I'm looking at a speaker kit that costs around 400 euros per speaker.
A "high quality" kit is also offered for about 80 euros more, changing the filter components to more expensive ones.
Knowing the specs and values for these components I've managed to do even better sourcing all components myself - slightly more expensive caps, baked coils - while still keeping the costs slightly under their "high quality" version.
But will this actually make a noticeable difference with speakers in this price range?
On one hand I really want to support diy kit sellers, but on the other hand I know I'd be bothered everyday by the thought I could have had better components in my speakers for the same amount of money.
How much of a difference does it make to get slightly upgraded components?
I'm thinking of mailing them with my dilemma to see if they can match my version..
A "high quality" kit is also offered for about 80 euros more, changing the filter components to more expensive ones.
Knowing the specs and values for these components I've managed to do even better sourcing all components myself - slightly more expensive caps, baked coils - while still keeping the costs slightly under their "high quality" version.
But will this actually make a noticeable difference with speakers in this price range?
On one hand I really want to support diy kit sellers, but on the other hand I know I'd be bothered everyday by the thought I could have had better components in my speakers for the same amount of money.
How much of a difference does it make to get slightly upgraded components?
I'm thinking of mailing them with my dilemma to see if they can match my version..
..it depends on the rest of your system, but generally the most determinative location for premium components is with respect to your tweeter's filter, and with respect to components: *resistors (if used) are the most determantive of quality, followed by capacitors, and distantly - inductors.
..you do need to make sure that resistance values are substantially similar for inductors when comparing them. Typically the exception to this is the inductor on your final low freq. driver (woofer) where lower resistance is almost always beneficial.
*note that resistance is often a function of the inductors in the high-pass filter (without actually requiring a resistor).
IF you actually have resistors in your network, consider looking at expensive ones.
TonyGee's site has a nice break-down of subjective cap quality and their specific "sound" on his website:
Humble Homemade Hifi - Cap Test
Note: when you start looking at a "range" of cap quality say 8's and 9's - pay MORE attention to the subjective effect than the quality rating.
..you do need to make sure that resistance values are substantially similar for inductors when comparing them. Typically the exception to this is the inductor on your final low freq. driver (woofer) where lower resistance is almost always beneficial.
*note that resistance is often a function of the inductors in the high-pass filter (without actually requiring a resistor).
IF you actually have resistors in your network, consider looking at expensive ones.
TonyGee's site has a nice break-down of subjective cap quality and their specific "sound" on his website:
Humble Homemade Hifi - Cap Test
Note: when you start looking at a "range" of cap quality say 8's and 9's - pay MORE attention to the subjective effect than the quality rating.
I have measured very expensive caps that are objectively worse (have a higher dissipation factor) than the bargain varieties Dayton/Erse/Axon. That said, poly caps (in contrast to electrolytic units) in general have very low dissipation so the differences are insignificant -- despite subjective claims to the contrary.
With regard to inductors, the worry is saturation at very high power levels (if iron core). So with iron-core inductors to be used at high power, there is value in using an expensive variety. For air core at a given DCR, there is no important difference between inductors. If you can find a difference its a physics discovery.
I am not aware of any improvement in resistors to be gained over the usual noninductive Mills/Dayton units. They behave as functionally perfect circuit elements. Really flat response over the audible FR range. Unless, of course, you exceed the power rating. Then you will have a problem.
The issue in all cases is to pay close attention to deviation from rated value. It is possible that an inductor or capacitor will deviate "audibly" from its stated nominal value if used in a sensitive circuit location. For inductors this is often easy to correct by unwinding. For capacitors this may justify purchase of a 2% or 3% unit rather than a 5% if the crossover topology is sensitive to it.
When a component is put is a sensitive location I generally measure units to very high precision with a frequency-dependent test protocol. The manufacturer name on the L/C/R element does not concern me, but the actual L/C/R value does concern me.
Poetic justice is somebody buying a very expensive boutique cap for $100 that is out of spec by 5% to replace a $5 Axon cap that was only out of spec by 2%.
Last edited:
Pygmy, my viewpoint on the matter is there is very little difference you might hear between modern components these days.
Typical resistors used in passive crossover networks have very little inductance and their inductive reactance at audio frequencies is negligible. As long as the resistor value is within 5% tolerance of what is required there is no audible difference.
Audio quality capacitors within 5% tolerance of the prescribed value are much the same as resistors, there is very little difference in sound quality between brands. ESR can vary between types somewhat but this is generally small enough not to have a significant audible effect. Paralleling capacitors to make but a value of capacitance can reduce ESR a little but is hardly worth the effort for most designs. Try to avoid electrolytic bipolar capacitors if one can unless it is needed for a very low frequency filter where polypropylene capacitors would be too expensive to make up the value. Electro bipolar can have significant ESR values and generally have a wide tolerance margin of the stated value.
