Thursday, January 28, 2016

Introduction to The W. Marshal Leach Jr. Memorial Speakers


Remembering Dr. Leach


Sometime around 1983 I was fortunate enough to audit a class at the Goergia Institute of Technology given by the late Dr. W. Marshal Leach Jr. (1940 - 2010). That man literally taught me all that I know and all that I have forgotten about loud speaker design.  Though I was far too young and mathematically inexperienced to follow the class numerically I gained enormous knowledge and lifelong interest in the electronics and mechanics of music reproduction in the home.

Dr. Leach's paper, Build a Low TIM Amplifier published in Audio Magazine in February, 1976 was a turning point in solid state amplifier design. It is not unfair to call it the single most successful solid state amplifier design ever if we may include all the manufacturers who took the ideas presented if not verbatim at least included each of the sections with some twist here and there. From Bryston to Yamaha. No design in the history of discrete solid-state amplifiers has ever been produced by as many  manufacturers.

So in these pages I will offer up my heartfelt thanks to Dr. Leach and the Georgia Institute of Technology for allowing me to steal what I did like a giddy thief.  It is partly in gratitude, partly in regret I was never able to thank him properly. To paraphrase, poorly, Harvard University president Drew Faust:

Universities have responsibilities and capabilities far beyond being degree and job creating machines

What follows therefore is my own humble example and thanks for that truth and the part that Dr. Leach played for me. 


The Speaker Plans

I humbly present the plans for a series of DIY speakers, the Leach Memorial series, free of charge. The series is comprised of two matching designs:

  • LM-1 - Classic 2-way bookshelf using a 1" ring radiator matched to a 5 1/4" woven fiberglass mid-woofer in a ported Dayton cabinet. Small speakers suitable for intimate listening and surrounds. About $470 per pair.
  • LM-1C - The MTM (mid-tweeter-mid) version suitable for main, center and surround duties. Also integrates well with most subwoofers. About $350 each.


Sound Quality

The speakers all share a frequency range  from around 60 Hz to flat past 20 kHz.  I really like about the Peerless mini-woofer line. The liveliness of the midrange in addition to a butter-smooth response is outstanding in the price class. It's pretty expressive without being forward, so they sound closer to really good three way speakers.  The ring-radiator tweeter makes the treble sound easy and smooth, but the tight dispersion gives clarity at the expense of a big, diffuse sound field which some may miss without more diffusion in the room.

This is a really open and neutral sounding speaker, free of gimmicks or hype.  Of course, I would say that as the creator, but hey, it's not like I'm making money here.  Take my views as the rantings of a proud designer.

Please build, learn and share these designs freely, and if you find this information helpful or inspiring please give a little thanks to Dr. Leach.


Disclaimer

Except as noted above, I am in no way affiliated with Dr. Leach or the fine Georgia Institute of Technology.  All the technical mistakes I will undoubtedly make are mine and mine alone.

Further Reading

As of January, 2016 several books published by Dr. Leach are still available directly through Kendall Hunt Publishing. in the "Higher Ed" section but not easily from Amazon. I believe they are of general audio interest (amplifiers, speakers, op amps, etc.) as opposed to being specifically about loudspeakers.

Dr. Leach's resource pages are still maintained at the Georgia Tech website here, including his notes on audio related things here.

More specific to loudspeakers and relevant to all the discussions about speaker measurements is Dr. Joseph D'Appolito's often cited Testing Loudspeakers.

Saturday, January 16, 2016

Speaker Crossover Modification - 102

Real Capacitors

Real capacitors are actually pretty complex.  We like to think of them as being purely capacitative but they actually behave like capacitors, inductors and resistors.  Of the  component types you'll use they are probably the most complicated in the real world.

If the goal of a designer may be to maintain the Farads, while reducing the inductance and resistance then the answer is usually to use multiple small capacitors rather than one large one. Modern capacitors are much much better and smaller than they were even 15 years ago though, so the benefit with high quality caps and at audible frequencies is a hotly debated issue.

For those of you who prefer to try for themselves, I offer you this pair of experiments. The first simple and cheap, the second simple and painfully expensive.

