Intermountain F units: MU and sound install

[Cajonpassfan]

I've done a Loksound Micro install in a powered IM F unit before, using a small Knowles speaker and I am reasonably happy with it. For this ABB F-7 project, I've decided to try the bigger speaker and enclosure that come standard with the ESU Select Micro decoders for better, deeper sound. This involves turning the middle unit into a dummy, milling out the frame to accommodate both the speaker and the decoder, and driving the two adjacent units' motors from this decoder. The decoder seems to have plenty of power to drive two N scale motors under full load, or so a presume. I'd also like to interconnect the pickups from a at least two locos for a more steady power supply.

Here's what I've done so far: milled out the frame, and thinned the enclosure sides on a small disk sander to 17mm, and thinned the inside of the shell by about .4mm on each side by filing, to make for a snug fit.

The next challenge is to find appropriate four pin connectors to place between the units for both power pickup and motor control. I have some beautiful Richmond Controls connector sets, but the wire is so fine that I'm concerned whether I'm asking for a problem....is the wire heavy enough to carry the needed power?

I also found Micro connectors at the TCS website which look a bit heavier but don't have specs for either.
Knowledgeable thoughts or advice would be appreciated.
Otto K.

[jagged ben]

Why power both adjacent units from the sound decoder?  I don't see the advantage in that.  Seems to me the third unit could have it's own decoder and you could speed match as necessary.   I see the advantage in having extra pickup for the sound decoder, but not driving two units from one decoder.  Conceivably two motors could also confuse any BEMF functions.

You could still wire jumpers between all the units for reliable track pickup, regardless.

As for the wire and connector size, I think it would be fine.  Those TCS connectors come with 32 gauge wires, but I believe that's the same gauge that comes with all their decoders, i.e. they're telling everyone that's fine for N scale motor leads.   I would probably test everything extensively without shells on, just to be sure.  The nice thing about this model is that the couplers are attached to the chassis (not the shell), so you could fully test it pulling a full train.  If there's a problem it should show itself within an hour of pulling a full train on grades.   Most likely it would be okay.  It's just that if something heats up and melts you wouldn't want it to affect the shells.

[jagged ben]

BTW, thanks for pointing out the connectors.  I was unaware of them and can definitely anticipate some uses.   For example, I cannot stand the idea of ditchlights or beacons that are installed to the shell without an easy way to disconnect them from the mech for servicing.

[Cajonpassfan]

Why power both adjacent units from the sound decoder?  I don't see the advantage in that.  Seems to me the third unit could have it's own decoder and you could speed match as necessary.   I see the advantage in having extra pickup for the sound decoder, but not driving two units from one decoder.  Conceivably two motors could also confuse any BEMF functions.

Thank you for your thoughts Ben. Other than not having to spend money on yet another decoder (this decoder "hobby" is getting expensive) I thought that not having to match different brand of decoders in the two powered units would be preferable. And, since the units will be semi-permanently wired anyway for power pickup, why not run two extra wires to power the motor as well.... the motors draw next to nothing and the decoder is rated at 1amp.

Your comment about possibly confusing the BEMF functions is interesting, and something I haven't considered. I'd like to learn more about this issue. My experience is the other way around: I have a set of Kato PA's with each powered unit having its own decoder and I had to turn the BEMF off because they weren't playing nicely together. Granted, these are older Lenz decoders I hardwired in a long time ago and perhaps the technology is better now....

Thanks,
Otto K.

[jagged ben]

Don't know if the motors will draw next to nothing, especially when you consider they'll be pulling the dead weight of the dummy loco.   I'm pretty sure you have grades on your layout, no?   Locos going up a grade with a train draw more amps.

Do you have a way to test the amp draw of the two locos in DC, with the dummy and full train?   If it were me I would at least want to know.  .5 amp for an nscale loco is not completely out of the realm of possibility.

[peteski]

Otto, if you install the ammeter between the DC throttle and the track it will measure the current draw of all the locos on the track.  The LED headlights add just few extra milliamps of draw (or you can unplug them for the test to just get the current consumption of the motors alone).  Most modern N scale motors (like Kato and Atlas) consume around 250mA at full throttle (12V) and under load (but not stalled).

I also like to comment on your choice of speakers.  The reason those small sugar-cube speakers (like Knowles) is because they are very efficient.  They have very softly-suspended and long-throw "cone" and a very strong magnet.  The older larger speakers (like the one shown in your photos) have fairly stiff suspension for their cone (so it doesn't deflect as easily).  If you could cram 2 Knowles speakers in the same space as the speaker you are using now (as long as you have a tight enclosure) you should end up with a sound superior to the large oval speaker you are planning to use.  If the decoder can handle 4 ohm speaker load, hook those speakers in parallel - otherwise hook them up in series. Just pay attention to the proper polarity or the sound from each speaker will be canceled by the other one.

[jdcolombo]

Hi Otto.

Agree with Peteski here - I don't know how much room you have in that slot, but my speaker experiments have shown that a 13 x 18mm Knowles Grande will outperform a 14 x 25 or 16 x 30mm oval speaker. 

