I have just completed my first crossover modification with surprising results. Would love to get some feedback to help understand what I'm hearing.
So, I have modified the crossover in one of my B&W DM602S3 speakers. My goal was to tame the "fatiguing" nature of these speakers and I've partially applied The changes discussed here https://www.diyaudio.com/community/threads/b-w-dm602-s3-crossover-mod.328314/post-6826338.
Essentially, I have converted the first order LPF to a second order, putting a 4.7 uF capacitor and 1.5 ohm resister in parallel with the driver ( hopefully that is the right terminology). My understanding is that this second order change would cause the midrange driver to drop off at a steeper rate, hopefully reducing some unwanted higher frequency resonances.
Well, I wasn't sure what the change in audio would be but I did not expect this. The sound is markedly clearer, perhaps even brighter, although I am not prepared to say more or less fatiguing yet. The unmodified speaker sounds ... "Compressed" in comparison. Not sure how else to describe it. Definitely lacking the clarity I'm getting from the modified speaker.
I have to wonder if it's the circuit change or the quality of parts that has made the most difference. I did not spend much on the jantzen caps and miles resistors but I'm betting they are better than the stock parts.
Anyway, would love to know what the experts think about my experience. I for one think I might be a believer in quality caps now.
So, I have modified the crossover in one of my B&W DM602S3 speakers. My goal was to tame the "fatiguing" nature of these speakers and I've partially applied The changes discussed here https://www.diyaudio.com/community/threads/b-w-dm602-s3-crossover-mod.328314/post-6826338.
Essentially, I have converted the first order LPF to a second order, putting a 4.7 uF capacitor and 1.5 ohm resister in parallel with the driver ( hopefully that is the right terminology). My understanding is that this second order change would cause the midrange driver to drop off at a steeper rate, hopefully reducing some unwanted higher frequency resonances.
Well, I wasn't sure what the change in audio would be but I did not expect this. The sound is markedly clearer, perhaps even brighter, although I am not prepared to say more or less fatiguing yet. The unmodified speaker sounds ... "Compressed" in comparison. Not sure how else to describe it. Definitely lacking the clarity I'm getting from the modified speaker.
I have to wonder if it's the circuit change or the quality of parts that has made the most difference. I did not spend much on the jantzen caps and miles resistors but I'm betting they are better than the stock parts.
Anyway, would love to know what the experts think about my experience. I for one think I might be a believer in quality caps now.
If guessing is what you want, then I'd suggest the reduction in breakup or possibly an improvement in the cross itself may be creating a more consistent and smooth output throughout the room. You also suggest it may be brighter, which is not unlike the type of comment you might get when adding a second order filter or possibly just changing the balance around the cross.
I have heard these speakers in various incarnations, and they are generally bright and fatiguing to my ears...
Kevlar 6" 0r 7" bass, and a 4 ohm metal tweeter with a large ultrasonic resonance.
Rutcho's plan was to roll off the bass more steeply to reduce the 4-5kHz typical Kevlar breakup, which you can see on Rutcho's waterfall plot for the bass, and reduce tweeter level by increasing the resistor at the front of the tweeter filter to 3.3R and changing the 4.7uF in the tweeter filter to 3.3uF which raises the crossover point a little.
https://rutcho.com/tweaks/01_bw_dm601s3/bw_dm601s3.html
Kevlar is notorious for breakup, unlike polypropylene. It sounds quite nasty.
http://www.nutshellhifi.com/library/speaker-design2.html
Endless room for experiment really. I would be using a simulator with similar drivers to see what happens to impedance because excessively low impedance is hard on amplifiers.
But with these simple mods, I wouldn't expect anything to be worse.
Best, Steve.
Kevlar 6" 0r 7" bass, and a 4 ohm metal tweeter with a large ultrasonic resonance.
Rutcho's plan was to roll off the bass more steeply to reduce the 4-5kHz typical Kevlar breakup, which you can see on Rutcho's waterfall plot for the bass, and reduce tweeter level by increasing the resistor at the front of the tweeter filter to 3.3R and changing the 4.7uF in the tweeter filter to 3.3uF which raises the crossover point a little.
https://rutcho.com/tweaks/01_bw_dm601s3/bw_dm601s3.html
Kevlar is notorious for breakup, unlike polypropylene. It sounds quite nasty.
http://www.nutshellhifi.com/library/speaker-design2.html
Endless room for experiment really. I would be using a simulator with similar drivers to see what happens to impedance because excessively low impedance is hard on amplifiers.
