During the final review of the NiWatt drawings for the first production chassis run, we revisited all of our assumptions.
With this retrospective, we thought it would be a good opportunity to document the process – from initial concept on through production. There’s quite a bit to cover, and we’ll do this in a multi-part series.
After composing this first installment, it could easily have been titled “The NiWatt Design Manifesto”.
One of the big takeaways for us (no surprise) is that a design can’t be all things to all people. The good news was, that as we refined the design, we really didn’t have to make any compromises.
I’m visualizing this series as covering the following areas, although additional topics will likely surface as we explore this:
- Today’s post: general approach, requirements, testing, and user acceptance criteria, and stakeholder identification (aka the design manifesto)
- Breadboarding – proof of concept & early prototypes
- Pre-production prototype (layout and parts refinement)
- Ongoing refinement & beta testing
- Production chassis design
Requirements gathering, testing protocol and acceptance criteria along with stakeholder identification steps are key every project’s success. Paying too little attention to any of these elements is one of the key reasons projects fail.
Think of requirements, testing and user acceptance as reflections of each other. You can’t test something you haven’t specified, and you need to specify requirements that meet your target user’s needs (user acceptance criteria).
As these requirements were identified, a design philosophy emerged:
Requirement: it has to have that DHT magic – tonal beauty that words fail to describe. This is non-negotiable.
Requirement: it cannot sacrifice speed and dynamics in search of tonal beauty. We’re looking for the response: “I didn’t know a 300B could do this”. Also non-negotiable.
Requirement: it should be able to drive speakers with a minimum sensitivity of 94dB in a large room (18′ x 30′ x 10′).
This is somewhat arbitrary since no two 94dB speakers are alike (we’re assuming tube-friendly speakers, of course). Similarly, acoustically live rooms behave differently from dead ones, but broadly speaking, these are reasonable targets to aim for.
Requirement : it should have minimum implementation complexity.
Many users don’t have room for a 4-box system (dual mono amplifiers with two external power supplies). We decided on a dual-monoblock configuration with each chassis containing its own power supply.
An alternative two-box consideration would have been a stereo signal chassis with an external power supply connected by an umbilical. Given that both of these options are two box implementations, we prefer the dual mono option in order to shorten cable runs.
External power supplies present additional challenges, and implementing an on-board power supply in each mono chassis neatly side-steps them. There’s no sense creating extra problems to solve when the potential benefits are questionable.
Requirement: it should fit on an audio rack – a 19″ maximum width with a “reasonable” depth (i.e. 15″).
Requirement: it should have user friendly hookup architecture & shortest possible signal path.
Mirror-imaged amplifier pairs address these requirements – a user friendly cable location, shortest possible cable run as well as shortest possible internal signal path.
Domestic friendliness is a consideration. We’ve seen amplifiers that correctly locate the input jack close to the first amplification stage. In many cases, the resulting location of the jack is in an awkward place for the user, such as on the front of the amplifier. Since many users have thick, stiff interconnects which do not gracefully make 180 degree bends, this can add two feet to the length of the interconnect, while being unsightly in a domestic environment as well as being a potential tripping hazard.
Requirement: it should be low maintenance and require no expertise on the part of the user to change tubes.
Requirement: it should be a conservative design with low stress on components (long component life).
Select operating points that don’t stress the components and which don’t mandate “factory certified” replacement tubes.
Too many high-end tube companies have created the false impression that vacuum tube products are unreliable. Consider that military communications during World War II were vacuum tube based and they were reliable. Sound engineering practices result in reliable components.
Requirement: tubes should be readily available.
We selected the 300B power tube for several reasons:
- It’s extremely linear and like most directly heated triodes, it takes on the characteristics its driver tube (more on this in subsequent posts). Hearing and living with Lynn Olson’s personal Karna amplifiers convinced me that the 300B can convey remarkable tonal beauty while at the same time not being syrupy and slow.
- The 300B hits a sweet spot in terms of output power and complexity of power supply design.
- The 300B offers the largest number of choices in current, high quality production tubes (Emission Labs, Sofia, Elrog, etc.).
- If you drive a 300B correctly, and you’ll have all of the delicacy and nuance favored by lovers of lower power (and easier to drive) tubes like the PA4 and 45. Driven correctly, a 300B is not syrupy and slow sounding.
Later on in prototyping, we replaced the 5AR4 rectifiers with damper diodes. Damper diodes were not only a significant improvement, but they are in plentiful supply in new old stock at bargain prices because they’re not very popular.
We’d have loved to have found a less popular replacement for the 6SN7 (input and driver tubes), but have found nothing as suitable to the task.
Requirement: it should have an elegant, understated design aesthetic. A product in the 5-figure price category should not compromise aesthetics.
Requirement: it should be compatible with a wide range of line stages/preamplifiers.
- Develop and test grounding methodologies to minimize the possibility of ground loops.
- Allow for balanced components to interface with the amplifier.
- Provide enough gain for those who opt for “passive” line stages.
Requirement: construction should be robust.
The amplifier should easily withstand the abuse of shipping. All components should be properly supported and anchored. We look to the durability of guitar amplifiers – products designed to remain in service for decades while withstanding repeated movement and the abuse of roadies – not that you’d ever intentionally abuse your amplifier.
Requirement: it should be easy to service. Does anyone remember the early Toyota mini-vans? Changing the spark plugs required that you drop the engine.
Stakeholders are not always obvious, but you need to do your best to identify them and address their concerns.
The Customer: their interest is clear. They want a “best of breed” product with an elegant visual design which drives their speakers at a fair price. They want after sale service from their dealer – or in the case of direct sales, from the manufacturer.
The Manufacturing Team: the amplifier should be “easy” to build. The construction layout should be intuitive in order to simplify the build process (minimize errors) which in turn shortens the QC process.
Servicing Technicians: the layout should be clean & intuitive & and easy to service in the eventuality that it is needed. Hard-wired construction is a benefit for long-term durability (as well as reducing stray capacitance). We touched on difficult to service automobiles above. If there are any unique assembly/disassembly steps, they should be documented in the owners’ manual.
Vendors and Fabricators: parts should be readily available while of course, meeting sonic criteria. The chassis design should be easily manufactured to high mechanical and cosmetic standards.
The Shipping Department: it should be easy to design robust shipping packaging for it. This is not as great of a challenge with amplifiers as it is with turntables, but shipping should always be taken into account.
Compliance Groups (CE, ETL, etc.): safety first! We’ve seen current production vacuum tube products which actually omit the safety ground in order to circumvent ground loops! There are approved ways of eliminating ground loops without compromising the customer’s safety.