Speakers' Corner
                                                                                     
Back to intro page    Read Speakertalk 2008-2010

02-02-2012
January and February for sure are diy speaker months! The cold winter months call people to the workshops cutting sheets for cabs, reading diy websites and making all sorts of plans. Well, it also tells most diy business is taking place on the Northern hemisphere - as least based on website statistics and mails received.
What I have noticed over the last 2-3 months is that I get more and more mails I can't answer properly. Questions about strange speakers I've never even heard of or how I think a given speaker will interact with a given room. The latter is really impossible to answer. Whether you buy a pair of speakers or build them yourself you always run the risk of total failure due to poor room acoustics. Too much bass or too little bass are the typical problems and I'm mostly short of suggestions as the only information is have is basic room dimensions. 
Quests for building finished speakers is another frequent topic and having a full-time job this is out of the question. I can hardly cope with my own plans and have ideas for years to come. Sorry!  
ScanSpeak tweeters have kept me busy and I have launched some measurements and an article on a waveguide for ScanSpeak D2904/710003 tweeter. This waveguided tweeter was tested in the DTQWT construction and for the first time found suitable as an alternative for Audax TW034. Better? Hmm... different and in the end a matter of taste - and definitely more expensive.
Out of curiosity I also bought a line stage kit from diyhifisupply and I can fully recommend this product, which can be had for down to 300 USD + two good coupling caps. Not a bad entry into the valve world and it should be trouble-free for years and years. The single tube is likely to be stable for decades. This line stage together with an e.g. JungSon Duro class A amp is an unbeatable combination that can be realised for close to 1000 USD. Today we can have high-performance gear at price that was unthinkable a decade ago.
Last but not least someone lend me a pair of ScanSpeak 18WU/8741T00 drivers and what was planned as a quick review of one of the Illuminator series, turned into weeks of crossover tweaking. Eventually a pair of stand-mount monitors was the result. Being used to DTQWT and Jenzen speakers it took some time to get used to a small "6+1" again. 150 square centimeters of pulp really can't move a lot of air, thus numerous crossovers were made before I was satisfied - and it gave ideas for future experiments. The Illuminator drivers are not as smooth as former Revelator drivers and extra care is needed in crossover design, but these sandwich membranes deliver details former drivers may have been short of. Value for money? Doesn't make sense. You're buying a specialty product with specific design features and you have to pay. And never judge a product based on measurements alone. It may takes many constructions and endless crossovers to get to know a product's full potential.

Check out this interview of Neil Young talking about the inadaquacy of MP3 files and low-fi! Do you know what kind of media Steve Jobs listened to when at home? Thanks to Max/Norway for sending the link!

05-12-2011

November seems a busy month on speaker building. The number of mails have been overwhelming. Hope some of you got sensible and helpful responses. I know some of you didn't because of what is described here: crossovers and this choices but that's how it is. I can't tell if drivers I don't know from own experience will make a good system and most of you writing appear not having measuring equipment to get the crossover right and even the best drivers can be a waste of money without a good crossover design. 

Launching the Jenzen Accu wasn't easy. This is a seriously expensive speaker and I can't help wondering if someone building the Accu will think it's worth the money. So, I had to ask myself if it feels right for me and after several weeks' audition and drivers breaking in I felt it was right. I've missed it every time I had something else running and it makes me condident it's worth while. 
With a speaker like this you have to get past "new-speaker" honeymoom to know if it is right. I always seriously doubt "break-in" as most break-in is taking place in out brains, but I think the Accutons need some break-in. A little rough and edgy at first but smooth and homogenous after many hours of playing. The AudioTechnology bass driver is everything I'd hoped for in the TL cabinet. Articulate and dynamic, even at low levels.
I recently visited a guy having a pair of small stand-mount ScanSpeak 18W/8531-G00 + D2904/700005 speakers in nice black piano lacquer finish (Gamut). He paid more for these speaker than the total cost of the Jenzen-Accu. Now, the Jenzen Accu does take several times the space, but comparatively the Jenzen kit is a steal for the money if you can handle the woodwork.
With the Jenzen Accu I will also start a practice of writing something about the limitations of my speaker constructions, which may discourage a few builders, but limitation is part of every speaker design. I have done this before on very small speakers. Based on response I can see that people often overestimate the soundstage a given speaker can produce. Some people think a small and inefficient speaker just needs a huge amp to get it going - and it may certainly help on some parametres, but it doesn't change the fact that a given speaker is limited by the amount of air it can move as a result of the total membrane area. A midrange driver really shouldn't move much, in fact cone movement should hardly be visible even when played loud. Same thing for bass drivers. Visitors are always surprised seeing - and feeling - the cones of the two 10" bass drivers in the DTQWT when played really loud. They barely move +/- 1 mm - and they shouldn't. 
I would really have liked to launch the Jenzen Accu with the final front panels but the Jenzen cabs will see more drivers before they'll end up as the final Jenzen Accu, because the Accu is likely to be the final statement of the Jenzen series. This is kind of a problem because the incitement to make more Jenzens is diminished by these drivers. How can I make something better, because we always want to make something better? Well, the AudioTechnology version may throw a few surprises too. I hope to have a 18H52 built for midrange with a suitable impedance not needing any attenuation at all. The Accuton C173-6-191E didn't and it pays off.
There won't be more speakers launched this year and I think the Accu is a nice finale to a fruitful 2011 in trying out stepped baffles and simplified crossover topologies improving overall performance.

