Genuine Amplifier

I bought a genuine amplifier for my loudspeakers, QUAD. The loudspeakers are QUAD ESL57, electrostatic loudspeakers designed 50 years ago, and the amplifier is fairly old QUAD 606, which is relatively new, but born to drive electrostatic loudspeakers. I did not expect a lot, because its design is not modern. They designed the amplifier as small as they can make, with minimum power consumption. It is completely different style from true A class amplifiers from PASS lab., Mark Levinson, KRELL and other high-end amplifiers, which are concentrating on the quality of the sound and do not care, (or more precisely, emphasizing their passion by neglecting the other specifications on purpose). They are expected to drive any loudspeakers including extremely hard to drive electrostatic ones. But, QUAD 606 seems designed in completely different manner.

What happened is far beyond my expectation. It is designed just to drive nothing else but that specific loudspeaker. And I can understand that they do not want to listen to any of the audio equipments, but just to the music.

I am still working on the amplifiers, and some other components. I have to jump over this high hurdle from UK, to be happy.

Balanced Digital Cable (continued…)

There is a big improvement archived by the balanced digital cable with controlled impedance. I did not expect that significant difference, by modifying digital connection, but what happened was drastic, i.e. the accuracy of just 0s and 1s makes difference by improvement of the circuit that carries just 0 and 1 signals. Some people explain that by improvement of jitter.
However, there is another difference between digital and analog world, at least in my case. Analog circuit tends to change its sound over time as it ages, toward smother sound in most case, and that is clearly recognizable. In contrast, I cannot notice any “aging” in digital circuit, at least in my case.
The improvement is applying analog experience to the digital signal path, and the improved circuit also ages. I may expect “aging effect”, but what happens is different. I suspect that there should be some threshold of jitter that I can notice the difference.
If I have some time and funding to do the research, I would commit subjective test and statistical analysis on this topic. I am quite confident that there is significant difference in “aging effect” in analog circuit but not in digital circuit.

But, that is just a dream.

Balanced Digital (SPDIF) cable

I made a simple SPDIF cable using Category-5 cable, and I feel great improvement over an optical fiber or a coax SPDIF cable. It is a modification of Zuka’s cable driver. The input jack, which plugs into the digital-output device like CD player, is directly connected to the cable driver part. One CMOS inverter is used for threshold, then the signal is converted to balanced output by LTC1485. 　I designed that the power can be supplied from the local transformer, or from the DA converter side via unused pair of Category-5 cable. It is possible to share same power supply with receiver, but I am not sure it is good for the sound or not, and I use a transformer at the input side to power the cable driver.

I also implement a receiver. The first version uses just a pulse transformer at the receiver side. It worked, but the sound quality was far from what I expected. Even an optical cable sounds better than the first setup. Thus I put a 110R termination register. There is some improvement over the first version and sounds better than the optical fiber, but it is still not as good as very short coax SPDIF cable. I decided to get rid of the pulse transformer and use a differential receiver, just as seen in the datasheet of LTC1485.
The key improvement might be achieved by known-impedance cable and proper termination. Although the cable is outdated (category 5 is not commonly used now, and 5e or 6 are popular for 1G network), it works great.

Brick wall made of capacitors

I have made another power supply for the balanced-output preamplifier.
There are 2 major improvements:
(1) Electrolytic capacitors replaced with ceramic ones
(2) Connection of the audio transformers as chokes modified.
Some enthusiasts in Japan use surface mount ceramic capacitors instead of electrolytic ones for power supply. Each of the ceramic capacitor looks like a brick, and quite a number of chips must be piled up to archive sufficient capacitance, it is called “brick wall.”

In my case, it is parallel lanes rather than the brick walls. I have curved 2 trenches on cupper surface of PCB, and soldered over 100 chip of caps along each lane. Total capacitance should be around 700uF each, which is little more than 1/10 of the previous one. I am planning to increase the capacitance by making 3rd and 4th “lanes.”

Telecom transformer that I have used is about 1:1, primary and secondary circuits are connected to each of cathode (positive) and anode (i.e. plate and grid, for negative). DC component is canceled with this connection, and yielding more effective inductance.

