Garage 1217 Project Sunrise Thoughts and Distortion Surface: SINAD can suck it

I was reminded the other day that I had this amp in my care when @Luckbad popped up after an absence. @Luckbad actually wrote a comprehensive article on how to put this darn thing together here: https://www.superbestaudiofriends.o...project-horizon-iii-project-sunrise-iii.1478/



I had compared this amp to two hybrid other inexpensive amps a few years ago. I felt it definitely had its place back then and even now it's a contender. The downside to the Project Sunrise is that there's a bit more elbow grease with respect to having to set bias and deal with jumpers, The upside is that people who love to mess with stuff or like to roll tubes can have a lot of fun.

I know this has been asked: what about the Ember? Well the Ember sucks ass. Even @Elmer Danilovich has moved on form the Ember. I don't remember the schematics well, but I believe the Ember has an OPA541 chip in unity gain as output buffer. The means high open loop gain and probably 120db of negative feedback locally (assuming since the input tube will have decent gain - and headphones don't need it). The Sunrise has a single output device (MOSFET?) running single-ended. Downsides is that it's cap coupled on the input and output. I tried the Sunrise without the input caps by lifting the jumpers (downside is scratch pot), but I did not like the sound: too grainy. The input caps impart a slight veil. Also the treble sounded more grainy than what I remember**, but this could be matter of source and different bias setting (there is an ideal bias point, but ultimately it's subjective within certain a range).

What I was really interested in was a distortion surface measurement, which seems to have ruffled some feathers. I'm not sure why people are against this because we get a lot more information that better distinguishes gear. We've seen some results from transformer coupled tube amps already*, so I thought it would be interesting to show distortion surface from a hybrid (tube input and solid state output) amp such as this. Note that this amp isn't very powerful either (the Sunrise is the less powerful variant of the Horizon), so this would be another angle.

*I should provide a full set for the SW51, an output transformer coupled amp to compare.
**I found out why, I just came from the Pietius Maximus, so had to adjust.

 

I ran the following with a 300-ohm load. FWIW, I didn't see any worse results with a 32-ohm load with spot measurements, so I may not bother. It does mean that that 32-ohm load results won't be any worse or maybe just slightly worse than this. The amp doesn't mind 32-ohm loads. It's just the design of it.

G1217 Project Horizon Headphone Amplifier

• 300-ohm load
• 90kHz bandwidth
• Low input gain
• 1.5R output impedance
• Russian 6DJ8/6922 type 6V tube
• APx555 "cheat" HPSA mode on
• Volume knob of headphone amp was set for unity gain
• Gen level to amp inputs then set on the APx555 to arrive at desired output from amp

Second Harmonic Distortion Product


Oh noes! How horrible! What a piece of shit. In most instances, we will see distortion go down (because of increasing dynamic range) and then start rise. This just shoots up and up.

If Amir saw this, no doubt he would exclaim incompetence upon the designer of the Project Sunrise amp! (Except he wouldn't because the designer is a senior member at ASR. Amir would pull his punches.) The fact remains, the Project Sunrise is one of my Garage 1217 favorites.

Also, let's look at from a more reasonable perspective. Most headphones will be used with loud listening near -12dBu and normal listening well below that from -18db to -30db where things are yellow and green. It's absolutely fair to say that the Sunrise commits sins of commission, but it's not nearly as bad as we think. Note that this scale doesn't get anywhere near the 2Vrms used by ASR because 0dBu is 0.775 is like 100db with a Sennheiser HD650 and will make you deaf quickly. As @atomicbob says "2Vrms is not a good place to be" (for the vast majority of headphones).

Besides we also need to consider more harmful 3rd and higher harmonics. So the story is not complete yet!

Third Harmonic Distortion Product


Hmm, now this is interesting. This is what we would typically expect. It's not the lowest in the world, but it's decent, good actually. Let's check out H4-H20 next.

Sum of 4th-20th Distortion Product


COMMENTARY:

• It's fair to say that while the Project Sunrise is a high distortion amp, most of the distortion is in the second harmonic. The Sunrise sounds good although it's a bit veiled from the cap coupling. I wouldn't even say that it sounds pleasant because pleasant can have a negative connotation (too soft, gooey, tubey). The Sunrise sounds lively and immediate.
• It's been said that the second harmonic is less harmful because it the harmonic closest to the fundamental. Maybe there is something to this.
• The distortion is fairly even from lows to highs. It's solid-state output and there's no output transformer at the end which may raise distortion in the lows at higher levels.

Notes:
H2, H3, and H4+ above do not include the noise component
H2 and H3 tend to be dominant over the higher orders, additional graphs will be added if this is not the case
H2 and H3 do sound different from each other, as to how, that's another discussion
H4+, the higher order distortions can be considered to be "crap" factor
There is thought that higher order distortion may have an outsized negative affect on sound quality
THD+N does include the noise component as well as THD (sum of H2, H3, and H4 and above)
Distortion is signal that is not supposed to be there
These graphs tells us how loud the distortion is at that harmonic (freq) vs output level (dBu)
H2 or second order at 100Hz means there is stuff at 200Hz that is not supposed to be there
H3 or third order at 100Hz means there is stuff at 300Hz that is not supposed to be there
Distortion that lands in the ultrasonic region matters, if SOTA chip companies like TI care, then so should you
0dBu (0.775Vrms) output level on most headphones will cause hearing damage after a few minutes of exposure
0dBu output level may burn out the voice coil on sensitive headphones with extended use

 

Just a quick reality check on the spectrum analyzer to confirm high distortion even at low levels, but primarily second order, same parameters as above. AC mains can be an issue with higher sensitivity headphones.

1Khz -10dBu


-20dBu


1MHz bandwidth. Note high frequency spikes near 325kHz, 650kHz, and 975kHz. Some audio designers and audiophiles feel switching frequency below 1MHz needs to be cleaned up, the closer to the audio band, the worse it will have an affect on sound quality.


One day, I will test with Noise Nuke.

 

THD (total harmonic) and THD+Noise


At first glance, this may seem kind of odd. However, we already know that the second harmonic is "sticky" regardless of output level. As we go lower in output level, at some point we start hitting the noise component (noise floor and AC mains hum), hence why THD+N has kind of a U shape from low levels to high level. As noted previously, the distortion is fairly even from throughout the audio band.

--

The above is taking the playbook of overlaying THD+N over THD in 2D graphs, but applying it to the 3D distortion surface presentation. If anything, this gives us an interesting perspective:

1. Perhaps THD+N or SINAD is not the end-all-be all. According to purely THD+N, the Project Sunrise is defective and should be thrown into the rubbish bin. The designer is an idiot who should be tarred and feathered, and then maybe drawn and quartered.
2. Heck, maybe even THD by itself without the noise component, that is without further qualification and breakdown into more distinct harmonic components, doesn't tell us the entire story.

THD+N and even THD do tell us something, but it just doesn't quite work in this case.

In a nutshell:

THD+N / SINAD can suck my balls.

The Sunrise while it commits sins of commission subjectively (not as much as the above measurements would suggest) is a really really good sounding amp with high levels of engagement. Kudos and praise to the designer.

 

Fixed animated gif.

Yes, it has that saturation deal. A major difference from a transformer coupled amp is that the effect happens more in the lows with those. A minor difference is that the Sunrise distortion is very much second order.

Does anyone know the part number for the power switch? It's busted - so I jam it into the on position.