on their decoders shortly. I bought an SD70ACe from John, the prototype frame-milled unit for a bargain price. I tried operations with just a single 47uF cap, and found my GEVO was still stopping and resetting at bumpy track joints (not really bumpy either) and the Kato Unitrak had just been cleaned, the engine freshly reassembled and very clean and free-running. You can buy capacitors like the ones we used, in the same dimensions, up to 100 uF -those would cost just over a buck each. I'm going with those from now on. I like the prime mover and various auxilliary sounds very much. Horn is OK, not loud enough comp with the prime mover. Only one bell per prime mover, both sound electronic rather than mechanical, maybe that's the proto too. Brake squeal is WAY too short, it's like 1/10 of a second, coupler sound also too short, F11 (handbrake) sure sounds like wheel clack, dynamic brake sound is good. Missing a good auto brake squeal when stopping, as well as auto notch-down when applying dynamic brake. I just looked on Digitrax's website and I can't find hide nor hare of this decoder anymore. It used to be front and centre. I've asked them why, hopefully it's being "improved" with the new speaker and hopefully, an on-board cap.
I do not understand why they made this decoder with a speaker not able to go into a cab frame. I am very happy with what I have done. Bob.
The 10x18 speaker sure looks a lot easier to install then putting the original speaker in the fuel tank.
I now have the shell on my FVM GEVO, and I agree that the sound is quite loud, considering the tiny 10mm speaker. The SD40-2 sound is much better than the MRC 1806 SD40-2 sound, if that's what it really is in that decoder, but that is probably because there is no enclosure of the MRC speaker. The GEVO sound is very different, kind of clattery and irregular, with lots of pops and pumping sounds. Unfortunately, no speaker that is less than about 25 mm diameter can even touch frequencies under about 850 Hz, so we are hearing everything but the deep rumble of the prototype. Even though the horn and bell don't sound really great, they do sound like the real versions. I would assume that these are turbocharged but I can't hear the turbo whine at all. One thing I learned on this forum a few years back was a "trick" with Digitrax "3" series decoders, and I wanted to try that on this new one. It does work. If you set CV54 to a value of 1, and CV4 to 63, the F6 key becomes a brake; push it and the engine starts to slow down and will eventually stop, release it and it will accelerate back to set speed. If you set CV3 (acceleration) to a high value like 24, you will get a nice slow smooth acceleration back to your preset speed. Of course, you can't do this with manual engine notching since that also uses the F6 key; it works with auto or semi-auto notching. I haven't checked the manual for function remapping, it might be possible to remap manual notching to F12 which is not being used, then you could control the prime mover manually and still have the brake function on F6. Of course, the above special unadvertised feature works with all non-sound Digitrax decoders with the number 3, like the DN163K1B/C, DN163A0, DZ143, etc.
