aamacgregor- if you want to upgrade, consider some new xo components for the tweeter circuit.
Plenty of controversy here with mundoft resistors, but IMO they do seem to work better, and they are only $2.50 in any case.
https://www.madisoundspeakerstore.com/mundorf-10-watt/mundorf-2.2-ohm-metal-oxide-film-resistors/
https://www.madisoundspeakerstore.com/mundorf-10-watt/mundorf-10-ohm-metal-oxide-film-resistors/
Can't go wrong here either with Solen caps:https://www.madisoundspeakerstore.com/solen-capacitors/solen-3.3-mfd-fast-cap-400v/
https://www.madisoundspeakerstore.com/solen-capacitors/solen-22-mfd-fast-cap-400v/
Plenty of controversy here with mundoft resistors, but IMO they do seem to work better, and they are only $2.50 in any case.
https://www.madisoundspeakerstore.com/mundorf-10-watt/mundorf-2.2-ohm-metal-oxide-film-resistors/
https://www.madisoundspeakerstore.com/mundorf-10-watt/mundorf-10-ohm-metal-oxide-film-resistors/
Can't go wrong here either with Solen caps:https://www.madisoundspeakerstore.com/solen-capacitors/solen-3.3-mfd-fast-cap-400v/
https://www.madisoundspeakerstore.com/solen-capacitors/solen-22-mfd-fast-cap-400v/
Hi Steve, thank you very much for all of your hard work on this, I found your post really helpful and I really appreciate the time you have spent.
Ok, I'm going to have a go at taking the tweeter apart just to check that the coil is definitely the problem; if it isn't and it's just a failed internal connection, then I'll try and fix it carefully, but if it is the coil then I'll buy a replacement coil.
I would like to understand the LCR a little more, it seems to be there to remove the bump in the impedance, am I right? I'm not sure what effect the LCR has on the SPL output of the driver. If for some reason I'm unable to get hold of a replacement coil for the 9000 and decide to go for 9300's or 9500's, then I guess I'll need to tune the LCR to the impedance peak on the new drivers, plus perhaps add an L-Pad right?
Kind regards,
Andrew
Thanks for this, perhaps upgrading the quality of the crossover components is the right upgrade path for me to take. I've also heard that if I replace the metal external cross over case with a wooden one I will also get a really good improvement - apparently the metal case interferes with the crossover components.
Assuming the crossover sketched by system7 is correct (yours is surely not), the LCR is there because the filter is first order electrical, and the spike in impedance will have an effect in the filtered FR. See here: Series Notch Filter Designer / Calculator Help
If you want to swap tweeter you need also to change component values in the LCR notch (tricky). You don't need to add an L-pad because all 9000, 9300 and 9500 tweeters have the same sensitivity. If you want to try something you can try substituting the 2R2 resistor with a 1R8 (more SPL) or a 2R7 (less SPL).
Ralf
Edit: don't waste time in "upgrading" crossover components, there is nothing to upgrade (caps are film foils and coils are air core)
If you want to swap tweeter you need also to change component values in the LCR notch (tricky). You don't need to add an L-pad because all 9000, 9300 and 9500 tweeters have the same sensitivity. If you want to try something you can try substituting the 2R2 resistor with a 1R8 (more SPL) or a 2R7 (less SPL).
Ralf
Edit: don't waste time in "upgrading" crossover components, there is nothing to upgrade (caps are film foils and coils are air core)
Last edited:
I'd be very wary of anything metal near coils. The famous tests are by Troels Gravesen:
Placement of coils in crossover networks
Coils should be at right angles to each other ideally, far apart and away from metal plates.
I did some testing myself to assess if metal circuit boards with big ground planes might have issues at whatever angle you mount the coils. Indeed they do. A 20% change in value on this 1mH looks bad to me. The tweeter also picked up the 1kHz test tone of the inductance meter. It was audible as a whine.
Attachments
Mmmm... a metal box acts as a Faraday cage for what's inside. In this respect it is a good thing. The crossover is laid on a wooden plate, so not directly in contact with the metal. How much effect the box has on the coils, if any, I don't know without a measurement. But almost every hi-fi gear is made with a metal box.
Ralf
Ralf
It will be easy to tell. If the coil is fried you will know right away by the appearance and the smell. If it's a boken lead it can be fixed but stay away from coffee beforehand. You need a very steady hand and control over the heat. Not to mention stripping the enamel. Good luck with it.
