Rerurn to Romy the Cat's Site

In this thread I would like to talk a bit about the practice in DSET driven multi-way horns configuration.

I have expressed my view that the architectural configuration that Macondo and Super Melquiades end up to be I find is the most interesting  and the most capable among any another possible configurations of playback.  The “active” configuration when the naked channels are driver by DSETs has so many theoretical and practical advantages and offer so much flexibility to get the sound you need that I think one other ways might compete with it. I know, there are plenty Morosn who read my site and believe that feel so about DSET driven multi-way horns juts because I “accidently” end up with this configuration, but there are reasons why I call those people “The Morons”.

There are not a lot of people who runs DSET multi-ways and I am sure that all of them who do would confirm that in addition to great advantage this configuration might be pain in ass to maintain in proper operation order. I have expressed this sentiment multiple times, for instance in this thread:

http://www.goodsoundclub.com/Forums/ShowPost.aspx?postID=6512

When you have 12 active channels it means you have 12 amplifiers, 12 filters, 12 drivers, 24 cables and the perfect performance of your system is a derivation of a perfect performance of each and single individual element of the entire installation. In case of tube amplifiers and vintage drivers it is sometimes hard to maintain them perfectly calibrated, I can give you dozens of aspects that might go wrong and in my own playback I have seen one time or another all of them had happen. I had seen in my playbacks and I have seen in others playbacks when some of the channels wean to wrong place and in some case it was hard to detect that the installation is out of proper order.

In really those mistakes happens all time even with people who have very simply configurable installations.  I have seen many times people with a par or amp and a pair of speakers have reversed channels (R/L), not synchronized phases of the channels(R/L), importer phase of the individual channels within one or both speakers, disabled individual channels, running one of the channels absolute out of operational boundaries, failure of crossovers and many other situations. The same with amplifiers – there is so many that that might go wrong with them, particularly with tube amps. Do not be under suspicion that it happens only with inexperienced idiots – it happens with everyone. BTW, taking about the inexperienced idiots – do you know how much audio equipment is “professionally reviewed” by the industry reviewers while the components in completely defective state? If not the confidentiality considerations I would tell you plenty of stories who the “big names reviewers” reviewed the very famous speakers with reveres polarity between the channels, speakers with disconnected drivers, turntables that were spinning at 38 rotations per minute, the amplifiers that run 4 times more (!!!) plate current then it had to be, the phonostages that had -5dB crosstalk,  the cartridges that had 5 db difference between the channels, the digital elements the were miss-configured and down-converted any 24 bit to 16 bit and many-many other stories. I know that the person needs to be absolute idiot do not recognize it and the industry reviewers are a perfect match to this idiocy. Still, if the problem is minor then the problem might be tricky to detect, not to mention the Boiling Frog effect.

http://en.wikipedia.org/wiki/Boiling_frog

So, for years of using different amps and different multi-way acoustic systems channels I was trying to figure out a sort of an objective playback assessment test that world give an indication that envying in operating as it shall be. 

For quite along time I used my “Imaging Deviation Test”. The idea was that if you have a VERY imaging-complex test track and if you know HOW your playback present this complexity under a perfect condition then any variation of amplitude or phase of each channel would give you a slight deviation in imaging. I have a lot of experience to debugging playback juts by hearing it. The good part about the “Imaging Deviation Test” is that it tests EVERYTHING in playback. The bad part is that it not truly indicative to portray off what exactly went wrong – it rather shows off that “something” went wrong. Also, the way how playback images might change naturally and it forces you to keep a mental reference how your playback images in most resent setting. There are some other problems with this method…

The last 1-2 years I employ very successfully another that proves to be very accurate, very convenient and very easy to conduct. I do recommend this method for the folks who run DSETs with multi-way acoustic systems as the MOST indicative, faster and very accurate. Below are the steps that you need to take. Feel free to optimize it for you own installation but it will give you an idea. I call it the “Channels Blowing Test”

The Channels Blowing Test

1) Use, whatever means, tools and techniques, you use to make sure that your playback plays how as it shall. I do not specify HOW it shall be, remember: my objective is the maintenance or preservation of a reached reference level of sound.

2) Make sure that your L and R channels pass your “Imaging Deviation Test” from your listening position. (VERY important)

3) Using your phase meter :

http://www.goodsoundclub.com/Forums/ShowPost.aspx?postID=6573

…make sure that all your channels are set in proper polarity according to you intentions

4) Set your volume control in some kind of reference position.

5) Get any Sound Level dB meter. Here I need to stress that it might be ANY Sound Level meter. It might be $30 Radio Shake or $5000 Bruel & Kjaer - it absolutely irrelevant for this test. You do not need flat response all over, good microphone, an ability to cancel first reflections, run memories and do time averaging. You need a very small hand held devise that will show .1dB - .25dB resolution in very limited bandwidth. So, the contemporary non-calibrated, cheap, Chinese digital meters for $25 are your very best bet.

6) Attach the meter in the middle of any stick; I use 1M long wooden stick.

7) Put on your CD played a reference disk with calibrated signal and run let say 1kHz in loop. I use the Sansui Tu-1X’s internal generator that is 979Hz but it is irrelevant. You can use any generator you have handy, and make sure that it will be always available and will have a fixed amplitude and frequency for both channels.  

8) Play the 1kHz signal, go to your left speaker and place the stick with your meter on the mouth of your MF channel with microphone pointed toward to the horn throat. Slide the stick across the mouth and you get the max measurement and your microphone will be on the horn’s axis.

9) Stay on the side on the horn (not in front of it), holding the stick, note the max dB number. Write the number into your log.

10)  Do the very same experiment for each of the channels; writing the dB that you will get.

11)  You might need to run a special signal for your tweeters. Use a signal 2-3 lower then you crossover point.

12)  Do the same measurements for each channel of your right speaker. If your run a symmetrical configuration of your playback and you have more then .5dB differences between identical channels of your R/L speakers for horns and 1dB for direct radiators then correct the differences and then return back to the paragraphed #2 that you have failed.

13)  As the result you will have a log with reference acoustic pressure of each channel. In my case I have 2 x 6 measurements that indicate a proper Macondo/Melquiades performance and assure proper sound balance. 

14)  Sine you use just one frequency you will have not only the channel’s output but also dead points on transition slops for that will indicate the stability of your crossovers.

15) Pay attention you “blew” all your channels without disconnecting anything and the whole ceremony will take 10 minutes for a first time and less than 2-3 minutes for any subsequent tests.

16) Re-Blow your channels each time you change any tubes, change anything in playback or once a month if nothing was changed.

17) Do not keep your log as a computer file. Take a sheet of paper, drown the schematic of you multi-way acoustic system and write with big fat market the dB number from each channels. Then tape the sheet to the back of the speaker.

That is peaty much it. If you run multi-way with DSET or multi-amping and if you employ the Channels Blowing Test then you will be surprised how frequently the things get out of hands and how easy you will be able to detect the problems.

Rerurn to Romy the Cat's Site