hi, we have a 5” alu version in the works.Ok, thanks everyone. To clear up a few points and summarize then
Cabinets: 10L is as big as i can fit in the space i have available. I just happen to have these particular cabinet parts that fit that. if i were building from scratch they would still be 10L, maybe with slightly different proportions to accommodate a 6". In fact i could cram a 6" in these... just hadn't considered it wise. Maybe i will research this more as for some reason purifi don't do a metal cone in 5" and i did like the sound of the seas l12rcy/p (aluminum cone) units I made many years ago so there would be some advantage in going to the metal purify 6” beyond extra low end. FWIW they will be up against a wall, positioned either side of the monitor.
Bass.. well i said "as much bass as i can reasonably get" meaning reasonably within the other constraints. perhaps that was lost in translation out of my brain 😀 I'm not looking for room pounding here, i just want something accurate and low and exists
People are divided on whether 2-way is fine, my requirements aren't that extreme, a good quality mid-woofer and some eq/dsp at 1m will get me everything i need, vs 3 way (i'm counting everything with a sub as 3 way also) is really needed to get any bass without sacrificing mids. there may be some merit to this, even the purifi drivers come in different variants for mids and bass.
Other suggestions include fullrange 1-way (thanks but no thanks, i have zero experience of them and they seem like a can of worms, plus having all the existing mentioned drawbacks of 2 way) and cheap metal stamped drivers will be fine (again, thanks but no thanks, every stamped metal driver i have listened to has been a bit crap. Also I already have that right now, this is meant to be an upgrade.)
Tweeters appear to be a personal thing. People generally agree Bliesma are good but expensive, everything else there is disagreement about. which is interesting.
I might make up some test baffles out of scrap wood and do some experimenting there before committing wince I have heard good and not so good metal domes. miniDSP makes trying out crossover points easy 😀
I may do a build thread when it’s complete and see if I can get some meaningful measurements off them. Data on how Tufnol cabinets perform may be interesting.
cheers,
Lars@purifi
@lrisbo ooh that's interesting! I read the app note about using notch filters to cancel the cone resonance of the 6" aluminum cone. Am i right in thinking that if i wanted to run with an active crossover, I would still need a physical notch filter in front of the driver? there was a preference for series filters in the appnote to reduce current flow at the notch frequency, if that filtering were done prior to the amplifier, the amplifier would look like a low impedance t that frequency, which would result in more distortion if i'm understanding correctly?
yes, that is correct: a line level EQ notch would not have the same effect on the distortion. The high load impedance seen by the driver at the notch frequency reduces the current generated by the motor back EMF. This EMF contains distoriton from the motor generated at subharmonics of the notch frequency. The line level filter would not affect these subharmonic frequencies
If you have a passive crossover far enough from that frequency it will also produce a high impedance and have the same effect. And you need a notch in series of the driver - parallel notch doesn't work.
It's a good use for peaks that also show in the THD - not every peak does, but most.
It's a good use for peaks that also show in the THD - not every peak does, but most.
Just to be clear, this is the kind of notch filter that is needed.And you need a notch in series of the driver - parallel notch doesn't work.
With Notch filter at 10500 Hz.
I have heard this kind of filter is called a "parallel notch", I assume because the notch components are in parallel with each other. It goes in series with the driver.
this kind of filter is called a "parallel notch"
A parallel notch is the 3 paralleled components in parallel to the driver.
A series notch is the 3 paralleled components in series with the driver. This is what Jim illustrated.
dave
What is this simulation program?
That software is VituixCad.
@planet10 - my description is based on the VituixCad library of circuit elements. The library has a parallel notch as I have shown above. It also has a "series notch" which arranges the 3 components in series with each other. It seems to function as a low impedance load at the notch frequency. To my non-electrical-engineering frame of mind, this kind of notch filter makes the most sense if there is an upstream impedance between the filter and the amplifier.
In this example, the notch filter creates a 2 ohm path to ground at 10.5k. I needed a series resistance between the amp and the driver to create a voltage divider.
Dashed line without the notch filter
Dashed line without notch filter
So I don't really know what the exact terminology should be. I have often been confused by the terms "parallel notch" and "series notch", and I just go by the VituixCad way of naming things.
j.