Here's a video in which I compared a Knowles 13 x 18mm Dumbo (since replaced by the Grande) in various enclosures to a 16 x 30mm oval in a sealed enclosure made from lead sheet (the Dumbo was in an enclosure made mostly from .040 styrene, with lead sheet on the bottom).   If you go to the 3-minute mark of the video, you'll see the Dumbo in an enclosure that is 6mm high by 25mm long, which produces about 1 cubic centimeter of air space behind the speaker.  If you compare this to the 16x30mm oval that follows, I think there is just no comparison: the Dumbo kills the oval both at low frequencies and overall sound balance.  The key is to make sure you put the Knowles in a sealed enclosure that has at least 1 cubic centimeter of air space.  You can get there any way that works - for example, using an enclosure that has 8mm-high sides in a simple rectangle built around the speaker itself is fine; subtract the 2mm of speaker height and you have an enclosure that has an air space of 6mm x 13mm x 18mm = 1,404 cubic mm or 1.4 cubic centimeters.  The "long but thin" enclosure in my video gets you there too: 4mm (6mm height less 2mm for the speaker thickness) x 25mm x 13mm = 1.3 cubic centimeters.

So I'd stick with the Knowles Grand in this application. 

As for the wiring, I wouldn't worry about it.  Given the very short lengths of wire you will be using, 32-gauge flex ought to be just fine for up to 1-amp. 

Here's the speaker comparison video:

/>
Good luck!

John C.

[Cajonpassfan]

Thanks John, and Pete, for your comments; anytime you two agree on something I figure I better pay attention
Great video, too, it sure tells the story! Called DigiKey today and ordered ten Grand's for a grand total of twenty bucks and some change....this has got to be one of the better deals in model railroading!

One would think the speaker that comes with the excellent ESU decoders would be optimized for it, but it seems one would be wrong. That's why I bothered milling the frame. I'll do a comparo for the EMD sound once they get here.

Thanks again, Otto

[Cajonpassfan]

Arghh. Duh!
I just realized I did order some Knowles Dumbos a while back. So I built up an enclosure equivalent in size to the ESU 16x25 standard speaker enclosure. Here's the results, using the ESU EMD 1st generation decoder:


Well, it seems I'm unable to upload the two clips, but the difference is even more dramatic with diesel sound. Wow!
Now if I could only figure out how to upload the clips...
Otto

[peteski]

Thanks John, and Pete, for your comments; anytime you two agree on something I figure I better pay attention

One would think the speaker that comes with the excellent ESU decoders would be optimized for it, but it seems one would be wrong. That's why I bothered milling the frame. I'll do a comparo for the EMD sound once they get here.

I'm always happy to be able to gove some advice to a fellow model RR audiophile (even if that means agreeing with John).   

As far as the manufacturers optimizing speakers to match decoders goes, I don't think that they really do that.  It is more about what is available at the time and what is least expensive.  The Knowles-type speakers are fairly recent product and they are probably more expensive (on a wholesale level) than standard cone oval speakers (which are very old technology).  Plus the old-school speakers are easier to work with.  They have a well made mounting surface and solder pads for soldering leads.  Knowles-type speakers are designed to be installed in some sort of smart-phone assembly.  They do not have a well defined mounting surface or solder pads. They are usually just press-fit in the phone case and the leads touch metallic pads on a circuit board (solderless connection).

[jdcolombo]

The interesting thing is that now even ESU sells the cell-phone type speakers:

https://www.facebook.com/147086981998736/photos/a.156890394351728.28203.147086981998736/982810038426422/?type=1&theater

(I'm pretty sure that the speaker in the ad is the Knowles Donau).

The past few LokSound Select Micros that I have gotten didn't come with a speaker, which is just as well, since I threw about 30 of those things away . . .

John C.

[Santa Fe Guy]

Looks like they have been paying attention to all your great advice.
Rod.

[milw156]

Otto, if you install the ammeter between the DC throttle and the track it will measure the current draw of all the locos on the track.  The LED headlights add just few extra milliamps of draw (or you can unplug them for the test to just get the current consumption of the motors alone).  Most modern N scale motors (like Kato and Atlas) consume around 250mA at full throttle (12V) and under load (but not stalled).

I also like to comment on your choice of speakers.  The reason those small sugar-cube speakers (like Knowles) is because they are very efficient.  They have very softly-suspended and long-throw "cone" and a very strong magnet.  The older larger speakers (like the one shown in your photos) have fairly stiff suspension for their cone (so it doesn't deflect as easily).  If you could cram 2 Knowles speakers in the same space as the speaker you are using now (as long as you have a tight enclosure) you should end up with a sound superior to the large oval speaker you are planning to use.  If the decoder can handle 4 ohm speaker load, hook those speakers in parallel - otherwise hook them up in series. Just pay attention to the proper polarity or the sound from each speaker will be canceled by the other one.

it's a snug fit, but here's one way to get a pair of speakers (Knowles Donau's) into a Likfelike/Walthers Fairbanks Morse B unit and still keep the motor functional. and it is loud too!

[peteski]

it's a snug fit, but here's one way to get a pair of speakers (Knowles Donau's) into a Likfelike/Walthers Fairbanks Morse B unit and still keep the motor functional. and it is loud too!

Nice!!

[Cajonpassfan]

That is nice! Presumably, that should work for the FM Erie Builts, too? Did you have to mill or otherwise modify the frame to accommodate the speakers?
Otto