A retrospective on bad 2000s electronics

I grew up with electronic products from 25-30 years ago, which had a very distinctive aestethic and flavor of bad. Today, they are old enough to start attracting interest and collectors, so it's worth having a look back to remind ourselves what we are dealing with.

A short time ago, I tried using a set of 2.1 PC speakers that had spent their whole life in the retail box because of a broken cable. This had been forced by a temporary issue with my usual 2.1 Creative system, which is at least as old and was not much more expensive. I thought I was not picky about amplifiers, but oh my it was bad. Not even acceptable, just awful! The distortion was very audible with any combination of knobs, everything sounded far and indistinct, and I couldn't even reach an acceptable volume. I'm the kind of person who almost walks around with earplugs, so I don't think my volume needs are that high. The desire to find out how things could have gone so wrong led me to this blog and the work behind it.

Enter the Trust 2500P SOUNDFORCE

The speaker system in question was made by Trust. It seems like they are still around, but, at least in that period, everything they made was cheap and bad. This 2.1 is no exception. Two monitor speakers attach with a 3.5 mm plug to a modestly sized subwoofer and amplifier unit, from which connect the mains cable and a pea-shaped wired remote featuring the input cable, a power switch, and two tiny pots for volume and bass.

These are the detachable monitor speakers. So futuristic, yet so bad! I even checked and the tweeters are not fake.

My main curiosity was related to the placement of those controls in the circuit, which are basically a master volume and a subwoofer volume. Maybe a small fuckup ruined the headroom for the whole system. The teardown involved mostly extracting the circuit board from the subwoofer box, which revealed a 30W speaker inside. That seemed like too much, but I will find out why.

The circuit an the mains transformer. No capacitor plague here!

Shielding! Look, shielding!

Since there's not much to say about build quality let's get to the circuit.

Schematic

The trace above shows the spirit of the circuit, skipping over some details like the "phones" switch being part of the "phones" output jack and that there is actually some care put into separating signal and power grounds and decoupling. The only electronic components not shown would just be a couple of small electrolytic capacitors acting as high-pass for the vestigial tweeters. Overall, it doesn't look terribly wrong for a way to do this job on the cheap, using just two power amplifier ICs that will do, respectively, 1 and 4W on a good day with copious amounts of distortion. The input is fed to the first, stereo, one which powers the monitors and headphones output, then summed to mono to feed the subwoofer amplifier after some haphazard low-pass filtering.

My concerns about headroom don't seem to be supported. Both amplifiers follow an "attenuate first, gain later" design which is bad for noise but not for headroom. It is true that both amplifiers probably don't need as much gain as they have (about 40db and 45db respectively; they're equivalent to a noninverting op-amp with an internal 10k feedback resistor), but my main issues are neither noise nor subwoofer distortion; woofer/tweeter distortion seems to be completely controllable with the volume pot, so I guess the culprit is a combination of underpowered amplifier and inefficient speakers.

That said, there are two things that give me pause. The first is the filtering network right after the input. I triple-checked this, and it looks like you lose 2db at 10kHz already, because the series capacitor does not compensate fully for the capacitor to ground. Removing both is something worth trying, but it won't help with the distortion, just with the dullness.

The second thing of note is that the subwoofer is powered in bridge mode but there are no output capacitors. I tallied the capacitors on the board to check this, but it appears to be the case. I guess a 30W speaker being asked to do at most 2 can probably handle it, but still.

Conclusion

I was not really surprised that I couldn't find an easy fix to make notoriously bad computer speakers good, but curiosity got the best of me. I guess, in the end, the only thing to blame is the optimism with which these things were put together.