"Tools, Not Toys"
Q&A with Bob Gallien by Tom Bowlus
We all greatly enjoyed hanging out with Bob Gallien at the Winter NAMM Show, but I couldn’t resist a few more follow-up questions to share with our readers.
Q: The Fusion 550 marked your first tube preamp design. Why tubes, and why now?
A: Market demand. Bass players have been asking us for years to bring out a tube version of our amplifiers. They wanted the punch and grind that our power amp design delivers, but with the warmth and compression of a tube preamp. Interestingly, the Fusion is well received by tube lovers, but the solid state guys still prefer our other models. There definitely is demand for both types of amplifier.
Q: Your design employs three tubes, and uses all six valve stages, whereas most other “hybrid” bass heads use one or two. And you are running these tubes at a more robust plate voltage than the competition. Why six valve stages, and why the higher plate voltage?
A: Amplifiers with only one tube and 15V supplies are not really getting the benefit of what tubes offer. With low supply voltages like that, you can only run the tube in the cathode follower configuration, which is simply a unity gain buffer stage with no tube transfer characteristics what so ever. To get real tube sound, you have to take advantage of the grid-to-plate transfer function, which requires you to run the plate at about 160V and requires a tube supply of around 300V. If you really want a tube preamp, then all the gain and EQ stages must also be tubes, so in our case that requires six tube stages. We make tools at GK, not toys.
Q: Your heads use a serial equalization section. Why is this important?
A: When I started (1967), bass amplifiers used a very simple three-control passive “T” tone stack popular in guitar amplifiers. This tone stack works great for guitars and is the main EQ system in use to this day for guitar. I realized early on that to solve the problems bass players were having (power, punch, and clarity), an entirely new approach was required. A different power amp topology and more detailed equalizer was the answer. The typical way then to get more EQ was to use a graphic equalizer. The problem with this type of equalizer is that it is designed to reduce room resonances, and because of the parallel band configuration, they introduced resonances of their own. These things were never intended to equalize instruments, and for the most part only made the problems worse. I worked on the problem mathematically for several months (over a hundred pages of loop equations, all by hand) and came up with the serial equalizer solution we
use to this day. This equalizer requires a lot more parts, but provides the maximum usable response modification without introducing resonances and nulls of its own. It is popular to test equalizers by sweeping them with a signal generator and showing the response. This type of testing does not really show the important characteristics of an equalizer. In the real world, an equalizer is presented with thousands of frequencies at the same time and in this condition the parallel equalizer will produce many nulls and resonances
that make things sound false. The GK serial equalizer does not have this problem. The proper test method for equalizers is to use square waves for the signal. Square waves contain thousands of frequencies and will accurately show how an equalizer actually behaves. Square waves will show severe ringing in high-Q parallel equalizers, reflecting the artificial character of the resulting sound they produce (as well as all the other undesirable anomalies). These undesirable responses are absent in the GK equalizer, which means it provides useful equalization over the full range of its controls.
Q: I notice that you set up your EQ section “backwards” from the norm, from high to low. Why is that?
A: My first amplifiers that used the serial equalizer (300B, 600B) were hand-wired with terminal strips. In order to reduce noise, a serial equalizer must have the high frequency bands first, so with handwiring the high frequency pots had to come first. I can always tell when the competition has been looking at my designs, because for no reason at all they run their equalizer high to low. To my knowledge, GK is the only company using a serial equalizer, therefore the only company with a reason to run the equalizer from high to low.
Q: I have heard you mention that intermodulation distortion (IM) is more important than total harmonic distortion (THD). Can you explain?
A: Harmonic distortion consists solely of harmonics of the fundamental. For a 100 Hz tone, the harmonics would be 200, 300, 400Hz, etc. These harmonics are already present in the signal from the instrument, so for low-order and moderate levels, these harmonics tend to make the signal sound different, but not necessarily distorted. In fact, low-ordered, low,level harmonics can make a signal sound fuller and more pleasing if in the proper mix. Intermodulation distortion is entirely different and very easy to hear. It results from two frequencies mixing together and coming out with new, nonharmonically related, frequencies. For example 400Hz mixed with 1kHz will produce the harmonics of 400Hz and 1kHz, but also weird frequencies, like 1400kHz (the sum) and 600Hz (the difference). These new frequencies are not present in the original signal and stand out like a sore thumb, even in small amounts. Any time you have nonlinearity in an amplifier, you will get both harmonic and intermodulation distortion.