2023/01/23 created

The crossover network CD-211B A-NET, which is a bit unstable, has been replaced by Behringer CX3400. (Feb. 2022)
Since CX3400 has signal delay feature, time alignment adjust is done with the feature. In addition, a power distributor, TASCAM AV-250S, has been installed.

Configuration

I repleced the crossover network, CD-211B A-NET, with Behringer CX3400, since CD-211B is a bit unstable. It's my first time to use a manufactured crossover network in Gaudi system.
With CX3400 the user can adjust the timing of the signals, because it has a signal delay feature (for only the low band).

I made special racks for the tweeters of SS-309A to place the tweeters just above the squawker drivers. That enables time alignment adjust between the squawker and tweeter. I call the rack 'Tweeter Rack'.

The amplifiers are the old ones I used to use in the old Gaudi system (Ver.1). I believe the amplifiers but the crossover network are not so influential for the overall sound quality and functions. So I still use the old amps.

---

Layout

The component layout is as shown on the right (click to zoom up).

The new crossover network (CX3400) is placed not at the spot where CD-211B was placed, but on the TV rack (J-8010) where the power amplifiers are installed.
In this revision, the power distributor (AV-P250S) is also employed. I couldn't find any other good place to install it, I placed it on the top of the TV rack. And, I placed CX3400 on it.

I put Tweeter Rack on the squawker horn (H400) and put the tweeter (T925A) on it.

The following figures show the layout of the whole system.

---

Cabling

The cables used in Gaudi II are as shown in the table below.
The all inputs and outputs of CX3400 are balanced type and the connectors are XLR jacks/plugs. I made cables for CX3400.

From	To	RefDes *1 (model#)	Wire	Length	Sender connector	Receiver connector	Note
ADP (SL-1200GR)	MC head amp (HA-213)		Low capacitance unbalanced wire (1-core shielded)	1.5m	RCA plug	RCA plug	SL-1200GR's accessary Tied with earth wire Stereo pair
MC head amp (HA-213)	Preamp (PA-210)		OFC balanced wire (2-core shielded) BELDEN 8412	1.5m	Teflon insulated RCA plug Cold-GND connected	Teflon insulated RCA plug	Unbalanced Stereo pair
DAC (UD-301)	Preamp (PA-210)	CB1	Quad-core shielded wire Canare L-4E6S	1.5m	RCA plug Canare F-10 Cold-GND open	RCA plug Canare F-10 Cold-GND bridged	Unbalanced Stereo pair
Preamp (PA-210)	DAR (MR-2000S)	CB2	Quad-core shielded wire Canare L-4E6S	2m	RCA plug Canare F-10 Cold-GND open	RCA plug Canare F-10 Cold-GND bridged	Unbalanced Stereo pair
DAR (MR-2000S)	Preamp (PA-210)	CB3 (Canare RC018)	1-core shielded (Canare GS-6)	1.8m	RCA plug (Canare F-10)	RCA plug (Canare F-10)	Unbalanced Stereo pair
Preamp (PA-210)	Network (CX3400)	CB5 (Canare RC03-B2)	4-core shielded (Canare L-4E6S)	3m	RCA plug (Canare F-10)	XLR pulg (Neutrik NC-MXX) COLD-GND bridged	Unbalanced Stereo pair
Network (CX3400)	Tweeter amp (MA-208)	CB12 (Classic Pro CXX050)	2-core shielded	1.5m	XLR jack Cold-GND bridged	RCA plug (Canare F-09)	Unbalanced Stereo pair
Network (CX3400)	Squawker amp (MA-215)	CB13 (Classic Pro CXX050)	2-core shielded	1.5m	XLR jack Cold-GND bridged	RCA plug (Canare F-09)	Unbalanced Stereo pair
Network (CX3400)	Woofer Amp (DAD-M100pro)	CB14,15 (Classic Pro CXX050)	2-core shielded	1.5m	XLR jack Cold-GND bridged	RCA plug (Canare F-09)	Unbalanced Mono x2
Tweeter amp (MA-208)	Tweeter (T925A)		OFC speaker cable Audio Technica AT365S	2.8m	Directly soldered	w/o plug Titan oil applied
Squawker amp (MA-215)	Squawker (D1405+H400)		OFC speaker cable Audio Technica AT365S	2.8m	Directly soldered	w/o plug Titan oil applied
Woofer Amp (DAD-M100pro)	Woofer (FW305)		0.65mm ETP copper single wire 47 Lab Stratos	1.8m (Amp to woofer terminal)	w/o plug Titan oil applied	Directly soldered	Two wires twisted at 30cm pitch

*1: Some of the line cables are given reference designators (CB1-3,5,12-15). They are specially designed for Gaudi II system.
For details, see the PDF below:
[Design document on line cables (LineCable_Design.pdf)]

---

AC Power Supply

The figure on the right shows the AC power connection (click to zoom up).

