rafi

What follows is a description of some of the work that went into restoring Q Division’s 8068. While this isn’t a short article, it’s just the tip of the iceberg of the immense effort that went into the project. The end goal, a highly functioning and stunning sounding console, is now a reality.

Q Division’s relocation to Rindge Ave afforded us ample time to work on the Neve. The console was certainly functional, but it was long overdue for a thorough servicing – and Studio A was too busy to afford that much downtime.

Originally a quad board made in the mid 70’s, Q acquired the console in the 90’s from EMI Brazil. After being restored by Fred Hill, it was installed at Q1.0 on Albany street (~1993-1999). At Q2.0 it was in daily use from sometime in 2000 to July 2021 when we powered it down and packed up the Highland Ave Location. We joked that the console had been on longer than some of Q assistants have been alive. To oversee the process of restoring the console, we tapped Andrew Lypps, Associate Chief Engineer at Berklee College of Music, and long time friend who has provided us invaluable tech support over the years.

Maintaining the sonic integrity of the original Neve design was paramount, though functional improvements that reflected the way the world records music now, as well as the specific workflow at Q Division, were on the table. We looked at our service records and discussed recurring issues. Q’s Neve had been modified by Fred Hill in the early 90’s, before it was installed into Q1.0 on Albany Street. His modifications were very clever, adding layers of functionality and routing options. But, some of his modifications were not clearly documented, making service and repair difficult and confusing.

Restoration would fall into four basic categories

• Electronic (improving sonics and functionality).
• Mechanical and cosmetic (improving physical issues with switches and knobs, faceplates, legends, illumination etc).
• Examining modification and improving usability where possible.
• The last piece of the restoration effort that Andrew was vehement about was improving the documentation. Our documentation was in disarray – old schematics were damaged from being reproduced so many times, and some information was missing entirely – especially as far as modifications were concerned – so this last part was no small task.

Either as part of Fred Hill’s restoration or before, the original Neve spec tantalum capacitors had been changed out for electrolytic, perhaps because the tantalums are more prone to catastrophic failure. We did an A/B test and decided to restore the original Neve design. A total of 57 capacitors per channel had to be replaced; the console has 32 channels, for a total of 1824 capacitors. There are capacitors in a few other sections that were also replaced.

Faulty and/or dirty switches – rotary pots, stepped switches, push switches, swots (pots with built in switches) – are common issues on Neve’s of this age. All these mechanical and electrical components had to be carefully cleaned, repaired, and in some places replaced entirely. Many of the components are no longer manufactured, but can be made as special orders in large runs. Since we only needed a handful of components, ordering parts could get complicated. Sourcing parts in general is tricky – some parts are so rare we had to 3D print parts to repair or replace extinct components, or we adapted available parts by hand.

To clean all the pots and switches, we proceeded with a thorough sonic bath. Before a sonic bath, the complete disassembly of every module is required, including the removal of cards, knobs, and switch caps. Then a three part cleaning process takes place – including the sonic bath itself, a rinse cycle (also in an ultrasonic tank), and finally a drying cycle. After that three part process all moving parts are lubricated and carefully exercised (failure to do so could result in total failure and damage to contacts). In addition to the moving parts that had to be cleaned, there were many points of contact that needed to be cleaned – card slots, edge connectors, relays, pins and the like. ALL of those were wiped and rubbed, lubricated, wiped again, and dried. Edge connectors on the console itself had the tensions of the pins readjusted.

We found ourselves in a difficult situation as we took the modules apart and prepared for recapping and cleaning. We long knew the console had an idiosyncratic anomaly that we didn’t quite understand, and which even at this point we didn’t realize was a known fault. Q’s Neve had something we called “goo,” which is affectionately referred to as “Neve Honey” amongst Neve users. It’s a sticky substance that permeated some modules, leaking everywhere, including into sensitive connections. We had figured out this “goo” was leaking out of the transformers, but we didn’t understand how or why it was leaking, and we weren’t sure how we’d solve the issue. There’s no real consensus in the Neve world about this issue, but we figured out that when transformers heat up, this substance oozes out of them – and the insides of a console that’s always on, is quite warm. The substance is incredibly sticky, and it follows gravity – so it was inside edge connectors, module frames, wire looms, on the floor of Studio A at 363 Highland Ave, and all sorts of spots in between. We had to get it off of all the surfaces! Since it’s so incredibly sticky, we chose elbow grease and isopropyl alcohol, heat, and tens of thousands of q tips of various kinds and sizes. There were some solvents that worked on this substance – but they were toxic and only made the task slightly easier. We re-sealed the leaking transformers using glue, and we tested this solution under heat lamps – the leaking stopped. We’ve tested and listened carefully and found no difference between transformers that have leaks and those that haven’t, and the improvement to signal flow and ease of service is massive.

