Tube Amp soundstage


I hope everyone is safe and healthy during these strange times. I wondering if someone could explain to me the reason my tube amp has a deeper soundstage than my SS amps? Two years back I built an Elekit 8200 which puts out around 8-10w/c in ultra linear mode, depending on the power tubes. I usually run KT88’s or 6L6’s, and less often EL34’s. It powers a pair of Tekton Enzo 2.7’s which are quite efficient at a claimed 98db. The SS amps I’ve used with these speakers include a vintage NAD 35w receiver, a Musical Fidelity m3si @ 85w, a Rogue Sphinx @ 100w and a Hegel H80. Now granted, non of these amps are what I would consider high end audio, but no matter what, the little tube amp always seems to have a deeper, more 3-D soundstage and the SS amps sound a little flatter. Same source, same DAC, same speakers and cables. There are things I appreciate about the SS sound, such as tighter, better defined bass and an effortless ability to play louder (which I do less and less), but every time I rotate the little tube amp back in, I hear a slightly more organic sound and that deeper soundstage.
dtapo

Showing 2 responses by atmasphere

lots of feedback kills image depth, easy test with an SPL meter, genius ( mindful designer ) Roger Modjeski designed RM-9 with variable feedback :-) ears are required
A bit of a correction here- feedback, properly implemented, does not kill image depth. **Improperly** implemented feedback does that and other things (brightness, harshness) as well.


For feedback to really do its job, it needs to be really a lot. Like over 35dB or so. Less than that and you have those problems above. One thing about sound stage BTW- to do it right you need zero phase shift in the audio band. This is why in our tube amps we've had 2Hz to 200KHz bandwidth, since you need to go 1/10x or 10x the cutoff frequency to prevent phase shift components in the passband - in this case the audio band of 20Hz to 20KHz.

But you can get around this with a class D implementation, since you can apply a lot of feedback with class D- so much that the amp goes into oscillation. The trick is to use the oscillation as the switching frequency. Then the amp might only have good bandwidth to 20KHz, but it can have so much feedback that its able to correct for phase shift that might otherwise be present on account of the filter at the output of the amp. Class D distortion is caused by errors in the the encoding and by something called 'dead time' (topic for another thread) but both tend to cause lower ordered harmonics. Because of this, a class D amp with a lot of feedback can have as much depth and stage width as a tube amplifier.
There are solid state amps that can make a sound stage as good as a tube amp, but most of them are class D.

The differences you hear between tubes and traditional solid state has been about distortion for the last 60 years. Solid state amps have higher ordered harmonic distortion that is more audible than that of tubes, and so tend to be bright. They also can have more distortion at lower power levels, and that can mess with low level detail. But on paper, they appear to have less distortion. But 'less' should be taken with a grain of salt, since the ear reacts differently depending on the kind of distortion being produced.


This stuff is pretty audible; its measurable too but understanding what the measurements mean, and more importantly getting the right measurements is a bit of a trick. One trick that is used commonly in audio is to measure the distortion of the amplifier at 100 Hz. Any solid state amp will measure well at this frequency. But the problem is that such amps have feedback to control distortion, which goes down as frequency increases, because the Gain Bandwidth Product on which the feedback relies is insufficient. So distortion rises with frequency. That is why its measured at 100 Hz, to cover up this problem.

The has been going on for so long that many people, including people that test amplifiers, don't think of it as a problem. But if you measure that amplifier at 3KHz things start to look different- you can see how its distortion is increasing.

So you have two ways to avoid this- either no feedback at all, or so much feedback that the amp is able to compensate for the distortion added due to the operation of the feedback. Feedback does this, if in **insufficient** amounts, through a process called 'bifurcation'. When there's **enough** feedback, this process is suppressed. But that takes north of 35dB, and traditionally to do that at 8-15KHz hasn't really been possible with most of the semiconductors available to designers in past decades. So we still have tubes, which are fairly linear even with no feedback, although they will make more 2nd and/or 3rd harmonic.

The lower ordered harmonics, the 2nd, 3rd and 4th, get treated by the ear the same way- they add a little 'bloom' or 'warmth' (audiophile terms for this) which is relatively innocuous compared to the higher orders, to which the ear is keenly sensitive. If the 2nd or 3rd is present in enough quantity, it will mask the presence of the higher orders and the amp will sound smooth. They also allow you to hear a greater perception of the sound stage- this is the part that most people don't get. The better class D amps have a similar distortion signature (although at a much lower level than tubes, but nevertheless mostly lower orders) and so they can sound quite tube-like (i.e. musical), including the wider deeper soundstage.

The thing is, you might think this to be an error of amps like this, but if you've had the opportunity to hear what the actual musical performance sounded like, you find out that its helpful, because in this way despite the distortion, the sound stage is presented much closer to the original than amps that don't have this property!