"The six TimbreLocks are bias pots. Most mistake the scheme for basic bias adjustments. That's not it. It's rather more far reaching. Obviously these pots adjust the load line of the tubes and range from 78mA to 92mA. But this also affects the load in the associated output transformers and the operating points of the magnetics where we get 90% of pentode power efficiency and the curves of a triode. That’s quite the trick. We obtain 15—almost 17 watts—from a single-ended KT88 in class A1 but with a triode transfer function due to how the series-parallel connected six amplifiers operate. The taps in our transformers are different, the way they connect to the tubes is different. There is a screen tap which used to be called ultralinear and another winding for the cathode for a type of local feedback. A Stockham paper from the 50s showed how a triode can be modeled as a pentode with feedback by removing the grids. When you take the grids out of a pentode, you get inherent feedback between plate and cathode. It’s simply a matter of view. Even though six single-ended amplifiers tied together only at the outputs running strategically dissimilar output transformers are more complex than the two series-connected amplifier halves of the Model 88, their behaviour of a perfect high-power triode is actually more precise. The yellow or factory/start setting shown below is 10:00, 11:00 and 2:00 o’clock for both banks. The left and right tube of each trio should remain fixed while the center ones are the wild cards or kinky valves. Their output transformers are dissimilar unlike the other two matched pairs. The OPTs of tubes 1 + 2 are the same, the OPTs of tubes 5 + 6 are the same yet dissimilar from 1 + 2. Only tubes 3 + 4 have different transformers. That makes for a total of four different OPTs to require matched tube pairs. All six tubes don’t have to be matched but the individual pairs do.

"I should add that the output power of 100 watts—130 watts from KT120— was a fringe benefit but not the primary design goal. The primary design goal simply relied on six rather than two variables to more accurately approach my theoretically modeled ideal triode behaviour. TimbreLock of course also accounts for tube aging. We can squeeze a lot of life from these KT88. I have run tubes for 2'500 to 3'000 hours without any decline I could detect. The way one detects small changes from tube aging is to swap in brand-new tubes. That’s what I did. I kept aging a complete set in the amplifier which I left playing 24/7. This set I compared at intervals to brand-new tubes. I tracked how many hours I could get from the original set before any changes became noticeable. It took more than 2'000 hours before I thought I could detect any. That was admittedly running them without interruption. If you turn the amps on and off for individual sessions, this number could change.


"Back to Brazil, our production line really works well now. But what would never work is state-of-the-art finishing and overall international customer support. I tried and spent huge amounts of money on it in the past to never attempt it again. If something broke, I’d send out another amp while the original was never returned. Instead it would eventually show up in the used market to have another repair requested. I just couldn’t keep track. The logistics involved were a nightmare. You could perhaps also say that I didn’t have exactly the right head for it and at the time lacked the proper partner who did. What technically could go wrong with the Model 88 at the time was the power transformer. This wasn't because it was badly designed or poorly made. It was due to an ergonomic design flaw. If you remember, all the TimbreLock LEDs then were red. The last (right-most) LED would light whenever current was 96mA. That said, you could keep turning the pot to reach 120 or even 130mA. The light would continue to show red. It never occurred to me that people might use the amps in this fashion, i.e. way well outside their intended operational parameters.


"After the MkII version of the Model 88 in 2005/2006, international business pretty much came to an end. Now I made all the LEDs green and only the last one red. The instructions in the manual were to never use the red LED. Actually it was perfectly safe if you set it where it just lit up and didn’t keep raising bias current beyond it. So this was a design flaw of ergonomics that didn't foresee how a customer could interface with the device to eventually take out the power transformer. That was basically the only problem we ever had with the Model 88. With the MkIII or M88 A3 as we call it, I went to three colors. The TimbreLock LEDs are green, the last one is red and the preferred/initial setting is yellow. This proved to be altogether more intuitive and user-friendly. It simply works. You have the factory setting of the Model 88 at 10:00 and 2:00 for the left and right TimbreLock, you have the ‘don’t go here’ red one at the end and all the personal adjustments in-between. The MkIII version also got even beefier power transformers which now are toroids. Those have advantages and disadvantages but I worked around the latter. Now the operational temperatures are lower and there is much more power headroom. In Europe the new Model 88 monos retail for €19'000/pr., the matching Model 5 preamplifier with remote volume goes for €9'000. A set shipping from Brazil is €28'000. We eventually plan on stock in Switzerland to expedite European shipments but currently we deliver all shippable models from Brazil within 30 days and accompanied by a 3-year standard warranty.

"The Model 5 comes fully into its own when it replaces a very good preamp in an otherwise non-Audiopax system. A good/right speaker for the Model 88 would be 90dB or higher, with an impedance above 6 ohms (5.5 minimum) and phase angles not over 40°. If those factors are met, the combination should work very well. The Mosfet version of the Model 88 which we currently don’t make behaves in exactly the same manner. It isn't intended for more challenging loads either. Sonically the minor tube microphonics give them slightly greater soundstaging capacity from their quasi dither. But my sound really isn’t about any particular output device, be it tube, transistor or hybrid. The moment such a device is coupled to a transformer, everything changes. Of course you couldn’t just use any transformer or tube. You must choose one that is copasetic with your concept so you can manipulate it to the desired point. It needs to be twisted in a kinky way to produce the real results.