According to the system design Ver.2.3, the power distributor, TASCAM AV-P250S has been employed.
The power distributor has replaced two power strips. Now only the two wall outlets behind the audio and TV racks, which are hospital-grade, are used for Gaudi system.

AV-P250S has sequential power-on/off feature.
The ten outlets in the rear panel are divided into three groups: TIMING 1 (three outlets), TIMING 2 (three outlets) and TIMING 3 (four outlets). After the power switch is turned on, the three groups are turned on sequentially at about 2.5-second intervals in the order of TIMING 1 ==> TIMING 2 ==> TIMING 3. After the power switch is turned off, the groups of the outlets are turned off sequentially in the reverse order.

CX3400 is fed from TIMING 1, the power amplifiers are TIMING 3.
CX3400 generates a big noise at the power-on/off transition. But the sequential power on/off feature prevents the noise from being radiated from the loudspeakers.

The preamplifier PA-210 Simplicity isn't fed by AV-P250S.
PA-210 is a tubed amp and it takes about 20 seconds to warm up. The tube may generate a big noise shortly before its temperature reaches the steady state. The interval between TIMING 1 and TIMING 3 is about five seconds, and not long enough to prevent the noise from being radiated from the loudspeakers.
When I start the system, first, I turn on PA-210 , second, the player, finally, AV-P250S.

The power cable of the MC head amplifier HA-213 is too short to connect to AV-P250S. HA-213 is plugged to the power strip OCB-1 as in the previous revision.

---

System-level Tuning

It's my first time to make system-level tuning of Gaudi system using a microphone and an instrument.
I already experienced it with my subsystem Kinglet. I utilized the experience and made fine tuning of Gaudi II.

The new crossover network CX3400 has LOW DELAY feature, with which I made time alignment adjustment.
Without this adjustment, the frequency response could be bumpy due to comb-filter effect, but I haven't done it since I built Gaudi system.

I carried out the following steps:

• 1. Measurement of frequency response in quasi-anechoic method, and adjust of the gains of each band
• 2. Measurement of the delay times of each band with one-cycle sine wave, and time alignment adjust
• 3. Measurement of frequency response and waveforms at the listening position
• 4. Confirmation of sound quality by trial listening

It's necessary to repeat this sequence of the steps several times for fine tuning.
I repeated it three times to get satisfactory result.

In the first practice, the crossover frequencies are set to the values specified in the sytem design Ver.2.3: fc1=1,800Hz, fc2=6,000Hz. But both the data and SQ (sound quality) were not good.
Second, the crossover frequencies were set to fc1=900Hz, fc2=7,000Hz (the dial of CX3400 indicated 8kHz).
Third, I changed the method of time alignment adjust.

From here, the outputs of each band of CX3400 are denoted by LOW, MID, HIGH.
The suffixes '-L' and '-R' denote the left and right channels, respectively (e.g. LOW-L, MID-R).

Quasi-anechoic Measurement

I once tried quasi-anechoic measurement with Gaudi Rev.2.1 for training (April 2021), though, it's my first time to do it for system-level tuning.
For details of quasi-anechoic measurement, see the section "Frequency Response (1m away from the loudspeaker)" of the supplemental information "Method 2: Measurement Methods".
I set the parameters in the near-field measurement as follows: the distance between the woofer and microphone, d=13mm, and the merge frequency, fm=460Hz.

At first, I didn't understand how CX3400 worked and got puzzled. As I went on measurement, I understood how to use CX3400 (for details, see this).

1st measurement: I wasn't used to CX3400, and the crossover frequencies turned to be fc1=2kHz, fc2=4kHz.

The figure below is the frequency response after the gains were adjusted.

The wave-like shape between 600Hz-2kHz may be a problem.

2nd measurement: the crossover frequencies were changed fc1=810Hz, fc2=7kHz, and time alignment was adjusted.

Toneburst sine waves whose frequencies were equal to the crossover frequencies were used for the time alignment adjust.

I forgot to save the data. I recall it was like the third measurement.

3rd measurement: the crossover frequencies were the same as the previous ones, time alignment adjust was done in a different way.

The line cable connected to the outputs of CX3400 was customized in this measurement; COLD and GND were bridged (recommended by Behringer).

In the time alignment adjust, I used toneburst signals whose frequencies were the middle of each band; LOW: 200Hz, MID: 2.5kHz, HIGH: 16kHz.

