As Lojzek says, reverse the polarity and adjust the delay to get the deepest null, change back to the correct polarity and check. One uses reversed polarity and cancellation because it is a much more sensitive test than trying to gauge a slowly varying degree of flatness.
What you and Lozjek write makes a lot of sense but shouldn't this have already been taken care of in the original design? I also notice that the PMC designs using these drivers seem to have the same distance and alignment as the Prestige. I'm definitely going to try what you suggest though.
The original design uses a passive crossover. Different delays are usually introduced in passive crossover by switching the polarity of the drivers and fiddling about with different slopes and their different phase changes. The ability to simply dial in the required delay is one of the many advantages of (most) active crossovers.
I would also recommend using a notch filter to remove the peak at around 800Hz - 1kHz. I always have the same problem with the ATC mid caused by a peak in the drivers response and amplified by baffle diffraction. It greatly improves the overall sound. (Speakers no longer shouting at you)
Aha, this makes a lot of sense now. Would I be right in thinking that the delay required would be around the 0.5 mS region at this frequency, just roughly speaking? I am hoping to get one of Rod Elliot's designs for active analogue crossovers, I don't think this incorporates any sort of delay tailoring. Is there any way to easily implement the required delay in such a system, or do I really need to go to a digital system?
The time delay required is the difference in path length from the listening position to the acoustic centre of the two drivers divided by the speed of sound. The slope of a baffle and how deep the drivers are mounted will both affect the required delay.
All-pass filters can provide delay in both passive and active circuits. According to google Rod Elliot talks about them on his web pages but whether he provides them in a kit compatible with what you want to do I do not know.
DSP is lot more flexible than active analogue circuits but requires some knowledge to use effectively that can lie outside what some audio hobbyists want to play about with.
If I were to design an active filter, it would have been done digitally. You can see
on mini dsp web how they explain to time align the speaker units. Basically you
decide where you want your mic be placed and then either by trial and error
dial the rough value of wanted delay in or you measure it, for instance, like
this: https://app.box.com/shared/ouxjjsx0m8bs00cil5iq
Roughly the offset of 6,5" woofer and dome tweeter on a flat baffle would be
1" which translates to 0,074 ms. Different mic positions have different offsets.
Which brings me to another variable in the Prestige design. If the delay is altered by relative positioning of the drivers, am I right in thinking that a crossover designed for a sloping baffle will not work correctly (in this aspect of time alignment) with a baffle that is vertical? The reason I ask is that the original Prestige design had a sloping baffle and the new boxes do not.
What new boxes? Mine bought last year and the Prestiges in their showroom and on their Website all have sloping front baffles
Hi
I use an LCR series notch filter (5mH + 6uF + 9 ohms) connected in parallel with the mid driver. This provides a notch at around 850Hz. You can adjust the R to increase or decrease the notch level. The R value includes the resistance of the inductor so you can use a low cost inductor because the R is high. (e.g. If the inductor measures 2ohms then use a 7ohm resistor to make a total of 9ohms etc..) (Typical value for Rtotal is 8 to 15ohms - the higher the R the less notch effect)
This circuit should also work for the 16ohm driver. (Typical value for R is 8 to 15ohms)
It seems that they tilt back by 6.5 degrees (for the earlier version anyway). I hope this means that the time alignment was done in this way by the designer and therefore doesn't need further adjustment. I can't think of another speaker, using these drivers, that has a sloping baffle.
There was an original design which I've never seen so don't know if that was flat. Otherwise the later design has the sloping front and angled sides.
That's the nub of my question: what is the supplied crossover doing in terms of crossover points and time alignment.
My speakers sound much more natural with all three drivers in phase
Had a reply from WA. Dropped them in last week as posted. They were very kind and helpful during my visit. At least they took the time to look at them for me:
Hi,
We have checked them thoroughly and I can confirm that they are made correctly and the circuits test as they should.
As mentioned, the crossovers are set at 380Hz and 3800Hz and as this is a tried and tested design, there is no reason that we can see, why it should not work prefectly well - assuming that the drive units are wired correctly.
So do you want us to send them back to you ?
Unfortunately, given I have tried these passive crossovers with both speakers and different driver configurations it appears there is no solution to the terrible freq response (I suspect a consequence of using a crossover designed for a 16ohm ATC driver with a different 8ohm volt driver).
So I will be moving on to a digital system for this project. The analogue active system is OK, but as I have previously said, given the challenging freq response of these drivers and issues with time alignment I believe the best outcome will likely be from a high quality digital implementation.
Looking at xilica crossovers as they have digital input and have excellent DACs and analogue outputs. They are highly thought of in the pro-industry.
I will likely be posting no more in this thread as I will move my questions to the digital line level forum.
Hi,
We have checked them thoroughly and I can confirm that they are made correctly and the circuits test as they should.
As mentioned, the crossovers are set at 380Hz and 3800Hz and as this is a tried and tested design, there is no reason that we can see, why it should not work prefectly well - assuming that the drive units are wired correctly.
So do you want us to send them back to you ?
Unfortunately, given I have tried these passive crossovers with both speakers and different driver configurations it appears there is no solution to the terrible freq response (I suspect a consequence of using a crossover designed for a 16ohm ATC driver with a different 8ohm volt driver).
So I will be moving on to a digital system for this project. The analogue active system is OK, but as I have previously said, given the challenging freq response of these drivers and issues with time alignment I believe the best outcome will likely be from a high quality digital implementation.
Looking at xilica crossovers as they have digital input and have excellent DACs and analogue outputs. They are highly thought of in the pro-industry.
I will likely be posting no more in this thread as I will move my questions to the digital line level forum.
Last edited:
Can you post a link to the continuation of your project using the digital crossover so it's easy to find? Or post the title of your new thread?
That is not an acceptable response from a company that has supplied a faulty crossover. The mid/tweeter crossover is clearly wrong and to claim the woofer/mid crossover frequency should be 380 Hz is obviously incorrect for a midrange driver with an Fs of 400 Hz. At least the actual hardware has a correct looking woofer/mid crossover at 500 Hz where it should be. I don't know what is going on at Wilmslow but it doesn't look healthy.
The fix for the crossover is almost certainly just to use the correct component values in the mid/tweeter section. This may require the replacement of only a single component depending on what the fault actually is and since it is the mid/tweeter section the components are not going to be expensive.
I've already said what is wrong with the original KatieandDad and bushmeister crossover. And how to fix it.
http://www.diyaudio.com/forums/multi-way/210627-wilmslow-audio-prestige-platinum-59.html#post4307163
You only have to look at bushmeisters +++ measurements to see a phase problem and a slope problem in the mid to tweeter transition. Really quite simple, IMO. But waffle on by all means. Me, I'd fix it.
http://www.diyaudio.com/forums/multi-way/210627-wilmslow-audio-prestige-platinum-59.html#post4307163
You only have to look at bushmeisters +++ measurements to see a phase problem and a slope problem in the mid to tweeter transition. Really quite simple, IMO. But waffle on by all means. Me, I'd fix it.
Attachments
I've already said what is wrong with the original KatieandDad and bushmeister crossover. And how to fix it.
http://www.diyaudio.com/forums/multi-way/210627-wilmslow-audio-prestige-platinum-59.html#post4307163
You only have to look at bushmeisters +++ measurements to see a phase problem and a slope problem in the mid to tweeter transition. Really quite simple, IMO. But waffle on by all means. Me, I'd fix it.
Shouldn't Wilmslow fix it?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.