Tang Band speakers


Recalculating my measurements would mean I need to remove roughly 2.5mm of material rather than 4. In either case, I think it will be too much effort to do that for every hood locomotive that needs it. Too bad…f0 of 105 is much nicer than f0 of 150.

And wow…that monster in every F unit? Gonna be interesting to see that consist in person.


Dave F-

I made a near-field recording of the 1828SD module playing the ESU 567B/C 76311 sound file. The speaker was open air, held so the passive radiator was about 1/2" above the table top, module facing up, as it will be mounted in an F-unit with the passive firing down at the track. I did not simulate car body sound apertures through the dynamic brake side and roof grills.

This recording begs for a good set of external computer speakers for playback. I will need a lot of restraint to keep the volume turned down during operating sessions.:grinning: Anyway, it does show that much higher quality sound can be had in O-scale.




Thank you very much for the research and data. I just ordered several of the different sized modules to experiment with. I’m really looking forward to hearing the results.



I’ve been doing a lot of research on sound in model railroading lately, since watching these Tang Band speaker videos. I’ve been measuring frequency responses of different speakers, models, and scales - including the prototypes. Obviously we know that it is very difficult to get true low bass frequencies due to the small relative size of our models, but what does “bass” really mean?

I’ve analyzed the low frequencies of several prototype diesels and steam engines. In general most diesel prime movers have a low rumble right around 40-50 Hz. This is true bass that you can feel as well as hear when you are standing near the engine. Steam engines also have many different sounds that drop below 80Hz.

So what does this mean for our models? In general terms I have found a large discrepancy between “average” and “best” in the different scales. Example, it seems that most casual listeners make comments about “good low end” when they hear tones ~500Hz.

The interesting point about that is 1. 500Hz is obviously far off from what the prototypes are doing and 2. 500Hz is actually achievable from O all the way down to N. Factory speakers in O will get to ~400Hz pretty commonly. Factory HO speakers like Intermountain uses with LokSound decoders get to ~500Hz then roll off. Factory N scale cell phone type speakers, believe it or not, are not far off of HO in the 500-600Hz range.

What about “best” type install? I have been able to get down to and below ~100hz with Tang Band speaker modules in O scale, as have others. This is significantly better bass than many aftermarket installers are getting with “high bass” speakers. The best HO installs I’ve found have good response down to ~300Hz which is close to factory O installs. Those installs tend to use cell phone type speakers, which happens to be what the best N scale installs use as well. So I guess it shouldn’t be a surprise that the best N scale installs I found use two cell speakers and get down to ~350-400 Hz.

So what does this all mean? I think it shows that there is a lot of room for improvement in bass response in all scales over the factory speakers. It also means that if you can basically “move up a scale” by upgrading. I was most impressed by O with Tang Band and N with cell speakers. N can really be made to sound about as good as HO. Of course if you want the best relative sound O scale wins. There is a big difference between sub 100Hz and 300Hz on the low end.
I’m working hard to see how close to 50 Hz I can get :slight_smile:


Care to share some of your frequency response plots?


Sure Bob. These aren’t scientific results. They are taken with my iPhone 6 microphone, but they are useful for relative comparisons.

Smallest Tang Band module w/LokSound:

TB 2008:

I don’t have a chart for the larger TB module in the S4 yet. I will get one soon, but my ears tell me I’m on the right track :slight_smile:

Stock speaker:

With TB:


Jonathan -


Just a clarification here for the readers. The frequency response plots from Tang Band and those measured by Larry are just that. The speaker sound level is measured while being driven by a sine wave of constantly-increasing frequency, using an individually-calibrated measurement microphone.

Unless I missed something, the iPhone plot shows the sound pressure level while playing a decoder sound file. Because the sound file itself does not contain equal energy at each frequency (or is not a pink noise source) we see the product of the decoder’s sound file and the speaker’s frequency response. It is an apples-orange comparison.

Regardless, your measurement is quite interesting, and the sound comparison “speaks” for itself. I’m pretty sure you would enjoy the sound even more if you disconnected the pancake OEM speaker because its midrange resonance can still be heard.

Was that the 2008 or 1925 module?


Hi Bob,

Here is my latest install of a 2008 into an S scale SW1:


Fabulous :slight_smile:


I can’t wait to start working this into my Santa Fe PA’s of course with Alco appropriate audio files… Bob… as usual. Just incredible :blush:



Erik; ATS shoe on that PA?


This is a 1949-early 50’s era detailed set #62L - A and #78L research by Gary Schrader and company imported in 2010 by Key Model Imports. They know Santa Fe. I weathered lightly and plan on following Bob’s lead on DCC and sound also with my L-105 and M68. Thanks for the reply.


Really. Good to know.
I was surprised because the ATSF installed ATS on the line through Raton in 1947, and I know that these frequented there, thus limiting them to 79mph without them.

I’ll still install them on my PAs!


Nice illustration/photos, Erik!

The PA should have plenty of room for a large speaker install unless the height from the rear drive shaft to the roof is limited. If it is a “tank drive” then no worries.

I haven’t been inside many O scale steam engines. There should be plenty of room for a gutsy TB speaker in the tender, but in O scale our ears can easily home in on a single speaker location. Stack talk is just that, from the stack. Is there room in those engines in the front of the boiler? Are the stacks open to the boiler, or sealed? Post some photos of the guts with dimensions.

ESU L V4 decoders have two 4 ohm speaker outputs, which could be used for a smaller driver in the smokebox and a larger one for bass in the tender.


Now Bob…we just have to figure out how to blow smoke at someone if they look down the stack for the speaker :slight_smile:


Craig, a one-time-only option for smoke is possible, but requires a new decoder every time you want to use it again! har har


Craig -

Or something! For some reason my mind wanders to an old 3 Stooges movie in which Curly orders clam chowder and finds out that the clams are “really fresh!”:grinning: Sometimes stem engines would burp and send a jet of water up the stack. Evidence of that could be seen on the UP steam shop drop-ceiling until new management removed the building’s insulation.

I haven’t opened up the N&W Mollie boilers just yet, but there may be a chance to install a second speaker in the smokebox to fire up the stack. The V4L decoder has two amplifier outputs (a stereo class D amplifier chip) so driving a second could be possible. Or the sound could fire rearward from the smokebox and exit around the gear tower. This would still move the sound stage a little bit forward.

ESU is one of the last decoder brands to support a cam, but their version uses tiny magnets and a Hall-effect sensor (no cam bounce!) I can forgive mis-quartering in HO because everything is small but in O scale the chuffs need to start at 0 degrees, not 45 degrees, and stay in perfect sync.

Frolin -

Been there, did that…


Hi, guys!

I wanted to share with you first that work is progressing on the Tang Band install in one of my F-units.
Here’s the fuel tank that had been machined to accept the T1-1931S.



Nice! Please share more photos as the install progresses. I’m sure you will love the sound.