The responses of the squawkers are not good. As frequency increases, the SPL decreases. The left squawker's reponse decreases more. This deviation between channels is a problem.
There still is the dip below fc1.

Time Alignment Adjustment

For details of time alignment adjustment, see the section "Time Alignment Adjustment" of the supplemental information "Method 2: Measurement Methods".

1st adjustment
I adjusted the delay times of each band at the crossover frequencies, fc1(=810Hz) and fc2(=7kHz) to make all the delay times the same.
The delay time of the MID (squawker) is the reference. The delay of LOW (woofer) is adjusted by LOW DELAY of CX3400, and HIGH (tweeter) is by moving the tweeter back and forth.

The maximum value of LOW DELAY of CX3400 is 2msec. I found out it's not long enough. Longer than 2msec is necessary.
The calculated value is about 0.7msec (=0.23[m]/345[m/sec]x1000) where 0.23m is the on-axis distance between the woofer's and squawker's diaphragms and 345m/sec is the sound velocity.
I suppose this misfit is caused by the phase shift at the crossover frequencies.

At first, I placed the tweeter just above the squawker driver, but it was not enough. The tweeter must be withdrawn further.
I moved Tweeter Rack backward as possible.

Before adjustment		Tweeter Rack at the back end of Woofer Box	Tweeter Rack at this position doesn't good look (doesn't fit w/ H400)

2nd adjustment
I adjusted the timing of each band by using single-cycle sine waves whose frequencies were in the middle of each band: LOW: 200Hz, MID: 2.5kHz, HIGH: 16kHz.

The LOW DELAY turned to be 0.7msec as calculated.
I don't why but the position of the tweeter was the same.

I noticed a big issue when I observed the waveforms during the adjustment. The sound wave from the squawker was followed by very long ringings. It must be a problem of the squawker driver (to be described later in "Self-evaluation").

Characteristics at Listening Position

The chart below is the final data.

The curves in treble look almost the same as quasi-anechoic measurement.
In bass and midrange, the chart exhibits dips unlike quai-anechoic measurement. Especially, the dips just over 100Hz are deep. The curves gradually fall under 50Hz.

The chart below is the result of RTA.

RTA exhibits the poorer treble response than sine wave sweep. The attenuation of SPL from 2kHz to 6kHz was -8dB (L-ch) and -2dB (R-ch) in the sine wave sweep. They are -13dB (L-ch) and -7dB (R-ch) in RTA.
The bass response is flat in RTA un like the sine wave sweep (the response at 20Hz or lower is an error in measurement).

The RTA result also exhibits deep notches just above 100Hz. These notches are caused by the room acoustics because they didn't appear in quasi-anechoic measurement.

I think the curves by sine wave sweep better correspond to my perceptual frequency response than RTA's.
On the other hand, I realized RTA is more useful to adjust the gains of the crossover network.

Trial Listening

1st trial listening
I was surprised by terrible sound! I couldn't stand it.
I changed fc1 from 1.8kHz to 900Hz.

2nd trial listening
Gaudi R2.11 still didn't sound good. Something was different.

3rd trial listening
SQ improved, though, Gaudi R2.11 still sounded worse than R2.1 of which crossover network was CD-211B.
As for stereo imaging, I think R2.11 was a little better than R2.1. Localization was stable.

---

Room Acoustics

The room acoustics is the same as Rev.2.1. I continued to use the acoustic goods I had employed for Gaudi with some exceptions.
For details, please see the page "Problems in Gaudi".

---

Self-evaluation

I struggled to install the product (CX3400) to my Gaudi system, just as I spent much time and labor to introduce PC audio to Gaudi Ver2.1. I had thought it easy to install a product because all I had to do should be put it in the rack, connect cables and make a little adjustment. But it isn't true.

It is recommendable to make measurement of the product before installing it. By doing it, you can learn the strong and weak points of the product, and sometimes find a big issue. This is a sort of precaution to prevent unwanted troubles.

Sound Quality

I said above that the SQ of Gaudi Ver2.11 was so bad I couldn't stand it. Howerver, I continued trial listening for another month, as I always do for my DIY works. Listening a lot is necessary to evaluate SQ.
The SQ turned to be better gradually. Finally, it is almost as good as Gaudi Rev.2.1 whose crossover network was CD-211B . I've never experienced such amplifier that its SQ changes so much over time. My NOBODY amps except tubed type never change in SQ.

I noticed one more thing; CX3400 needs warm-up.
Shortly after power-on, CX3400 sounds like a mono amp. The sound stage is so narrow. About 10 minutes later, CX3400 (and Gaudi R2.11) comes to sound normally. It's my first experience too.

