I started work on designing Trinity in March of 2015.  Today I feel confident in saying I am done prototyping and ready to move into production... for real.  I said this once before, but it never happened.  I realized then that I rushing things for no reason.  I decided to take my time and create something I could be super proud of for the rest of my life.  Today I feel so massively accomplished.  It has been the most complicated journey of my life, but I'm a better, smarter man as a result.  I'm looking forward to sharing my masterpiece with the world. 

There have been several massive hurtles on this journey.  The biggest was noise, which I have done a great job of limiting (-90dB at full volume, drop the input volume to 50% and the noise is -110dB).  Second was the EQ section.  The EQ was strangely way more difficult to get perfectly integrated than I expected.  It would cause self oscillation at extreme settings and with the knobs centered the response was not flat.  There was also an 18kHz roll off regardless of knob positions.  I tried so many different EQ circuits with no success before finally getting smart.  After finally getting some software to run EQ simulations I was able to design an EQ that is perfectly flat in response when the knobs are centered, offers a good amount of boost and cut, and does not cause self oscillation at any setting.   I started with the "James" EQ, which is a passive version of the Baxandall circuit.  That circuit is not perfectly flat with the knobs centered, so I had to modify it.  The circuit still oscillated at extreme settings, but thankfully I came up with a mod that resolved this issue.  I got it worked out really good.  The other element of the EQ equation is the amplifier section that drives it.  The EQ has its own tube, which is out of circuit when the EQ is bypassed.  Its a 6922 cathode follower circuit with some negative feedback around it to create a super clean low impedance output to drive the EQ.  It sounds stupendous!  The EQ adds a bit of THD when engaged, but its not over the top.  It sounds very warm and musical.

Next week I'll do some thorough tests measurements for the Trinity page.  Next step is to build a production model.  Thanks to the high cost of prototyping it will have to wait a little bit.           

Without a doubt, this project is the most significant thing I've worked on to date.  Its a complicated analog device that needs to perform like a space ship (minus the whole going to space thing).  Lots of requirements like low noise, flat frequency response throughout and beyond the audio band, low THD, decent amount of gain, all controls work perfectly, no self oscillations at any setting, and tons of flexibility.  Basically it needs to do a ton of stuff really well and offer a sound that's highly desirable.

That all said, I went back to the drawing board about a month ago or so, with good initial results.  Its better now.  The current prototype has no EQ in the circuit yet.  I wanted to get everything else performing perfectly before designing the EQ section.  I'm writing this blog post today because I'm finally really satisfied with the amplifier.  It sounds absolutely stunning.  High resolution and pretty low noise with a decent amount of gain.  The sound is what its all about and it sounds nothing short of amazing.  The current topology is a single ended ultra-linear pentode amplifier driving a custom output transformer.  THD is 0.185% with a 2kHz sine wave feeding the input with 80mV and the volume knob on the channel strip set to 100%.  At that setting the noise floor is measured at -90dB A-weighted RMS and the gain delivered is 48dB.  Turn the volume knob down to 75% and the noise floor drops to -100dB.  Kick on the turbo and the gain jumps from 48dB to 57dB!  With the turbo engaged, the THD is just over .3% with the volume knob cranked.  With those settings, the A-weighted RMS noise floor is -80dB.     

There's still a lot of work to do, but at least now things are looking better than ever and most importantly, it sounds absolutely wonderful!