Hi Ralf, ok, I'll admit it again... I got one of the resistor values wrong... very wrong in my diagram, but the rest was correct... or at least I think it was. 🙂
I'm a little curios, the link you provided suggested that I could calculate the values for the notch filter if I know the Fs, Qes and Qms values; is there are reason why you have put "tricky" in brackets? Is there more to it than the link suggests?
Thanks Steve, I'm going to move the crossover to a wooden box, since as you suggest, the metal box is causing an issue - I noticed that in the following reviews someone mentioned they noticed a big improvement when they changed theirs:
QLN Signature Splitfield Floorstanding Speakers reviews - Audioreview.com
and in addition, QLN themselves mention that the later revisions had a wooden box instead of a metal one.
History of the Signature speaker | Qln
Anyway, I guess it can't hurt to give it a try. 🙂
Also when I transfer the crossover to a wooden box, I'll try re-orientating the coils so that they are all at right angles (I'll lay one flat, one upright front to back and one upright left to right).
Thanks Cal, I'll try to be careful. To be honest, the person that soldered the terminals onto the tweeter seems to have made a bit of a mess and melted a fair amount of the plastic around the terminals - looks to me like their iron wasn't hot enough and they left it on the terminal too long in order to melt the solder; I'm guessing my speakers were soldered by a junior. Either way they made a mess - it did half make me wonder whether they might have damaged the coil so that it would be more likely to fail. Anyway, I'll let you know what happens when I take it apart. 🙂
It does act like a Faraday cage, but it also will make the coils act as if they are not completely air core - since the coil will effectively act like an electromagnet and will interact with the metal case. The metal case around electronics in general helps shield the components in the box from external interference, but in this case the shielding effect is probably outweighed by the interference of the box with the coils. I guess that is why people say that the QLN's sound better if the box is wooden.
When doing that type of soldering, I like to wet the tip of the iron with solder, use paste and do a 'cold" solder where only the solder touches the material and you release the iron as soon as the solder flows. You don't preheat the material. It's a way to avoid overheating when working with the delicate stuff. Remember to clean the paste off afterward.
I've never used paste before. I tend to very quickly tin the surface and the wire with a thin layer of solder using a wet iron and then bind the connection in what sounds like a similar way to you with (I don't let the iron touch the materials when binding the connection, only the solder - it all happens very quickly). That way, the materials don't get too hot and I get a good connection. I do tend to use a very hot iron though which means the contact time is very short resulting in less heat transferred to the underlying materials. I'll look into using paste though, it sounds fun.
No need. If you are able to do it your way stick with it. I'll tell you a secret. If it's only one joint, I will not bother with an iron, I'll use a gun for speed sake. By using paste the whole thing happens in a flash. It may be old fashioned but hey, I guess that's because I'm old. 🙂
So, it's been a while and well I failed to obtain a new coil for my D2905/9000. As far as I understand the issue was that ScanSpeak believed they had stock of the D2905/9000 and so vendors thought they could order the part, but as it turns out, it seems that ScanSpeaks manifest for that item number was incorrect and they don't have any left. So I guess that means saying good bye to the good old D2905/9000's (I'm not a fan of purchasing such things on ebay).
In the end I decided to go for the D2905/950000, I realise that the D2905/930000 is closer in profile to the D2905/9000 (on both FR and impedance), but I've heard of a number of people saying that the D2905/950000 works really well with the 18W/8545K and sounds a bit more transparent than the D2905/930000.
So, as well as buying the tweeters I also figured that I should change the RLC filter. Using following values from the tweeter specs and putting them into an RLC calculator:
Fs = 550Hz
DC Res = 4.7 ohm
Qes = 0.6
Qms = 1.04
the result is:
R = 7.412 ohm
L = 0.82 mH
C = 102.64 uF
Back converting the above gives the following:
Fs = 548.559 Hz
Damping factor (series) = 1.31116
Damping factor (parallel) = 0.190670
Bandwidth FWHM = 1438.61 Hz
Modifying the RLC values a little to more obtainable values we end up with:
R=7.6 ohm
L = 0.84 mH
C = 100 uF
resulting in the calculated values of
Fs = 549.137 Hz
Damping factor (series) = 1.31112
Damping factor (parallel) = 0.190676
Bandwidth FWHM = 1439.97 Hz
Given the above, I'm going to get rid of the current 1.5 ohm, 1.8 mH, 22 uF RLC filter and replace it with a 7.5ohm (2 x 15 ohm ceramic in parallel), 0.84mH (air core), 100uF (Solen) combo.
