This topic comes up a lot! Someone will email me with, “I want to build {insert project here} from your web site. But…”. Sometimes it’s just that part availability varies. Sometimes people want to use parts on hand which are slightly different. Sometimes people want to substitute tubes, transformers, or whole assemblies. And sometimes people want to actually change the topology! So at what point is the amp being built no longer the design on my website?

The catch here is that there is no simple answer. However, depending on the change and its magnitude we can make some generalizations about what matters and what doesn’t.

Let me first say that no one is required to build an amp like I do. It’s your amp. You can build it however you want. It usually boils down to the question of how much substitution can be made in a design before an amp is no longer the amp I designed and built. This matters because the amps I post on my website have all been designed and built with very specific requirements and desires in mind. And these requirements and desires color virtually every decision I make in a design and build. Sometimes modifications are in concert with my intent and sometimes they aren’t.

I’d like to break down the discussion by starting with the small changes which really make little or no difference in the design. Then I’ll build to those more significant changes, and finish with those changes which yield a completely different amplifier (and amplifier performance) than what’s posted on my website.

Chassis and Layout

The most common substitutions I get asked about have to do with the chassis. All of my chassis are hand built and obviously are not commercially available. A person can duplicate what I’ve built if they so desire (I’ll even provide what drawings I have) but it’s not required. Usually another commercial or hand built chassis in a different configuration makes no difference to amplifier design and performance. Moving around different elements (switches, controls, tubes, transformers, etc) if done with care and thought, don’t matter to amplifier operation.

I also personally believe that the chassis is an important part of the “individualization” of a project. I actually encourage everyone to really think about how they want their projects to look and how they’ll fit into their own unique environment. And this is not just esthetics. Is there space behind the amp for connection? Do connections need to be made on the sides (or top)? Where are the controls placed? I designed both the 6EM7 vertical amplifier and the 6CY7 V2.0 “Zebrawood” amplifier so that they could fit on narrow tables or shelves. Other projects like the source selection preamp were specifically designed to be placed (or stacked) with other equipment. The 6DJ8 headphone amplifier was specifically designed to sit on my desk at work in a specific location. These are just a few of the considerations which may go into chassis choice and configuration.

And this leads to the discussion of some specific layout decisions. I have a series of things I do with most of my projects simply because I personally like them. Some, like toggle style power switches, jewell power indicators, IEC style mains connectors, RCA style input jacks, and 5-way binding post for speakers are really just a matter of personal preference. These can be changed at will. Others, like heavy suspended signal ground busses, single point signal/chassis ground ties, twisted pair filament wiring, and ground lift switches can all have (under some circumstances) major implications for amplifier or total system performance. However, if layout decisions are made and implemented with thought, care, and attention to detail, changes can be made here without impacting performance. As an example, if someone really prefers single point star ground configurations that’s perfectly fine. It’s still the same amplifier and the performance should be nearly identical.

Resistors

In most instances the choice of resistor technology will have virtually no impact on amplifier performance. The only exception to this is that the use of carbon composition resistors in small signal stages can (but not necessarily do) raise the overall background noise level in the amplifier. I know that this is likely to start flame wars across the DIY sites. However, most of the on-line arguments about resistor technology (and especially resistor self inductance) are arguments about low level factors which have no impact on audio amplifier performance. So long as the resistors are of the proper value (within about 10%) and rated for the power dissipation required by the design, then use whatever resistors you want. It really doesn’t matter.

Personally I tend to use metal film resistors except where the power requirements require wire-wound types. This is solely to keep the noise floor as low as possible. And I personally like to use 1% tolerance resistors, but that is solely my preference. I also tend to choose resistors which are physically large enough to easily see and handle. Often there are many places in a design where 1/4W, 1/8W, or even 1/10W resistors could be used. However, I find such resistors typically so small, and with such fine lead wire, that they are difficult with which for me to work. So I generally limit myself to 1/2W rated resistors that are at least 10mm in length so that I can see the markings and manipulate them in my hands without undue issues.

Capacitors

There is more sophistry, marketing, misunderstanding, suspicion, lying, and pure unmitigated BS concerning capacitors than any other electrical component. This miasma of misinformation is promulgated by component manufactures, component resellers, amplifier manufactures, various “experts”, audio magazines and websites, various authors, and vast army of internet forum bloviating know-it-alls. I don’t know what it is about capacitors which raises the ire of the phony “audiophile”. However I can assure everyone reading this, just as there is no such thing as “signal path” in the amplifier, there is also no such thing as the “magic” capacitor. You may have deduced from the tone of my comments that I do not suffer such foolishness gladly.

The truth of the matter is that most good quality, well made capacitors from reputable manufactures will produce excellent results in your amplifiers. In fact, these capacitors will generally produce better results than the vast majority of the boutique components sold for dozens to hundreds of times more.

