No, I'm referring to the edge resonance distortion which is present in drivers with excellent motors and independent of drive level. This includes Scan-Speak, Seas, SB/Satori and others who have very good motor designs. Metal cones have a better termination with the surround so they deal better with this.
Hi,
Sorry but you have got it completely wrong.
W15 :
Distortion :
Note how the the 8+KHz peak tracks distortion peaks to 4+KHz
for 2nd, 2.7+KHz for 3rd, 2+KHz for 4th and 1.6+KHz for for 5th.
Its pretty obvious for optimum distortion you x/o below 2KHz,
though if you don't its only third harmonic that really gets worse.
5th if it was generally higher would be a real problem, it is not
great by any means, hitting the same as 2nd and 3rd at 1.6+KHz.
Supression of the 8+KHz peak in the x/o has no effect on distortion.
Its the x/o response roll off that reduces the drivers distortion.
rgds, sreten.
Last edited:
Indeed but they don't tend to excite anything considerable other than the 2nd harmonic which is subjectively quite harmless and they are literally impossible to xover out of the design. The issue with the w15 is quite manageable with an appropriate tweeter and well chosen xover point/slope.
Third harmonic distortion is always 3x higher than the fundamental. If the fundamental is 2700 then the third harmonic would peak at 8100 which is not being shown in the graphs. Breakup does not infere it downward resonates eg IM.
I find this discussion quite interesting. I have a similar peak in response and rise in third harmonic which correspond to a typical 6.5" midbass. This rise in distortion occurs ~1450. The upper peak in response occurs around 4700, a little 2dB rise that's nothing to compensate for. Exciting this 1450 third harmonic does not excite the peak at 4700, nor is there a distortion peak at this point. Exciting the third harmonic fundamental of ~484Hz creates the peak in the third harmonic @ 1450Hz.
Btw this peak is -43dB down. Is repeatable with Arta and REW using two different cal'd mics.
Sreten agree with you on the fifth harmonic.
I find this discussion quite interesting. I have a similar peak in response and rise in third harmonic which correspond to a typical 6.5" midbass. This rise in distortion occurs ~1450. The upper peak in response occurs around 4700, a little 2dB rise that's nothing to compensate for. Exciting this 1450 third harmonic does not excite the peak at 4700, nor is there a distortion peak at this point. Exciting the third harmonic fundamental of ~484Hz creates the peak in the third harmonic @ 1450Hz.
Btw this peak is -43dB down. Is repeatable with Arta and REW using two different cal'd mics.
Sreten agree with you on the fifth harmonic.
It's important to point out that not all of the leading experts in the loudspeaker field agree on the importance and audibility of distortion in relation to other aspects of performance. I'm not saying this to infer that distortion doesn't play a role - as I'm sure we've all noticed it at various times from different speaker designs.
I would be one whom subscribes to this idiology. Tho do believe the distortion masks low level detail and negatively affects DDR.
Since there have been some questions regarding the overall design I decided to give some history on the speaker. This also might help the owner should he decide to make changes in the basic design.
Many of my designs are started with ideas from customers where they have specific goals for cost, size, parts used, etc. and I try to accommodate them (provided that I feel that it's not something that will degrade the sound). That said, it's not always something I would do for myself but I have to work within the constraints that are given to me. Most DIY builders tend to ignore this approach because you are pretty much free to do what you want (though most still have to deal with the limitations on cabinet size, budget, availability of drivers / materials, and so forth). Your goals might be different too - such as high output / sensitivity, an impedance more suitable to your amp, etc.
The RC4 came about when I was contacted by a DIY who wanted to help a friend put together a home theater. He had chosen the drivers to use with a Hiquphon tweeter, Seas W15, and dual Seas W22 woofers. Outboard IB subwoofers were going to be used to extend the bass from the main speakers.
A friend of the theater owner liked the speakers but couldn't add subwoofers so he asked me if I could integrate them into the design. I looked at available drivers that were within his budget and also for a plate amp to drive the sub. Given the space available (about one cubic ft.) I chose the Peerless based on it's upper end bandwidth, decent excursion capability, and good sensitivity. I adjusted the amplifier settings to have a flat bandpass response with the inclusion of a second order boost filter to extend the bass. This turned out to work very well; in fact, I later sold several separate subwoofers using the same design.
This second set (like the first) had no high pass filter on the W22's because both owners had crossovers available in their front end equipment. The third set changed to what the OP here has, with the addition of a high pass filter on the dual 8" section and substitution of a Fountek ribbon tweeter. In this system the front end had no active high-pass available so I added a passive filter to reduce the modulation on the W22's. I also switched to a newly available plate amp which had more power and greater flexibility with the addition of a notch filter for correction in rooms with a significant bass peak.