Last but not least: I've already had a number of mails asking if a 20-30 watt PP/SET valve amp can drive the Jenzens speakers. To my mind they can't. On average we're some 6 dB below e.g. DTQWT sensitivity and we need at least 4 times the amount of power comparatively. The 80 watt Jungson does really well and generally I recommend 100 wpc, being valve or solid state for the Jenzens. Maybe a really good 60-70 watt PP valve amp will do well too, but I don't have such a thing and can't guarantee.
I also want to stress the importance of quality of the amp, in particular for the Accu version. Only a few days ago I had a visitor bringing in an Advantage S150 solid state amp from Bladelius. Now, this is a nice and powerfull amp, easy on the ear and with a wide soundstage but it seriously missed the holographic, reach-out midrange magic. The Accus can reveal much more than this amp is capable of doing and the JungSon outperformed it over and over again - despite costing a fraction of the price. (I only use the power amp section of the Jungson amp).    

01-11-2011
Loudspeakers we don't come across every day! One with limited dynamic capability and one with almost unlimited dynamic capability.

I've had the opportunity to review the Manger MSW transducers and my expectations were high. I had hoped these one-of-a-kind drivers had the answer to phase distortion and impulse correctness, but I was seriously disapointed. These drivers have a high magnitude of nonlinear distortion and fortunately I found another article confirming this, because after decades of praise I was a bit nervous to rock the boat. I don't say these drivers are useless, but they have serious limitations on terms of portraying the depth of an acoustic scenario (lack of dispersion) and they don't go loud before distortion becomes clearly audible.

---o-0-o---

Who would think that a few kilometers from where I live there would be a JBL Paragon loudspeaker fully functioning? The owner bought these back in 1975 and has only had new surrounds added to the legendary LE15 bass drivers. A friend of mine set up the arrangement and we had an enjoyable hour in front the Paragon.
These speakers - or more correctly this speaker - is no less one-of-a-kind compared to the Mangers. I think the 4" 375 midrange compression driver is the most powerful ever produced and mated with the 075 ring radiator and 15" bass driver, this speaker is - if anything - a clear demonstration of what converting  electrical energy into acoustic energy can mean. 

For those unfamiliar with the Paragon design, the above sketch tells what's going on. The most unusual thing about this speaker is the midrange horns pointing towards a huge curved front panel meant to disperse sound into the room; this quite opposite to the tweeters located inside the bass horns and pointing forward. Now, sitting in front of the speaker - on the floor - because vertical dispersion from the tweeters is limited, the usual quest for soundstage depth, spatial information and three-dimensionality is far off. Forget it! This speaker makes music in a room and it does it like few. Time alignment, phase coherence? Forget it! The shocking thing is that when energy transfer is optimised way beyond what we are used to, we may get away with grossly neglecting all the other parametres. This doesn't mean the Paragon is the way to go, because it isn't, but we can go home and seriously think about ways to improve dynamic headroom and reduce distortion. The inability to perform appropriate energy transfer from our rediculously small direct radiating 5-6" middrivers is pinned to the wall.

14 Oct 2011
Jenzen SEAS ER and Jenzen NEXT constructions launched. Pics of the finished cabs have been added. If you have the time and skills in woodworking, you may be in for a treat from modest priced drivers (SEAS ER). The NEXT version doesn't come cheap, but delivers the goodies.
Response has been overwhelming and three ER kits were sold before the paper was dry, so there seems to be quite an interest in large speakers and I feel confident going on with the work developing further Jenzen variants.

