The reality of a transformer

[SUNRAY]

Very easy to understand (albeit there are several different ways of 'looking at it'). However, unlike your microphone transformer scenario, in this case one usually will want to use an (impedance) 'matching transformer', since one wants to optimise power transfer to the speaker.

Fair enough but you are only talking about how one marks the transformer, not what it is doing (which always is 'matching' the source impedance to the speaker impedance, to optimise power transfer) - the only important thing being that the 'power' figures specified only apply to a particular voltage (which I will assume is 100V RMS at full power) (and which I might have expected would be indicated on the transformer!).

Once one knows the source voltage at a particular maximum power, it is arithmetically trivial to determine what source impedance it equates to, and hence what impedance ratio (hence also turns/voltage ratio) is required to 'match' the source impedance to any given speaker impedance.

Assuming 100V RMS at full power (and also assuming I can do arithmetically correct in the middle of the night!), the following tabulation shows that for the example you illustrated. In other words, marking the transformer with powers (with an assumed voltage) and speaker impedances is just one way of doing it. The same information would still all be there if the transformer were specified in terms of turns, voltage or impedance ratios - but perhaps some of those who use them would not be capable of understanding that?

View attachment 280533

Kind Regards, John

Apart from not actually 'matching' the impedance you are correct (although it can of course be used that way). It is commonly called, and even listed as, a matching transformer. However it's usually used to tap power off a line rather than 'match' the power, in exactly the same way as a mains transformer for an ELV 10W bulb.

100V output amplifiers are rarely run into a matched loudspeaker load.

I've blocked out the information in your chart that is not required or considered or useful in 99.99999999999% of applications of such a device and as you say; not even understood by a very high percentage of the installers of such a device.

Some of that information may come on a data sheet with the transformer or a device containing it

or may even be printed on a device containing one

But the reality is in the main the the only 'really useful' information is the power.

As you say the calculations are trivial and similar to some I do on a regular basis for bespoke situations

 

[JohnW2]

Apart from not actually 'matching' the impedance you are correct ....

I don't understand why you continue to resist the concept that the transformer is 'matching' the source impedance to the speaker's impedance because, if it is designed optimally, that's precisely what it does.

I obviously don't know the turns ratios of the transformer you illustrated. However if (as one would hope) it were designed for optimal power transfer to the speaker then, if one had a speaker connected to the appropriate secondary tap, then what one would see (could measure) at one tap of the primary of the transformer would be an impedance equal to the source impedance appropriate to that primary tap.

To give an example. If you connected a 4Ω speaker to the 4Ω secondary tap and then measured the primary impedance at the 10W tap, if the transformer were perfect and optimally designed, you would get an answer of about 1,000Ω - which is the source impedance implied for a source which provides 10W at 100V RMS. I don't see why you do not like calling that 'matching'.

Mind you, maybe I've been misunderstanding, since you go on to write ...

100V output amplifiers are rarely run into a matched loudspeaker load.

... which begs the question in my mind as to why one would want to ('unnecessarily') to feed a speaker in a sub-optimal fashion.

Out of interest (so I can determine to what extent such transformers result in 'mismatch') do you possibly have specs for any speaker transformers which, as well as indicating primary powers and speaker impedances, also indicate the turns (voltage) or impedance ratios for each combination of primary and secondary taps?

I've blocked out the information in your chart that is not required or considered or useful in 99.99999999999% of applications of such a device and as you say; not even understood by a very high percentage of the installers of such a device.

Fair enough but, as I implied, you could 'block out' different columns (including the two you have retained) without 'loss of information'. In fact, one only really needs one 'column' - either the turns/voltage ratio or impedance ratio to provide 'full information'.

As I think you agree, we are (now) only talking about how the transformer is marked/labelled.

The downside I see of marking the primary with 'powers' is that those figures are only correct for one input voltage (100V RMS in your examples). The transformer could obviously be used with any line voltage, but the marked powers would then be incorrect. As I said before, I'm surprised that does not seem to be made clear in the labelling of the transformer you illustrated.