All three iterations have their own set of tradeoffs depending on your goals. Obviously the W15 is the main constraint on sensitivity; however. the overall system is 86db and it will play pretty loud and clean without a huge amount of power. The narrow bandwidth of W22's of the passive version could easily be changed by omitting the low pass and shifting down the crossover point. With a 2-channel system and no active high-pass available then you could let the W22's roll off naturally at 60hz and cross over to the sub at that point. The tradeoffs are greater IM distortion / excursion on the W22's and losing some loading due to the drivers being further from the floor than the subwoofer.
The W22's have some limitations because of suspension and magnetic restrictions on excursion. This could be an issue for those that favor high levels for listening. The same is true for the ribbon, in which case those who like to play things loud might consider a 1" dome with less compression at high output levels.
Depending on budget and customer requests I also have used other subwoofer drivers in the same format (Scan-Speak 23W and 26W units). Each have their own tradeoffs with the 23W having the best bass extension at the penalty of much lower sensitivity.
Many of my designs are started with ideas from customers where they have specific goals for cost, size, parts used, etc. and I try to accommodate them (provided that I feel that it's not something that will degrade the sound). That said, it's not always something I would do for myself but I have to work within the constraints that are given to me. Most DIY builders tend to ignore this approach because you are pretty much free to do what you want (though most still have to deal with the limitations on cabinet size, budget, availability of drivers / materials, and so forth). Your goals might be different too - such as high output / sensitivity, an impedance more suitable to your amp, etc.
The RC4 came about when I was contacted by a DIY who wanted to help a friend put together a home theater. He had chosen the drivers to use with a Hiquphon tweeter, Seas W15, and dual Seas W22 woofers. Outboard IB subwoofers were going to be used to extend the bass from the main speakers.
A friend of the theater owner liked the speakers but couldn't add subwoofers so he asked me if I could integrate them into the design. I looked at available drivers that were within his budget and also for a plate amp to drive the sub. Given the space available (about one cubic ft.) I chose the Peerless based on it's upper end bandwidth, decent excursion capability, and good sensitivity. I adjusted the amplifier settings to have a flat bandpass response with the inclusion of a second order boost filter to extend the bass. This turned out to work very well; in fact, I later sold several separate subwoofers using the same design.
This second set (like the first) had no high pass filter on the W22's because both owners had crossovers available in their front end equipment. The third set changed to what the OP here has, with the addition of a high pass filter on the dual 8" section and substitution of a Fountek ribbon tweeter. In this system the front end had no active high-pass available so I added a passive filter to reduce the modulation on the W22's. I also switched to a newly available plate amp which had more power and greater flexibility with the addition of a notch filter for correction in rooms with a significant bass peak.
All three iterations have their own set of tradeoffs depending on your goals. Obviously the W15 is the main constraint on sensitivity; however. the overall system is 86db and it will play pretty loud and clean without a huge amount of power. The narrow bandwidth of W22's of the passive version could easily be changed by omitting the low pass and shifting down the crossover point. With a 2-channel system and no active high-pass available then you could let the W22's roll off naturally at 60hz and cross over to the sub at that point. The tradeoffs are greater IM distortion / excursion on the W22's and losing some loading due to the drivers being further from the floor than the subwoofer.
The W22's have some limitations because of suspension and magnetic restrictions on excursion. This could be an issue for those that favor high levels for listening. The same is true for the ribbon, in which case those who like to play things loud might consider a 1" dome with less compression at high output levels.
Depending on budget and customer requests I also have used other subwoofer drivers in the same format (Scan-Speak 23W and 26W units). Each have their own tradeoffs with the 23W having the best bass extension at the penalty of much lower sensitivity.
Hi Mike,
it does peak at 3x the frequency, only it is plotted at the excitation frequency
so with each higher order of distortion graph, the peak moves more to the left side.
There is an AES paper on "How To Graph Distortion Measurements" presented
at the 94th Convention 1993 March 16-19 Berlin which covers the subject briefly.
Getting very busy at day job so not much to update.
Rick has advised on some more measurements to make.
If time is available over the weekend, plan to do some of the following
1. Impedance measurements
2. Mid cabinet volume adjustments
3. Update measurements, also do sub-woofer response
3. Driver only response on baffle
Rick has advised on some more measurements to make.