I know all the arguments about such audio speciality components. Pure silver foil capacitors, with oxygen free copper leads, hand rolled on the thighs of Japanese virgins under light of the full moon, cryogenically treated and burned in, blessed by mystics, carefully hand selected, and personally tested and matched by golden ear audiophile experts. Capacitors designed by gurus of the audiophile universe, produced in magical places, and shipped under armed guard straight to your door. Components said to be worth every penny of the king’s ransom charged for each one.

Truth be told, it is highly unlikely that you will be able to tell the difference between these well marketed capacitors and the boring, mass produced 716 series orange drop capacitors sold to manufacturers world wide. As such, I recommend such good quality (if somewhat pedestrian) capacitors regardless of the position. For coupling capacitors good quality polypropylene film capacitors (like the 715/716 series) are my “go to” component. For some smaller value capacitors in frequency matching circuits (like RIAA circuits or tone controls) I like good quality silver mica capacitors. And for power supply filters and cathode bypass, quality aluminum electrolytic capacitors are the norm. There is no need to ever pay more than a handful of dollars for a capacitor. And most should be no more than one or two dollars. Even the high voltage, large value electrolytics should not be more than ten to fifteen dollars a piece. And there is no reason to use a “non-polarized” capacitor in place of these large electrolytic capacitors.

More than once I have been working with someone who will go to great lengths to save $20 on an output transformer. But will then think nothing of spending over $100 on a pair of coupling capacitors. This is one of my definitions of insanity.

And now for the real heresy, the actual value of the capacitor is not that critical. You may have noticed that there are lots of very specific resistor values available. Specifically 96 different values for each decade of resistance at the 1% level. But for capacitors there are only 24 standard values for each decade of capacitance below the 10,000pf (0.01µf) level and only six per decade above the 0.1µf range. Yes, just six. The reason is very simple, the values are not that critical to the operation of most circuits. So if the schematic says 0.47µf then 0.33µf and 0.68µf are not going to materially change the performance. Yes the exact rolloff frequencies may change a little but in most instances not enough to be audible under real world conditions. I make such substitutions regularly based on component availability and most people never know the difference in performance.

Inductors

The case with inductors is far different than capacitors or resistors. In audio amplifiers, the vast majority of inductors used are relatively large value chokes as part of power supply filtering. For the most part the exact values of these inductors are not that important. For every amplifier I have a B+ ripple requirement. This is usually about -90dB. But in general I tend to overshoot by 10 to 20dB simply because of the inductors available and the topology I like. As an example, a 10H inductor followed by a 100µf capacitor provides about 55dB of ripple reduction. Reducing the inductor to 6H still provides over 50dB of ripple reduction. So if the ripple was about -35dB coming out of the first capacitor (this is kind of a typical value) the overall B+ ripple would be -90dB with the 10H choke but only -85dB with the 6H choke. This 5dB difference is actually quite small. Since I almost always follow the primary filter with channel specific filters providing more ripple reduction, either case will likely be acceptable. Often times the selection of filter chokes has much more to do with price and availability than actual inductance value. So long as the inductor in question can handle the supply current, and the overall filter topology meets ripple requirements, the values are not that critical.

The one exception I will address is that for choke input filters. These are the filter topologies where the rectifier is directly followed by a large inductor. In this specific topology it is critically important that the inductor be large enough to conduct current through the entire 360° ripple cycle. If the inductor is of insufficient size, the B+ voltage can increase significantly with large spikes at the primary ripple frequency. Often large enough to damage the amplifier. On the rare instances where I would use such a filter, I will clearly note the minimum inductance value for the first choke.

Power Transformers

The most commonly swapped out piece of iron in any design is the power transformer. The reasons for this are many and varied. Most having to do with availability, mains voltage differences, cost, and size. Most substitutions are acceptable so long as a couple of characteristics are checked. Secondary voltage and current need to be sufficient to supply the amplifier. This includes the B+ secondary as well as the various filament supplies necessary for all the tubes. Isolation is also a necessary condition if the design uses directly heated signal or power tubes. Generally a separate winding per directly heated tube is required in these special cases.

Even higher secondary voltages than required are generally acceptable so long as the power supply design is modified to include an appropriate dropping resistor. In general, if the specified power transformer is not available, just finding a substitute which is “close” is acceptable.

Output Transformers

The most important characteristic of the output transformer is the reflected primary impedance. And this really boils down to turns ratio. Everything else is secondary. The one caveat on this statement is that you need to be using a transformer designed for the job at hand. Single ended output transformers and push-pull output transformers are fundamentally different. They are not interchangeable.

People often ask about a whole myriad of secondary transformer characteristics; primary inductance, inter-winding capacitance, core configuration, and a host of others. None of these are as important as the turns ratio. This is because first and foremost the purpose of a transformer is to “transform” the impedance levels between the two ports: primary and secondary. This is the primary function. Everything else is about second order and tertiary effects. The transformer makes the low impedance speaker look like a much higher impedance to the plate circuit of the power tube.

