[Kevin Haskins]

I've been prototyping some active designs.   I'm pretty happy with how everything is falling in place from the design standpoint but I need to get some feedback on chassis design choices.

First off, the way I have these designed the amplifier and crossover go in a traditional monoblock type enclosure.    The dimensions are not finalized but they should be around 6" wide, 3" tall, 14" deep.     I'm machining them out of extruded aluminum stock so the enclosure is substantial.    The way I have the monoblocks organized is that I have them with true balanced inputs via a THAT Corp. 1200 line receiver.    I then run the signal into the active crossover and out through the two amplifiers per channel.     The speaker is hooked up to the amplifier through a standard 4-pole Speakon connector, using standard Cardas Quandlink speaker cable (you could use any 4-stranded cable choice).    This setup looks completely like a normal passive setup with the only difference being that there are two amplifiers internal to the monoblock, one driving the tweeter, the other driving the midwoofer.

It interfaces with upstream equipment exactly like a traditional amplifier.   It is modular, in that it can be used with 2-channel or multi-channel setups by just increasing the number of speaker/monoblocks.     

The limits of course are that it locks end-users into my amplifier design.    That has traditionally been unpopular in the high-end market.    People have all different kinds of ideas about what is the best amplifier to use on a given speaker.    My idea is that the designer has a better vantage point, and a better chance of getting the choice right if they can fix that part of the design.    It also gives me the flexibility of using active networks which allow for several design tricks that I cannot do with passive designs.    My logic for going this route is simple, I want better performance and I can get it with an active design.

I'm open to any feedback though.   I know I'm swimming against the current in high-end audio by developing an active speaker design.    They have traditionally never been successful in high-end audio.    I'm hoping to change that.    It has certainly allowed me to improve my design and for what I consider a marginal price difference.   Once you consider what most people typically spend on amplification the price difference is not only negligible, but it favors the active design.     I'm estimating price points on something like the Kepler in active form @ $2K/pr.     I've also extended the bandwidth by another half octave, decreased distortion through better filtering design, time aligned the drivers, and improved their overall on/off-axis response.   

Any comments?   

[jimdgoulding]

Hey Kev, I wouldn't presume to have suggestions but a cool thing about my Meridian actives (each is bi-amped) is that they have active crossovers with dip switches on the back that give me some options for tailoring the sound to my tastes or environment.  One of the things my speaks can do is punch way above their weight.  DAC's with pre's or disc players with a variable out can sure reduce the number of boxes and IC's in a system.  Once upon a time I had a Sony CDP with a variable out that delineated the holy bejesus out of instruments in space.  Talk about precise n' clean.  Got a friend with a EMM DAC/pre who was over but we only listened with the DAC in my system I'm so biased bout what my tube pre does for me.  I can see applications where an active system might be just what the doctor ordered (i.e. space limitations, budget, and let's not forget sound if done right something I am reminded of nitely).  I'm thinking magazines might be more receptive, even anxious, to review some new actives.  Best wishes, dude.     

[Carl V]

Objectively/rationally your premise makes sense & it may be
the 'best route'.  However, objective & rational aren't 2 adjectives
often associated with audiophiles.  Maybe Recording engineers,
maybe dedicated HT installers, maybe professional sound installation
guys who do churches, conference rooms, small theaters in museums...
but for the average joe on these boards or reading S'phile----NO!

Paradigm made some nice actives at one time....briefly.  Meridian hasn't
taken over the market nor have they really made great strides in market
share.  PMC has had nice actives for years.  Genelec isn't the 'go to'
product either.

So your active desing sounds good....make it available with amplification
and one with just an active XO, where the end user could use their own favorite amps. 
Because not everyone agrees that hypex is the 'best choice'.

Good luck & keep up the good work.  I thought they sounded good at VSAC

[Kevin Haskins]

Hey Kev, I wouldn't presume to have suggestions but a cool thing about my Meridian actives (each is bi-amped) is that they have active crossovers with dip switches on the back that give me some options for tailoring the sound to my tastes or environment.  One of the things my speaks can do is punch way above their weight.  DAC's with pre's or disc players with a variable out can sure reduce the number of boxes and IC's in a system.  Once upon a time I had a Sony CDP with a variable out that delineated the holy bejesus out of instruments in space.  Talk about precise n' clean.  Got a friend with a EMM DAC/pre who was over but we only listened with the DAC in my system I'm so biased bout what my tube pre does for me.  I can see applications where an active system might be just what the doctor ordered (i.e. space limitations, budget, and let's not forget sound if done right something I am reminded of nitely).  I'm thinking magazines might be more receptive, even anxious, to review some new actives.  Best wishes, dude.   

