Sonic Frontiers SFS-80 Tube Amplifier: A "new" benchmark to beat from 1992
Up next on testing gear that would be impossible to ship safely... the Sonic Frontiers SFS-80 tube amplifier. 70 lbs, all concentrated in the transformers.
View attachment 232618
View attachment 232623
Believe it or not, Sonic Frontiers is a company we all know very well... today, they're known as Anthem AV
. In the late 80's or early 90's, a Canadian part distributor began to ship DIY kits for tube amplifiers. They began to sell their finished products under the Sonic Frontiers brand and rapidly grew. As home theater and the DVD era was introduced, Sonic Frontiers established a "budget line" known as Anthem. Generally, Sonic Frontiers focused on tube-based products while Anthem focused on solid-state. Fast forward to 1998 and Sonic Frontiers/Anthem were purchased by Jerry VanderMarel and Scott Bagby, the founders of Paradigm speakers. The 2 channel tube line was retired and Anthem focused exclusively on solid-state products and has become the company we all know and love today.
One of the questions asked is if tube amps are just intended to be distortion factories? Some of the worst measuring devices here are the pricey Carver tube amplifiers. We've seen questionable modifications to the Dynaco ST-70 which measure poorly, though potentially attributable to "questionable" tubes. The SFS-80 is built around the tried-and-true Dynaco ST-70 platform but with over-engineered transformers and power supplies to handle up to 80 watts per channel. The circuitry is a Joe Curcio design. What's great about the SFS-80 for tinkerers and hobbyists is that the front meter is NOT a VU meter which is pretty but not all that useful. Instead, this is a bias meter which allows you to fine tune the bias of your power tubes by turning a potentiometer. Modern designs have gone with self- or fixed-bias designs or "fool-proof" turn the knob until the LED light turns on or off. With an actual bias meter, you can really tinker with the setup.
Today we'll be looking at a specimen that has never been restored, oxidized bindings posts left as is, but does have the following tubes which are different from factory
1) Joint-Army-Navy GE 6550A tubes, biased at 40mA and 50mA
2) Joint-Army-Navy Sylvania 6DJ8 tubes
View attachment 232620
View attachment 232621
Unlike McIntosh tube amps, the SFS-80 does not have multiple binding posts for different output impedances and you're going to have to de-solder/re-solder the output transformers to change the setting. You can go with 3.5 ohms or 8 ohms, and my testing load is 4.48 ohms.
These tubes were bought off eBay and were tested on a B&K 707 Mutual Transconductance tube tester to be "according to spec" (according to the booklet). Joint-Army-Navy genuinely means that these tubes were designed for the US military and met the specs mutually agreed upon by the Army and Navy. This was the terminology developed during World War II and provided an un-classified designator for electronic components. Sometimes JAN tubes were superior to the mainstream consumer products. Other times, they came from the same production lines and since every tube met the military specifications, the tubes sold to the US government were printed with the JAN nomenclature. These tubes don't carry any special audiophile fairy dust. They weren't matched specifically nor are they the "super musical" Mullards or Gold Lion. What they do represent are tubes when tubes were used for scientific instrumentation, radars, and other military gear. You know, when you were trying to bring down distortion as much as possible.
The test setup is:
Panasonic UB9000 -- balanced-2V --> Marantz PM-11s2 as pre-amp --single ended--> Sonic Frontiers SFS-80 --Vishay Dale 1% NH-250 4.48 ohm resistor--XLR-in-> E1DA Cosmos ADC
I have decided to ground the E1DA to the pre-amp.
Based upon my Marantz amplifier testing, my test setup is as much as 9 dB worse than @amirm 's APx555.
Here is the 2W and 5W output at 40 mA bias, where I normally keep the amplifier and think it sounds better. This is typically what an EL34 tube amp would be biased at.
sView attachment 232605
View attachment 232606
And here are the results at 50mA which would be typical for a KT-88/6550 tube and will net more peak power before clipping.
View attachment 232611
View attachment 232612
Whew! I'm glad this tube amp measures well.
Before measuring this amp, I'd my tell audiophile friends and family that this SFS-80 really does sound great even against expensive solid state gear as long as I keep the volumes low. I really don't think it's a distortion factory like everyone says. Thankfully the measurements correlate with my opinions.
Obviously the appeal of tube amplification besides the nostalgia/anachronism is the presence of non-linear distortions which anecodatally sound great. Looking at these measurements, I wonder if tubes produce the effect of analog dither with all of the noise spikes at the -100 dB range that continue well into the critical mid-range. However, when tube coloration can also measure this well, you certainly don't have to worry so much about feeling guilty of sighted bias. We also see evidence that UNLIKE solid state where a bit more power can improve SINAD as you get out of the noise-limited region, this tube amp performs better at lower power and in fact better at the lower bias. One missing detail from these measurements is that when playing the 1 kHz test tone, the output transformers started to hum at 1 kHz. I don't think these transformers, which were made in Canada at the time, were encapsulated in epoxy resin like McIntosh transformers/autoformers are. The noise would have been masked by actual musical content, but I wonder if there are microphonic effects that may be non-linear as well. Last, it's not obvious to me that there are differences to the even or odd harmonics. It's been explained that topology has a bigger difference than the actual use of transistors or tubes, and this data seems to confirm that. Last this tube amp has minimal negative feedback -- you potentially could get better results if a stronger feedback loop was applied.
Remember that my measurements are as much as 9 dB worse than what Amir can get with his carefully designed setup. Still, my 5W SINAD of 74 dB is pretty impressive. It means I'm beating the Lyngdorf TDAI-3400, Anthem MRX520, Pioneer Elite VSX-LX504, and every tube amp tested to date at ASR. I don't know if we'll really get the full 9 dB when measured on a proper setup, but I'm very confident that this unit would break past the mean amplifier SINAD of 78. If this amp can break into the 80 dB range, that really would be saying something about the types of audio gear being passed off today as "audiophile premium products" wouldn't?
I actually think the PM-90 and SFS-80 sound very close to my ears, at low volumes. I remember thinking to myself that the PM-90 measured pretty well at ASR, but this tube amp sounds like it's in the same general performance category. At least with my measurements it does in fact measure very similarly. The SFS-80 is just a plain old amplifier whereas the PM-90 is a full integrated amp with preamp, phono amp, etc.
Today, Anthem is known as a company with premium room correction technology and a premium audiophile manufacturer that still produces some of their products in Canada. Thirty years ago, they were hobbyists working with tubes and it seems as if this tube amp really showcases how the ethos of science and engineering can be applied to tubes. You can have all the nostalgia of tubes and North American manufacturing along with whatever positive colorations exist with tubes while simultaneously getting measurements that remain "on par" with today's solid state electronics. Since tubes cannot match the performance of solid state electronics, offering tinkers the ability to fine tune the bias is a worthwhile feature in my opinion. It's much easier to dial in the bias with an actual meter like this as opposed to having to bring out a multimeter or rely on an LED that turns on/off.
What isn't clear to me is if the performance showcased today, which sets a new benchmark for tube amps, is a reflection of special tubes or just an amplifier design that had the goal of "making tubes as accurate as possible." All I can say is that this combination of Canadian electronics and US-manufactured tubes is another 1992 Dream Team that sets the new benchmark to beat when it comes tube amplification.
View attachment 232626