Search
  • chaddavidkelly

Reopening old wounds... LOL... finally revisiting the WA-87

This article is one that I've put off writing for years... The development and release of this mic, being my first experience dealing with China engineering across the sea and the communications barriers, differences in expectations, etc. to this day leaves me with a very frustrated feeling. It was a learning experience for sure, and one that I vowed not to repeat (hence both of the companies I manage today utilize fully American engineering resources that we own the files and legal rights to, and more importantly are made to our exact wishes). For years, however, people have asked me why I have not listed a modification/upgrade for this microphone, being the first released microphone that I worked on as a product developer. I haven't yet, nor have I talked much about it publicly; but I have finally decided to address it (both in writing, and in offering up a modest upgrade). Prepare to take a very deep dive into mic geekdom.


Sometime soon after a 15 year career at a larger pro audio manufacturer (PreSonus), I had the opportunity to help develop a line of condenser mics for a small company called Soundengine. We developed a tube, a solid state, and a pencil microphone. They turned out quite nice (in fact there is one record that was done with them); but in hindsight were simply doing what many people have done, which is to take a Chinese mic design and 'augment' it with a few better parts like Cinemag transformers and nicer tubes, etc. I suppose that's a lot better than the many companies who import cheap Asian mics and 'don't' even bother to do that, and still charge a premium for them... but I digress! Sadly, this venture didn't work out because both myself and the company's founder had to move out of state at the same time, me to Texas and him to California, and so these plans were put on ice.


Shortly after moving to Texas and taking a position as head of operations and product development at Warm, I set out trying to improve the existing product line as well as spearheading the next generation of them. I worked on the WA-2A (TEC Award winner), EQP-WA, WA-4x12, mk2 preamp lines, a number of things that never came out, and of course the microphone line. To be quite honest, my modus operandi was simply to repeat the work I'd started at SoundEngine and 'augment' some decent Asian microphones that we had access to. 797, whom we used as a primary contract manufacturer, already manufactured microphones for BLUE, Studio Projects, and a number of other manufacturers, so things came together rather quickly.


The 87 was my first task, and we followed the same engineering path that we'd used to make the WA hardware up to that point, which was to rely heavily on the engineering resources of the contract manufacturer themselves; because hiring (or renting) the services of an engineering firm was just simply not something the company founder had any interest in paying for. This surprised me because this was something of a taboo at my former employer, where we designed and owned all of the documentation to our product designs all of the time. Otherwise, you find yourself at the mercy of a factory in China who own all of your mechanical and electrical files and data, and can decide what to show you, how something is going to be designed, and what files and secrets they are willing to give up and which ones they will hold onto to ensure you must continue to do business with them. This is, in fact, the pickle that WA basically exist in as matter of their business model. This is all simply to say that when I first asked the factory to make boards for the original u87 model, I didn't exactly have full control over what they were going to give me. My job, as it would be for most of the next 2 years, would simply be damage control and to put the best possible facelift on what we were being given.


What we got back, ultimately, was something loosely based on the original 87 schematic, biased to a different FET due to the unavailability of the classic parts such as the K105 and 2n3819, and with a lot of the polystyrenes and tantalums sort of substituted for more ubiquitous parts (off brand WIMA style film caps and other generic parts). As often happens, time was running out on this project. I was able to 'air my grievances' with the factory on certain things and get one more round of prototyping done before we had to put a pin in it, as they say; but after the second prototype, that was it. I got some things done to ensure it worked well; but never got to dictate the BOM (build of materials) ingredients, what FET was used, biasing instructions, and so on. The factory held a lot of that information close to their chest, because they ultimately want to do what is best and most cost effective for them, not for us. Of course, we made sure it had a great transformer (I auditioned at least six parts, from Crimson, Jensen, AMI, Cinemag, and others.) and a really good capsule (3U's K87 is possibly the best capsule they make from their own designs, outside of contract manufactured capsules from other designers, and it has gotten better over the years since then, in fact). But the stuff 'in the middle' bothers me to this day, because I just didn't get enough of a say in any of it, and I felt the result was somewhat decent, very 'vanilla' and workhorse sounding condenser mic... but just NOT an 87. It did not have the weight and bottom end depth, the upper midrange sizzle, or just any of the hallmarks of what the classic mic sounded like. It bothered me, and still does more than 5 years later. It bothered me when people didn't like it, and it even bothered me more when people did like it and claimed that it sounded like the real McCoy (because I knew darn well that it didn't). It sounds OK, but sounds processed and plain, a bit less open and articulate than it should be, and just doesn't sound 'real', if that's a good descriptor.