So when substituting output transformers the first rule is stick with the primary impedance rating. The power stage has been designed to operate with a specific impedance in the plate path. Change that and all bets are off the table. The other two major considerations are the power handling capability of the transformer and the DC current carrying capacity of SE output transformers. The power handling capability relates to flux density in the core before saturation is reached. The DC current carrying capacity relates to how much flux the core can support and sill transmit the rated power between the ports.

All the other characteristics will affect how the transformer does its job, but not in a significant way. Bandpass response my be slightly different and out of band stability if feedback is applied are the only real things affected by all those second order and tertiary characteristics.

And here’s the real heresy, you probably won’t really be able to hear the difference in most cases. If your transformer is rated for a bandpass of about 40Hz to about 18kHz, that’s all that is required for musical reproduction. Remember “20Hz to 20kHz” was a marketing scheme designed to sell transistor amplifiers. It was never a requirement for “high fidelity” reproduction. The general rule is, if you want to use a different output transformer just get the impedances correct, make sure it has suitable current and power handling characteristics, and buy from a reputable manufacturer.

Tubes

Tubes are critical to the operation and characteristics of an amplifier. Changes of tube types necessarily become serious points of discussion. Every tube type is different. Even tubes which share the same pinout can (and usually do) have vastly different characteristics and performance parameters. The short answer is tubes types are NOT interchangeable entities.

However, these facts do not dissuade the people who routinely ask if they can simply substitute one tube for another. This usually happens with driver or preamp tubes but sometimes with power tubs as well. Usually it’s because of one of two reasons. Either the inquirer already has the tubes on hand or because they “believe” that some particular tube is “better” than what is called for in the design. The simple fact that each design is built around the characteristics of a particular tube seems to not matter.

Granted, I do have a few designs which are compatible with several different tubes. Both the Marblewood and the Spalted Alder are in this category. However, in both cases I am careful to describe on the appropriate project page exactly what tubes will work in what positions. Why anyone would think a KT88 would work in a 6V6 amplifier (yes, I have had people ask that very question) is beyond me. Different tubes have different characteristics, require different operating conditions, and different circuits to work properly. When the tubes change, it’s not my design anymore. Even if all the other components are the same.

Topology Changes

By “topology changes” I mean anything that changes the schematic. These changes can be deceptively simple to very complex. They are also things that yield a totally different amplifier with wholly different operational characteristics.

Some of the more common questions I get are about “switch mode” power stages. This is when switches are used to change between triode, ultra linear, or pentode mode operation. This is a very bad idea. Each amp on my website has been designed around a specific topology. Changing the topology is such a significant way virtually never yields good results. In general the entire power stage design must be changed when the topology is changed. Seldom will the circuit values in the power stage yield a good bias point or operational characteristics if the topology is changed.

The next most common question is about feedback. My amplifiers are NOT designed to utilize feedback. Contrary to what you’ll read on a lot of forums, feedback requires a completely different set of characteristics to operate properly. This includes different bandpass characteristics, different rolloff characteristics, and carefully designed operation outside of the audio band. Most people who simply implement feedback attempting to “improve” a design get marginal results at best. Anyone desiring to implement feedback in a design needs to be able to read Black’s original paper ( H.S. Black, “Stabilized feed-back amplifiers”, Electrical Engineering, vol. 53, pp. 114-120, Jan. 1934) and implement the design and stability requirements stated therein. Feedback design is not something you can do by the seat of your pants. It takes knowledge, analysis, and no small amount of mathematics.

Other topology questions I get range the gambit from simple to complex. People asking about removing capacitors, changing biasing methodologies, using all manner of tube configurations including various followers, active loads, buffers, and all manner of inappropriate circuitry. Usually this stems from some discussion on anther website. Usually along the lines of “But BigTubeGuy over at PompusForums.org said that I have to use this configuration or my amp will sound like crap.” The bottom line here is that anyone can build anything they want. But my amplifiers are designed to meet specific requirements and if you change the topology, then it’s not my design anymore. No disrespect to “BigTubeGuy” but if he can’t clearly articulate the reasons for his opinions (from an Engineering and mathematical perspective), then he’s likely just blowing smoke.

Conclusion

I understand that this has been a long and rambling post. But I think many people would be shocked to find out at how much of my time is spent carefully explaining why what people want to do to my designs won’t yield the results they think they will.

There are always changes which can be made; for any number of reasons. Some of these changes have no effect, or minimal effect, on the functioning of the amplifiers I’ve designed. Others will have larger effects but will still be manageable. And there are a lot of other changes which simply won’t work or will yield a design which is fundamentally different. And when the resultant circuit no longer operates in the same manner, when the operational characteristics are fundamentally different, it’s not my amp. In these cases my capability to help you navigate these uncharted waters will likely be limited at best.

When you sail off the edge of the map, “Beware! Here be dragons!“

As always, questions and comments are welcome.