Good idea and one I've kicked around.    Essentially what you are adding is a "tone" control which I'm not opposed to.     You have to ask yourself what you want them to do.    For loudspeaker designers, one of the difficulties is bass voicing.    You make assumptions about how the loudspeaker is used.    A stand mounted monitor may be designed with enough baffle step to sound good stand-mounted and you would design that differently than you would one designed to be placed on a wall.      Those are things that I could adjust for via dip switches for different mounting configurations.     

The upper frequency stuff is a little more of a head-scratcher.    It becomes more of a tone control.     Not a bad thing necessarily but something that audiophiles avoid.     The thing is, you wouldn't necessarily need to add additional circuits, and extra active components to achieve it.   You could do it in the existing active crossover so you are not adding additional electrical complexity. 

There are good design rules when designing some of these circuits that I tend to follow strictly and one of those is to keep feedback loops short and tight to prevent parasitic capacitance/inductance and RFI pickup.   I'd prefer to keep the DIP switches to set those values on the board, close to the actual feedback loop so that you have ZERO chance of causing problems.    I've spent endless hours trying to optimize my PCB design and I don't want to mess it up adding a feature.   So the actual DIP switch would be internal rather than external to the chassis design.     

The other thing I have on board now is a couple optional notch filters.   At this point, they only allow someone to populate the board and use them but I'd like to have a service whereby we could have a socket to add notches.   The customer would still need to get a room measurement and that would require that I come up with a repeatable method for doing so but we could add an option so that we could allow for a couple LF notches to deal with room issues.    This is only added to the LF device so you don't complicate the HF device with additional active components.   

I was thinking of loading something like this free Praxis script on a mini laptop and have a rental unit for room measurements.   I'd calibrate the mic and have a package people could easily use to get accurate measurements.    If the consumer took the measurement, and sent me the file I could easily design and populate the notch that would go on a riser card and could be inserted into the board.   

http://www.libinst.com/SynRTA.htm

http://www.dell.com/us/en/home/notebooks/laptop-inspiron-10/pd.aspx?refid=laptop-inspiron-10&s=dhs&cs=19&ref=lthp








   

[Kevin Haskins]

Objectively/rationally your premise makes sense & it may be
the 'best route'.  However, objective & rational aren't 2 adjectives
often associated with audiophiles.  Maybe Recording engineers,
maybe dedicated HT installers, maybe professional sound installation
guys who do churches, conference rooms, small theaters in museums...
but for the average joe on these boards or reading S'phile----NO!

Paradigm made some nice actives at one time....briefly.  Meridian hasn't
taken over the market nor have they really made great strides in market
share.  PMC has had nice actives for years.  Genelec isn't the 'go to'
product either.

So your active desing sounds good....make it available with amplification
and one with just an active XO, where the end user could use their own favorite amps. 
Because not everyone agrees that hypex is the 'best choice'.

Good luck & keep up the good work.  I thought they sounded good at VSAC

I agree... your right about the market not going for the concept.    I look at it as one of ignorance and that can be overcome with a little education.        I don't need the kind of sales that a typical company would need.   I'd prefer to design a niche product and sell a moderate to small volume.   
 
Imagine if a company came out with an amplifier that drastically improved the loudspeaker hooked up to it.   Imagine it gave an extra half bandwidth of bass, decreased distortion, improved imaging, soundstage and PRAT!    It would be the talk of the audiophile community.   Reviewers would trip over each other trying to review it and customers would swoon.   That is what I'm doing.  I'm making a better amplifier, one optimized to the loudspeaker it will be driving.   


