jdcolombo you've inspired me!

[jdcolombo]

Hi Tony.

Since I don't have an H16-44 to look at, it's a little hard to make suggestions.  But in looking at the photos of the frame on Spookshow's database, it looks to me like this would be possible.

Speaker in the center, where you were going to put it.
LokSound in the rear. 

Now I understand that the H16-44 has two LED's, one for the main light and a red/green LED to light the marker lights.  I think what you could do is cut the rear portion of the circuit board off (the section with the LED's and maybe any surface-mount resistors that connect to the LEDs), and tape it ON TOP OF the LokSound, then wire whatever you need to wire to it (yellow and blue?  Do the two LEDs operate independently, or do they both come on when the loco is in reverse?).  It looks to me that there is enough length on the rear shelf for the LokSound - though you could get some more by cutting down the rear of the top of the center part of the frame (to just above where the motor mounts are - but be careful, because you don't want to crack the frame here), so that the thin part of the LokSound fits there, and the thick part fits behind on the shelf.  You'll need to do some measuring.  If the space is there, then you can keep the front part of the circuit board for the front lights and to use for power pickup for the decoder; you can run wires underneath the speaker.  You might also have room in the front part of the loco to mount the keep-alive cap. 

Another possibility is to mount the speaker in the fuel tank area - don't know if you have enough room there or not.  Or mount the speaker in the cab area of the shell - that's what I did with my VO-1000.

John C.

[tony22]

John, I'll take some measurements but I'm not sure this would work. The problem in the light shields for front and back. Here's the link to the Atlas exploded drawing for the H-16

http://www.atlasrr.com/pdf/PartsPDFs/NParts/NH1516Loco.pdf

Note parts 520226 and 520227. These slip between the LEDs on each end, to keep one LED from bleeding its light into the marker lenses meant for the other. When the engine is moving forward both the main headlight and the Green markers are lit. Those shields are built so that the PCB could not be offset any lower than where it sits in the frame.

Thanks for the suggestions, however. Gives me some areas to looks at that I may have ignored so far.

[jdcolombo]

Hi Tony.

One last thought.  You don't have to use the Atlas light shield.  Think about alternative solutions: - paint a piece of .010 styrene black and position it between the LED's.   Or, glue a surface-mount white LED to the main headlight and then put a piece of black electrical tape over that to block out other light, while keeping the marker light LED in operation from the board (this is probably what I would do).  There may be more - but in general, if it's just an issue about how to light up lights, it usually can be solved.

John C.

[PGE_Modeller]

Hi Tony.

_ snip -

Now I understand that the H16-44 has two LED's, one for the main light and a red/green LED to light the marker lights.
 
- snip -


John C.

Not intending to hijack the thread, but something to remember is that, on a diesel locomotive the only "marker lights" are the rear ones when the locomotive is running "light engine".  The lamps to the front of the locomotive are "classification lamps" or "class lamps".  They are unlit for scheduled trains, show white to indicate an extra train or green to indicate that one or more sections of a scheduled train is/are following.

Ngineering have a very nifty item listed as a "steam locomotive class light simulator" that would be equally appropriate to use in a diesel.  It converts a single funtion output from a DCC decoder into a three-state function that cycles through 1) no lamp, 2) white lamp, or 3) green lamp.  By connecting green and red LEDs instead of white and green, it may also be used to simulate correct  marker lamp indications for cabooses:  not lit when not on a train (or otherwise required by rule to not be lit), red when the train is occupying the main track or green when the train is in a siding clear of the main track.  Two units would be required if running on double track to allow independent control of left and right markers to permit the correct display of markers when standing or running against the current of traffic (red to the outside, green between tracks - Rule 19 in Canadian Uniform Code of Operating Rules, 1951 to 1962 editions).

Cheers,

[tony22]

Thanks guys. I've been a little busy at work but between last night and just tonight I finished chopping the frame for one of the H-16s. I have to clean it up a bit before fitting the other pieces into it.

I'm sure you who have done this already know, but anyone else who wants to try and grind a metal frame down for this kind of thing - wear safety goggles and a respirator! You have the equivalent of metal dust flying around. I was thinking as I was doing it that it would be a lot safer if I had some sort of grinding booth that kept all the dust enclosed but still allowed access for a Dremel and my hands.

....

Nooooooo! I ground down one of the frame halves too closely. It cracked when I pulled off the tape that I used to keep the dust off the rest of the frame. Oh well. Off to Atlas for new parts.

