彙整 | 二月, 2007

全音 vs 多路分音 之四 : Micro Utopia 的 烏托邦

24 二月

在過去二十幾年的HiFi經驗裡,令旺記折服,又在射程範圍内的喇叭不會太多,其中使旺記没有猶豫便落鎚的是 JM Lab 的 micro utopia。在 micro utopia 之先JM Lab 先出 grand , mini 和 mezzo。Grand 和 Mezzo没有想過,Mini utopia 聽過後沒有太大亢奮,但 micro utopia 一出,旺記聽過一輪,相信 micro utopia 攪對了 phase coherence,立刻陶腰包付錢搬貨。

若果要Micro Utopia出低頻,旺記衹能說 no comment,因為它在五十週以下是行人止步。不過, Micro Utopia的強強強強強項,絶對是 phase coherence。

找個好的人聲錄音,加些千絲萬僂的伴樂, 立刻可以聽到Micro Utopia的相位是 “做對了"。誰説人對相位不敏感?

找來 Micro Utopia的 step response。甚像一個全音單元。

Square wave response 我都測過。好到不敢相信。所有測試都已是七年前的事,當年旺記手上的全音單元,統統都不是它的敵手。

後來手緊,賣了。Wrong decision。

我的 Micro Utopia上過 MJ。現在留下的唯有個體回憶。旁邊的是Lowther,衰聲。

新的 Be 好不好? 旺記不敢説不好,賣貴50%,可以怎說?

Q factor 的聯想 ( 3 )- 再論 class D damping factor

22 二月

過年前又往北京出差了幾天,凍得要命又括大風,我坐的港龍搖滉得很厲害,機翼的 flaps 不受控地上下拍動,如此強勁的 sheer wind 我是首次領教。結果飛機要試了三次才能着陸,機裡的小朋友都嚇喊了。

記得在西安與國内同事閒談北京,他劈頭便說 : 那些並不是甚麽好地方 …

農曆年假和朋友吹水 hifi 又講到 damping ,朋友說開始明白 class D 並不太適合推全音單元。


Class D 輸出阻抗在零點幾 ohm(若是 filterless 可能得零點零幾),接近 contant voltage source. 用 constant voltage source 直推單元,隨時 overdamp 也。現代多路單元預咗要用分音器,仲要係 fourth order 的,加上分音器後 system response 增送給你幾個 pole,加埋分音器的損耗,實際的 Q 值就會大許多,system response 甚至有 overshoot的情形。所以並非直推便一定好聲,但要玩直推(包括電子分音)或簡約分音如一介分音則應選 Qts高,而更重要係 Qms高的單元。


Interview with Joachim Gerhard of Audio Physic

By Lars Mytting
29 Jul 1996


– Many “old" paper woofers still sound astonishingly good compared to modern drivers?

– Oh, yes. We have not always went to the better. What many driver manufacturers have done the last years, is to increase the damping to make the frequency response more flat. But some old drivers, like the famous 6,5″ paper woofer that Jan Paus at Seas made several years ago, (The Seas CA 17 RCY, ed. note) was optimized for low loss. So they made a compromise between frequency response and sensitivity. This driver was very good, and was used by Wilson Audio for many years. Later, in the 80’s, manufacturers started to add more mass, they added more damping, and they made surrounds with high loss. That gave an extremely flat frequency response, but also a lot of energy storage. This compared, the old drivers were much quicker. They had some resonances, but you could get rid of that in the crossover. It was this run for flat response that gave a lot of modern drivers this dull, uninteresting sound. And you can also measure higher second and third harmonic distortion in some of them. If you compare the on-axis response between an old and new driver; you will see that the energy in the treble is far higher than in the new drivers. These so-called “modern" drivers often has a Qms of maybe 0.8 or 0.6. The old drivers had Qms values of maybe 5 to 7! We found that drivers with a very high mechanical Q sound more open, more clean and dynamic. And when you look at it, you find it is very simple, because they have less loss. The surround is easier to move, the spider is better constructed, they have better air flow, higher sensitivity. So a high mechanical Q is a very good indicator of energy storage behavior. This is one of our secrets. One of the many!

下次再同大家分享 system response 的 overshoot。有 overshoot 就有好聲,冇老點你,stay tuned.