With both approaches we show you below it's important to keep the total capacitance to vary no more than 1% of the original value, or you leave simple capacitor upgrading for the realm of crossover redesign.

Multi Capping

One thing to try after your first mod is multi-capping. Multi-capping is an accepted practice to reduce high frequency inductance and to get just the right values. It's not in and of itself an esoteric practice reserved for the lunatics. What makes it esoteric or crazy is that we bother in audible frequencies. That's what drives the engineers mad.

A universally accepted use of this technique is to get the right capacitance. For instance, if your design calls for a 7.5uF capacitor, only one audio capacitor company I know of (SCR/Solen/Axon) makes them.  If you don't have access to them, or don't want to use them, you could create an equivalent capacitor from almost any other brand by combining a 6.8uF capacitor and 0.68uF, yielding 7.48uF, a difference of 0.3%, well within most tolerances.

For the lunatics however, who want to reduce ESR or some other magical benefit, we do this where it appears to be otherwise unnecessary. The benefits of multi-capping is, in theory, more important with larger caps since they can become the most inductive. I generally don't bother to multi-cap anything under 5uF.

Here's an example:

As you can see, the 10uF  of C1 has now been broken up among C1, C2 and C3.  The total new capacitance is 10.08uF, less than 1% difference.


When looking to try this, keep price in mind. Some brands are sold strictly on uF, so the cost with one or 3 caps will be very similar, while other brands have a hefty minimum price, which can really increase the final cost.




Bypassing

Bypassing is just a term-of-art I use to differentiate from multi-capping.  While in multi-capping you use the same capacitor type for all the values, with bypassing you may use entirely different types of capacitors. At it's simplest, capacitor bypassing is to take a larger, often less expensive capacitor with high storage density (i.e. small size for the uFs) and bypass it with a film or foil capacitor of a much  smaller value, making a more ideal capacitor than with just the larger capacitor.  This is a common practices in many types of electronic designs outside of audio especially in power supplies where a large, electrolytic is used for it's storage and small size, while a tantalum or film cap will ensure the effective impedance at high frequencies remains near-linear. The designer saves money and space compared to using an all-film solution. What engineer wouldn't love that?

It is a reasonable, and common, option in speaker crossovers to take a bipolar electrolytic capacitor and bypass it with a nice film capacitor. 

Sadly, audiophiles leave reason behind here by bypassing a perfectly good film capacitors with another perfectly good film capacitor. The main goal of bypassing film caps for us in the crazy world of audiophilia is to color, or remove coloration of sound from existing capacitors. Engineers will state, correctly, that there's no accepted measurements for doing this, and little proof of it's merit at audible frequencies besides actually changing the total capacitance and ESR.  To which I say simply, listen for yourself and spend no money your ears don't compel you to.

Doing capacitor by-passing in speaker crossovers well almost always forces us into the land of expensive and exotic capacitors. It's definitely not for the faint of heart or bargain hunters.  Some would argue that at best it's psychosomatic, at worst charlatanism. It seems like Homeopathy, a practice completely discredited in some parts of the world. I think of it much more like layering in cooking. Regardless of whether you agree with the practice, I'll explain the basics. It's ultimately your ears and wallets that should determine the value.


Now, let's get into some examples.

 Here we have bypassed C1 with an 0.1uF capacitor. It's important NOT to bypass with more than 1% value of the original.  So if you have a 3.3uF cap you'd need an 0.033 uF bypass.

Next, don't bypass crappy caps.

I've had really good luck using copper foil bypass caps like the Audyn True Copper, but their minimal size is 0.1uF and largest is 2.2uF.  A sadly limited range for such a good sounding capacitor. Their smallest size is too big for anything less than 10uF, and the largest size is still too small to use as a prime cap most of the time.

Tweeter caps often run in the range of 2.2uF to 6.8uF. That's fine but to get 1% of 2.2uF we'll need 0.022uF which practically forces us into the land of expensive exotic capacitors, and this is why I don't recommend bypassing as the first thing a new modder should try.