Maybe I'm behind the times, but I have to say that I'm used to things like 'audio output transformers' being specced in terms of primary and secondary impedances (hence, by implication, impedance ratio). To take a (physically) tiny example, I must have used many dozens, if not hundreds, of LTT700s over a period of decades, and they are specced in terms odf impedances, even today :

Kind Regards, John

 

[SUNRAY]

I don't understand why you continue to resist the concept that the transformer is 'matching' the source impedance to the speaker's impedance because, if it is designed optimally, that's precisely what it does.

That's almost what this type of transformer does, ie it transforms the low impedance speaker impedance to a higher impedance

I obviously don't know the turns ratios of the transformer you illustrated. However if (as one would hope) it were designed for optimal power transfer to the speaker then, if one had a speaker connected to the appropriate secondary tap, then what one would see (could measure) at one tap of the primary of the transformer would be an impedance equal to the source impedance appropriate to that primary tap.

To give an example. If you connected a 4Ω speaker to the 4Ω secondary tap and then measured the primary impedance at the 10W tap, if the transformer were perfect and optimally designed, you would get an answer of about 1,000Ω - which is the source impedance implied for a source which provides 10W at 100V RMS. I don't see why you do not like calling that 'matching'.

Yes correct

Mind you, maybe I've been misunderstanding, since you go on to write ...

... which begs the question in my mind as to why one would want to ('unnecessarily') to feed a speaker in a sub-optimal fashion.

Out of interest (so I can determine to what extent such transformers result in 'mismatch') do you possibly have specs for any speaker transformers which, as well as indicating primary powers and speaker impedances, also indicate the turns (voltage) or impedance ratios for each combination of primary and secondary taps?

No I don't have any such spec's but I imagine they exist if I could be bothered to look on line.

Fair enough but, as I implied, you could 'block out' different columns (including the two you have retained) without 'loss of information'. In fact, one only really needs one 'column' - either the turns/voltage ratio or impedance ratio to provide 'full information'.

As I think you agree, we are (now) only talking about how the transformer is marked/labelled.

One could but it would be totally pointless to remove the useful information and replace it with something that would necessitate a lot of complex calculations for the designer/installer of a system

The downside I see of marking the primary with 'powers' is that those figures are only correct for one input voltage (100V RMS in your examples). The transformer could obviously be used with any line voltage, but the marked powers would then be incorrect. As I said before, I'm surprised that does not seem to be made clear in the labelling of the transformer you illustrated.

I dare say it's written on there somewhere, it may have been the wrong one to illustrate with perhaps this is a little better:

Maybe I'm behind the times, but I have to say that I'm used to things like 'audio output transformers' being specced in terms of primary and secondary impedances (hence, by implication, impedance ratio). To take a (physically) tiny example, I must have used many dozens, if not hundreds, of LTT700s over a period of decades, and they are specced in terms odf impedances, even today :

View attachment 280595

Kind Regards, John

Yes that is indeed a very standard device and spec, and very relevant to the designer of a circuit.


Here is a sound system in a pub:

Using the 4W transformers shown feeding 7 8Ω speakers how would you calculate the requirements?

Doing it the standard way with tranformers marked in power it's 3x4W + 2x2W + 2x1W = 18W

Chances are an 18W amplifier is hard to source so one would aim for say a

40W with built in tuner, that allows for some extra speakers to be added if required, especially for when the summer allows for people sititng outside.
Now it's easy to see these are not 'matching' anything, they are presenting 3x2.5KΩ, 2x5KΩ & 2x10KΩ total 555.6Ω to an amplifier rated at 250Ω and very far from matched.

In exactly the same way as if they were a selection of 4, 8 & 16Ω speakers each connected to a 4Ω transformer, although it's perfectly normal to still refer to them as 'matching transformer' and in exactly the same way as 4, 2 & 1W LED's running in individual 12V mains transformers but not a single person would ever think to use the term 'matching transformer' there.