I expect that this 2 transformer, tube rectified, choke-input and non-electrolytic filtering is the combination of best available. But, its performance is not as good as expected, I have observed considerable ripple (about 500mV peak to peak), and sound quality is not as good as the electrolytic one, electrolytic one is more “stable” & “confident” sound
than new one, and ceramic one has somewhat “weaker” sound. But I expect that would be better with 2 other streets and as it burns in.

「壁」を作りました
部品が有ったのでもう1つ真空管整流の電源を作りました。今度は「壁」を作り、チョークの配線を変えました。生基板にカッターでスリットを2つ切り込み、積セラを片側100個少々づつ並べて半田付け。容量は片側700uF位、電界コンデンサ版の1/10強です。外観上は「壁」ではなく平行線で、近々に3丁目、4丁目を追加する予定です。他に、電話用のオーディオトランスも、1:1なので正電源では、カソードごとに、負電源ではアノード（プレート＋グリッド）ごとに1回路を割り当ててDC成分をキャンセルするように配線しました。これで、実効のインダクタンスは増えているはずです。
正負別トランス、真空管整流、チョークインプット、壁平滑は「最強」になる予定だったんですが、さにあらず。リップルが500mv(p-p)ぐらいあり、音も電解バージョンは安定してどっしりとした感じに対して、新作は弱々しい感じ。エージングして容量を増やせばまともになるのではないかと考えているのですが…
(すいません。だんだん日本語がおぼつか無くなってきているようで、日本語のスペルチェックが引っかかっていますが…)

Hand woven cable

I made a pair of line cable between balanced pre-out to balanced power-input. The cable consists of 3 braid cables because the balanced line requires 3 lines. It is unusual because,

• Unshielded line cable
• Line cable made of split low-end Monster　speaker cable
• Looks like fancy Kimber cable, but it is not.

I have not tested it because the power amplifier is still under construction. But, it should be okay. Its signal level is little more than the power amplifier’s output level and the cable is driven by low impedance operational amplifiers.

more on power supply

The circuit is really simple (see figure).
These tubes are two of 7370 (40V version of 5687WA). Heater is connected to AC input along with 270R (for 117V AC), or 150R (for 100V AC).
It is strongly recommended that average output current does not exceed 20mA, because the average grid current is rated 6.6mA/unit (maximum).

Power Supply for Discrete OPamp preamplifier

I made a vacuum tube power supply for semiconductor (aka Kaneda #168) preamplifier.
I have surplus dual triode (7370). It seems it is tough enough to use for a small output rectifier with Grid and Plate connected together. Thus dual triode performs like a full wave rectifier. 24V-0-24V transformer is rectified with it, yielding 35V output (no load). I put a small telecommunication transformer as a choke and current limiter to protect the tube from the rush current to the capacitor (2 of 3300uF).
Two almost identical 35V power supplies produce +/- 35V w/o load, and around +/- 25V with 2 discrete opamps.

This power supply is for a balanced output preamplifier to drive modified dual-Zen9, or two-SEWAs. I am planning to use 2 SEWAs par channel with this preamplifier. One advantage of the bridge-connected class-A amplifier is that it consumes constant current, independent from the music signal.

The housing is an emptied ATX power supply. It has many slits for cooling, so I do not have to poke any hole. Driling is a pain for me.
Modern ATX power supply now contains vacuum tubes, and used for an analog amplifier. Isn’t it cool?

金田式＃１６８モドキのプリアンプ用に、真空管整流式の電源を作りました。余分にあった7370という双三極間、かなり丈夫そうだったんでグリッドとプレートを接続して全波清流間として使っています。電源トランスは24V-0-24Vで、無負荷で35V、コンデンサ（3300μFを2本パラ）へのラッシュカレントを防ぐ目的を兼ねて、電話用のオーディオトランスをチョークの代わりに入れているので、チョークインプット（モドキ）として動いている予定です。
ほぼ同じ回路を2つ作り、無負荷で+/-35V,オペアンプ2個の負荷で正負約２５V になります。
この電源、アンバランス入力からバランス出力になるようなプリアンプのために作成しました。Pass Zen V９やSEWA（７Watt amplifier、定電流負荷のA級1トランジスタのソースフォロアパワーアンプ）を2台、BTLでつなげるためのものです。

電源のケースはATX電源の流用。冷却用のスリットがたくさんあいているので手間のかかる穴あけは1つもしないで済みました。