Hello IamJF, could you show us a cross-sectional drawing of that special geometry?And when you add the special geometry of the T25A
As I don't own a laser or water jet cutter ... that's hard to do 🤓Hello IamJF, could you show us a cross-sectional drawing of that special geometry?
T25A (and T34A) has a very high dome shape which gives it better stability (higher dome break up) and wider off axis radiation as normal alu domes. It really stands out. BUT - this also causes reflections in a waveguide, they are pretty bad in normal WGs. You would need to design a special one for wide radiation - not sure if that makes a lot of sense.
... majority of the time I see the L‖C‖R->+ called a parallel notch filter in series with the driver. The notch frequency is ~5khz/sqrt(mH*uF).
In practice I've found it easier or more intuitive to apply than the other kind where R should be pre-fitted to actual driver impedance; although to be precise, the parallel notch R should be post-fitted as well, but could often be skipped. (@hifijim could you please compare removing the 16ohm etc.)
In practice I've found it easier or more intuitive to apply than the other kind where R should be pre-fitted to actual driver impedance; although to be precise, the parallel notch R should be post-fitted as well, but could often be skipped. (@hifijim could you please compare removing the 16ohm etc.)
Last edited:
Jim, when I started out this terminology was very confusing. Partly, because I can find the terminology used both ways! I did do a lot of research and can tell you what is the most common. Comments below...
This (above) is most commonly referred to as a PARALLEL NOTCH FILTER just like you stated and per:
Martin Colloms may have the best approach as he doesn't use either ambigous term, but rather the above is a "damped shunt resonant circuit" which of course shunts to ground (and is parallel to the driver) while the below is a "series resonant circuit" which is in series with the driver.
The below images is most commonly referred to as a SERIES NOTCH FILTER.
People can argue all they want about what they "should" be called, but all I'm stating is what they "are" called most commonly. Don't shoot the messenger.
Just to be clear, this is the kind of notch filter that is needed.
With Notch filter at 10500 Hz.
View attachment 1277112
I have heard this kind of filter is called a "parallel notch", I assume because the notch components are in parallel with each other. It goes in series with the driver.
This (above) is most commonly referred to as a PARALLEL NOTCH FILTER just like you stated and per:
- VituixCADs library
- Jeff Bagby's Passive Crossover Designer (PCD)
- Vance Dickason's Loudspeaker Design Cookbook
- Ray Alden Speaker Building 101
- A Google search looking at the first 10 to images
Martin Colloms may have the best approach as he doesn't use either ambigous term, but rather the above is a "damped shunt resonant circuit" which of course shunts to ground (and is parallel to the driver) while the below is a "series resonant circuit" which is in series with the driver.
The below images is most commonly referred to as a SERIES NOTCH FILTER.
People can argue all they want about what they "should" be called, but all I'm stating is what they "are" called most commonly. Don't shoot the messenger.
That software is VituixCad.
... my description is based on the VituixCad library of circuit elements. The library has a parallel notch as I have shown above. It also has a "series notch" which arranges the 3 components in series with each other.
View attachment 1277263
.
we could say a parallel LC tank in series with the load or a series LC tank in parallel (shunt)with the load
"Series notch filter" is used here for a notch filter in series of the DRIVER. How the notfilter itself looks like depends on the function you need - in this case 3 parallel components.
Vituix Cad probably calls it that way cause the programm don't knows where you put the 3 components - in series or parallel or totally funny somewhere. But electronic people call a filter in series to the device - SERIES X Filter. In our case - notch filter.
Vituix Cad probably calls it that way cause the programm don't knows where you put the 3 components - in series or parallel or totally funny somewhere. But electronic people call a filter in series to the device - SERIES X Filter. In our case - notch filter.
Thanks for all the clarification, but to me it sounds like there is still some ambiguity in the terminology. And that is OK. I just need to remember that when someone says "try a parallel notch filter", I should ask "in series with the driver?" to be certain we are on the same page.
That R in the notch filter sets the maximum impedance of the filter, so it sets how deep the filter cuts and the Q. Making the R from 16 Ohm to infinite causes the notch to be very high Q and about -20 dB.(@hifijim could you please compare removing the 16ohm etc.)
- Home
- Loudspeakers
- Multi-Way
- Driver recommendations for small high quality nearfield