As we removed knobs and switches from the modules it became clear some of the faceplates were quite damaged and would require re-screening. This cosmetic repair would require that each faceplate would be removed, stripped of existing paint, repainted, and re-screened. The signature Neve paint color had to be matched (originally known as Royal Air Force FAA Extra Dark Sea Grey, it was not something you can just get at a paint store), graphics for the screens had to be made, silk screens had to be manufactured, and a workshop that could sandblast, paint, and screen the panels had to be found. Nicole Anguish at Daykamp Creative was able to reproduce the original graphics at the correct size and even added a small Q to the module number. A company in western Mass was able to connect us to a silk screener, and took care of the cleaning, painting, and re-screening.

We knew we’d need to replace some knobs and switch caps – many Neve knobs are available from Hook Studios, on the west coast, though at no small cost. We replaced over 250 knobs on the console – worth noting that all of these knobs have an inner brass fitting and are fastened with metric grub screws that had to be sourced and replaced, then the knobs carefully calibrated when reinstalled. Some filters are dual concentric knobs, and their shafts were damaged; new shafts were hand made by Lawrence Fine at his workshop in Hudson Mass. Some switch caps either do not exist at all, don’t exist in the correct colors, and certainly do not exist with the correct legends. Here Andrew came to the rescue by 3D printing parts and etching switch caps. Switch cap legends that had faded were repainted with black or white paint, and a clear coat of protective paint was applied at the end. Some faulty switches and their brackets are impossible to find, and require careful redesign of existing materials to make them look and function the way they’re supposed to.

The console had many incandescent bulbs, but we wanted to replace them all with LED’s. In addition, existing LED’s were either dim or were in various colors and levels of brightness, so all existing LED’s were replaced with new LED’s. The housing for many of the old LEDs had decayed and required replacing; of course, not all LEDs are the same size, so a variety of LED housing had to be sourced (some trial and error there). Choosing colors and brightness was a process of trial and error, and some locations where incandescent lamps existed required painstaking research and even more trial and error before we found lamps that worked, fit, and looked correct. In more than one instant we found products that were perfect, only to learn they had been discontinued by the time we were ready to order them. In one case a company agreed to remanufacture the bulbs they had discontinued, if we ordered them in a large enough quantity! The total number of LED per channel is 9 (multiplied by 32 and that’s 288 on the modules and channel strips alone); there are 31 bulbs for meters, 32 bulbs for channel mutes, 14 bulbs in the center section, 6 incandescent bulbs on the power supplies were replaced with LEDs, 2 different LEDs for solo indicators in the center section and meter bridge, and 12 LEDS were retrofitted to two banks of Speaker Select and Playback Select switches that were not illuminated. That’s a total of 355 lamps! LEDtronics were instrumental in sourcing some of these hard to find bulbs.

Two VU meters were missing entirely and would need to be sourced – on the original console one was a phase meter, and another was a digital clock! We were able to source a phase meter, an item that is very difficult to find, but the digital clock is unobtanium. Instead we added a meter to act as a solo bus meter (more about that modification later on).
As we worked through the recapping process we needed a way to test modules and cards w/out putting them back in the console. Q’s house tech Michael Healey custom built a test jig that can accommodate almost every component, including amps and relays; this jig is both impressive (it’s etched using a CNC mill), and super useful to have while the main frame was powered down and every inch cleaned and straightened.

Having a test jig meant we could really spend time restoring the console’s frame; for instance, almost all the plastic rails that held modules had broken or were missing. We were able to find the company in England that holds the original designs and has a limited stash of newer rails (in the two sizes we needed). Those were shipped to use by our dear friend Mark Dyde as the company does not ship products overseas.

It will come as no surprise that various screws had gone missing over time and that threads were stripped in all sorts of places – on modules and the frame. We learned that Neve used three different screw thread conventions – imperial, metric, and British Association threading, making replacing screws a strange scavenger hunt. Almost all screws were replaced and match the original spec, a painstaking process. Some parts of course cannot be replaced, and had to be retapped or carefully manipulated to line up and secure modules.

As mentioned, the documentation we have on hand for the console is vast but in various states of readability. Some sections are missing entirely, and some modifications Fred Hill made are only notated via description and rudimentary sketches – almost back of the envelope sketches. Andrew Lypps not only reverse engineered and figured out all the modifications, but drew beautiful new schematics where the original are incomplete or missing. In addition, he drew up a comprehensive wiring schedule that we used to check every single wire and connection on the console frame. During that process we removed old wiring that was no longer necessary (some of which was sticky with “Neve Honey”).