Inductors are generally within a 5% tolerance but do have a resistance value that can vary according to guage (AWG) and the core type that it is (air, iron core etc). DC resistance of the inductor where ever used can have an audible effect throughout the audio range with the exception possibly in the HF range where it is generally in parallel with the tweeter.
Always remember when considering components the manufacturing tolerances of the drivers can be as wide as 20% of their stated specification. So when evaluating what improvement may have been made by changing components note that the drivers themselves are likely to undergo change too.
C.M
Typical resistors used in passive crossover networks have very little inductance and their inductive reactance at audio frequencies is negligible. As long as the resistor value is within 5% tolerance of what is required there is no audible difference.
Audio quality capacitors within 5% tolerance of the prescribed value are much the same as resistors, there is very little difference in sound quality between brands. ESR can vary between types somewhat but this is generally small enough not to have a significant audible effect. Paralleling capacitors to make but a value of capacitance can reduce ESR a little but is hardly worth the effort for most designs. Try to avoid electrolytic bipolar capacitors if one can unless it is needed for a very low frequency filter where polypropylene capacitors would be too expensive to make up the value. Electro bipolar can have significant ESR values and generally have a wide tolerance margin of the stated value.
Inductors are generally within a 5% tolerance but do have a resistance value that can vary according to guage (AWG) and the core type that it is (air, iron core etc). DC resistance of the inductor where ever used can have an audible effect throughout the audio range with the exception possibly in the HF range where it is generally in parallel with the tweeter.
Always remember when considering components the manufacturing tolerances of the drivers can be as wide as 20% of their stated specification. So when evaluating what improvement may have been made by changing components note that the drivers themselves are likely to undergo change too.
C.M
One way around wide driver tolerance is to order several (say at least 4 or more) of each kind of driver, measure the pertinent parameters of each (offhand these would be resonant frequency, DC resistance, sensitivity) and use the best matched pair between them. The relative importance of each parameter, and how to calculate "best matched" using them might be topics of another thread.
It's also easy enough to match crossover components to 1 percent by measuring components and parallel or series-connecting them to come up with precise values. (Use a cap that's about 95 percent of the desired value, measure it to get the actual value, then use a smaller cap to parallel it that's closest to the remaining value needed - likewise with inductors in series, but also keep the windings at 90 degrees). I have a BK-878 LRC meter that has 1 percent accuracy and probably 0.2 percent resolution (4999 count), it would be ideal for this.
But I agree, 5 percent might make 0.1dB difference, so this could be overkill. Then again, with a high-order (third or higher) filter, a small value change could make a large difference.
Member
Joined 2009
Paid Member
I mentioned it in my response but I will say it again: among the worst performing poly capacitors -- in terms of both dissipation factor (loss angle) and deviation from rated capacitance -- were the most expensive. This was based on very-high-precision testing I did about 7 years ago. But, because poly caps are so good (by design), the dissipation factor is generally very tiny (about 1e-3, or 100 times smaller than an electrolytic cap).
jCandy,
Not sure what you measured, but I can see where you are coming from.
I've measured Mundorf MKP, several Clarity and Audyn caps, they were all pretty much spot on. Early Clarity's were like consistently 1% below marked, but well within spec, but they have fixed that. I also tried some eye-watering expensive Duelund silver foil bypass caps. Same. At least spec wise I had nothing to complain about recently.
Not sure what you measured, but I can see where you are coming from.
I've measured Mundorf MKP, several Clarity and Audyn caps, they were all pretty much spot on. Early Clarity's were like consistently 1% below marked, but well within spec, but they have fixed that. I also tried some eye-watering expensive Duelund silver foil bypass caps. Same. At least spec wise I had nothing to complain about recently.
Last edited:
In the MPP group, I measured 14 capacitors. 8 of the 14 were within +/- 1% of specified capacitance (below 10kHz). A more difficult measurement is the dissipation factor, since its is very tiny for poly caps. With the exception of one terrible inexpensive cap, the two most expensive caps (5x the cost of the budget caps) had the highest dissipation factors.
I am happy to PM people the results if there is interest.
I am happy to PM people the results if there is interest.
I really cant be bothered with crossovers or expensive speaker kits.
My favourite speaker has no crossover and is a £20 1980 Fane 12-50WRMS speaker in a £30 for plywood etc. W bin.
I would recommend a full range Fane 12-250TC or two in a sealed box with no crossovers. Much less money but with good results.
My favourite speaker has no crossover and is a £20 1980 Fane 12-50WRMS speaker in a £30 for plywood etc. W bin.
I would recommend a full range Fane 12-250TC or two in a sealed box with no crossovers. Much less money but with good results.