But with these simple mods, I wouldn't expect anything to be worse.
Best, Steve.
Interesting and trying to wrap my head around this. Logic would tell me that if a driver is rolling off quicker at high frequencies then I would expect a diminished "brightness". But what I don't know could fill many books.
Yes, I have been happy with them for many years but perhaps my tastes or sensitivities have changed and I am now feeling the fatigue. I see a new set of speakers in my future but until then, I'll have some fun experiments with what I've got.
My changes are obviously based on Rutcho's design, however, the midbass driver and crossover network for my DM602S3 varies a bit from Rutcho's DM601S3. I have not yet touched the tweeter network other than removing the 1.5 ohm resistor in parallel with driver. Rutcho replaces the 4.7 uF cap with a 3.3uF. The DM602S3 stock cap is 6.8uF so I wonder if dropping all the way to 3.3 is the best choice... might try a 4.7 if I change it at all.
Regarding the impedence, I'm taking it on faith that I'm not doing anything too outlandish. By any chance, would someone know of similar drivers (with frd and zma files available) that I could use to simulate the changes?
And thanks for the linked articles, @system7 .
Consider that a second order filter might resonate to create a peak. Naturally I don't know in this case, that's up to you to check.
Another issue is when your upper woofer response is taken away, the treble stands alone. This might create a brightness.
Similarly, breakup creates inconsistencies.
Another issue is when your upper woofer response is taken away, the treble stands alone. This might create a brightness.
Similarly, breakup creates inconsistencies.
This is a good way of controlling breakup at 5kHz in this case:
It's called a tank notch, and with 1mH coil you'd want 1uF and 22R for 5kHz. Really changes the sound entirely!
The resistor is important because without it you'd get a dead short at high frequency with a shunt capacitor across the bass.
It's called a tank notch, and with 1mH coil you'd want 1uF and 22R for 5kHz. Really changes the sound entirely!
The resistor is important because without it you'd get a dead short at high frequency with a shunt capacitor across the bass.
An LCR in series with the woofer followed by a zobel across the woofer (both after the LP filter) will solve just about any bandwidth and amplitude of breakup. If the impedance dip is too low, an inductor will need to be added in series with the zobel, turning it into a second notch (reacting with the first series notch).
The most effective and impedance friendly notch filter is a combination of both series and parallel notch. It keeps the speaker impedance response from skewing the target FR response of the LP filter.
Sometimes you can get away with a single notch filter in series if the woofer impedance is very linear. The zobel fixes this in most cases, but sometimes a second LCR is unavoidable.
The most effective and impedance friendly notch filter is a combination of both series and parallel notch. It keeps the speaker impedance response from skewing the target FR response of the LP filter.
Sometimes you can get away with a single notch filter in series if the woofer impedance is very linear. The zobel fixes this in most cases, but sometimes a second LCR is unavoidable.
In my experience this can be due to excess Xmax on the tweeter, due to an inadequate filter slope, feeding it too much bass. The excursion causes IM distortion which is very tiring.
Maybe the spec-sheet for the tweeter will give some clues to this.
Recommendation - get a cheap measurement mic and have a look what's going on here. Even better - measure the impedance! You can destroy your amp when something goes wrong with the filter network ...
You seem to cut some midrange - more "clear" sound. Exciting first ... listen longer. You will change the phase behaviour when you change the filter order, it's likely you have some chancellation.
Get your mic, check it out.
You seem to cut some midrange - more "clear" sound. Exciting first ... listen longer. You will change the phase behaviour when you change the filter order, it's likely you have some chancellation.
Get your mic, check it out.
Lots of information here for me to process.
Since my last post, I completed Rutcho's mod and have found the highs to be overly attenuated (in my subjective opinion as I am not able to execute measurements yet. Mic is on the way).
I'm going to make some minor tweaks and try to find a happy place (aurally). Then I'll learn how to do some proper measurements and follow up on these latest suggestions.
Many thanks to you all for the feedback!
Since my last post, I completed Rutcho's mod and have found the highs to be overly attenuated (in my subjective opinion as I am not able to execute measurements yet. Mic is on the way).
I'm going to make some minor tweaks and try to find a happy place (aurally). Then I'll learn how to do some proper measurements and follow up on these latest suggestions.