I actually thought of announcing a mod for this mic 2 years ago, using the approach of throwing out the entire 'guts' of the mic and putting in a new set of boards that worked with the NOS FET part and got much closer, ingredients-wise, to the original design. I did six of these for WA customers and people LOVED it. But I realized that this approach would be wasteful, time consuming, and punishingly expensive to the customer because the cost of the mod would come close to the cost of a used specimen on eBay. That approach just does not fall in line with how Signal Art approaches mic mods. We try to be budget conscious and work to pragmatically improve a mic to be the best 'for what it is', without spending a fortune. If you have that kind of money to spend, just go buy a whole new and better mic! So what we determined the best approach to be was something similar to what we are offering now as a premium mod to the WA-47... to literally just strip the PCB bare and clean it down to the bone, and then meticulously re-populate it with only the best parts and the RIGHT parts to get the most out of the circuit. I'll dig into this further.


Thankfully, as the developer on this product, I own one of the actual working prototypes, and have been able to play around with every possible permutation of the board, and gain a better understanding of where things went wrong. What I finally determined was that this product really sort of suffered a 'death of a 1000 knives'. It was not any one thing. The transformer is great. The capsule is really good. The board design, at least in terms of the schematic, is close enough. The FET is decent (It is, in fact, an 'actual' NOS Fairchild part, just not the part usually associated with the 87 design), and the voicing is decent. It's just that when you cumulatively put a few dozen generic parts in series with one another, and then on top of that assemble it VERY messily, with a lot of glue and gunk and mess, you have this sort of 'bucket brigade' loss of tone that is cumulative, and audibly evident.


What I do on the mod is to essentially strip the board bare. it is de-soldered, stripped of components, and scrubbed down with anhydrous alcohol. All of the glue and gunk is removed as well. I completely re-populate the boards using only the very best parts. WIMA audio grade capacitors, audio grade polystyrenes, and over-spec'd tantalums. Many of the film capacitors are upgraded to polystyrene. Many of the electrolytic capacitors are upgraded to tantalum. The polystyrene capacitor count goes up from one to NINE! The tantalum capacitor count goes up to five! The electrolytic capacitor count goes down to zero, which is what I want it to be. And we use an audiophile grade adhesive on the polystyrenes to reduce microphonics, as opposed to the copious amounts of 'hot glue' used by the factory, which is more prone to rigidity, to coming undone, and to chemically/electrically interacting with the components and boards. Upgrading the polystyrenes, particularly in the high-z section, really opens up the mic and adds some top end clarity, which some reviewers have described as the key difference between this mic and the newer Warm Audio 87 mk2. ...but we don't stop there. Every single resistor is replaced. All generic metal film resistors are replaced with either carbon film or carbon composite resistors, 5% or better tolerance. With Carbon composite resistors, I personally measure and toss out any that are not at least 4% tolerance or better. Many argue that carbon film and carbon composite resistors have a sonic advantage, and some don't believe this; but what is undeniably true is that they are the historically correct parts that the original unit was designed with in mind. I do believe there is a subtle but cumulative way that carbons interact in the circuit, esp. after it has heated up and been on for a long time, and I also personally believe that carbons are less prone to RFI and EMI (high frequency radio and electrical interference) because carbon is far less receptive 'antennae' than metal film. There is a long history of 'carbon cables' preferred by guitarists for exactly that reason. Lastly the high value radial resistors (this is in reference to things in the meg ohm and gig ohm range, not their cost) are replaced with much more rugged, higher wattage rated, axial resistors. These stock high value radial resistors used by the factory are just cheap and unreliable, and notorious for breaking because their pins are soldered onto solder tabs and then soldered to the main boards with minimal clearance, often causing a failure at the tab juncture of the resistor. if you've ever had a WA-87 that was intermittent, or only worked after being on 'for a while', I can almost guarantee you these resistors are the culprit. I throw them all away (there are 4 or 5) and put in bulletproof, reliable parts that will never fail on you.


A decent case against the WA-87 mk2...


To be honest, one of the reasons for finally making this mod available now is because I realized there was an opportunity here to possibly save folks some money. A lot of people are 'eBaying' their original units and buying the new mk2 version. I have heard the differences between the two, and they are subtle. The mk2 has a touch more clarity and crisp articulation, all other things being equal; and I am certain that my mod gets you to that place and then some, with the original unit, but also brings it up to such a higher build quality and reliability factor.


