Hey everyone!
Little teaser intro - I changed the response of my Seas coax tweeter from the Red trace to the Green trace (not actual level, separated for clarity).
In 2022 I purchased the Seas MR18rex/rf which is a coaxial mid-range and tweeter. Sadly, I was disappointed by the performance of the tweeter. The on-axis response is horrible and you need to go a decent way off-axis before it smooths out. The on axis response is what you see in the red trace above. Rutcho found similar results*
So, I decided to have a go at making my own tweeter for it.
There are a limited number of tweeters small enough to fit inside the voice coil of the mid-woofer and I tried quite a few. I settled on the OX20SC00. It doesn’t fit as standard because its too deep and also unstable on the pole piece. Luckily I have a cnc milling machine, so I made a jig to hold the tweeter safely and modified the body.
I then took a casting of a section of the mid-woofer cone and carefully modelled the profile in CAD. This allowed me to draw up nine variations on interface parts between the tweeter dome and the cone. I 3D printed all nine and yesterday I tested each one in my anechoic chamber. One or two worked surprisingly well! big smiley face
The response is below. Traces top to bottom with 5dB separation:
0 degrees (absolutely dead on-axis)
5 degrees
30 degrees
60 degrees
Additionally, below is the distortion at ~90 dB* I applied a 1st order DSP filter at 4000Hz which produces an almost perfectly flat frequency response from 1KHz up.
As typical for large roll surround tweeters 2nd harmonic is slightly raised, but >3rd harmonic is very low. For a small tweeter, this thing could cross low when used in a waveguide like this. It brings to mind the Buchardt S400 with their 19mm dome in a 6” waveguide.
I’ve looked at all the coax drivers on the DIY market and this seems to be the best tweeter performance of any of them. Seas original, SB Acoustics, Tangband all have some nasty diffraction and dips going on in their tweeter response. The mid-range cone from Seas is excellent, of course, since that is their area of expertise.
So here is my question – If you are in to coaxials, would you be interested in this tweeter upgrade as a product? I was thinking about selling the modified tweeter and waveguide piece as a plug and play kit for the Seas MR18REX/RF. It would probably be about $50/piece.
If there is interest, I also considered talking to a distributor like Sound Imports. Seas told me they can sell the MR18 coaxial without the tweeter for a lower price but there is a prohibitive MOQ for someone like me. Working with a distributor who already has a relationship with Seas, they might be able to get a tweeter-less MR18 and offer it with my tweeter upgrade at a price similar to the original MR18 giving better value. However, I’d like to gauge interest.
Footnote:
I simulated the design in AKABAK. As can be seen in the measured response the traces bunch up a little between 5KHz-10KHz. This is due to the cone profile of the MR18. I have put some thought in to designing my own coaxial from the ground up so I can do an even better job, but again MOQ and non-refundable costs are prohibitive and scary for one guy like me. At least offering this upgraded MR18 could ‘get my feet wet’ and perhaps the future would hold more exciting things!
SEAS MR18REX/XF H1699-08/06 Loudspeaker Measurements and Information (rutcho.com)
*My SPL meter died. So I had to use my phone to set SPL for distortion measurment)
Little teaser intro - I changed the response of my Seas coax tweeter from the Red trace to the Green trace (not actual level, separated for clarity).
In 2022 I purchased the Seas MR18rex/rf which is a coaxial mid-range and tweeter. Sadly, I was disappointed by the performance of the tweeter. The on-axis response is horrible and you need to go a decent way off-axis before it smooths out. The on axis response is what you see in the red trace above. Rutcho found similar results*
So, I decided to have a go at making my own tweeter for it.
There are a limited number of tweeters small enough to fit inside the voice coil of the mid-woofer and I tried quite a few. I settled on the OX20SC00. It doesn’t fit as standard because its too deep and also unstable on the pole piece. Luckily I have a cnc milling machine, so I made a jig to hold the tweeter safely and modified the body.
I then took a casting of a section of the mid-woofer cone and carefully modelled the profile in CAD. This allowed me to draw up nine variations on interface parts between the tweeter dome and the cone. I 3D printed all nine and yesterday I tested each one in my anechoic chamber. One or two worked surprisingly well! big smiley face
The response is below. Traces top to bottom with 5dB separation:
0 degrees (absolutely dead on-axis)
5 degrees
30 degrees
60 degrees
Additionally, below is the distortion at ~90 dB* I applied a 1st order DSP filter at 4000Hz which produces an almost perfectly flat frequency response from 1KHz up.
As typical for large roll surround tweeters 2nd harmonic is slightly raised, but >3rd harmonic is very low. For a small tweeter, this thing could cross low when used in a waveguide like this. It brings to mind the Buchardt S400 with their 19mm dome in a 6” waveguide.
I’ve looked at all the coax drivers on the DIY market and this seems to be the best tweeter performance of any of them. Seas original, SB Acoustics, Tangband all have some nasty diffraction and dips going on in their tweeter response. The mid-range cone from Seas is excellent, of course, since that is their area of expertise.
So here is my question – If you are in to coaxials, would you be interested in this tweeter upgrade as a product? I was thinking about selling the modified tweeter and waveguide piece as a plug and play kit for the Seas MR18REX/RF. It would probably be about $50/piece.
If there is interest, I also considered talking to a distributor like Sound Imports. Seas told me they can sell the MR18 coaxial without the tweeter for a lower price but there is a prohibitive MOQ for someone like me. Working with a distributor who already has a relationship with Seas, they might be able to get a tweeter-less MR18 and offer it with my tweeter upgrade at a price similar to the original MR18 giving better value. However, I’d like to gauge interest.