I'm not using Hypex amplifiers in these.   Not because of sound quality, but I needed something that I could more easily integrate into the board and I needed the flexibility to do it in small volume (the OEM modules from Hypex require volume purchasing).     The amplifiers are Class A/B designs using the LM3886.   A single LM3886 for the tweeter, bridged units for the midwoofer.  Trying to gauge what is going to be "popular" in the audiophile community is an act in futility so your suggestion makes sense from a marketing standpoint.   It would add a lot of complexity in that I'd need to design the output stage to drive a range of amplifiers, and I'd have to add a circuit with gain to match amplifiers.    It adds more complexity so it would be something I'd have to have a very good reason to do.     

[mbolek]

Kevin,

Good luck on the design.   Kinda bummed that you won't be going with Hypex.   As previously noted, maybe an option with just an active crossover will be attractive to us with custom Hypex amps 

Either way, looking forward to what you come up with.   I have held off on the Keplers, waiting to see what was next from you.  I hope you look at a full range solution also (ala Von S. VR4s).

marty

[Kevin Haskins]

Kevin,

Good luck on the design.   Kinda bummed that you won't be going with Hypex.   As previously noted, maybe an option with just an active crossover will be attractive to us with custom Hypex amps 

Either way, looking forward to what you come up with.   I have held off on the Keplers, waiting to see what was next from you.  I hope you look at a full range solution also (ala Von S. VR4s).

marty

I've not done all the listening trials yet but my take on the situation is as follows:

I really like the Hypex amps for use with traditional loudspeakers.    They tend to not care what kind of load is attached to them and they have tons of power.    I also like their efficiency, as I'm keen on environmentally responsible devices.   

I designed a PCB based upon the National Semiconductor design notes for the LM4780 several years ago.   It was very good, much better than it had any right to be for the price.   It also was bulletproof, with little DC offset, no nasty noises at startup/shutdown, and it really sounded pretty darn good for as easy as it was to design around.     When doing some listening test with my loudspeakers and comparing it to the Hypex, the Hypex solution won by a small margin at normal listening levels.   When you pushed them, the Hypex won by a larger margin.     The Hypex amps (UcD400s where used in comparison so it wasn't apples-apples) are just a better solution for a wider range of high-end loudspeakers.

I've also used that little LM4780 amp at a couple audiophile events.   In one case, it replaced a well respected amp that was >$3K and almost everyone thought it sounded as good, if not better than the expensive amplifier.    Of course that was a limited trial, in one situation with one set of loudspeakers.   

The LM3886 (the LM4780 is two LM3886s in one package) as far as I'm concerned, is a high-end solution when used within the context of their design limits and designed right.  They don't drive difficult loads as well as the Hypex amp.  The high-power solutions require specialized care for heat removal and they should never be operated into their clipping-protection mode.     When kept within those parameters, they are an excellent solution and they are very difficult to differentiate from other well executed designs.   Certainly you are at the point of serious diminishing returns when you implement them driving an appropriate load.   

In that context, they are actually an ideal solution for active designs whereby they are only called upon to drive individual drivers.   In the tweeter the amplifier is only called upon mWatts of power and having separate amplifiers used on the midwoofer (bridged they deliver about 100-120W, more peak) you effectively improve the amount of system power.    I'm using approx. 150-180W per speaker active which doesn't directly correspond to the same watts driving the passive design.   It is hard to do a fair comparison but you could use a 1.2-1.5 multiplier and not be too far out of line.




   

[doug s.]

i am w/carl - go ahead and package an amp/x-over that you have chosen for your speakers.  but offer the option of customers to choose their own amplification.  even their own x-over solution, if they desire.  active dsp x-overs are only going to get better and less expensive...   

best,

doug s.

[jimdgoulding]

That makes three.  Btw, the dip switches on my speaks specifically allow for the user to reduce the highend output from flat to -2db at intervals of 1/2db at 400, 800, and 2kHz.

[Kevin Haskins]

i am w/carl - go ahead and package an amp/x-over that you have chosen for your speakers.  but offer the option of customers to choose their own amplification.  even their own x-over solution, if they desire.  active dsp x-overs are only going to get better and less expensive...   

best,

doug s.

I agree that DSPs will be more prevalent with time, especially in HT because they already have the DSP to play with.

The biggest advantage of a DSP is the fact that you can dynamically set notch filters.    The rest of the hoopla over them is marketing and they are not necessarily an upgrade over an analog filter.   Otherwise I'd use the AD units which are pretty easy to design with.