6L6 comparison

[g1]

I came across this on the CE website in their tech articles, I hadn't looked through there before. I have seen the similar comparison for 12AX7's, but did not know they had done one for 6L6.
A couple things stood out to me. First was that the 'guitar player impressions' seemed to have zero correlation to anything they had analyzed. Or worse, sometimes seemed to contradict the findings.
Second, the 3 frequencies they used to 'generalize' seemed maybe not the best choice for guitar amps? (60Hz, 700Hz, 6KHz)
Third was the uniformity of the frequency response graph (near bottom of page). Yes, there were up to 3.5db variations in gain, but they were right across the board. If you could normalize or zero out all of them at 400 or 500Hz, I think the graph would show remarkably little variance.
It was not clear to me whether they re-biased for each tube, so that may account for much of the gain variation.

I guess the bottom line is that perceived differences in brands have little to do with frequency response as measured here (so nothing new then ).

Link to article: https://www.cedist.com/tech-articles...ent-made-tubes

[Helmholtz]

I guess the bottom line is that perceived differences in brands have little to do with frequency response as measured here (so nothing new then ).

^^^^That!

Tube parameters don't depend on frequency, so any tube is able to produce a flat frequency response from DC up to RF.
Provided the circuit supports a flat response.

Tubes differences in gm, Ra and shape of plate curves can somewhat influence the f-reponse of a given circuit.
That doesn't mean that tubes have a sound of their own. Different circuits will give different results.

A frequency response only makes sense for small signals ("linear range").
With distorted signal a f-resonse is not interpretable/comparable.

The response measurements in the article give rise to some questions:
- What is the f-reponse of the OT?
- Was the B+ supply stabilized?
- Speaker load or dummy resistor?
- What was the signal level/distortion?
- What the frequency response of the measuring equipment?
- How good is repeatability? With same tube and using different samples.

The small variations with frequency along each reponse don't look real. They could be caused by PS ripple.
I don't see such effect with my own measurements.

Anyway, looking at averaged responses the main difference between the tubes seems to be in gain (due to gm variation up to maybe 40%, which is considerable).
That is if the measurements can be trusted.

I think that sound differences between tubes are mainly due to different harmonics structure, not showing in f-reponse.

The whole article is rather questionable.