How do I Increase Speaker Impedance

[LMSFan72]

Hi. I have been severely modifying my class 37 to fit an EM2 speaker. This is very large at 65mmx29mmx14mm so having cut away and re-strengthened the chassis there isn’t much space left! Now, the decoder requires 8ohm impedance and the speaker is 4 ohm, 2W peak and 1W RMS. Therefore, I want to increase the impedance so I don’t overload the decoder. Normally I would have added another 4ohm speaker in series but I can’t find one small enough to fit! Therefore, I wondered how else to increase the impedance. Can I use a resistor circuit, and if so what power rating and resistance would work? Many thanks

[Rog RJ]

You could use a 4 ohm 1 watt resistor in series but this will reduce the volume.  The best thing to do would be to get a correctly rated speaker.

[LMSFan72]

Thanks Rog. The whole point is to get the bass output, so reducing the volume may not be an issue. I have another decoder rated for this speaker and the sound is excellent but I am intrigued what this top notch speaker will do to a TTS chip. Reading on the web I am pretty sure there is a circuit that has one resistor in parallel and one in series that will work and that will also tune out noise.....maybe! Just need to work out values..

[Fishmanoz]

Don’t think any series/parallel is going to work for you.  I’ll let someone else give the formula to figure out multiple values of series parallel that will give you 8 ohms but the simple one is a 12ohm in series and 16ohm in parallel.  Now you have only a quarter of the volts on the speaker and half the current which sounds like only  one eight of the power from the amplifier getting to the speaker. 

 

Can‘t think of any other simple way apart from the 4ohms in series.

 

PS.  The formula for impedance using a resistor in parallel Rp and series Rs to give a combined impedance of 8ohms is

8 = 1/(1/Rp + 1/(Rs +4)). This of course has multiple solutions for various values of Rp and Rs as we have 2 unknowns and only 1 equation. 

[LMSFan72]

Thank you Fishy. After Rog suggested it I ordered a couple of 4ohm 1W resistors to try. I am going to compare the resistor with it to it with another smaller 4ohm speaker in series witth it with the body off to see how different it is..... failing that I'll pop the zimo decoder in.

[Chrissaf]

In pure physics, energy cannot be created, only converted. If one can create energy, then you will have discovered the 'Holy Grail' of Nuclear Fusion (as opposed to Nuclear Fission, as used in bombs and power stations).

.

OK, so if energy can only be converted from one form to another, then consider how this applies to putting a 4 Ohm resistor in series with a 4 Ohm speaker on the output of the decoder.

.

Let's say that the decoder amplifier is capable of delivering 1 watt of power into 8 ohms, and your 8 ohms (very simplisticly) is made up from a 4 ohm speaker and a 4 ohm resistor. I say very simplisticly, because the speaker is an inductive load and has an Impedance of 4 ohms and Impedance is NOT the same as the resistance of a resistor. Impedance varies with frequency, resistance does not.

.

The quoted 4 Ohms is a nominal value, usually measured at a specific frequency in the speaker supported range. But the 4 ohm resistor is a fixed ohmic value throughout the whole dynamic frequency range of the Amplifier output. Thus the volume will vary with frequency because the power split between the resistor and the speaker will vary too, due to the changing speaker Impedance. But ignoring the complicated physics of this, let's just assume a basic equivalent resistive circuit.

.

So the decoder Amp is pushing out 1 watt of power. Half this power will be dissipated in the speaker coils and converted into sound energy. The other half of the power will be dissipated in the resistor and converted into heat energy, and since you can't hear heat that potential speaker energy is completely wasted and lost.

.

The speaker volume is therefore at best halved, but due to the complications of mixing resistance (the resistor) and Impedance (the speaker) in the same circuit. The actual volume of the speaker may potentially be more impaired than just loosing half volume.

.

Fishy is correct, a serial / parallel resistor circuit will only make matters worse as more resistors mean more heat dissipation, and more resistor heat losses means less power is available for the speaker to convert into sound energy. Remember in physics, you can only convert energy from one form to another, you cannot create OR destroy energy. The Universe 'Bang Bang' 46 Billion years ago excluded. But then based upon current scientific knowledge, the 'Big Bang' does not easily fit in with current physics theory.

.

In an ideal world, it is always best to match the speaker Impedance to the Impedance required by the Amplifier.

[96RAF]

Early TTS decoders had a 4-ohm speaker but it was just too loud even on master volume setting 1 and to protect the amp I suppose this was changed to the current 8-ohm speaker. If you could identify the TTS amp chip you could look up its OEM data sheet to find out if a 4-ohm speaker would be outside its handling parameters.

