Bourns again?

[Fonotec]

The LS-76 was designed in the mid-'70s, and, at the time, it was apparently not practical to store data for more than 3 seconds in a standard TTL circuit. In order to compensate for the timing of the lathe's responses to the inner-wall, Left Program modulation, which are a turntable later, when it makes room for the the right channel's groove modulation, a copy of the (real-time) Left Program signal is sent to the lathe, along with the right and left advance signals, as well as an extra copy of the right advance signal. Even though it can only hold the left channel audio peak values for around 3 seconds, that's longer than necessary, since, this way, it only has to hold onto them for 1.8 seconds (the duration of each turntable revolution at 33.333 RPM), whereas, had it received the left channel audio that's used for feed expansions in advance of the cut (as it must do with the right channel audio), it would have 'forgotten' those values before it could use them after 3.6 seconds elapse (which would exceed the ~ 3-second memory limit). {Holding onto peak measurement data for much longer would be possible only four years later, as seen in the 1980 AES Convention debut of the Sontec Compudisk CD-80, Zuma II, and Neumann VMS-80 disk computers, using microprocessors.} But for 1976, using the copy of the left Program signal is an elegant solution with an only-somewhat-awkward requirement for those who work normally with just 2, or possibly 4, channels in disk-cutting consoles...

Even though it only needs to be able to process (with eq and dynamic correction) four channels of audio (the advance and program stereo signals), an LS-76 studio's stereo mastering console needs to be able to send six channels of ('all-analog') audio - Left and Right, and another Right (copy) of the Advance repro (head stack) signals; plus Left and Right, and another Left (copy) of the Program repro (head stack) signals.

{The depth automation circuit uses the Left and Right Advance signals, and the carriage-feed automation circuit uses the Left Program (copy) and Right Advance (copy) signals.}

In order to generate these analog copies and to be able to fine tune the levels of each channel to the millivolt (e.g., +4 dBu @ 1 kHz = 1.228 vRMS), I've been using a modified ATI MLA800 ('active' balanced output version), which has 8 channels (and am simply not using 2 of the 8). However, I replaced each of the op amps (which mercifully are already socketed) with Burr-Brown SoundPlus types and replaced the capacitors with mostly Wima polypropylene and the rest with Nichicon of the 105ºC tolerance. (Am using the Burr-Brown chips at Unity Gain, fwiw.)

Since this was no longer the ATI MLA800 that it used to be, I spray-painted the faceplate of the first one I modded silver and have been referring to the chassis as the Silver Saddle. Its two, stereo inputs (with jumpers connecting to the inputs of the copied channels) are located right after two, stereo, bridged H-pad attenuators (each having a 2.5-k-Ohm input and output impedance), so the Silver Saddle acts to buffer the signal between the passive attenuators and the various destinations for the calibrated signals - the cutting amp inputs and the lathe automation audio inputs.

Initially, I replaced the original 10 k-Ohm thumbwheel pots with 12-turn trimpots by TT Electronics.

TTEpot.jpg (25.42 KiB) Viewed 3194 times

Unfortunately, their pins were not in the same locations as those of the thumbwheel pots, so I've been using soldered jumpers to connect from the PCB through holes to the pins of the trimpots. Without any means of attaching the trimpots to the chassis, I resorted to affixing them to the faceplate with daubs of hot glue (...yes, many sides have been cut with compliments for LP pressings with the signals having gone through hot-glued pots - the secret to 'analog warmth"? ); The hot glue holds for a long time and doesn't interfere with the trimpot's use, but the jumpers kept getting disturbed when the glue would finally break (due to the disturbance caused them by my daily tweaking), and I'd try to glue some more on (like a more-on?);

Finally, last week or so, when I was trying to adjust one of the hot-glued, little trimpots, one of its three jumpers (in/out/wiper) broke off, along with one of the trimpot's pins. There was nothing but a stump to which no jumper could be soldered anymore.

Of course I was out of those TT Electronics trimpots, too.
{This didn't take me 'off the air' (as Len might say), since I have an identically-modified, additional ATI MLA800 that has a black front panel (so I call it, the 'Ebon Saddle'), but I wouldn't simply rely on it, since it uses the same, hot-glued trimpot-mod (for now) as was on the Silver Saddle...until just now.

But this was the very crisis I'd been waiting for, because it gave me the gumption to solve this fiddly issue at last (meaning after many years of just getting by with the precarious pots and their not being well attached, even though I could always eventually get each channel to behave for cuts, I'd, at last, sort this out properly).

On ebay, I found a small trove of 12 N[ew] O[ld} S[tock] Bourns TRIMMIT series, avionics-style, sealed pots. They were shipped swiftly from Ontario and arrived in good condition - all of them working perfectly. What's more, their size is perfect for the existing through holes of the face plate so that each trimmer can go in its own panel hole, and the ring-nut that came with each trimmer perfectly hugs the front of the trimmer onto the face plate. I just installed it and ran some test tones through it. Calibration was very simple, and the many turns of each trimmer allow for very precise adjustments without the huge jumps in level caused by the tiny trimpots made by TT.

Bourns_again.jpg (2.72 MiB) Viewed 3194 times

You might say my console is 'Bourns again'. );

Silver_Saddle_2Ø23.jpg (354.49 KiB) Viewed 3194 times