Many thanks to you all for the feedback!
This seems very plausible based on my listening experience so far. Now, I did reverse the polarity on the tweeter, as suggested in Rutcho's mod.
Hope to take some measurements soon and get a clearer picture of what is happening.
I would think a microphone is a complete waste of time and money unless you know what you are doing with one.
And then you need to master simulators and filters. But don't let me stop you.
I can usually hear what is wrong with a speaker and given a schematic I know how to fix it....
THIS is your B&W 602 S3 Speaker. It's nothing like the 601. I did this schematic work myself years back by wading through the manual.
7" Kevlar bass which I assume is 8 ohms nominal, 4 ohm metal tweeter. Looks like there is room on the PCB for R3 and C3, but I'd try and keep C3 physically small like an NPE type.
B&W Support is down at the moment, so I am a bit limited here:
http://bwgroupsupport.com/
Seems this speaker dips to an alarming 3 ohms at the top end but is described as 8 ohms nominal.
I ran a very rough sim to see how it works. The 8.2R is just a modelling device for an 8 ohm tweeter I used here. The dotted lines are with the B&W original filter.
I found the shunt RC doesn't actually do much. The RC tank cuts breakup around 5kHz considerably. Replacing the 0.47R with 2.2 at the tweeter input improves impedance at the top, whilst reducing level 2dB.
I see I used 1.8R in the shunt, not your 1.5R, but that is no matter.
Make of it what you will. It is what it is. But definitely room for improvement IMO.
Best, Steve.
And then you need to master simulators and filters. But don't let me stop you.
I can usually hear what is wrong with a speaker and given a schematic I know how to fix it....
THIS is your B&W 602 S3 Speaker. It's nothing like the 601. I did this schematic work myself years back by wading through the manual.
7" Kevlar bass which I assume is 8 ohms nominal, 4 ohm metal tweeter. Looks like there is room on the PCB for R3 and C3, but I'd try and keep C3 physically small like an NPE type.
B&W Support is down at the moment, so I am a bit limited here:
http://bwgroupsupport.com/
Seems this speaker dips to an alarming 3 ohms at the top end but is described as 8 ohms nominal.
I ran a very rough sim to see how it works. The 8.2R is just a modelling device for an 8 ohm tweeter I used here. The dotted lines are with the B&W original filter.
I found the shunt RC doesn't actually do much. The RC tank cuts breakup around 5kHz considerably. Replacing the 0.47R with 2.2 at the tweeter input improves impedance at the top, whilst reducing level 2dB.
I see I used 1.8R in the shunt, not your 1.5R, but that is no matter.
Make of it what you will. It is what it is. But definitely room for improvement IMO.
Best, Steve.
Looks like you have also added the 8.2 ohm resistor in parallel with the tweeter. If my new knowledge is correct, this is to level the impedance, correct?
Hello all.
Maybe this is a bit of a noob question thats been handled before but, does the voltage of a cap in the filter have any effect on the sound of the speaker?
Reason i ask is. I just replaced the old 50v elcap caps in my speaker for 400v Jantzen caps. All values are correct but im missing a lot in my mid and low now. Only thing that seems to be better is the tweeter.
Maybe this is a bit of a noob question thats been handled before but, does the voltage of a cap in the filter have any effect on the sound of the speaker?
Reason i ask is. I just replaced the old 50v elcap caps in my speaker for 400v Jantzen caps. All values are correct but im missing a lot in my mid and low now. Only thing that seems to be better is the tweeter.
Capacitance is (in a good capacitor) independent of breakdown voltage.
If the new cap is film, and the old one is electro, perhaps the speaker was tuned for an old electro and now the new capacitor is simply too good.
It may settle down though, if not, maybe add a small (1-2 Ohm tops) series resistance to the new capacitor, to simulate what they may have used?
If the new cap is film, and the old one is electro, perhaps the speaker was tuned for an old electro and now the new capacitor is simply too good.
It may settle down though, if not, maybe add a small (1-2 Ohm tops) series resistance to the new capacitor, to simulate what they may have used?
I think that might be it. After searching on interner i found
https://www.falconacoustics.co.uk/a...capacitor-nonpolarised-reversible-series.html
They claim to be one on one with the old Elcap. Ill try these and see.
https://www.falconacoustics.co.uk/a...capacitor-nonpolarised-reversible-series.html
They claim to be one on one with the old Elcap. Ill try these and see.