All pictures by Rob Tavaglione
In 1999, I purchased a 24-channel Soundcraft Ghost analog mixing desk. It has turned out to be the only true fixture in my busy, little mid-level studio. Although I’ve been very satisfied with its performance, that performance has slowly slipped over time, despite a well-controlled environment, plenty of maintenance and periodic repair. I knew that it was time for another round of such repair, and I prepared myself for long nights with bloodshot eyes, a soldering iron, a magnifying glass and handfuls of pan pots, input trims and jacks to coax into this affordable (approximately $6k) modular console’s innards.
I’d heard of companies that prided themselves in the refurb’ing and modification of mid-level analog mixers such as mine; by upgrading certain key components of these consoles, one could achieve increased performance at a fraction of the typical costs attributed to products with such performance.
After some basic research, I found Nashville-based Creation Audio Labs — makers of some very clever, unique audio products such as the MW1 Studio Tool — which devotes an entire section of its website to servicing of Soundcraft Ghost desks.
Further research revealed many testimonials lauding the improved performance as well as increased durability and stability of a “modded” Ghost. I then had enough encouragement to take the plunge, realizing that this process could reveal valuable information to not only Pro Audio Review readers who own Ghosts, but also to those who own similar analog mixers and are considering further investment into them, plus those who have seen the attractive low prices and abundance of used analog desks these days and are considering getting “out of the box.”
An underside view of Tavaglione’s Ghost, open and on the bench.The Soundcraft Ghost
First marketed in 1994, the Ghost is a comparatively old and resilient product, one that it is still available new today through Soundcraft dealers, although the console has recently been officially discontinued. New 24- and 32-channel Ghosts retail for about $6,700 and $7,600, respectively, and used models are typically available with widely variable pricing based on condition, location, etc.
Although the desk is modular — with one large and one small PCB for each channel — these channel strips are only accessible by removing the Ghost’s bottom panel. Such modularity was the crucial feature considered in my original purchase of the Ghost, but the console also offers the full feature set I required: an inline design (with mic/line input and tape return on the same channel strip), full meter bridge, four mute groups, a HPF and a direct out on each channel, defeatable EQ with two shelves and two fully parametric bands, eight auxes total (two of them stereo), four stereo returns, eight subgroups, an oscillator and facilities for multiple 2-tracks, monitors and cue mixes.
My Ghost has a small onboard computer for SMPTE sync, mute automation, MIDI controllers and machine control, but the Ghosts still available today are generally the LE models (the “lite edition,” if you will), without such computer and features. Frankly, I only use the mute groups for convenience now that I’m all DAW-ed up. My Ghost was still largely functional, but diminished in its sound quality from years of heavy use. Over the years, a number of noisy pan pots, scratchy tape return input trims and failed quarter-inch channel insert jacks had been replaced. The replacement insert jacks provided by Soundcraft were now Neutrik gold-plated contact models, but their important significance was unknown to me. My stereo imaging was now a bit mushy, noise floor was higher than ever, I had a few channels with low levels and subgroup 8 was out...so I needed some repair as well as modification.
(L-R) socket closeup; cleaning pots; CPS275 power supply.
Modding the Channels
After a six-plus hour drive from Charlotte, I arrived in Nashville to meet Brandon Arora of Creation Audio Labs with my Ghost and its power supply crammed in the back of my car. That trip saw a “whole lot of shakin’ goin’ on,” if you will, but was certainly less traumatic than a bumpy trip through the multiple hands, trucks and conveyors of UPS or Fed Ex.
Upon getting the patient up on the bench, I told Brandon of my problems and the troublesome channels. He explained that these were almost always the product of failing insert jacks. It seems the older, non-Neutrik/non-gold connectors slowly corroded over time and that even burnishing/cleaning was rarely enough to get proper contact and not slightly increase resistance, level loss, adding noise and affecting frequency response. A quick cleaning with Brandon’s favorite Techspray Contact Cleaner II (quite expensive, but quite excellent) revealed he was right: My channels and groups were OK, and my insert jacks were not.
