Once again, Don M provided organization and design concepts with Bill R adding his expertise in power supply design to the project. Our goal was a bit loftier this time around. We would offer preamplifiers to be built where the builder could choose any mix of options for features such as remote control, phono preamp boards, balance control, stereo/mono etc... Another nice feature was that Chuck B would again provide his expertise in scoping our preamplifiers using his recently constructed distortion magnifier so that we could adjust and control even-odd harmonic distortion and phasing. This eventually proved priceless as the sonics of the preamps definitely improved after the final adjustments were made. 

As with our F5 project, a full write-up by Mr Pass is available as a thread on the DIYaudio website. We sourced circuit boards from DIY audio. We scrounged for parts from many sources and once again were assigned the task of sourcing the ever increasingly rare Toshiba transistors. Rather than spend hours fashioning chassis to accommodate connections, we drew a CAD file so that Par-Metal could pre-stamp the rear panels for us. The small additional cost was well worth it as the time savings were significant and the results were professional.

Our first task was to come up with a low noise regulated supply. There are many DIY options for this. Bill and Don settled on a CRC high filtering supply board. Remote control boards were given their own dedicated small toroid transformer. Phono stage boards were also given their own dedicated power supply boards which were the same as those used for the BA-3 gain stage boards. We used Antek Inc. toroid transformers mounted within heavy shields. The shields had a slightly higher cost than the transformers! The transformer outputs went to a chassis mounted bridge R followed by 1500uf filtering per rail. This part of the circuit was placed near the toroid. A thin permalloy chassis divider was placed so that this transformer and initial power supply stage was somewhat isolated from polluting the remainder of the chassis boards and components with EMI. Our meters showed that this additional shield lowered EMI exposure at the BA-3 board by about 15 milligauss.

The power supply circuit board is based on the adjustable LM317/LM337 regulators. We installed 0.47ohm resistors on the front end of the power supply boards which added a bit of isolation and noise filtering when multiple PS boards were used off of the same transformer. Some tinkering with cap values on the regulator feedback loop and with the front and back end filtering amounts enabled our supplies to have the noise ripple below the resolution capability of Bill's o-scope. The best results we have seen in literature is to get similar LM317/337 regulated circuits to the minus 90db level where any residual noise is only measurable at the high end of the audio band above 10khz. We were in that same ballpark. Lower noise power supplies are possible, but none that we were aware of at reasonable cost and complication required for a mass DIY project. 

Mike P offered his shop as a convenient central location with the added benefit of a clean environment and nice heating for our winter get-togethers. Seventeen members were building preamplifiers and there were about 12 varieties. This presented the challenge to keep everyone on task and at a similar pace. The BA-3 pcb is well thought out and can be populated and soldered in about an hour. As with the F5, more time was spent mounting chassis parts, customizing front plates and running wires than was required to populate and solder our supply and BA-3 pcb's.

Most of our remote controls were preassembled boards by Lite-Audio. Attention had to be paid with cable routing and grounding arrangements. We cut two traces to eliminate ground loop paths on the boards and added a lead to terminate at our chassis star ground. This is not the fault of the design, but these alterations are suggested for use in a single ended application. The remote board also drives input selector LED indicators. We fashioned our chassis fronts to accept the LED's and the remote IR receiver light pipe. Holes were drilled for controls as well.

At one extreme, Mike built a purist preamp. It had a rear mounted BA-3 board with rear mounted stepped attenuator and input selection controls which had shaft extrensions to the front panel. At the full featured end, we had units with remote control for volume and relay controlled input switching, a phono stage, and an output circuit for driving headphones. The BA-3 circuit makes an excellent headphone amplifier capable of driving cans with impedance as low as 32 ohms with no issues. A pre-built opamp based phono board was the most popular for that option. It is very similar to the popular Bugle phono stage with the only topical difference in how much feedback is applied globally verses locally. Most builders that applied these phono boards chose to swap out mylar caps for better grade polys and some even opted for swapping to higher performance op amps 

and slight changes to gain structure.

Once all units were completed, we had a session at the Pavek Museum where Chuck supplied us with a PC scope and software to enable us to analyze the noise and distortion characteristics of each preamp. Power supplies were set to match rails, BA-3 boards were pre biased a bit low. Units idled to come up to temperature where bias could be dialed in and DC output minimized. At that point, we could dial in the distortion characteristics with the P3 control on each channel. We favored to minimize third order below second order. A couple units had noise levels a bit too high, so some rework was required. We found that keeping wire size at a minimum to enable tight twisting of the conductors and power supply wires to be a key point. Twisting wires aids rejecting EMI/RF noise of the conductor pairs. It also aids in suppressing noise emission by AC lines. You can understand why some high end preamplifiers have power supplies carried in separate chassis. 

So how does it sound? I allowed my unit 100 hours of break in time before final adjustments. I have bias set to spec and odd order distortion set to minimum allowed which leaves a small bit of mostly 2nd order. My reference preamp had been a tube unit I built based on the Marantz 7c gain circuit, but using modern parts, boutique caps and various wonderful old stock 12AX7's along with a tube rectifier. I considered it the flagship component in my system. It didn't take long to realize that this BA-3 sounds better in my system. The biggest difference is in the sound staging. That is probably due to our ability to custom set each channel for exact balance and best/lowest distortion characteristics. I quickly decided to build a second BA-3 preamplifier, but this time with the purist approach of opting for a minimal featured unit with rear mounted input selection and stepped attenuation. Both BA-3 units I have built are now occupying places in my two systems. My once coveted tube preamp sits on the shelf collecting dust. 

Don M