Like everything there is a law of diminishing returns. For capacitors I tend to use Supersound caps as they are a decent sound for a reasonable price. They are significantly better than more inexpensive caps and I prefer the sound to similarly priced caps.
Buying Duelund at ten times the price isn't going to get you ten times the quality though. Differences will be small, but they will be there (with suitable partnering equipment).
Buying Duelund at ten times the price isn't going to get you ten times the quality though. Differences will be small, but they will be there (with suitable partnering equipment).
Dissipation factor is not directly related to distortion. A high dissipation factor can simply mean the capacitor has a higher ESR and ESR itself is not necessarily a source of distortion. The main source of distortion is the capacitance changing versus current or voltage. This results in both harmonic distortion and inter-modulation distortion where the capacitor is acting as attenuation in a filter. Imagine you have a capacitor acting as part of a lowpass filter in a woofer with fc=1khz, but everytime a big bass transient comes along the voltage across the capacitor causes it to drop to 80% of it's original value and your lowpass filter moves up to Fc=1.2kHz. Now all the treble signals around 1kHz are going to be amplitude modulated by bass transient as the effective lowpass filter is being modulated between Fc=1kHz and Fc=1.2kHz -> now you have intermodulation distortion. The same effect happens as a single tone at 1kHz causes the capacitance to change the corner frequency and this results in harmonic distortion.
Thankfully, the types of caps used in speaker crossovers (plastic film and aluminium electrolytic) do not have significant capacitance-voltage and capacitance-current relationships so they do not create appreciable amounts of distortion. Some capacitor dielectrics such as Class II and Class III multilayer ceramic caps have terrible capacitance-voltage relationships hence why they are not used in signalling and instead find a use as power supply decoupling caps. Electros are also somewhat less stable with temperature so if it becomes exceptionally hot or cold their capacitance may change slightly. Speaker drivers themselves are pretty unstable with temperature so this shouldn't be a major concern.
The main concerns with using an electrolytic cap is longevity and power handling. Capacitors self heat due to their ESR. Heating is bad for electros because it dries out the wet electrolyte which causes the ESR to increase, which causes even more heating and so forth until they fail.
If anything you shouldn't place electros in the woofer filter where currents are higher. Kind of ironic since for some reason people seem to put ultra expensive poly caps in the tweeter filter and neglect the woofer filter. As long as electros aren't subject to excessive currents or stored in very hot ambient temperatures they should still give 10-20 years of service life or more.
If you need very good temperature stability, exceptionally high power handling or you simply want the crossovers to last 50+ years without degradation then a plastic film cap might be warranted. You have to ask yourself if the speaker drivers themselves will last that long. Other than those concerns, electros are fine.
Thankfully, the types of caps used in speaker crossovers (plastic film and aluminium electrolytic) do not have significant capacitance-voltage and capacitance-current relationships so they do not create appreciable amounts of distortion. Some capacitor dielectrics such as Class II and Class III multilayer ceramic caps have terrible capacitance-voltage relationships hence why they are not used in signalling and instead find a use as power supply decoupling caps. Electros are also somewhat less stable with temperature so if it becomes exceptionally hot or cold their capacitance may change slightly. Speaker drivers themselves are pretty unstable with temperature so this shouldn't be a major concern.
The main concerns with using an electrolytic cap is longevity and power handling. Capacitors self heat due to their ESR. Heating is bad for electros because it dries out the wet electrolyte which causes the ESR to increase, which causes even more heating and so forth until they fail.
If anything you shouldn't place electros in the woofer filter where currents are higher. Kind of ironic since for some reason people seem to put ultra expensive poly caps in the tweeter filter and neglect the woofer filter. As long as electros aren't subject to excessive currents or stored in very hot ambient temperatures they should still give 10-20 years of service life or more.
If you need very good temperature stability, exceptionally high power handling or you simply want the crossovers to last 50+ years without degradation then a plastic film cap might be warranted. You have to ask yourself if the speaker drivers themselves will last that long. Other than those concerns, electros are fine.
Last edited:
This is the kit I'm considering:
Strassacker, Komponenten: Lautsprecher, Frequenzweichen, Bauelemente
Do you think it's worth going the extra mile on the "high end Weiche" (high end filter
) ?
Strassacker, Komponenten: Lautsprecher, Frequenzweichen, Bauelemente
Do you think it's worth going the extra mile on the "high end Weiche" (high end filter
) ?
You do know how to speak kindly of caps
A year ago i cheapened out buying a set of ESA for a crossover. Could not stand them from the very first minute but still let them play for a month before discarding them. Not sure if even Wima are that bad.
MR, otoh, very, very nice. Annoyingly, the pricing in this case has a strong correlation to perceived quality. Perhaps boutique cap makers, unlike others, are not deaf
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.