I think CD-211B is a bit better than CX3400 in SQ. But it could be thanks to the inaccurate operation of CD-211B. Some circuit in CD-211B is untstable. In addition, the slope of the filters are 18dB/oct and not as steep as CX3400 (24/dB/oct). These may cover up the shortcoming of the squawker driver D1405.

Shortcoming of D1405

In fact, I knew D1405 has the shortcoming when I bought it in Oct. 2011. As the catalog of D1405 says, it has a poor treble response. The peak of SPL is around 1kHz, and SPL constantly decreases as frequency increases. The SPL at 10kHz is 10dB lower than at 1kHz.
This driver isn't suitable for a Hi-Fi system. But I still wanted it, because it is a product of Fostex, which I believe to be one of the best loudspeaker manufacturers, and it was as expensive as 30,000 JPY (the list price was more expensive). Moreover, I had longed for horn-loaded squawkers since I was 17. I was so excited and lost my critical attitude. I thought it to be a minor problem.

The catalog of D1405 offers a typical application where 12" woofer W300A, wood horn H400 and tweeter T925A are used together with D1405. In the network, 2nd-order filters (12dB/oct) are used for woofer and squawker, but for tweeter 1st-order filter (6dB/oct) is used. In this scheme, the sound emitted from tweeter overlaps much with that from squawker, and this makes up the poor treble response of D1405.
I don't like this method, because it spoils the merit of multi-amplifier system.

When you use an LS unit whose frequency response isn't flat, it's hard to adjust the gains of each band properly.
When it comes to D1405, the following two settings are applicable.

Smooth crossover Excess bass and poor treble	Balanced frequency response Crossover isn't smooth

I chose the left one, because I focus on the smooth crossover.

I don't have any solution for the ringing of D1405 but replacing the driver.
The ringing gets longer as the frequency decreases. D1405 seems to lack damping.
However, I didn't notice such terrible ringing when I made measurement of Gaudi II in 2017. My D1405 could be broken!?

Appearance

I placed CX3400 and AV-P250S on the top of the TV rack, which is the most conspicuous position. But they aren't unpleasant to the eye, contrary to my anticipation.
I was afraid many LEDs in the front panel of CX3400 would be offensive to the eye, and the AC outlets in the front panel of AV-P250S would also appear ugly. But I was wrong. I quit placing half-transparent aclyric panels in front of them, unlike the system design.
By the way, I will move AV-P250S to the bottom of the audio rack AR-421 when the rack is completed, as specified in the system design Ver.2.3.

Tweeter Rack doesn't look good. I should have well-designed it. It's finish is no good as usual (I'm not good at painting).
If CX3400 is replaced by a digital crossover network, Tweeter Rack is no longer necessary (the digital crossover can delay signals in treble too). I determined to use a digital crossover in the system design Ver.2.3, so I'd like do so in the near future. Then, I'll remove Tweeter Rack.
However, there's a possibility I would change my mind and use an analog crossover (maybe CX3400) once more, if I wouldn't be happy with the digital one. I won't throw away Tweeter Rack. I'll re-paint and save it for the future use.

Thanks to the power distributor, the number of the wall outlets used for Gaudi II has been reduced from four to two (every wall outlet is dual type). Now only the outlets behind the audio TV racks are used. The cords of the power strips are hidden from the eye.

The cord of the power strip (for ADP) was seen in the bottom left corner		The power strip has been removed
The cord of the power strip (for TV, STB and BDP) was seen on the right of the loudspeaker		The power strip has been removed

User-friendliness

Thanks to the power distributor, turning on/off of the amplifiers became easy. As mentioned above, only the preamp is still fed from the power strip. I'd like to replace it with a semicondctor amp and connect it to the power distributor as soon as possible.

---

Wrap-up

SQ has gradually improved as I continued to use Gaudi Rev.2.11. But I think Rev.2.1, where CD-211B was used, was a bit better.
Rev.2.11 is a temorary revision. I'd like to improve Gaudi to the next revision as soon as possible.

I believed Fostex D1405 is the best squawker driver for many years, but now I can't.
I am planning to employ another driver in the next revision.

I re-recongnized during the trial listenings that the sound of vinyl has something different from digital source. It has a strong healing effect. I have a strange feeling when I listen to it as if it vitalizes cells all over my body. It affects not only the state of mind but physical condition. For example, the analog sound heals fatigue. I even experienced it cured my constipation. Even Hi-Res digital doesn't have such effect.
The system design Ver.2.3 specifies using a digital crossover network. I wonder if the digital crossover spoils that healing effect of analog source. If it does, I'll reconsider an analog crossover to replace the digital one.

---