Now the thing is, I ordered all of the parts this afternoon, but I just want to check that what I have done is correct?
I realise that I haven't accounted for the resistance due to the inductor (which is apparently going to be around 0.47 ohms), but figure that I can always run a 150 ohm resistor in parallel with the two 15 ohm resistors to compensate if needs be.
I'm hoping I haven't, but have I missed anything?
In the end I decided to go for the D2905/950000, I realise that the D2905/930000 is closer in profile to the D2905/9000 (on both FR and impedance), but I've heard of a number of people saying that the D2905/950000 works really well with the 18W/8545K and sounds a bit more transparent than the D2905/930000.
So, as well as buying the tweeters I also figured that I should change the RLC filter. Using following values from the tweeter specs and putting them into an RLC calculator:
Fs = 550Hz
DC Res = 4.7 ohm
Qes = 0.6
Qms = 1.04
the result is:
R = 7.412 ohm
L = 0.82 mH
C = 102.64 uF
Back converting the above gives the following:
Fs = 548.559 Hz
Damping factor (series) = 1.31116
Damping factor (parallel) = 0.190670
Bandwidth FWHM = 1438.61 Hz
Modifying the RLC values a little to more obtainable values we end up with:
R=7.6 ohm
L = 0.84 mH
C = 100 uF
resulting in the calculated values of
Fs = 549.137 Hz
Damping factor (series) = 1.31112
Damping factor (parallel) = 0.190676
Bandwidth FWHM = 1439.97 Hz
Given the above, I'm going to get rid of the current 1.5 ohm, 1.8 mH, 22 uF RLC filter and replace it with a 7.5ohm (2 x 15 ohm ceramic in parallel), 0.84mH (air core), 100uF (Solen) combo.
Now the thing is, I ordered all of the parts this afternoon, but I just want to check that what I have done is correct?
I realise that I haven't accounted for the resistance due to the inductor (which is apparently going to be around 0.47 ohms), but figure that I can always run a 150 ohm resistor in parallel with the two 15 ohm resistors to compensate if needs be.
I'm hoping I haven't, but have I missed anything?
I know it's been a while since I last wrote anything on this post, and I want to thank everyone who added to this subject and helped me.
So, I wanted to give an update, as mentioned, I changed the tweeters (to the D2905/950000) and replaced the old 1.5ohm, 1.8mH, 22uF RLC circuit with with a 7.5ohm (2 x 15 ohm metal oxide in parallel), 0.84mH (air core), 100uF (Solen) RLC circuit.
So, do they sound better or worse? Well, they sound different, that is for sure. On the one hand, they sound more transparent, cleaner and that feeling of roll-off that I had with the D2905/9000 is no longer there. Voices and guitars other classical instruments sound more in the room with an increased level of detail. On the other hand, the QLN's used to have a mesmerising ability to cause you to lose yourself in the music, albeit if I'm really critical, with that annoyingly slight roll off at the top end. What I wanted was to maintain that mesmerising ability whilst removing the top end roll-off and adding a little more transparency. What I have got however, is no top end roll-off, extra transparency and detail, but that mesmerising ability, though it is still there has been slightly impacted (not massively, but enough to notice, and enough for me to want to try to correct it).
Having listened to them for a while, I feel the issue is that that the tweeters are a fraction too loud overall and this is causing certain instruments at times (e.g. cymbals) to sound a fraction thinner than they used to (or should) which is a distraction. It doesn't sound like there is a particular frequency that is louder than others, but just that the whole upper frequency range is that fraction too loud. I'm thinking that I might play with the L-pad to bring the tweeters down 1db to see if that helps or even fixes the problem. They are so close to giving everything that I'd hoped! Hopefully a minor play the the L-pad will fix that feeling that the tweeters are too loud.
That said, letting my ears guide me is one thing, but to know exactly what is going on, I think that I need to invest in some measuring equipment that I can hook up to my computer. Any suggestions on a good calibrated USB microphone, or even one that requires a phantom power supply and amplifier that can be plugged into a computer?