[jdcolombo]

Hi Tony.

Sorry to hear that.  I did this with my first RS-11 and had to order new frame parts for it.  It sucks.  It's also weird, because it seems that different Atlas models are made with different frame material.  The RS-11's are very brittle, but the recent run of GP7's are different - they seem to have more "give" to the metal.   Same for the SD9's, which seem to be made of the same stuff that the GP7's are.

I also think that perhaps one could use something like JB Weld to fix the cracked frames.  I still have the cracked RS-11 frame, and I might try this.  If it's just a single crack on one end of one side of the frame, then most of the integrity is there, and all you really need is to make sure the frame doesn't droop and throw off the mesh between the worm and trucks.   I've used JB Weld to fix all sorts of metal things around the house - like the broken handle on my Weber gas grill - with good results.  I just don't know if it would work in this scenario or not.

The real solution is a milling machine.  If I were going to do many more of these sound installs, I'd buy one.  I came within a hair of doing it a couple of months ago when Micro Mark had a big sale going on, then pulled back.  But a small milling machine sure would make all this a lot easier.

And yes - Goggles for sure.  I didn't use any kind of respirator, but I didn't grind the frame as much as I cut it using a diamond-based cutting wheel in a Dremel.  I'd just make  slices where I needed them and then chunks of metal would come off where the cuts were made.  Too bad someone doesn't make a precision laser cutter for home use!

John C.

[tony22]

Yes, I wish I had a milling machine too for this kind of work. I also have JB Weld and thought about using it to repair the damage, but in this case I don't think it would work. It cracked at the top right where one of the nibs that the PCB slides into is located. That nib flew off. I found it, but I don't think even JB Weld would guarantee that the bonded nib would be able to stay in place once the PCB was slid into it. I ordered two pair of frames from Atlas. Hopefully they'll be here by Saturday.

The real problem was I was really trying to keep out of the way of the dust as I was grinding. I have the Dremel diamond wheel but it seemed to be having a hard time cutting lengthwise through the frame. I started making slot cuts across the top and then I went back and "pushed" the face of the wheel through each piece between the slots. That cranked up a lot of dust! Trying to keep from getting covered with the stuff threw off my concentration.

So for the next go around I built myself a little "grinding shield". It may or may not work. It's a piece of clear Lexan "hinged" to a piece of wood with Duct tape. I'm attaching a pic of the good frame half that I ground down and the shield. You can see on the frame how close to the nibs I had to go to create enough space for the LokSound decoder.

I don't know if this will work. I uploaded the pics to one of my other accounts and linked the urls here.

http://www.audiocircle.com/index.php?action=gallery;area=browse;image=96649
http://www.audiocircle.com/index.php?action=gallery;area=browse;image=96650

[tony22]

Hmm, dumb question. Anyone know on this motor if the orange wire goes on the bottom tab or the top?

[peteski]

Orange would be the side which was originally touching the right side of the frame (red decoder lead).

[tony22]

Orange would be the side which was originally touching the right side of the frame (red decoder lead).

Which turned out to be the bottom, which is good since that's what I expected.

Thanks Peteski.

[tony22]

Okay. The decoder is installed and everything is reassembled (sans body shell). I didn't hook up the speaker yet because I wanted to make sure the basic decoder functions work first. Let me tell you this installation was a b!tch. I had to grind away more of the frame than I thought and routing the wires required a bit of imagination. I'll be posting pics in a bit.

So it runs beautifully and all the lights come on in the right colors in the right directions. But the Green marker at the cab end (when the cab end is in the FWD direction) is very very dim. The White LED on the bottom is as bright as the other end (nice and bright), and when the long hood is moving FWD the Red at the cab end is also nice and bright. It's only the Green at the cab end. The fact that the White LED and the Red in the R/G at the cab end are the correct brightness tells me the voltage coming off the Blue wire at that end must be okay. I checked the connections and it all seems okay with the multimeter, but the Green voltage at that end is less than 1 volt. At the long hood end it's 2.1 volts - what I expected. Too tired to troubleshoot tonight; that's a recipe for disaster.

[tony22]

It's done! It's done! My apologies to this site but I'll have to do like I did before and link to pics that are on my Trainboard account. It's too bad this site doesn't allow direct uploading of pics as attachments. Unfortunately I don't have an account on any site where I can post a video, so the whole point of seeing (and hearing) the outcome is lost. Anyway.