26 Aug 2011

It's with great satisfaction I now lauch the Yamaha NS1000M Up-grade Kit. This completes the range of up-grade kits covering the most noticeable speakers of the Sixties, Seventies and Eighties, the Tannoy MG15, the JBL L100 Century and now the Yamaha NS1000M. All three speakers were made in considerable numbers and many are still in service today. 

 

This kit will not be particularly cheap as making good things better takes serious means. The kit is a "closed kit" like the JBL L100 kit, meaning no crossover component values is given here, it comes with the kit. 
I feel confident NS1000 afficionados will find the investment worth while. Substituting the old electrolytic caps makes most of the improvement - and most of the cost. Smoothing tweeter roll-off and re-designing the bass low-pass section provides an overall LR4 topology, which was targeted in the orginal crossover, but not fully accomplished. Adjusting attenuation resistors will allow the same tonal balance as before, only with enhanced transparency and ease of listening.
The beryllium middome is what really makes this speaker, and it's so good it deserves the best. The NS1000 is a special speaker, not to everyones taste, but for those you have the front gear that can deliver the level of resolution these speakers can help come through.

The first Jenzen speakers, Jenzen SEAS ER is close to being finished and I only need to paint front panels to make them ready for presentation, hopefully by the end of September. As I have SEAS CA18RNX drivers on the shelf - and they fit the routings - it would be silly not also to include this version in case someone should have these drivers available. Only the mid crossover needs minor tweaking to fit this driver.
Following the Jensen SEAS ER and Jenzen SEAS CA, the Jenzen NEXT is next in line. All the nextel units fit current driver routings and I can borrow the bass drivers (W26NX002) to supplement the W18NX001 and T25C003.

Since last update I've had requests for speakers below 200 €/pair to 4000-6500 €/pair! Not much in between. Sorry to say that none of these requests can be fulfilled. I have two kits below 200 €/pair (I think), here and here. SEAS and Peerless 5+1 inch speakers. If cabs are made from MDF and not too expensive capacitors are used, I think these can be made from around 100 €/speaker. Most speakers I build are based on curiosity. If I don't have a feeling I can learn something new, motivation is low. Now, 6500 € is another story. First of all, developing speakers this size is a serious investment in terms of time and money. Secondly - and based on history -   nobody builds it, not even the person who asked. Thirdly, I can't get rid of it again - unless I sell it for nothing. Last, but not least, I have the speakers I need, the DTQWT and If I had the space, the DTQWT-12. Regarding seriously expensive speakers, some diy'ers think it's a matter of making "some" cabs and throwing in expensive driver - and then I have to make it all work. It doesn't come that easy. Often a range of cabs have to made to get it right and sometimes the tweeter - despite its price and praise - just doesn't blend well with the middriver and another one has to be found. Cloning the Wilson Audio MAXX3 or Marten Coltrane 2 is not an easy task and in both cases OEM (Original Equipment Manufacturer) drivers are used, and finding substitutes may be impossible.
If you're not suffering economic depression or have lost it all on Wall Strett, two of the Jenzen speakers to come will set you back serious money by using Audio Technology and Accuton drivers - if I can make it work. Buying expensive drivers is gambling, sometimes it works, sometimes it doesn't. These speakers will not be 6000 €/pair, maybe half and the reason I can make it happen is "re-cycling" and borrowing. The Accutons I had to buy though. Not cheap.


28 June 2011

After the 8008-CENTER speaker, the Jenzen speakers are next and I recently cut sheets to get started on the cabinet work as can be seen to the left. "Jenzen" because they are based on the Jensen 1071 project, thus the name Jenzen to make a distinction.
The Jenzen-2011 projects may develop into quite a range of speakers starting with two constructions based on SEAS drivers. As announced previously, SBAcoustics, ScanSpeak, Peerless and Audio Technology drivers may find their way into these cabs too.

DIY is about making quality without paying fortunes and what we can do commercial manufacturers cannot, is totally spoil the drivers with good cabinets. We can spend endless hours making cabs and we don't have to charge anybody, and we can even afford decent components for the crossover. Something that is a serious limitation to a commercial speaker, where working hours and components' cost have to be kept low. 
We can make rock-solid cabs for almost nothing because the cost of MDF is low and even exotic Baltic birch as seen to the left is moderately priced. I paid some 300 USD for all the good 23 mm BB seen on the photo.

