DiscoCAT Speaker Design

[Fatmarley]

Hope you don't mind me linking these here S-man - More caps for sale in the classifieds...

Polypropylene capacitors, Mundorf, Audyn, Jantzen

I'm selling all my capacitors for half price, plus postage (usually around £4). Let's start at the most expensive, and my favourite capacitors. If only I could design a speaker around a single 2.7uF capacitor. There would be no way in hell I'd sell these. Not cheap, but by far the most natural...

pinkfishmedia.net

 

[S-Man]

No problem.

 

[S-Man]

Not surprising, reactance is 796R at 20kHz! For comparison 6.8uF is 1.2R.

Yehbut...

Humble Homemade Hifi - Cap Test

Humble Homemade Hifi - Very high quality loudspeaker kits, components, upgrades, modifications and custom solutions.

www.humblehomemadehifi.com

"
Cornell Dubilier 940C / 0,01uF / 3000VDC bypass - 10% tolerance

Technical specifications (according to manufacturer): "Type 940 round, axial leaded film capacitors have polypropylene film and dual metallized electrodes for both self healing properties and high peak current carrying capability. This series features low ESR characteristics, excellent high frequency and high voltage capabilities; high dV-dt for snubber applications; capacitance range 0,01 to 4,7µF; capacitance tolerance ±10%; rated voltage 600 to 3000 VDC (275 to 500 VAC, 60 Hz); operating temperature range with ripple 55ºC to 105ºC / full rated voltage at 85ºC - derated linearly to 50% rated at 105ºC; insulation resistance more than 100.000 MO x µF; test voltage between terminals @ 25ºC 160% rated DC voltage for 60s; test voltage between terminals & case @ 25ºC 3kVac @ 50/60 Hz for 60s; life test 2000 h @ 85ºC, 125% rated DC voltage; life expectancy 60.000 h @ rated VDC 70ºC; case material UL510 polyester tape wrap; resin material UL94V-0 epoxy fill; terminal material tin plated copper."​

Sound: Using the 3000VDC Cornell Dubilier 940C as a bypass capacitor in the value 0,01uF really is the "icing on the cake" when you are looking for a low cost bypass capacitor that gell's everything into one coherent sound. Best bang for the buck! Over the years I have used these in many different combinations and everytime they just added more to the overall sound. More coherency and more ease. Especially when you have been "cooking" to find your favourite mix of capacitors and only need to cement the different characters together. Adding the 940C creates an overall more mature sounding speaker. And no, they are not better than the Duelund or VCap bypass capacitors that excel's in naturalness and richness of tone and in transparancy and air. The Duelund and VCap are in another league, not only in sound quality but also in price ;-) Compared to the Jupiter Copper Foil Paper & Wax used as a bypass, the Jupitor is has richer overtones and has a tonally slightly warmer balance. The Vishay MKP1837 used to be my favourite low cost bypass capacitor but now it is the Cornell Dubilier 940C / 0,01uF / 3000VDC. It surpasses the Vishay with quite some margin ​

Verdict: use them everywhere possible!"

I was hoping the above might be true. Although I suppose I haven't used them everywhere

 

[martin clark]

I've always found the idea of 'add 10nF of magic, just in case' to be abs bobbins in any use case, as universal advice.

Whatever it is - just use one, good/ appropriate cap.

 

[S-Man]

Well I might have to eat humble crumble!

The magic bypass cap does seem to have an audible effect . Bypassing the Bevenbi caps seem to slightly improve the clarity and calm things down a little. Anyway, the Jantzen caps arrived and were duly fitted along with the magic bypass caps. I was distinctly unimpressed with this combo – poor dynamics and a recessed upper mid range, so I disconnected the bypass caps and things are now much better.

More listening required, seems like the whole thing is a bit of a minefield.

 

[S-Man]

Talking about humble homemade hi-pie. I don’t agree with Tony Gee.
Don’t use bypass caps everywhere, maybe not even anywhere!

The DCs with the Jantzens are sounding wonderful. There is a definite increase in refinement without losing any fun factor. The tweeter sounds like a much more expensive one now. I will have to wait to see if it equals the ATC one, but I definitely would not swap the overall sound for SCM7s - good as they are.

 

[James]

The DCs with the Jantzens are sounding wonderful. There is a definite increase in refinement without losing any fun factor. The tweeter sounds like a much more expensive one now.

You're welcome.

 

[James Evans]

So much of this is subjective and room/system dependent, and sadly is very much buy before you try

 

[S-Man]

Yes, although it worth getting advice from competent and rational forum members.

I’m always a bit suspicious of gurus. How anyone can rate 10s of caps over a period of years on a points scale is beyond me! I’m always concerned when price and performance have a direct, albeit logarithmic, correlation.

 

[Fatmarley]

I’m always concerned when price and performance have a direct, albeit logarithmic, correlation.

I'm not sure they always do. I like the cheap mundorf mcap (white ones) and supreme version, but thought the much more expensive silver/gold/oil were terrible.

 

[S-Man]

I'm not sure they always do. I like the cheap mundorf mcap (white ones) and supreme version, but thought the much more expensive silver/gold/oil were terrible.