The first picture shows how much of the frame I had to chop away to fit the decoder. Note also the small bits I had to cut away where the two halves meet, below each set of circuit board nibs. This was to create a channel so that I could route the wires under the decoder instead of on top of it. If I ran the wires above the decoder the body shell would not have fit.



The next pic shows the small bit I had to grind away to clear the resistor I added below that circuit board. I had to do this because, while I used Brad Myers' great instructions for putting DCC in the H16-44, leaving that small section of the short hood board with the 1 KOhm resistor wasn't going to work. I had to cut that part away and replace that resistor so I decided to put it on the bottom of that board. Here's the frame pic. The part I had to grind away is a little hard to see bit you should be able to make it out behind the front worm gear opening.



and here's the short hood PCB



You can see Brad's original intention here

http://n-scale-dcc.blogspot.com/search/label/Loco%3A%20Atlas%20H16-44

Now this will be a little hard to explain, but because I cut away that resistor on the PCB, the white wires going to the tail end of my replacement resistor also required a small jumper added from that end to the back end of the diode on that board (otherwise the Green LED will not light). I slipped a very small length of wire into the test point hole at the top of the board (in my picture above). That test point connects to the back end of the diode.

So I said I channeled out the frame so I could run the wires under the decoder. Here's the way I dressed the wires before doing any cutting. I had to try this a few times before I got an arrangement that would flow under the decoder and allow for it to be set into the cutout without any stress on the wires.



The next shot shows the decoder and boards wired up before I soldered the connections to the motor. I wanted to do it this way so I could test it all the boards to see that it all cleared. It was tight. I honestly don't think I could have added either a box for the speaker or a keep alive cap. At least not the way I did it. BTW, this pic does not show that little jumper wire I talked about above. I discovered I needed that after the first electrical test.



Here's the final setup in the chassis and the test showing the lights operating correctly. The back of the speaker is protected by two layers of Kapton tape and there is an additional layer of Scotch double-sided tape holding it down to the PCB.



A couple of observations:

If I knew how tough this one was going to be I'd have picked a different engine for my first try. Trying to do an installation like this while keeping the original guts and lighting arrangement was interesting, to say the least.

After the decoder was installed and I ran the engine I noticed that it's generating a lot more noise than it used to under DC control. Not mechanical noise - electrical motor noise. Like the frequency used by the LokSound is causing the motor to make a lot more noise. At high speed it overwhelms the engine sounds coming out of the speaker.

Which brings me to my last observation. While I am thrilled with having done this sound installation, I have to say the three samples of the SCP-08B speakers that I tested all have a slight "buzzing" quality, most easily heard when the air horn sound is played. This is either in the engine, tested as standalone with the decoder wired directly to the PowerCab, or even in an enclosure. Maybe the master volume is too high, but it really isn't all that loud, especially compared to the level of engine noise.

I'm happy it came out well and works, though. Wish I could figure out how to get you all to see a video.

edit...

I changed the Air Horn to a Leslie S-3 and now there is no buzzing (or practically none) from the speaker. The only bit left that's concerning is the motor noise at higher speeds. The loco tops out at about 70 smph (about right for a H16) at speed step 126. At speed step 70 the motor noise masks out the LokSound engine sounds.

[tony22]

Hey! I just found out that now that Yahoo owns Flickr I automatically have a Flickr account since I also have a Yahoo account. Here's a video. Sorry it's out of focus but I was too excited to get this posted. I can take more later.

https://www.flickr.com/photos/121313115@N07/13390700353/

[robwill84]

I would replace one of the led's and board with a smaller LED and SMD resistors, remove the black plastic piece, and build an enclosure in the front. You've gone to this much work, you might as well do it right. It will sound so much better with a proper enclosure!

[tony22]

Rob, I did build a .156" high enclosure for the speaker. That's about what will fit in that area. While it helps a little bit, I will be left with a problem up front. Those black pieces front and back are there to make sure the Red/Green LEDs don't "bleed" into the white headlight LED (and vice versa). If I pull either one I wind up with an off color white headlight; I'd have to replace it with something else. The difference with such a short height box was there but pretty small (IMO). With the body shell over the speaker it winds up creating a bit of a resonant cavity anyway.

For my second H16 if I were to try something different I might go with John's suggestion of putting the speaker in the channeled out midsection (which was my original plan) and then doing the LED replacement with SMDs like you said. If I do that I think the decoder board will fit over the long hood PCB.