Both cabs features transmission line loading and the midcab some 46 mm thick side walls to keep cabinet colouration low. First driver set-up will be SEAS CA26RFX/ER18RNX/T25C003 as seen above. This reed/paper pulp cone midrange is something! Smooth response and resolution on par with even magnesium.

5 June 2011
So, when are we building loudspeakers? If visits to my website hold any relevance to the number of speaker projects initiated, then February is the month where we go to the workshop cutting MDF/plywood sheets. According to Jantzen Audio, this was also the month where most DTQWT kits were sold ever. 


Website visits relative to max.

I admit I have to pull myself together really hard to get started cutting sheets, but once started I get caught up in the process and even think of the next project during the process. Making cabs takes peace and concentration to get it right and every single time I have finished a pair of cabinets, I know how I will make the next pair to perfection. However, the new pair is always another speaker with different cabs and takes new measures to get it right.
Right now a cab (yes, one cab) is made for an 8008 center speaker featuring a small 4" middriver and the Audax TW034 for treble. This time the Audax is used without the waveguide. The quest for a center speaker to supplement the DTQWT is becoming more and more frequent and despite good weather the table saw was taken to the back yard. Hopefully the speaker will be finished before the end of July. The crossover looks easy on simulation and if reality follows prediction, this will be a serious powerful center speaker.

Month of May was mostly spent making client crossovers for a 2-way Audio Technology speaker and two 3-ways based on SEAS and ScanSpeak speakers. The ScanSpeak speaker had its SS 15M exchanged for a SBAcoustics 6" driver and although not as smooth as the SS 15M driver, the sonic results were good. At around 100 USD each these SBA drivers deliver excellent sound, but the break-up nodes in the 4-11 kHz range takes special precaution. A single notch filter can make most of it, but ideally two are needed to make a smooth roll-off. It also has some "rubber" (?) problem around 1.5 kHz, which is not apparent from driver specs. The SB17 is a small 6" driver featuring only 120 cm^2 membrane area, where a SEAS 6" mostly makes 126 cm^2 and a ScanSpeak 6" revelator makes 150 cm^2, thus also claimed a 6½ inch driver. True 7" drivers are rare, but Audio Technology 6I52 is an example with some 165 cm^2 radiating area.


SBAcoustics SB17RNXC35.

The other SEAS project featured twin M15CH002 mids on a wide curved baffle. The SEAS 10" nextel bass driver and Crescendo tweeter completed the set-up. Now, this speaker can obviously be compared to the Cyclop and the main lessons to be learned are these: Use minimum 130-150 cm^2 membrane area for the midrange if crossed at ~250-300 Hz - and if we want a "big" speaker that can play loud. A wide curved baffle allows the middrivers to work all the way down to its point of crossover without any baffle step loss. I render this important and have become more and more concerned with baffle step loss and baffle edge diffraction as -mostly - necessary evils. When we can avoid it, there's a premium. Last but not least both speakers had an all-pass filter to the tweeter allowing true LR2 topology. Stepped baffles may had prevented this, but this was not within design parametres and the benefits outweighed the added complexity.   