Footnote:
I simulated the design in AKABAK. As can be seen in the measured response the traces bunch up a little between 5KHz-10KHz. This is due to the cone profile of the MR18. I have put some thought in to designing my own coaxial from the ground up so I can do an even better job, but again MOQ and non-refundable costs are prohibitive and scary for one guy like me. At least offering this upgraded MR18 could ‘get my feet wet’ and perhaps the future would hold more exciting things!
SEAS MR18REX/XF H1699-08/06 Loudspeaker Measurements and Information (rutcho.com)
*My SPL meter died. So I had to use my phone to set SPL for distortion measurment)
Last edited:
Furthermore, I attach the impedance response and FR files if you want to play. I also include a suggested passive filter that simulation says will give LR-24 at 2.3KHz.
Last edited:
Goes to show how badly most coaxial units are. Most are just an afterthought.
Great results 👍
It's not entirely clear to me, but the measurements are from the drivers with the green plastic bits on it and NOT just the driver with tweeter (2nd photo when holding in your hand)
Great results 👍
It's not entirely clear to me, but the measurements are from the drivers with the green plastic bits on it and NOT just the driver with tweeter (2nd photo when holding in your hand)
Yes absolutely, all measurements original and modified are taken in that test box. Unfortunately I killed the original tweeter so could not measure in my anechoic chamber that I built more recently. The original tweeter response is measured nearfield in-room, but same test box.
Nice project but what does the off-axis response look like in comparison with the original?
The off-axis response tends to be more important than the on-axis for room reflections because the on-axis area is small and contributes effectively nothing. Somewhere off-axis is also the listening axis with most coaxials and, indeed, many conventional speakers. Sharp diffracting edges causing an irregular on-axis response may be improving the off-axis response. This may well be a good trade in terms of engineering though perhaps less so in terms of marketing. The DXT lens is an example of this sort of thing.
The off-axis response tends to be more important than the on-axis for room reflections because the on-axis area is small and contributes effectively nothing. Somewhere off-axis is also the listening axis with most coaxials and, indeed, many conventional speakers. Sharp diffracting edges causing an irregular on-axis response may be improving the off-axis response. This may well be a good trade in terms of engineering though perhaps less so in terms of marketing. The DXT lens is an example of this sort of thing.
I see your point. You can find off axis in the Rutcho link above. Sadly I didn't save my own off axis measurments for the original.
As shown, even 15 degrees has a 5dB dip in the response. (Note 10dB scale)
Personally I don't like to sacrifice on-axis response, I think it's one of the most important performance criteria. I also find diffraction sounds bad, those DXT tweeters are a good example - hashy and tiring to my ears.
As shown, even 15 degrees has a 5dB dip in the response. (Note 10dB scale)
Personally I don't like to sacrifice on-axis response, I think it's one of the most important performance criteria. I also find diffraction sounds bad, those DXT tweeters are a good example - hashy and tiring to my ears.
Nice work. I'm currently having a play with the Kef Q100 coax and also I have the earlier uniQ driver that came sans tweeters so I did a DIY on that using the SB Acoustics SB14 tweeter. It was the smallest one I could find. Not measured yet mind 
Feels lame that seas thought this was even an acceptable product to put out. Incredible improvement you've made.
Seems like a practically necessary upgrade for anyone with intent to use the drivers.
Seems like a practically necessary upgrade for anyone with intent to use the drivers.
If you find this lame, look at the more expensive drivers, not just from Seas. 😄
It's an interesting approach.
Curious to see if some other coax drivers could be improved by just printing some additional parts and/or fitting a different driver.
It is not possible to cure the original issue with DSP, because the frequency response changes at each listening angle.
I think the MR18 is the best candidate because it has that nice smooth surround which means we don't need to fight even more issues.
Brain picking time - is there a way I could modify the shape of the cone? I was thinking maybe closed cell foam or something overlaid. Could be a lower risk way to make a better waveguide from the driver without dropping thousands on a custom driver not guaranteed to perform.
Could these kinds of waveguided be beneficial?
First is from
https://www.trendyol.com/netex-aqou...nge-800-watt-maximum-power-2-adet-p-742061345
Second from yelitech.com
Last is named QP-353F
Idea taken from @Randy Bassinga s similar thread
Cheers!
First is from
https://www.trendyol.com/netex-aqou...nge-800-watt-maximum-power-2-adet-p-742061345
Second from yelitech.com
Last is named QP-353F
Idea taken from @Randy Bassinga s similar thread
Cheers!
Attachments
Last edited:
I've played with these Seas coax units before and discovered some small tweaks which make a huge difference without the need to custom machine parts. It does mean dealing with the stock tweeter but they're not so bad if you know what to do with them.
The main thing is to remove as much of the overhanging plastic rim at the surround area and space the tweeter forward with aluminum washers under it. The next thing is to trim as much of the overhanging VC former as possible to clean up the radiating path to the cone, reducing diffraction.
You still have the air gap resonance between the tweeter body and VC edge to deal with, but it cleans up the entire FR and removes most of the cavity resonances you get from the tweeter's plastic edge. The Seas tweeter element isn't that bad and it copes with a reasonably low crossover point. I dont have measurments for you, but the improvements are pretty big. The few millimeters of distance the tweeter moves forward can be compensated with offset crossover filter slopes.
The main thing is to remove as much of the overhanging plastic rim at the surround area and space the tweeter forward with aluminum washers under it. The next thing is to trim as much of the overhanging VC former as possible to clean up the radiating path to the cone, reducing diffraction.
You still have the air gap resonance between the tweeter body and VC edge to deal with, but it cleans up the entire FR and removes most of the cavity resonances you get from the tweeter's plastic edge. The Seas tweeter element isn't that bad and it copes with a reasonably low crossover point. I dont have measurments for you, but the improvements are pretty big. The few millimeters of distance the tweeter moves forward can be compensated with offset crossover filter slopes.