high vs low beta transistor comparison in the x0xb0x

In the x0xb0x (mods page) we're introduced to the idea that high beta transistors give a "more pronounced" resonance, which of course started a holy quest by the tweakers on the x0xx0x/x0xmods forum... Here we'd like to show semi-scientificly what kind of effect different betas have on the x0xb0x.

Transistors were tried on 3 x0x PCB locations: Q8, Q9, and Q10. These are the 3 locations "the german" says should matter when choosing high beta valued transistors to create a "more pronounced" resonance in the x0xb0x.

Transistors tested

733k:
Q8: 412, Q9: 400, Q10: 399  (beta/hfe)

733p (high beta):
Q8: 340, Q9: 320, Q10: 319  (beta/hfe)

733p (low beta):
Q8: 249, Q9: 250, Q10: 252  (beta/hfe)

x0x knob settings tested

saw 0 100 100 100 100 100
saw 50 75 65 100 100 100
saw 100 100 65 100 100 100
sqr 0 100 100 100 100 100
sqr 50 75 65 100 100 100
sqr 100 100 65 100 100 100

NOTE: Values are transistor beta (gain measured by radioshack multimeter).

Qualitative results:

733k and 733p(hb) sound very similar (difference of 60-80 beta)
733k and 733p(lb) sound different (difference of ~150beta)
- - 733p(lb) produces a higher resonant tone (or watery chirp) than with the 733k

At this point it would be good if someone spoke up about what exactly "more pronounced" resonance means (http://www.ladyada.net/make/x0xb0x/mods.html) but in simple listening tests, the higher beta 733k seem to make the x0x's resonant tones _lower_ in frequency than with a low beta 733p. It's possible "more pronounced" means louder, but if there is a difference in loudness, it's very subtle. We'll use the spectrograph to see:

Comparison of the sequence (knobs:100 100 65 100 100 100 100) using each type of transistor - 733k, 733p(hb), and 733p(lb).
As you can see here, as beta increases (top row) the resonant frequency drops

Comparison of the sequence (knobs:50 75 65 100 100 100 100.jpg) using each type of transistor - 733k, 733p(hb), and 733p(lb).
As you can see here, as beta increases (top row) the resonance loudness seems to stay stronger. With the low beta (bottom row) the resonance appears to become quieter faster.

Quantitative results

Looking at spectrographs (above), the resonance appears slightly louder for higher beta transistors.
We can also clearly see the correlation between beta and resonant frequency.

WARNING: Each .wav file is around 3MB

Conclusion

There is a subtle correlation between resonance loudness and transistor beta value where the resonance component becomes _louder_ as transistor beta rises, which while hard to hear, we can see on the spectrograph. While it was hard to hear a difference in loudness, it does seem to be there, and we suspect it may show it's benefits when using overdrive efx to strengthen the higher resonant harmonics (such as the high watery resonance the 303 has).

We've also seen a very clear correlation between resonance freqency and the transistor beta value - however this may not matter much, since resonance frequency can be tweaked with a trimpot.

So let's assume we have "tuned" the resonance the same between different transistor configurations (using the res calibration trimpot). A good question at this point: so what is "the best" beta value to use? Do we prefer the stronger resonance afforded with higher beta transistors? Or the faster tapering resonance found with lower beta transistor? Or does it really matter either way? Discuss amongst yourselves... :) I think I'll leave the 733k's in...

Future Work

Run some tests with overdrive to see if the louder resonance (with higher beta) shows up more, giving a "more pronounced" resonance.

Find a way to "dial in" this "more pronounced" resonance from low-beta to high-beta. Might be an interesting mod. Is there something other than transistors that could be tweaked to give a higher gain.

Could we simply boost the resonance a little (i.e. the resonance boost mod) to compensate for low beta transistors??

It's possible only one or two of the Q8/Q9/Q10 locations affect this resonance boost due to transistor beta value, it would be good to narrow down which socket is the cause.

A similar test should be done replacing all 733p's with highest beta (~400) 733k's, and then again with lowest beta (~250) 733p's... To see if there are other differences dues to beta. Remember "the german" predicts Q8/Q9/Q10 are the important ones... It would be interesting to see the data on this to back it up. It may turn out to be a lot of work for little result, but it would be nice to know.