audio-talk

simon wrote: ↑Sun May 23, 2021 11:27 pm
Is there any reason why I can't use a cheap 24V LED lighting SMPS with 2 no. 12k resistors between the output rails with the centre of the resistors tied to ground? I don't see why not but it's rather a while since I've played with magic smoke containing vessels so thought I'd ask before doing something I might regret.

Interesting question Simon.

I'd have gone for two 12v units tied together, but what do I know.

So, is this not supplying the treble or upper midrange ?

If this has been brought on by listening to my tt simon, dont think theres an issue with your speakers

pre65 wrote: ↑Sun May 23, 2021 11:42 pmI'd have gone for two 12v units tied together, but what do I know.

I don't have two or even one 12V units so it's rather academic

IslandPink wrote: ↑Sun May 23, 2021 11:48 pm So, is this not supplying the treble or upper midrange ?

Not sure what the question is Mark. At the moment I have cheap eBay class D modules with 180Hz low pass filters built in for the SB subs, which "handover" to the 208s with GM70s. And another cheap class D for the supertweeters, for now.

I need the active crossovers to limit the SB subs to around 65Hz so I can bring the 285s in above them up to somewhere between 400Hz and 600Hz, then bring the 208s in.

Ant wrote: ↑Mon May 24, 2021 6:37 am If this has been brought on by listening to my tt simon, dont think theres an issue with your speakers

no not listened to it yet Ant. I just need to get my finger out!

Personally I would avoid the LED supplies, I would expect they would not be noise free. There are circuits on the net to create an active centre tap that would do better than a couple of resistors.

simon wrote: ↑Mon May 24, 2021 8:30 am

Ant wrote: ↑Mon May 24, 2021 6:37 am If this has been brought on by listening to my tt simon, dont think theres an issue with your speakers

no not listened to it yet Ant. I just need to get my finger out!

Easier said than done with your daily schedule !

simon wrote: ↑Mon May 24, 2021 8:28 am Not sure what the question is Mark.

It was along the lines of what Nick has subsequently said - re. quality/noise of the supply.

Okay thanks chaps, I'll not bother with the SMPS then. I've an overkill toroid with 2 15V secondaries that I could use with a first resistor for "choke" input, and drop some more voltage with a second filter stage if needed. I was just thinking of avoiding that with the SMPS.

By way of light relief from the 45 amp I'm well overdue an update on this.

Since building the speakers around 18 months ago I've been using cheap class D 3116 amps below 180Hz with built in low pass filters with the SB subs, then the 208Es above with the GM70 amp with internal highpass filter, and the T90As above 5kHz with another class D 3116 amp and a simple second order filter. The 285GMFs haven't been in use at all. And I've really enjoyed it. So much so that I haven't exactly got my finger out with the active crossovers to get the 285s running.

The 3 way Linkwitz-Riley boards are built and wired up (and yes, I did use a nasty LED lighting SMPS to get me going) and feed the 6 channels of 3116 class D amps. Probably around £40 to £50 all in for 6 channels including PS. They are on the face of it not bad at all, though there is some hiss. But they have that slightly unrelaxing quality that means they wouldn't stand the test of time. But they were only meant to get me going too, not a long term solution.

So I have:

SB subs with a low pass at 65Hz
285GMFs 65Hz to 415Hz
208EZ highpass as 415Hz with natural roll off around 5kHz
T90A with a simple second order passive filter off the 208s at 5kHz

I established before that as the room is quite small there are probably a lot of reflections and room issues in the plots, and it's clear I'm not a measurement engineer, but here are the averaged plots for the listening position and from 1m away. After a little listening to tune the sound in by my ears this is what I measured Initial listening was interesting. There have been a lot of changes so it's a little tricky to work out what's changed, but I've heard detail in familiar recordings I've never heard before, and that was from mp3s. And there's a proper weight to instruments like a grand piano.

I think the next job is to try with some decent valve amplification and see what happens then. Then perhaps think about some passive filters to replace the active crossovers. And maybe some room treatment.

I'm interested in people's thoughts on the plots, those far more experienced than me at looking at these things.

Thoughts:

- Good start overall. <20Hz LF extension is always nice to have.
- The raised 3.5kHz-8kHz region is probably responsible for the extra detail. Suspect the novelty will wear off on that.
- Odd that you seem to have lost some level >12kHz when everything's combined.

I'll have a bit of a ponder and come back later. For now, I've got some old mics to reassemble.

Chris

Thanks for your thoughts Chris. I suspect if I repeated the measurements they'd be different again - they did seem quite variable as I was doing them. The combined plot is quite different from the individual measurements taken at the listening position between 1k and 3k. Maybe I'll try again and see how they compare.

I'v e taken a load more measurements this afternoon. To start here's a comparison between REW's averaged plot based on 8 measurements against my own manual version of 9 measurements. I've noticed difference in the plots in the past sometimes.

The x8 REW average has a centre reading right in front of my face with 7 in a circle with c200mm radius. The x9 self average is a 3x3 grid c200mm between points with the middle point in front of my face.

All these measurements are of the left speaker only, taken from the listening position approx. 2m from the speaker.

First the SB sub, very little difference:
Next the 285GMF, some difference:
Next the 208EZ, more significant difference perhaps:
And finally the T90A. Or it would have been but it looks like I can upload max three attachments?

Here's the T90A, a greater difference again: