Analog Obsession
Developing the LA-3A – The Return of the Clones
by Dennis Fink

The original UREI LA-3A was designed by Brad Plunkett in the late 1960s and was still in production when I started working for him as a design engineer in 1977. The LA-3A was positioned as a smaller, lower-cost solid-state version of the LA-2A, and was designed around the same T4 optical attenuator. Brad is one of the best analog designers I've known, and so it was pretty cool to be involved in re-creating his product at UA. Here's some of the story about developing the UA clone of his LA-3A. (For those of you who worry about this stuff, no stem cells or sheep were used in this cloning.)

It seemed that in every product meeting, someone would bring up the LA-3A. We were getting requests for its return since UA came out with the re-issues of the 1176 and LA-2A; it was the natural course to see the LA-3A make a return as well. Finally in early 2004, we had a product brief and started to define our approach. We pulled out the original schematic, and got a few original units of our own.

What should we do?
We discussed all manner of "enhancements" and modifications to the original unit. We talked to a number of well-respected studio types (a Bob here, an Alan there) and decided on creating an exact replica, with three small exceptions. We added XLRs for input and output, but preserved the barrier strip for back compatibility in existing installations. We used an IEC power line connector so UL wouldn't get cranky. We included a means to switch in the most popular "gain mod" that our advisers said was most important. With this feature, users who have implemented the gain mod in their own units could add new units with the same configuration by flipping a switch, instead of soldering on the PCB.

In every other way, our LA-3A would be an exact replica of the original. We have a tradition of staying true to the original product with the UA 1176LN and UA LA-2A, so our LA-3A carries on in this way.

Where do we get the damn parts?
One of the first challenges was to find the same components that were used in the original units. Some of the discrete transistors that were used were no longer in production...bummer. While I was attempting to qualify available substitutes for the transistors, our materials guy, Sean, did his magic and located some sources for the obsolete devices, and we were under way.

“It seemed that in every product meeting, someone would bring up the LA-3A. We were getting requests for its return since UA came out with the re-issues of the 1176 and LA-2A; it was the natural course to see the LA-3A make a return as well.”

Then there were the magnetics. At UREI we rolled our own transformers, so we had no off-the-shelf or old stock sources to raid. We did have some vendors who had cloned the transformers for the UA 1176LN and LA-2A, so we went back to them for this project. Like a turkey on thanksgiving, one of our vintage LA-3As was carved up to provide samples of the original input, output, and power transformers, as well as the autoformer used to drive the T4. Another part of the carcass will be used later when we get to PCB layout.

As for the passive components, we used carbon film resistors as in the original and got the original vendors for the pots. We already had some of the pots and meters from our previous cloning of 1176 and LA-2A, just as UREI used previous design components back then. Of course, we used our T4 from the LA-2A as well, having only to fab a smaller can for the 2U package.

Let's put it together
So while parts were being gathered and fabricated, it was time to create something to put them in. We drew a schematic in our CAD program, using the same reference designators as on the original schematic. It would help our PCB silkscreen look more like the original. When we got to the PCB layout, the ways of the old came back to mess with us.

Back at UREI, we used Bishop Graphics tape and donuts for layout by hand and Xacto knife. It was painful back then and would be painful now, but we wanted the PC traces to be exactly the same on our clone. No, we didn't go looking for some tape and donuts. (We did, however, have an Xacto knife.) Instead, we took a digital picture of the traces on the PCB from our turkey carcass, flipped it over, and imported it into our PC layout program. The PCB designer, Kevin, placed all the components and traces exactly as on the old board, fighting the orthogonal trace feature of the layout program. A picture of the top of the original was used as a guide to re-create the board's silkscreen graphics, too. We were cloning like an English geneticist with a sheep.

The remaining bits of the carcass were the chassis parts and front panel (kind of like the bones you use to make soup). A few hours with the calipers and we were ready to create the original mechanical parts in our CAD program. The chassis artwork was guided by another set of imported pictures. To tie everything together, we found the same cable and used the same wire colors everywhere.

The Results
A few months went by, Leon and Joe assembled the pieces, and we were looking at our first prototype of a UA LA-3A. We went out for beer.

Powering the beast up and debugging it showed a few component values were wrong on our schematic, and a few parts were in backward (beer?). That was easy to fix.

The PCB traces came out great.

So, Si and I then spent a bunch of time measuring with the AP and listening with our ears (our two favorite test instruments). The results were incredibly close to the original. Where there was a difference, it was very small, and in each case, the UA version was the better one.

We'd made our clone, and we don't have to keep it in a pen or shovel out anything. It's gonna enjoy roaming the range with our 1176 and LA-2A. Go buy one.