For my PA hire work I have something like a dozen 120W amplifiers (amongst others) and it's very common to run a single 15W speaker with one to cover an area.

 

[JohnW2]

That's almost what this type of transformer does, ie it transforms the low impedance speaker impedance to a higher impedance

I'm glad that, at last, you agree - so why do you have so much of a problem with describing that as 'matching' the source impedance to the speaker impedance?

Yes correct

Glad you agree. So, again, why don't you like that being called 'matching' the 4Ω impedance of the speaker to the 1,000Ω impedance of the source?

No I don't have any such spec's but I imagine they exist if I could be bothered to look on line.

Fair enough - I'll look for myself (and 'report back') - I just thought you might have the information 'to hand'!

I dare say it's written on there somewhere, it may have been the wrong one to illustrate with perhaps this is a little better:

Yes, since it mentions 100V, that's certainly better. Is that marking intended for colour-blind people (and, I wonder, why is Blue abbreviated to "BUL"?!) ?

Yes that is indeed a very standard device and spec, and very relevant to the designer of a circuit.

Indeed, but why, in this case (but not others0, are you happy for it to be effectively specced in terms of impedance ration, with no mention of 'power'?

Here is a sound system in a pub: ....... Using the 4W transformers shown feeding 7 8Ω speakers how would you calculate the requirements? ... Doing it the standard way with tranformers marked in power it's 3x4W + 2x2W + 2x1W = 18W .....

You've moved the goalposts a long way to a situation which is far more complex and way beyond any of my experience. I'll have a think and will respond separately in due course.

Now it's easy to see these are not 'matching' anything ...

Prior to do my thinking, I would say that, IF the aim is to maximise power transfer, then they are matching 'something' - namely matching the source impedance to the combined effective impedance of all the loads (each load being a speaker plus a transformer) ... but, as above, 'watch this space'!

Kind Regards, John

 

[SUNRAY]

I haven't moved the goal post at all, it's the same device I've been describing all the way through, It is not designed to match the individual loudspeaker to the source impedance (although it can be used in that mode), it is intended to 'tap off' a specific amount of power (specified at maximum volume) from the speaker line. The ultimate aim could be described as matching the whole speaker array to the source impedance but as I mentioned before it is not a common occurance.

One Emercency Voice Evacuation system I worked on consisted of a stack of 32 480W amplifiers; ~15KW total load of speakers on 30 amps; ~3KW or 20%. That does feel exceptional for a permanent fixed system but I have little experience of EVAC systems in large shopping centres.
Generally I'd expect something like 60-80% but 95% is not unusual.

Ironically my limit of use of LT700's has been to fit a remote 8 or 15Ω loudspeaker from amateur radio kit.

 

[JohnW2]

I haven't moved the goal post at all, it's the same device I've been describing all the way through, ...

They might be 'the same devices', but now there are six of them wired in three different ways (hence resulting in unequal loads), all being fed from a single source. In that situation it's not even obvious what 'matching' would mean. The meaning of matching A to B is straight forward enough, but matching A to several (not identical) things would be a very different concept!

As I said, you've dragged me into very unfamiliar territory, so I'm going to have to do some thinking before I can attempt to say anything remotely intelligent about it!

Ironically my limit of use of LT700's has been to fit a remote 8 or 15Ω loudspeaker from amateur radio kit.

The LT700 (and its sister the LT44) have been around 'for ever', I think well over 50 years (I'm pretty sure I was using them in the 60s - and certainly in the early 70s). They are very small, very cheap and very versatile, and I have used them for countless purposes (as I said, at least 'many dozens' over the decades), many widely divorced from their primary 'intended' purpose. If I remember correctly, the LT700 has a turns (hence voltage) ratio of about 9.5+9.5 : 1 (hence impedance ratio of about 90+90 : 1) whilst the LT44 has a turns/voltage ratio of about 6 : 1+1 (impedance ratio about 36 : 1+1).

Kind Regards, John