My ESU decoder test rig has a small built in speaker with a switch that can be set to OFF (i.e. use decoder speaker) or 16 or 100-ohms. I presume this is achieved purely by way of shunting resistors into circuit.

Rob

[2e0dtoeric]

I've just seen this thread, and as Chrissaf put, adding a resistor in series will more than half the audio volume by converting the audio amp output to heat. As the 'now' heater is going to be in a tiny enclosed space, it will not be able to dissipate that heat, so will get hotter, and hotter - - - - -

[Chrissaf]

I presume this is achieved purely by way of shunting resistors into circuit.

.

Not necessarily.

.

Impedance can be converted up and down via a wire coil transformer. I used to know all the formulas for calculating the winding coil ratios to Impedance conversion, but now long forgotten (40 years ago since I last needed to know them). I really do regret throwing out all my apprentice training course notebooks and manuals that had all this information in them. In my industry, we used to use Audio Transformers that had Impedance transforming ratios of 1:1 (for line isolation) and 1:2 (to convert 600 ohm Impedance audio transmission lines to 1200 ohm Impedance).

.

The transformers were designed to have minimal 'Insertion Loss', but even so some power would be lost due to heat generated in the winding coils of the transformer.

.

Not really a suitable solution for the OP, because 1) he has already stated that he has no room for a second speaker, so minimal room for an Impedance transformer, and 2) these devices are usually custom made for specific manufacturers product requirements, so would be difficult, if not impossible, to source as a purchasable component.

.

The ESU board could potentially have a small audio transformer on the PCB performing switch selectable coil ratios to provide different speaker Impedance matching with minimal power loss.

.

EDIT: However that said, just had a look at my ESU 53900 (later model to Rob's). There is no visible audio transformer on the board, and the speaker selection switch look to have tracks that lead to some SMD resistors. I would suspect that the speaker is natively a 100 Ohm speaker and that switched resistors 'shunt' the speaker to present a lower ohmic value to the decoder Amplifier, just as Rob surmised. This is completely different to putting a resistor in series with a low Impedance speaker to increase the ohmic value. Using resistor 'shunts' still isn't very efficient from a power point of view, but much better than trying to increase speaker Impedance.

[Fishmanoz]

Given we have a speaker coil (largely inductive), I chose to use the term impedance rather than resistance in my post but chose to ignore the varying effect with frequency, and the phase shift for that matter. To a reasonable first order, you’ll get half the sound power out of the 4ohms in series solution. And given that TTS are reputed to be overly loud if anything, that 3dB loss in power is unlikely to be a problem.

 

Running 2 x 4ohm speakers in series will restore the volume as all energy goes to sound except for losses.  However, do make sure you have them in phase so both speaker cones are moving in the same direction not opposite directions (unless you mount them half a wavelength apart, but let’s not go there).  But remember that actual sound performance will be highly dependent on mounting arrangements, with unsealed mounting being extremely inferior to sealed ones with drivers this tiny. 

[LMSFan72]

Well, thanks everyone for both informative answers and different suggestions. I think I may well have found enough room to "squeeze" in a second speaker with a cut-down sound chamber.. I will keep experimenting..!

[2e0dtoeric]

It may be an idea to invest in a cheap audio amp from e-bay or somewhere, to experiment with. Then if/when you blow it up, you haven't lost an expensive decoder!

Find a solution that works for you and THEN worry about getting it into the available space! 

[96RAF]

I have the manual for an old ZTC sound decoder and it instructs one to ‘trim’ the round speaker to fit the loco.

The diagram shows this to be by way of clipping off the sides of the speaker frame until they are within the loco limits.

Goodness knows what this does to the speaker performance, as no mention is made of being careful not to cut into the speaker cone, although from the diagram this must have happened.

Not something I would recommend. 

Rob

[Fishmanoz]

Given that theoretical optimum performance of a speaker can only be achieved by mounting it sealed into a baffle of infinite dimensions (so as to eliminate the possibility of the sound pressure waves just running around from the front to the back of the speaker cone and cancelling out), it is a wonder drivers used in locos ever produce any sound at all.  That then suggests the ZTC speaker after clipping is going to be approaching theoretical silence Rob.

 

PS.  Which reminds me of the old 1st April story about the person, worried about airport noise, who invented a spray that cancelled out pressure differences in air, thus completely dampening sound waves.  They tested it at Mascot in Sydney.  It worked extremely well.  They sprayed it over the runway and the next aircraft to take off ran completely silently down the runway and straight into Botany Bay without leaving the ground.  They shelved the idea.