I had nothing to do with the mk2, it was well after my time; but I'm a bit skeptical. To clear up a few marketing messages/misconceptions out there, neither microphones have ever had a 'custom wound' Cinemag transformer. The original has the Cinemag 2480, and the mk2 has the Cinemag CM-13113, which is their take on the classic T13 part, a smaller-package version of the original German BV-13. Both of these are catalog parts, and have been in production for decades. They are both quite good; however, I evaluated the CM-13113 back when developing the WA-87 and picked the 2480 deliberately. The 2 parts cost the same, and are nearly the same, physically; except that the 113 has an added humbucking coil and a slightly lower turns ratio, which in real-world application might give it about a .5 to 1dB higher output. I found the 2480 to be slightly smoother and more musical, and the 113 to be slightly harsher in the top end. I am confident that I would feel the same way today, so I would likely never consider that part an upgrade to what was in there before. On that note I should also add that I evaluated AMI's T13 that is sold by some retailers as an upgrade/mod part for the WA-87 also; and I have to state emphatically that this part does not perform any better than the Cinemag 2480. Personal tastes are, of course, subjective; but I felt that the AMI T13 was more 'closed in' and slightly less three dimensional. What's in there, on the original, is really the best part, IMO.


Now we come to the claim of making the mk2 with an 'NOS Fairchild 2n3819'.... and I honestly don't know where to begin here with that. I'm probably going to get myself in trouble; so let me just preface this by saying that I have not evaluated or dissected their part, and that this is purely my opinion and conjecture based on my experience. So, firstly I should state that for years I have sought out and purchased NOS Fairchild 2N3819's for 87 type builds as well as for an unrelated engineering project. Ultimately, I needed to obtain all of the remaining stock of this part that I could find, and I essentially did. I had a source in France that I completely cleared out (about 2000pcs). Later, I found one last remaining source in the UK being held at a military surplus parts vendor called Langrex and I cleared them out as well (about 2000 pcs). These were the last available Fairchild 2n3819's in any serious quantity out there, and I have them. They are being used in production of an unrelated product for another company I'm involved with; and eventually we will run dry and have to move to a new production part by Central Semiconductor (not the end of the world, that part is really good and I know how to offset the circuit with some other different parts to add some more vintage vibe back in as needed). Considering that WA probably manufacture this mic in 10,000 pc batches, it casts real doubt on whether that is a real NOS Fairchild 2n3819. Could it be, and could they have found an unlimited source that I somehow overlooked? Anything is possible; but I highly doubt it.. What is far more likely is that they obtained a surplus stock of something 'called' a NOS Fairchild 2n3819 that is likely a re-branded, counterfeit component. This is a notorious issue that plagues Chinese parts vendors, the counterfeiting of chips, capacitors, and other components. I've been burned many times myself and had to throw away hundreds of fake 2n3819's. In this situation, it may be not so much that the company was tricked as it may be that they 'wanted' to be tricked, if that makes sense. Because the goal may have simply been to be able to make this claim, and the truthfulness of that claim being secondary as long as the mic still performed well enough. Lastly, I have to reference their own photo of the 'NOS' Fairchild 2n3819 in the mk2.


I have bought these transistors from all over the globe, for years and years. Like most semiconductor and chip manufacturers, these have been made in various plants from Singapore to Thailand to the Philippines to the States to Taiwan to Mexico... and they all have a certain indicator in the plastic potting material that distinguishes them in terms of their manufacturing location. A trained eye can catch this, and its also used to sometimes identify the counterfeits. What I can say with certainty is that none of the ones I've ever purchased have ever looked like this. I've never seen one with the groove cutout on the right hand side as this part has, nor have I ever seen one so perfectly clean and with such clean silkscreening. Certainly this image is very Photoshop-enhanced; but furthermore, I believe might also be an indicator that the part is not an authentic Fairchild 2n3819. I have to stress that it's purely my opinion; but I highly doubt its an authentic part. And I don't think the company was fooled. I think the goal was more to fool the customer, and to bolster a marketing message. Again, I could be wrong, but knowing the quantities they manufacture and knowing what the state of the world's inventory of this part basically is.... and esp. as they continue to make more and more without any apparent worry of 'running out' of this rare NOS part... And again, ironically, the original 'mk1' mic actually does have a Fairchild FET.


This is certainly not intended to torpedo the mk2 version. I know that it works and sounds good, irrespective of minor FET and transformer quibbles. I realize many people will always want to have the latest and greatest of something. I am like that with many things as well. But what I do believe is that for all of those who feel left out by owning the original, or who can pick up the original mk1 on the used market for a song; I can now offer a mod that I feel will absolutely give you the best of both, and get you to a sound quality and build quality that surpasses either... for a more than fair price!


Thanks for indulging me on this trip down memory lane.

31 views0 comments

Recent Posts

See All