德國全音篇之 葵花寶典 (十)

9 二月



没有甚麽驚喜,可能看的 documentary 太多了。

倒是星羅奇佈的紀念品店子,數以千計的自篤導遊,煞有介事的區内電動車(set up 個售票店又賣車票,原來行程就係约一百米咁大把,仲話包來回),區外幾間大型的兵馬俑仿製廠…完全符合國情。


續 葵花寶典 (九)

“I love the 8 incher with matching super-tweeter. So you need that super-tweeter too, and evidently of the same good condition. And you want a second pair for stereo. I had to learn the hard way: Within those jolly good times a minor change in production, say getting another speaker to stuff it in, was never announced. You’ll see brand names (same year of production, same type number of radio) with different speakers, sometimes drivers were changed twice a year! And I can’t help you in finding out, you’re miles out in the cold and have to check yourself. Therefore it’s not so bad to hunt for those radios which once were spread out in tens of thousands. Thus you work your way through hamfests, garage sales. flea markets and the like, always carrying with you a collection of photos showing what you’ve already achieved, and of course you pinned down brand name, year of production and type number on each photo, didn’t you?


Old style radio speakers weren’t built in stereo pairs. To match them now measure two electrical quantities: DC resistance and inductance of the voice coil. A good match stays within 25%. Getting all this together the brand name doesn’t matter any longer – if you indeed got the speaker’s name, a sticker alone from the radio manufacturers name can mean nothing, as the speaker can be produced elsewhere, Talking of 8 inchers till now was neglecting reality a bit. Yes, it’s the size but not the shape, most of them aren’t circular, they’re oval, roughly 7 to 10 inches. It would be the audiophile way to start a fight on princjple – don’t do it. Go for those you get in good condition and good matching no matter if oval or circular. They’re fine anyway."

要執平嘢嗎?Elliptical的單元一點都不差過八吋的 counterpart,旺記甚至認為 elliptical的 tonal balance 更好。兩者市塲差價這樣大,完全是心理因素。

Q factor 的聯想 ( 2 )- 胆石之爭

2 二月

因為最近測試了幾款 Class D 放大器,感覺"大 力"放大器時代已經降臨。

且慢! 大力不等於低頻勁。

大力 是甚麼意思?


如果一部放大器的 輸出阻抗有零點零幾 ohm 的話,可以輕易有以下 spec :

50W 8 ohm

100W 4 ohm

200W 2 ohm

400W 1 ohm

800W 0.5 ohm ! ( provided that 唔會燒機)

不過,即使如此,不代表 低頻會谮得深。

上文已經講過,70年代的石機全部可直達 DC 。

影響 低頻表現的乃係 揚聲器的 Q 值也。

個別 揚聲器單元的 Q 值一般在 datasheet 上可找到,要測也不難,成品揚聲器則很少會提供,你知得越少,佢越着數。

不過大概一般人都没有留意,datasheet 上,或我們 测出來的 Q 值,乃係電壓轉移函數上的 Q。


Q 值越低的單元,阻尼越多,在 fo以下 roll off得很快。

最出名的是 Lowther ,Goodmans Axiom 80 。

因為 Q 值低,非用 BLH 放大低頻不可。

但係,如果 放大器輸出阻抗高,單元Q值就會提高。


Q值 其實分為 mechanical Q (Qm) 和 electrical Q(Qe)。

Qe 乃同音圈 磁塲成反比。放大器輸出阻抗高,輸出電流少,Qe便高。

要 Lowther出多D低音其實不難,增加 放大器輸出阻抗是也。石機先天高增益,要用負回輸, 輸出阻抗必然低,相反無負回輸的胆機輸出阻抗則高十倍八倍。

以前有前輩玩 Lowther 係用幾百瓦石機的 headphone jack 的。

旺記玩Goodmans Axiom 80 用高輸出阻抗推到拍邊都試過。當然不鼓勵。

上圖為曾登於 IEEE 某 transaction paper 的一張圖。 圖中可見胆機低頻比石機多好多。原文作者話此乃 揚聲器 reactive impedance 與胆機阻抗的相互影響。

計我話, 此圖中胆機阻抗高,揚聲器 under-damped也。

在同一 power rating 下胆機力水大過石機,旺記認為此乃主因。