There are a couple of things to keep in mind. These exotics are often physically HUGE! Make sure you look at the sizes before you purchase.  Also, in some cases they are so expensive you may be better off with buying entirely new caps.  For instance, the Duelund Precision Bypass caps, even at a discount are still around $100.  You may be better off buying a new Jupiter 100V Copper Foil or Duelund 100V RS as the base cap, and foregoing the bypass idea altogether.

Here's some brands that have values very useful for bypassing smaller tweeter caps:

 

Combining Bypassing With Multicapping

The astute reader will wonder though:

If bypassing and multi-capping are based on the same principles, shouldn't I just do both? 
Sound reasoning which only experience can challenge. My thoughts are that this introduces too many variables at once, but it's your money so go ahead.  You could for instance take the multi-capacitor example and replace C3 with an Audyn True Copper 0.68uF capacitor.  From the stand point of Farad math, that's just fine but most with experience suggest if you are going to mix your capacitor types to avoid mixing in more than 1% from a different type.  0.68uF / 10uF = 6.8% so it is definitely far beyond my recommendations here.  Again, your money.  More interesting and of higher value is for you to try both approaches and pick the one you hear sounds best, if either.

Additional Reading

One of the most often referenced, but not perfect, listing and review of capacitors is the Humble Home Made Capacitor review.  I particularly disagree with the rating of the Mundorf MKPs.  I think they are better than listed, but to each their own.

Speaker Crossover Modification - 101

While hanging out at the DIY forums we often get questions that go something like this:

I have a pair of (Wharferdale, Focal, Monitor Audio, Harbeth, Audio Research, etc.) speakers and I was wondering if I could upgrade them to get better sound?
The replies themselves go come in a range of suggestions. From don't change anything, to build your own speakers from the ground up.   When considering a crossover modification there are really two general approaches:

  • Examine the crossover design and drivers for opportunities of improving the design itself.
  • Leave the crossover design in tact but upgrade the parts used.

This blog will show you how to try the second approach and give you a number of suggestions.

If you are not already familiar with the components of crossovers such as capacitors, inductors and resistors, please visit my blog page on the subject, here.

Your Fist Mods

Congratulations, you are now fully informed of all the parts and topics needed to attempt your own modifications.  Your fist mod should focus on the tweeter. You'll replace any caps and resistors with high quality versions with the same electrical values.  Some ground rules:

  • Resistors - Replace with same or greater wattage, mandatory.
  • Capacitors - Replace with same or greater voltage, but really 100V or greater is overkill unless dealing with high power speakers.  That is, if you have a 600V capacitor, it's safe to replace it with 100V in a speaker if you don't plan on running it with more than 500 Watt amplifiers. 
  • Replace all resistors, but only caps in series are required. 

Burn-in Time

It's best to let your caps burn in for at least 48 hours of playing time.  I've heard all sorts of crazy things.  Like the Mundorf MKP's sound great at first, then crappy, then smooth out again.  So I strongly discourage you from rapid fire swapping of the parts. 

Brands

From my own experience, and that of modders I've made these suggestions to there are two brands I really want you to try first. If you can't hear a difference when you make these mods, then at least you know that your ears don't care and you won't join the ranks of those of us spending megabucks on hand crafted capacitors rolled on the thighs of Danish virgins inside a Tibetan monastery.  Be grateful if you can't!

On the other hand, many join us.  The point is, be a scientist.  Experiment on yourself and your speakers and trust your own experience more than mine. Ok, let's get started!

The first combination I'd like you to try is what I call the M&M combo, which is as follows:
  • Capacitors - Mundorf MKP  around $5 - $10 each
  • Resistors - Mills around $5-$10 each depending on wattage

Please, stick to these brands for your fist mod! This is a tried and true recipe for showing differences between mass-produced crossovers and the sound quality of high quality parts that are not outrageously expensive.

After you have tried this and want to roll your own my advice is to avoid SCR, Axon, and Solen when upgrading film caps. SCR makes so many OEM capacitors if you try to substitute a Solen for an original film cap you could very well be substituting one SCR cap for another and wonder why you bothered. 


Alternatives

Once you've tried the M&M combination, and become more curious about the benefits of high quality parts there are several other capacitor brands in the same price range you can experiment with.