Ns1000 attenuators output

Hi-Fi discussion
Post Reply
karma67
Posts: 222
Joined: Sat Sep 08, 2018 12:19 pm

Ns1000 attenuators output

Post by karma67 » Mon Apr 29, 2019 1:52 pm

After the posts about the mid and tweeter pots from Paul I thought I’d check mine out, I’ve been running them as he suggested, tweeters at about a 16th before normal and mids at an 8th below -3db, I like the sound so I thought I’d buy a dB meter to check.

Does this make sense, I turn all pots off and then set 1 mid to how I like them and then play a test tone making a note of the output and then bring the other mid up to match, the dB meter is resting on the wire grills. This seems like a good idea to match them both but I may be taking cobblers, what do you reckon?

sq225917
Posts: 999
Joined: Tue Jan 22, 2013 9:28 pm

Re: Ns1000 attenuators output

Post by sq225917 » Mon Apr 29, 2019 7:27 pm

All measurements at 1m from each speaker or at seated position.

Probably better trying a UMIK and REW, how PC savvy are you?
One more Paradise left to complete.

YNWaN
Site Admin
Posts: 3940
Joined: Wed Apr 27, 2011 2:02 pm

Re: Ns1000 attenuators output

Post by YNWaN » Mon Apr 29, 2019 7:58 pm

karma67 wrote:
Mon Apr 29, 2019 1:52 pm
This seems like a good idea to match them both but I may be taking cobblers, what do you reckon?
I guess you could do it like that. I think I would want to pick a test frequency that is near central to the frequency range covered by each of the drive units 1050 Hz for the mid and 8 or 9000 Hz for the tweeter. However, as Si implies, it's bound to be a touch crude.
Mark

karma67
Posts: 222
Joined: Sat Sep 08, 2018 12:19 pm

Re: Ns1000 attenuators output

Post by karma67 » Tue Apr 30, 2019 4:53 pm

cheers,yes crude but it seems to work reasonable well,or should i say better than not using one lol

ChrisOH1
Posts: 147
Joined: Tue Jan 02, 2018 9:58 am

Re: Ns1000 attenuators output

Post by ChrisOH1 » Thu May 02, 2019 12:34 pm

I had concerns when Paul was bulding the external cross-overs for mine that i would be losing the use of the L-Pads. I had not been a serial tweeker and what Paul found when he measured them probably explained why, as they were not really working well ie you needed to be below -2dB before really having much effect. I generally had them set at -2dB and left them alone. What he also found was that the L-Pads were shifting the cross-over point and interferning with phasing which in turn was messing up the balance of the sound.

I have limited experience of listening to lots of different speakers as a frame of reference but, fixing the cross-over and removing the L-Pads from the circuit (the dials are still there) was no loss as far as i can hear. I think the speakers imaging is dramatically better than before.

If I still had them and knew what i know now, i would be balancing them as you're proposing, ignoring what the number on the dial says as it may not be what is actually happening. In fact, i wish i had an unmodified pair to make comparisons to. Now there's an idea

Regards

Chris

reffc
Posts: 94
Joined: Thu Aug 30, 2018 8:40 pm

Re: Ns1000 attenuators output

Post by reffc » Fri May 03, 2019 3:56 pm

Unfortunately, balancing pots the way that's suggested won''t work properly.
Besides the driver overlap making meaningful driver spl comparisons nigh on impossible, an spl meter picks up room power response. The extended time domain of using test tones (they're really of very limited use) introduces all the reflections and room ringing that you don't really want from measurement, not just direct driver response, even from 1m. Even if you placed the meter at 1mm from the driver to gate out much of the room response, it tells you nothing about phase.

Half the issue with the attenuators is that they, strictly speaking, are not L-pads. The crossover, frequency, hence phase response, alters across the attenuators' level settings.

You really need to look at the driver balancing act with respect to the filter design and electro-acoustic response, not just the acoustic response. The impedance load presented by the driver plus filter varies with frequency (rising response) so to have the crossover point maintained, the impedance load of the filter, as seen by the driving amp, must remain constant at the crossover frequency, which is the whole point of an L-Pad. It maintains constant impedance so if for example your drive unit impedance at crossover is 8 Ohms (it's not but bear with it) whatever damping you apply must maintain that 8 Ohm load seen by the amp, otherwise a change in total impedance (driver plus resistor network) changes the effect of the filter by shifting the crossover point and ruining phase accuracy. There is a mathematical relationship between driver impedance at any given point, the amount of attenuation required at that point and the specific values of the L-Pad to maintain that crossover point and constant impedance.

If for example your driver impedance is 8 ohms at crossover and you want a 5dB attenuation to match the driver being crossed with, you would require a 3.5 Ohm series resistor after the filter and a 10 ohm shunt resistor. Using the relationship of the sum of paralleled resistive loads, you have total shunt impedance = 8 Ohms and 10 ohms in parallel...this equates to a summed 4.44 Ohm impedance. Add the series resistor value and you have 7.94 Ohms....ie, close on 8 Ohms (close enough).

The shape of the attenuation is dictated by the series resistor (ie how steep the attenuation curve) and the value of the step is dictated by the parallel resistor. Together, they offer the correct way to shape and damp output for the best fit response.

Hopefully, it's a little clearer (than mud?) now why spl meters aren't the answer as the starting point is an incorrect variable L-pad design (in this case).

karma67
Posts: 222
Joined: Sat Sep 08, 2018 12:19 pm

Re: Ns1000 attenuators output

Post by karma67 » Fri May 03, 2019 4:55 pm

slightly over my head sorry.
but in basic terms without paying someone like your good self to do the crossovers mods and working with what i have,ie a standard speaker surely to get both speaker drivers outputting the same using a db meter hasn't it got to be better than just setting them to approximate levels?
what you stated before about where to have the pots set is a big improvement (to my ears) than where i had them before,which was hf at half way between -3db and normal amd mids at 16th of a turn below normal. so much so i cant listen to them now like that,the mid setting at an 8th of a turn below -3db has the biggest impact on the over all presentation.

reffc
Posts: 94
Joined: Thu Aug 30, 2018 8:40 pm

Re: Ns1000 attenuators output

Post by reffc » Fri May 03, 2019 6:07 pm

The best bet if keeping crossovers as-is, would be a slight mod to replace attenuators with L pads of the values I provided in the build thread really. Failing that, for all the reasons above, it's really a waste of time and money to try and use spl meters to get things optimal, because as the design stands, it's more a case of making things "less worse" as it can't address phase or the shape of the response very well. You ought to be able to approximate the "less worse" position by ear without the meter. I think that the attenuator levels recommended are all that's really needed until such time as the bullet's bitten (or not) regarding proper L-pads. Hope that makes sense.

karma67
Posts: 222
Joined: Sat Sep 08, 2018 12:19 pm

Re: Ns1000 attenuators output

Post by karma67 » Fri May 03, 2019 6:31 pm

yep,thanks for explaining,id better start saving then..... :)

YNWaN
Site Admin
Posts: 3940
Joined: Wed Apr 27, 2011 2:02 pm

Re: Ns1000 attenuators output

Post by YNWaN » Fri May 03, 2019 9:54 pm

If you find yourself able to visit Sheffield you are welcome to hear my modified Yamahas - this should give you some idea of what to expect. I know that Roger (AKA Mr TigerPaw) was kind enough to say he found the experience ‘ear opening’ (in a good way I think ;))
Mark

Post Reply