2
\$\begingroup\$

I have a speaker switch which has been in continuous use since some time in the 1980s. It a QED SSU4. To my surprise, there are plenty of pictures on the net and even some for sale. Even more surprising, is that the prices are similar to what I paid new decades ago.

It connects up to 4 pairs of stereo speakers to one amp output. There are 4 nice rocker switches to turn each set off with no limitation e.g. all 4 off, all 4 on, and any other combination. (*)

I am interested in how it achieves this claim: "At no time will the load presented to the amplifier fall to less than 1/2 of the speakers in use. i.e. with 8 ohm speakers the minimum load will be 4 ohms which almost all amplifier are capable of driving." So, it is not simply connecting all in parallel. It is a passive unit (no power supply).

This seems to be true. It spent most of its life connected to a NAD 3020B which claims to cope with 4 Ohm but not 2 Ohm. At times, e.g. student parties, it was driven hard but never showed signs of failing. Now it is connected to a SMSL SA-50 which also seems happy but it is not run so loud any more as I am no longer a student.

I don't want to open it up and look as I may damage it. It is still doing an important job well.

I don't necessarily need to know how this specific switch is made but how could that function be achieved with 1980s technology?

(*) My maths says that is 16 combinations though the instructions claim only 15 and they are not omitting all off as all on and all off are explicitly listed.

Edit: further detail in response to comments.

Here is a link to a unit that looks just like mine. The packaging seems slightly different: https://www.canuckaudiomart.com/details/649374943-qed-4way-speaker-switching-unit-model-ssu4/images/1624887/.

As mentioned in comments, some combination of parallel and serial connections is a possibility but you might expect this to cause a considerable change in volume as the switches were flicked. There is a change but it is quite small. Nothing is currently connected to output 2 but 1, 3, and 4 are in use. No other set goes off at the wrong time as switch 2 is thrown. This suggests that output 2 is never in series with another.

\$\endgroup\$
6
  • 1
    \$\begingroup\$ A picture is worth a thousand words and a schematic even more. \$\endgroup\$
    – winny
    Apr 26, 2020 at 9:33
  • \$\begingroup\$ Google suggests that the insides might look like this i.warosu.org/data/diy/img/0009/56/1457185911393.jpg \$\endgroup\$
    – Jack B
    Apr 26, 2020 at 9:57
  • \$\begingroup\$ I haven't tried to follow those wires (if you were to turn that photo into a schematic and put it into your post, you'd increase the chance of getting a good answer) but it looks like it uses DPDT switches to build a combination of series and parallel connections. I expect than some combinations are up to 16 ohm. \$\endgroup\$
    – Jack B
    Apr 26, 2020 at 10:01
  • \$\begingroup\$ I am not sure of the easiest way to add my own images but I have added link to an existing image. Note that I am not necessarily asking how this specific unit was built but just how something with the same function could be built with consumer level 1980s technology. \$\endgroup\$
    – badjohn
    Apr 26, 2020 at 10:58
  • 1
    \$\begingroup\$ In case it is helpful, here is a photo of the QED SSU4 external labels: QED SSU4 label \$\endgroup\$
    – SSS SSS
    Apr 14, 2021 at 17:02

1 Answer 1

5
\$\begingroup\$

It has 1 and 2 in parallel with 3 and 4. Within each pair, it switches the speakers in one at a time or in series. There is a load resistor to keep the impedance at or below 220\$\Omega\$

I was curious enough to work this though, so I'm going to write an answer describing how I worked it out.

There are probably a number of ways you could wire something like this. The first clue is the specification that the impedance presented to the amplifier would be no lower than half the speaker impedance. This is a big clue that some of the time the speakers are in parallel. The second is that amplifiers put out quite a lot of power and this is a small box, so there is unlikely to be any resistors standing in for unused speakers or it would get very hot.

Google turns up some photos where someone has taken one apart. That is a big help. Cropped, and with labels added based on the other photos, we have: Photo of the inside of a QED SSU4, showing a tangle of wires Two things are immediately obvious:

  • The left and right channel are laid out exactly the same, and don't connect to each other anywhere. This is what you'd expect.
  • The wiring of speakers 1 & 2 is identical to the wiring of speakers 3 & 4 (except the resistors - I'll come back to them later), and connected in parallel. That tells us the first part of how this works.

These together mean we only have to bother tracing 1/4 of the wiring, which saves some time. Tracing out the wiring gives:

schematic

simulate this circuit – Schematic created using CircuitLab

Where I've drawn it out in all four combinations of switch positions, and highlighted the wires which are carrying current to the speakers. So that is the second half of how it works.

So What are the resistors for? Well when both switches are off, there is not path from amp+ to amp-, and not all amplifiers are happy driving an open circuit. They provide a 220\$\Omega\$ path through both switches to prevent that.

Why don't you hear a large change in volume from speaker 2 when you switch out speaker 1? Well, that's another question altogether. It will depends a bit how the amplifier reacts to changes to the load impedance, but also the human ear/brain combo has a very non-linear perception of loudness.

Finally, I though I would note down the impedance the amplifier sees with different switch positions (assuming all speakers are 8\$\Omega\$):

  • Any one speaker: 8\$\Omega\$
  • Speakers 1 & 2 or 3 & 4: 16\$\Omega\$
  • Any other pair of speakers: 4\$\Omega\$
  • Any three speakers: 5.3\$\Omega\$
  • All four speakers: 8 \$\Omega\$
\$\endgroup\$
3
  • \$\begingroup\$ My Googling found loads of images of the device but I did not find an image of it open; just poor Googling skills. So, it is simple yet ingenuous. Your guesses on the volume are probably correct. My typical use has just 3 sets of speakers attached. The main ones are in the room with the amp; these are connected to output 1. Two other rooms connect to 3 and 4. 2 is unused. This is because I had detected some drop in the level of set 1 when set 2 was connected. Since I am not usually in one of the other rooms when I change switch 3 or 4, the data is even less reliable than usual. \$\endgroup\$
    – badjohn
    Apr 26, 2020 at 16:00
  • \$\begingroup\$ It doesn't fit my memory but agrees with your analysis: turning the unused set 2 on kills set 1. Sets 3 and 4 may be on or off in any pattern without affecting set 1 (all quite common). The device was quite expensive for a box of wires with no active components but it seems that it was good quality. After more than 30 years it is still working fine with no noise when the switches are used. I don't normally immediately accept an answer but it is hard to see a better coming along. \$\endgroup\$
    – badjohn
    Apr 26, 2020 at 16:07
  • \$\begingroup\$ I just found this question while browsing for something else. Now I feel strongly compelled to build my own speaker switch box which I've been dreaming of for some 20 years now... (+1) \$\endgroup\$
    – Jahaziel
    Apr 14, 2021 at 17:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.