Our Neve is a 32 channel console – of which 8 of the channels had different faders and fader caps. In addition, the legends on most of the fader faceplates were pretty won out, and some were falling apart. We were able to source more of the original Penny and Giles faders, so that we’d be able to install 32 identical faders with easy to read legends on all of their faceplates. The original Neve fader caps are beautiful and intricate, but are impossible to source and complicated to recreate, so for the time being we’re not reinstalling those. Of course we cleaned all the faders and lubricated them, and thoroughly tested them.
The arm rest was rewrapped, as were a few other panels that are wrapped in fabric. New panels were manufactured where old ones were missing.

New Modifications:
Turning phantom power on and off was a feature that Rupert Neve felt wasn’t super necessary. Today, turning phantom power off is a very common feature of most mic pre’s and consoles. We felt it would be helpful to have a way to turn the phantom power off on occasion – especially during session set up and tear down, and while troubleshooting mics. Adding a phantom power on/off switch to each of the 32 channels would require significant modification to modules, faceplates, and other work besides, so that was out of the question. We agreed that a global on/off switch would suffice, and decided to add one near the oscillator panel, so engineers can quickly cut phantom power to the entire board.

Consoles often have a “dim” switch, allowing the control room volume to be attenuated as needed – usually by a variable amount that can be controlled by the engineer. However the dim pot on the Neve (the pot that controls how much the dim switch attenuates) is not only set behind the console, it’s also two separate pots – for the left and right. We’ve brought those separate left and right controls together as one pot, which is now installed in the front of the console, where engineers can easily reach it.
We’ve changed the functionality of one of the LED’s on each channel strip – changing a green LED that indicated a “channel active status” to a red LED that indicated a monitor CUT instead – the proximity of this LED to the cut switch and monitor fader made it make more sense this way, plus there are other ways to discern if a channel is “active.”

In the meter section we added a SOLO meter, allowing the engineer to see the level of the solo buss. As we were thinking about the meters and their functionality we learned that when Neve engineers ran test measurements they always took the meters out of the path, as they add a certain amount of distortion to playback. After much consideration we’ve decided to source the meters for the playback select section from a different spot along the signal path. This allows for cleaner monitoring, and some added flexibility to the functionality of the meters.

We’ve rebuilt the existing Playback Select and Speaker Select switches, adding illumination where none existed. Adding the illumination required 3D printing a bracket to hold the new bulbs, and running power where none existed. In the process of examining the switches we found out that a few would require replacement. We also made some modifications to improve functionality when assigning signal to a sub woofer.
Talk back functionality has been improved for workflows that are more common these days – i.e. individual headphone mix systems that take advantage of sends from pro tools. In those scenarios routing the built in talk back was sometimes tricky, but will now be easier with dedicated talk back patch points and both latching and non latching switches. The built-in oscillator has also been recapped and its use-ability and flexibility improved.

Power supplies had been restored a few years back when we ran into persistent issues, so those needed relatively little work, but we replaced incandescent lamps, and made new shorter cable runs from the power supplies to the console, and new power cables which should aid in stability and functionality.

A user “service panel” populated with multi pin ELCO connectors was added to the back of the patchbay, improving engineers ability to quickly print to tape, transfer to pro tools, and monitor either source. This is a workflow that’s common in certain sessions that previously required an array of patch cables (introducing numerous points of potential failure). And, because Q3.0’s live rooms are larger and there are more isolation spaces we’ve added functionality that allows engineers to choose which mic panel in performance spaces is normalled to the console’s mic pres, again cutting down on patch cables and potential points of failure.

To make using pro tools and screens more convenient and allow the engineer to sit in the center of the console while editing or mixing, we had a custom metal stand fabricated. It sits behind the meter bridge, and is bolted to the floor. The monitor arm is attached to that stand, allowing the arm to come down over the monitor bridge, where it can sit in the center section w/out causing too many any acoustic anomalies. Finally, Chuck built a tray that rests over the faders and the 2404 routing modules allowing a keyboard and mouse to be positioned anywhere on the console.

The team that made this massive project come together was lead by Andrew Lypps. Michael Healey and Chuck Hargreaves worked tirelessly on this project for many months, with incredible attention to detail and dedication. A handful of engineers “de-goo’d,” recapped, aligned, cleaned, dried, scrubbed, tightened, loosened, listened, and helped: Henry Soults, Megan Chase, Terrance Reeves, Colin Fleming, James Bridges, Joel Edinberg, Owen Smith (who did the lions share of wiring Studio A’s patch bay and panels!), and probably folks I’m forgetting.

I owe a serious debt of gratitude to this team. They faced all sorts of hurdles and surprises and dealt with them creatively, with humor, and perseverance. Making records on this console is now a real pleasure, and maintenance is much easier. The restored console will benefit musicians, engineers, and producers for years to come.

-Rafi Sofer, Head Engineer. Cambridge Mass, May 2023.