If time is available over the weekend, plan to do some of the following
1. Impedance measurements
2. Mid cabinet volume adjustments
3. Update measurements, also do sub-woofer response
3. Driver only response on baffle
I have an interest in how small speaker companies deal with their problems and particularly the inevitable early ones.
Vapor Audio for example had problems with poor finish and delivery in some early speakers which they handled unwisely and are still feeling the consequences. Some good lessons there on the pros and cons of starting with underpriced speakers, developing production methods during manufacture and some of the potential costs of handling customer complaints badly.
Wilmslow Audio in a current thread have delivered what looks like a faulty crossover which has lead to substantial poor publicity. The original purchaser expected everything to work perfectly and was unable to perform diagnostics beyond listening and saying it sounds rubbish. The inability to perform diagnostics beyond this may be the case for a significant proportion of their customers but can a company afford to take back and test a DIY crossover on this basis? The next owner of the crossover measured it and showed it was not performing as designed. Should the company take back and fix this now second hand DIY crossover given measurements showing it is faulty?
I didn't insinuate it was an unwise design. I stated it is was an unwise design and gave factual reasons why this was the case as did some other posters. What is very interesting to me is that you put forward arguments like the rest of your paragraph and it clearly worked for some readers. Not only that but these readers are likely to be the ones potentially in need of your services to help design speakers and crossovers. A good lesson there.
Thanks for clearing that up - know I know you weren't just insinuating 🙄
Clearly you want there to be "right" and "wrong" sides by chastising anyone who doesn't agree with you. Obviously you're a smart person but tend to ignore anything that doesn't fall into your belief system. You seem to suggest that only those who aren't as "experienced" as yourself will accept my opinion on this design. That's pretty arrogant and condescending, don't you think?
I am fortunate to have met three people who are considered by many speaker enthusiasts to in the top five of experts in the loudspeaker field. They all have very different approaches to speaker design and engineering. Does that make one of them right and the other two wrong? Not everything is black and white - there are often shades of gray involved.
I've not followed the Wilmslow story so I cannot really comment on that. I am somewhat familiar with the Vapor situation due to things passed on to me by customers and others involved. Some of it I addressed in another forum. As I noted there all companies (large and small) have problems arise from time to time and the key is how you choose to deal with them. I think some companies shoot themselves in the foot by raising customer expectations to a point that they cannot meet on a consistent basis.
How a company should respond? I cannot speak for the others but I always try to resolve problems to the satisfaction of the customer. That includes support of products regardless of whether the owner purchased from me or not. It's just the right thing to do.
Last edited:
Rick had pointed out errors in the impedance measurements.
Just disconnecting the driver for individual mid/high impedance measurement was not enough as the XO was still loading the common input.
I disconnected the mid/high sections at the XO inputs and measured again.
Posting them here to avoid confusion due to previously posted wrong Z-measurement results.
Just disconnecting the driver for individual mid/high impedance measurement was not enough as the XO was still loading the common input.
I disconnected the mid/high sections at the XO inputs and measured again.
Posting them here to avoid confusion due to previously posted wrong Z-measurement results.
Either you did not measure raw drivers or something is wrong with the
measurement setup. It ought to look similar to the ones made in free
air (manufacturer). W22 and W15 impedance peaks would have shifted
to the right side being in the cabinet.
measurement setup. It ought to look similar to the ones made in free
air (manufacturer). W22 and W15 impedance peaks would have shifted
to the right side being in the cabinet.
The drivers were measured with the crossover in the circuit. I compared to my software and the transfer functions are correct.
Why is it a concern? It looks fine and lines ups with the published measurements by SEAS. The slope downwards is due to the low frequencies progressively radiating in an omnidirectional fashion rather than mainly forward which they do at higher frequencies. At the very lowest frequencies the roll off is due to an overly large sealed chamber which is good in giving in plenty room to dissipate the rear radiation and since it is high passed has no adverse effect in use. The driver configuration and crossover should provide the "baffle step correction" to counter the downward slope. There will also be a bit of a bump/wiggle at the highest frequencies making the peak to trough greater than 6 dB. As suggested earlier, it only takes a few minutes to get a feel for all this for your driver and baffle combination using software like edge or similar.
PS Looking at how the response changes with your narrow baffle versus a wider flat baffle may also be interesting.
Last edited:
As Andy says here there is no problem. You can see from the impedance plot that the driver is tuned to a little over 60Hz. For a fully stuffed cabinet this puts the driver working in a box of around 7 litres. If the really tight hole around the driver was causing a problem the resonance peak would be significantly higher in frequency and/or highly damped (the peak would be flattened) as a result of the resistive nature of the tight space.
- Status
- Not open for further replies.