02-April-2011
Thanks for your concern! Yes, I'm still alive, only not publishing a whole lot of speakers every second month - mostly due to building amplifiers. That doesn't mean there aren't speaker plans in the drawer, actually many, many plans, but I do have a full time job and cutting sheets takes hours and I needed a break after the OBL-11 as mentioned in the Feb notes.
I have been selling a lot of stock drivers - and buying new ones, enough drivers for at least 6 new speakers. One that for sure will emerge some day is an Ekta mkII featuring HIQUPHON OW4 tweeters, stepped and sloped baffles, simplified crossover and solid cabs. The prototype sounds very promising and hopefully I'll pull myself together for the woodwork before too long - although springtime is ahead and usually calls for alternative occupation.
Next I'm gathering drivers for a range of Jensen-type speakers. Yes, big 3-ways with 10" bass drivers in transmission line cabs and 6-7" mids, stepped baffles and low-order filters. SEAS, ScanSpeak, Peerless (Nomex164 for mid), Ciare, SBAcoustics and Audio Technology drivers are on the shelf for these constructions, some very affordable, some seriously expensive. Please do not ask when these constructions will emerge.
Based on the last two year's experiences, building new speakers have become more time consuming due to new design criteria, and a lot of prototype work follows each construction. The all too common rectangular box is in many ways a bottleneck in making good sound from even seriously expensive drivers, where ensuring proper radiation area and proper time-alignment can make good sound from even low-cost drivers. What doesn't come cheap are good crossover components and I have a lot of reports back from people spending many hours on woodwork and drivers, but fail to spend more than average on crossovers. Even low-cost drivers benefit from proper components. It's like buying a Porsche with a 1.2 litre engine.  
05-Feb 2011
After finishing the OBL-11 things have been quiet speaker wise. Time to listen to music, attend concerts, reading magazines. I read e.g. Stereophile due to the thorough measurements on speakers done by John Atkinson. There are a lot to be learned from these reports and it's interesting to see how e.g. Wilson Audio speakers develop over time, in particular the Watt/Puppy, now Sasha WP, and the Sophia. From time to time I've had the urge to copy the cabinet design, but making a Sasha cabinet clone is extremely time-consuming and something that only a very few will follow. I'm sure the Jensen speaker will make up to any of these two speakers - and the cabinet is much, much easier. I'm pleased to see more respondents to this construction.
What has happened to the WP/Sasha, is a goodbye to the sub/sat set-up, where the mid/tweeter in principle is a full-range speaker supplemented by a subwoofer. The Sasha is a more common 3-way speaker with a point of crossover around 200 Hz. I'm not sure the Watt/Puppy really had a high-pass filter to the ScanSpeak 8545 derivative, but the new middriver, which, by the way, looks like an SBAcoustics driver, certainly has, and displays a smooth, almost 1st order decline, where the woofer more looks like 2nd order. This has worked for me in the Jensen, the PRELUDE and OBL-11. If we use a smaller midrange like the AudioTechnology 15H52 (PRELUDE) we need to raise the point of crossover an octave to around 350-450 Hz to enable the midrange to cope with required power handling. Most good 6" drivers, dependent on front panel dimensions, will do fine down to 150-250 Hz.
The "problem" of having a highpass filter at 150-250 Hz is the amount of microfarads needed for the high-pass filter. The quality of these caps is extremely important and 100 uF is serious money using super caps. 200 uF in total easily mounts to around 5-700 USD or more. For midrange I've experienced good results from
Obbligato Gold caps, and also the Obbligato Film Oil Caps do well for midrange, but not for tweeter; here we need better, and usually super caps are within reach as often only 3.3-6.8 uF is needed.

I've recently had the opportunity of auditioning - and measuring - the Eggleston Andra speaker and this was indeed interesting, not least being able to see if all the nice claims at producer's website really hold water, which they mostly don't. All drivers are standard off-the-shelf drivers (very good drivers has to be said) and the double Morel mids, which turned out to be Morel SCW 636 bass driver, and Dynaudio Esotar makes a seamless integrated soundstage, utterly transparent and dynamic. The point of crossover between mid and tweeter appear to be around 1.5 kHz. The two 12" Dynaudio bass drivers in isobaric configuration truly are outdated and the overall system would benefit from some modern bass driver with a higher mechanical Q compared to these relics. I have to admit I came from a 15"/open baffle bass set-up and found the Eggleston bass lacking punch and speed comparatively.
Now, an Eggleston Andra is serious money, 24 kUSD/pair, and for this we get spotless mids and treble, rock-solid cabs with an impeccable finish, and - electrolytic caps for midrange! I'd say, the most important part of the frequency range, and then electrolytics, albeit by-passed by a Solen cap, but nevertheless. Owner had replaced the entire crossover with really good components and it had paid off. I wonder what caps we will find in the Sasha and Sophia...
Eggleston claims 6.3 ohms minimum impedance; I measure 2.6 ohms! I guess they had a typo writing the numbers :-). They claim -3dB @ 18Hz; this is way too optimistic, -3dB @ 40 Hz seems more like it, not least considering the overall system sensitivity of ~88 dB/2.8V. Here we agree. The lesson to be learned is that DIY can easily compete with even the most expensive high-end stuff when i comes to loudspeakers - if you put all the working hours into it.

Tweaking my Audio Mirror amps for new 6C33 tubes has been truly rewarding, although not completely finished yet. Minor hum remains, but I feel sure I'll find out why some day. These 6C33 really are something. In the current set-up they deliver transparency that made me stay up all night going through my record collection - one of these rare occations that doesn't happen too often. With their 40+ WPC they can make my OBL-11 rock and kick butt like a solid state.

My JungSon amp has been modified to a power amp alone (bottom of page). The JA88D line stage is easily detached and the power amp input wires connected to the two XLR input sockets. The gain of the power amp section makes a perfect fit to my 6N6P line stage and adding a balanced output to the 6N6P is next on my to-do list.