Project 2845 is an ongoing internal skunk works project with the goal of developing the ultimate Single Ended Triode (SET) amplifier using a pair of 845 tubes in parallel.
It's no secret that at Bandwidth Audio, we love SET amplifiers. However, they only really 'work' when designed correctly. The output stage and power supply must be well matched for the power level and type of speakers the amp is expected to drive. If this is not the case (as with the majority of SET amps in our experience) , then expect to be giving up performance somewhere.
Pretty much all SET amps seem to have the midrange magic that triode freaks love, but how many are praised for their low end accuracy? Few to none! This is what I mean by giving up performance; you gain some detail and lifelike resolution in the midrange but sacrifice performance in the low end.
This is not surprising since most designs are copies of amplifiers from the the RCA Tube Manual or other DIY amateur radio manuals from the 1930's to 1940's. In fact, in old publications I have seen something like 50Hz described as (paraphrasing here) 'sufficiently low for good reproduction' or something to that effect. If we throw out the rule book and take advantage of modern components and truly engineer an amplifier from the ground up, we can do much better than this!
The 22A3 Amplifier was Bandwidth Audio's first (and currently only) SET amplifier. With only 5-Watts, most scoff at that printed number and never take a second look. However, wherever we take this amplifier to demonstrate (trade shows or local audio meet ups) no one has ever walked into our room and said, "that amplifier sounds small" or "this doesn't have enough power." In fact, most usually think it is a closed loop 2A3 push pull amplifier based on on the power level and accuracy it delivers, especially in the low end. Only when we tell folks it's a 5-Watt SET amp, we always get the same response: "Really?" The magic of the 22A3 amp is the parallel output stage designed for extremely robust drive capability and speaker control for a 5-Watt amplifier. We use the parallel stage for better drive, not more output power.
Since SPL only grows by +3dB for every doubling of power, pushing the 22A3 design from 5 Watts to perhaps 8-10W at the sacrifice of loudspeaker control, is the wrong design choice. The gain in headroom - and therefore the ability to drive less efficient speakers - is negligible.
Yes, the 22A3 is only a 5-Watt amp designed for efficient speakers (recommended minimum speaker efficiency of 93dB SPL 1W/1m for small rooms and 96dB SPL 1W/1m for larger rooms), but there is no reason a low output power amp needs to sound anemic in the bass department - even when driving 15" woofers.
This brings us to Project 2845.
Our goal is to take what we have learned from the 22A3 amplifier, and apply it to a parallel SET 845 amplifier with a goal of 35-40W of output power. This large high voltage triode will give us a significant and meaningful increase in output power compared to the 22A3 amplifier, when the tube is operated near maximum voltage, so we can drive much lower efficiency speakers (84 to 86 dB SPL 1W1/m) to appreciable SPL levels while still proving extremely good speaker control and bass presentation (which becomes even more of an issue at high power where woofer excursion grows).
With this goal, most anyone can experience what a SET amp is really capable of on their favorite loudspeaker, regardless of speaker efficiency.
On top of this, we will be designing this amplifier to Bandwidth Audio's standards for quality, reliability, and ruggedness (both electrically and mechanically) for years and years of trouble-free service. Electrical ruggedness and component longevity analysis are especially important in an amplifier with 1.2KV B+ voltage supply!
This amplifier design is more than just cobbling together off-the-shelf components. We will need to design in-house as a system:
Bespoke Power Transformer(s): 1.2KV B+ supply, Preamp supply, Filaments, Bias supply etc. (multiple possible)
40W capable 4/8/16ohm SE Output Transformer - completely designed in-house with our own impedance ratio, core size, winding pattern, Interleave specifications, and gapping requirements to meet 40W at LF with wide open loop bandwidth. This output transformer is critical to get the most out of our parallel output stage. Off the shelf won't do!
Preamp: Capable of 400V P-P signal swing with not only low distortion and noise, but flat THD vs Frequency while achieving wide bandwidth into the difficult capacitive load parallel 845's present.
Also, for those interested, we have started a video series on the design.
Project 2845 Video Series Part 1:
A few Thoughts on Chinese 845 Amps (or others for that matter)
In recent years there has been a flood of 845 tube amps from China. There are very few domestic US-designed (let alone US-manufactured) 845 tube amps.
From my inspection, most of these are very similar designs. In general, I would stay away from these for a couple of reasons.
They aren't built very well:
Poor build quality isn't a surprise to anyone, but build quality isn't just a nice-to-have for easy repairs. Looking at the inside of these Chinese amps, I have genuine concerns about safety. I have seen capacitors held down with tape and lots of exposed solder joints on heavy components floating in mid air with minimal clearances to other components or the chassis. With 1000+ Volts, safety in the design, circuit layout, and final build is a must and is no joke. Solder is for electrical connections, not mechanical. Heavy components held in place by solder or tape (think of what happens to tape in a high-temperature environment) is a genuine safety concern. If one end of the the B+ supply breaks loose and touches the chassis, you better hope the product is fused appropriately.
Cheap and often Counterfeit Components:
Cheap components and an often compromised design (to reduce the number of said components) leads to a very short useful life. Except for the transformers, the power supply in a tube amp is the most expensive electrical system. Quality capacitors are expensive, and you need to usually stack 3 or more in series to achieve 1-1.2KV voltage ratings. I have seen time and time again over voltage on the power supply capacitors and or a lack of appropriate voltage balancing across the stack.
If used in this manner, the capacitors wear out in a few 1000 hours! If you don't believe me, look at any manufacturer's lifetime rating of electrolytic capacitors. They are rated for only a few thousand hours at max specs. Overvoltage combined with high ripple currents from inappropriately designed rectifiers will over-heat and kill capacitors leading to hum, loss of output power, and a poor sounding amp.
In the end you get what you pay for... Always.
If you are interested in a Chinese amp, especially a low buck 835 SET amplifier, I would advocate for using that money on a conventional push-pull amp from a known brand. A good SET amp is expensive. A good 845 amplifier is really expensive. Even with 'mods', you can never turn it into something it's not. You can put a completely new circuit and the chassis to overcome a lot of issues, but you will still be limited by the off the shelf "flavor of the day" transformers that the manufacturer was able to source the day your amp was built.
If you are just getting into tube audio or the hobby, then a Chinese amp may not be a bad way to go, at least initially. You can at least get your 'feet wet' and get a feel for what it is all about (assuming you are not coming from a quality solid state amp). In that case, a Chinese tube amp will be a step backwards.
If you are considering an off-shore tube amp, I advocate you stay away from a SET amp and try a more traditional push-pull amplifier. Your dollar will go farther, getting you towards something that will be decent to start with, and provide "acceptable" sound.