So, I wanted to give an update, as mentioned, I changed the tweeters (to the D2905/950000) and replaced the old 1.5ohm, 1.8mH, 22uF RLC circuit with with a 7.5ohm (2 x 15 ohm metal oxide in parallel), 0.84mH (air core), 100uF (Solen) RLC circuit.
So, do they sound better or worse? Well, they sound different, that is for sure. On the one hand, they sound more transparent, cleaner and that feeling of roll-off that I had with the D2905/9000 is no longer there. Voices and guitars other classical instruments sound more in the room with an increased level of detail. On the other hand, the QLN's used to have a mesmerising ability to cause you to lose yourself in the music, albeit if I'm really critical, with that annoyingly slight roll off at the top end. What I wanted was to maintain that mesmerising ability whilst removing the top end roll-off and adding a little more transparency. What I have got however, is no top end roll-off, extra transparency and detail, but that mesmerising ability, though it is still there has been slightly impacted (not massively, but enough to notice, and enough for me to want to try to correct it).
Having listened to them for a while, I feel the issue is that that the tweeters are a fraction too loud overall and this is causing certain instruments at times (e.g. cymbals) to sound a fraction thinner than they used to (or should) which is a distraction. It doesn't sound like there is a particular frequency that is louder than others, but just that the whole upper frequency range is that fraction too loud. I'm thinking that I might play with the L-pad to bring the tweeters down 1db to see if that helps or even fixes the problem. They are so close to giving everything that I'd hoped! Hopefully a minor play the the L-pad will fix that feeling that the tweeters are too loud.
That said, letting my ears guide me is one thing, but to know exactly what is going on, I think that I need to invest in some measuring equipment that I can hook up to my computer. Any suggestions on a good calibrated USB microphone, or even one that requires a phantom power supply and amplifier that can be plugged into a computer?
Hi,
I am thinking to upgrade my tweeter to ScansSpeak D2905/9300 model. I like a warmish, laid back and slightly rolled-off presentation without fatiguing. After a lot investigating probably my choice seems good? Before I was hesitating between 9700, 9500, but the 9300 can have the most natural presentation. Does somebody has opinion, experience with the version 9300? Thank you.
I am thinking to upgrade my tweeter to ScansSpeak D2905/9300 model. I like a warmish, laid back and slightly rolled-off presentation without fatiguing. After a lot investigating probably my choice seems good? Before I was hesitating between 9700, 9500, but the 9300 can have the most natural presentation. Does somebody has opinion, experience with the version 9300? Thank you.
I've not tried the D2905/9300 so can't really help you there. However, having spent time with both the D2905/9000 and the D2905/9500, I will say that they are both warm, transparent and detailed, but the D2905/9500 is more so and definitely a step up (with correct crossover modification). Neither are in any way fatiguing. I found the roll-off of the D2905/9000 a little frustrating, it felt like something was missing. That problem/feeling has gone with the D2905/9500. So, basically, what I'm saying is, I think the D2905/9500 matches your description of what you are after. At least, I'd definitely recommend them.
What speakers are you thinking of upgrading?
What speakers are you thinking of upgrading?
Many thanks for your comments. I know choosing a tweeter is more hit and miss than a midrange. Well, I have an old Pioneer CS-R500 3-way speaker and this summer I made a full upgrade. Built a separate cabinet for the midrange, routing the baffle to flush mount both the midrange and tweeter and chamfering all the screw holes to nicely be flat with the baffle surface. Midrange is Seas MCA15RCY the tweeter is Seas 27TFFC. The tweeter has a lot detail, air but I never like the too much emphasized top octaves. Above 12-14Khz is too edgy and not well rounded sounding. I also tried the 27TDFC but here the lower treble is strong and missing the top octaves, the air that the 27TFFC gives. The 27TDFC's top end also edgy, so not free from it, but overall smoother. So the idea came from here to upgrade to the Scan-Speak 9300 at least give a natural ang globally neutral presentation. Well, reading older forums here there the people have different opinions and tells that the 9300 is the most natural sounding among the 2905 family. Although somebody describe too laid back or polite and prefer crispier like 9500 or 9700. I think this is a matter of taste, but I am 100% sure when I am tired or sleepy I prefer the the less detailed and warmer sound instead of bright. Also interesting that the 9300 and the 9500 has different sound where almost there isn't any difference between them according to specification.