That's my point: It makes me suspicious when there is a correlation.
I have been disappointed many times by expensive parts - generally boutique stuff.

 

[S-Man]

Thought I would try my V3 crossover again (as in post #16). Initially it sounded OK, quite clear and good detail. But after a while, I lost interest in listening. Reverted to the latest one... much better. Hard to describe why it's better, it's just more engaging and has a bit of magic missing in the V3 (and most other speakers).

So I'm running out of ways to improve the DCs now. The tweeter cap upgrade was very worthwhile and I'm still tempted to try to interior "pyramids" to reduce non-parallel surfaces.

What about resistors? I thought Mills were better than cheapo ceramic cased ones in the CAOW1s, but some say they sound a bit clinical.
Any thoughts on Jantzen Superes versus Ohmite Audio Gold? The former have a datasheet with almost zero info.

 

[Fatmarley]

Have you measured off-axis? I know it can be a pain to do, but it can reveal things in the frequency response that help make sense of things.

 

[S-Man]

No, I have been lazy in that respect. I have got as close as I can to ideal LR2 slopes, which ought to define part of the polar response. The other factor is of course the XO frequency. The later XO crosses lower, which should help the power response at the expense of distortion.

 

[S-Man]

Well, I got to compare the DCs to the ATCs again the other evening and the result was pretty similar to at my pal’s house. The DCs are slightly better everywhere except the treble. The ATCs are just a touch lean but do have a wonderful clarity and a distinct lack of any sort of coloration or distortion.

So I set about trying to get the best of both worlds. Unfortunately, I got distracted by a different active crossover which certainly red herringed me for a while.

I built another crossover V12 which looked excellent in simulations but was just not that engaging to listen to. Unfortunately, I tried V12 only with the red herring crossover.

So I’ve reverted to V9 crossover but taken out the Notch filter that “flattens“ the 10 kHz area. This is because I realised I had an error in my microphone calibration file and actually, the 10 K bump was caused by the microphone. Funny how I expectation-biased my listening impressions based on (incorrect) measurements!!

I’ve also done a bit of tweaking with tweeter level.

Anyway, I think I’m getting to the end of the development phase. They are sounding superb at the moment, and I think I’ve got the blend of clarity and body spot on for my tastes. I nickname this “edge to middle ratio“, like a good pizza it has to be spot on for maximum enjoyment .

 

[Toaster]

I'm intrigued by the base extenders you often mention S-Man. Looking back over some of your previous posts, I see you've referred to them as 'ZRB Bass Extenders', but a broader search beyond PFM doesn't seem to return anything relevant. Are you able to provide more details about these, or are they still potentially a commercial product?

 

[S-Man]

To be honest, I try to keep quiet about these, because nobody else can obtain any. However, in the context of this thread I felt there was no choice but to mention them because this had such a huge impact on the design choices for the DCs.

ZRB (zero resonance bass) came about after a very lengthy and in-depth investigation into the fundamental principles of drive units and how to control them, by a friend of mine. He invented a method of cancelling out some of the limitations and nonlinearities of drive units. The end result is that the driver can be driven below resonance and with an order of magnitude less distortion using his correction principles.

My side of the deal was to build two pairs of prototypes along with the PCB design for the correction circuit and finding amplifier circuits that were suitable. These analogue prototypes are still in use today 10 years later.

The principles have been commercialised in the form of the KEF KC 62 subwoofer. You can read KEF‘s interpretation of the technology and the benefits in their white papers. Personally, I think KEF missed an opportunity but the way they’ve marketed these. And for my purposes, I don’t want the delay of the DSP adding complication to the signal sent to the satellite speakers.

So I guess you could say that the ZRB is just a fancy subwoofer. The fancy part being that in a modest box (100 x 20 x 30cm) which doubles as the stand for each satellite speaker the bass can be extended pretty much as low as you want. I’ve chosen a first order rolloff at 10 Hz. Although this sounds ludicrously low, it provides the lowest possible group delay at audio signal frequencies. When I did the auditioning, I was shocked to find that I could hear the difference between a 20 Hz corner frequency and 10 Hz!!!

In summary: the ZRB is just a fancy sub .

 

[Toaster]

As described it sounds conceptually similar to motional feedback- I'll have a look at the white paper. Are specifics impossible to share because KEF now owns the design?
[Edit]: OK, so yes, KEF has patents pending on aspects of the design, and they describe the distortion cancellation as a combination of motional feedback correction by way of voice coil current sensing, and secondly the use of DSP to create an inverted model of the drivers electrical and mechanical parameters- or some of them- and use that to cancel distortion. From what you've said S-Man, the original design presumably used analogue electronics rather than DSP to do a similar job for the second aspect of the design?

 

[S-Man]

Yep although I think my pal is allowed to do his own system and market it if he wants to. Actually, he did want to and he wanted me as a partner, I was reluctant because I don’t want to turn a fun hobby into a bit of a chore.I was even more reluctant because I know how many hi-fi consumers like to model the very latest set of Emperor’s clothes .