audio-talk

IslandPink wrote: ↑Sat Feb 20, 2021 7:45 pm There's nothing on the Fostex trace around 300Hz, and nobody else seems to have measured it or at least posted a measurement. I did a couple of additional scans with the speaker pushed back further into the bay, and it's clear there's a lot going on in the room between 100Hz and 400Hz. Maybe there's a suckout from the room, where it was first positioned. Perhaps SimonC is out there lurking, with that model of the room that he ran up in the software ?

Yep, I'm still here. Not been doing anything on the audio front for the last few months so I've just been lurking in the background
Simon C

chris661 wrote: ↑Mon Feb 22, 2021 8:01 am The response curve of the horn looks nice to me - I've had good results with worse, but those needed some hammering into shape.
Could you check the distortion plots, and maybe response at the listening position, to see if we can figure out what's wrong?

Yes, i should do this. Give me 2 or 3 days.

vinylnvalves wrote: ↑Mon Feb 22, 2021 8:47 am Mark.. need to work back, what have you changed since the horns were sounding good? We’re you previously running then with passive crossovers? As you know the reflected loaded back thought the OPT’s changes the working line.

In the past I would generally have the ribbon tweeter on there as well. On paper it shouldn't affect how the horn sounds lower down, but maybe the extra finesse and harmonics higher up are masking the colouration.
It should be a good load line - I haven't applied any passive crossover to the horn - just run it open, moved from the chip amp ( test ) to the 45 amp ( listening ) , which is set up for 8 ohms.
The colouration is probably just because of running it open to the lower cutoff - but I hadn't noticed it as much before.
The 45 amp does have a high-pass of sorts - around -3dB at 175Hz due to the special output transformers - that will have a small effect at 500Hz, but that should be helpful.

I am also tempted to bring the Azurahorns down and try one with the other drive unit. I know they are a bit better at the lower end of the range, although I found them harsh if run up to 12 or 13kHz. I never tried a low-pass on them when combined with the ribbon. That might work - something around 4 or 5kHz.

The 45 amp does have a high-pass of sorts - around -3dB at 175Hz due to the special output transformers

If the transformers are being driven with a lower frequency signal, is it possible they are generating harmonics that will get involved.

Good point - I need to check the coupling cap/resistor.
When I ran it at Owston with the ribbon it was the same, but I didn't demo vocals then.

It's possible I've missed something, but it sounds like the compression driver is being run without a highpass filter, except whatever the OPTs are doing.

I'd strongly recommend adding the highpass before further evaluation. I'd also consider adding the ribbon - the rolled off HF won't be helping the vocal quality. That does make it a bit more difficult, though, to focus in on exactly what the horn is doing. In that case, I find switching that driver off and then on again (unplug the input from the amp, whatever) is a good way of refreshing.

Chris

Adding the ribbon is going to be a pain, since they both are integrated into my OB's for the time being.
However, I have managed to do another run here from just outside the horn mouth, and saved the distortion plot. I must admit I'm surprised at the levels of 3rd/5th harmonic, given this is probably no more than 90dB/1m from the horn. At this level, the 288 would be barely ticking over.
Do we know what the Behringer produces ?

The distortion on the interface should be vanishingly low. To test, plug the output into the input (loopback measurement) and see what you get.
You could also do a measurement with the amplifier in play. You'll likely need to engage the -20dB pad on the Behringer's input.

The harmonic distortion gets close to 100% at around 200Hz, so I wouldn't do any more sweeps without a highpass filter. The last time I saw distortion that high, I was hitting a 1" 8ohm compression driver with 70v RMS.

As a reference, the Behringer B3030a monitors I used to have here came in at a fraction of a percent THD across the range with a ~70dB (at LP, a few metres away) sweep.

Chris

I think he means the mike but still should be low.

No, I meant the test signal.

IslandPink wrote: ↑Mon Mar 01, 2021 11:42 am No, I meant the test signal.

Ohm ok, well thats down to whatever is generating the signal but I would expect at least -80dB levels of distortion.

Good. I can check the Radian 745 Be on the Azura to see if that's lower.

Mark,

Please heed my previous recommendation and implement a highpass filter. Harmonic distortion of that level indicates that something is very unhappy indeed, and you might be on the verge of breaking something.

I'd suggest doing a loop-through of the USB interface, and then a further test with the amp driving a dummy load.

That will help to track down the source of the distortion, and make sure everything is running properly before further acoustic tests.

I have seen numerous compression driver diaphragms shattered because they've been run without proper processing. High-power titanium units can put up with a bit of abuse, but aluminium diaphragms give up pretty quickly, and turn to (expensive) confetti.

Chris

Relax ! - I'm ahead of you on this one - already checked in 2014.
See comments in the middle of this page :
https://www.audio-talk.co.uk/phpBB3/vie ... &start=855

But I do need to check the tweeter SE amp and see if the bass is being taken out of the special output transformers. Will do that this evening.

edit : please bear in mind that the actually output level ( see earlier graph ) at 200Hz is very low.

OK, next point of importance - the distortion plot shown above was not done with the valve ( 45 ) amp - it was done with the 7297 chip-amp.
So whether the 45 amp's output transformers are saturating due to bass frequencies is of no relevance.

However in terms of listening to music, it did matter, and I checked the amp tonight. I had obviously made some consideration of this already - the connection from the drivers to the 45's is 0.0047uF/150k , which I believe is 225Hz high-pass. So the transformers should be working as intended.