We started the process by running a battery of bench tests to show the Ghost’s condition before any medical attention. Under PAR editorial director Frank Wells’ supervision, we ran tests measuring THD, noise, frequency response, output level, crosstalk and phase accuracy. As you can see in the accompanying bench test section, my Ghost’s performance was acceptable, but neither stellar nor up to original spec.
Brandon explained that possibly the most important aspect of the Ghost mod process was getting the power supply up to snuff; many mid-level consoles can sound considerably better simply with beefed up power supplies that can deliver the needed headroom and amperage at moments of peak power demand. The Ghost originally came with the PSM290 (and later the PSM300), which were Switch-Mode power supplies: transformer-less and lightweight, but inherently noisy.
Then came the first linear mode CPS250 power supply, which only had 160-volt amps available between the four power rails (+ 17.5 V, -17.5 V, 5V and 48 V) and was inadequate. Mine came with the CPS275, which offers 205-volt amps and is also still used with the K2 series of Soundcraft live desks. My power supply has failed twice previously, and Soundcraft techs were kind enough to advise me on the easy replacement of a bridging rectifier. Brandon recommended we replace that 25 A model with a 50 A one that can also handle more heat. We found that my bridging rectifier was unable to tighten up snugly to the heat sink and that a lack of heat dissipation may have been prompting the failures. We remedied this with some bonding cement for better heat transfer.
Left: Brandon Arora removing channels; right: Jon VanDoran depopulating the PCB.
Next up was removal of four channel strips that were to be souped-up with some better components. I had previously struggled with removing the 11mm nuts that retained each pot and was relieved to see Brandon had modified a socket wrench to grip the tricky little suckers, I’d soon be doing the same to a wrench myself. Brandon showed me about a dozen op amp (operational amplifier) chips that were to be replaced with high performance counterparts — parts that would hopefully lower distortion and increase phase accuracy with better THD specs and faster slew rates. The Ghost’s Texas Instrument chips had a THD figure of 0.003 percent and slew rate of 13 V/us typical, while the new National Semiconductor LM4562, Burr-Brown OPA134 and Analog Devices OP275 were more like 0.00008 percent and 20 V/us: cleaner by two decimal places and faster by a good margin.
THD & Slew Rate
Capacitors were also up for improvement with Panasonic FM caps going in, offering higher capacitance values and lower ESR values. This lower ESR value (a measure of resistance) makes for better performance, especially with 19 per channel. A pair of small caps was also added to the mic preamp section, improving noise performance. These new caps and the aforementioned op amps would improve performance but draw more power — another reason for power supply attention. Brandon and his team of Phil Gayfield and Jon VanDoran promptly de-populated, cleaned and re-populated the PCBs in fine fashion, making quick work of the desoldering and component removal (work that personally wears me out). The new chips were seated in sockets that allow future quick removal and replacement, without any soldering. Brandon further explained that many of the old components had pins exiting out the backside of the PCB that were bent downward, and this can make removal particularly time consuming and difficult, but these guys made it look pretty easy with their experience, custom tools and high performance de-soldering guns.
(L-R): Power supply bridging rectifiers; Channel PCB with the original ICs; Small PCB with a Neutrik gold TRS connector.
All pots were cleaned and lubricated, and all switches received a treatment of Teflon-based cleaner that lubricates. All critical points in the channel path were quickly re-soldered. The 100mm faders were lubricated, but not replaced: they hold up quite well and physically shield out contamination from the fader path. Last but not least, the all-important insert jacks were replaced with Neutrik goldplated versions.
Modding the Master
The center section of the Ghost contains large PCBs that handle multiple functions. Each board handles a pair of subgroups and either aux returns, aux send masters, talkback, oscillator, B-mix master or control room source selections. There are also a number of smaller PCBs that house the master jack field with nearly 20 TRS outputs: control room outs, mix out and inserts, group outs and inserts, aux sends and returns, etc. All of these jacks and the headphones jack (which normals for control room monitoring) were quickly replaced with Neutrik gold-plated ones.
Chips and caps were replaced as in the channel strips, with the requisite cleaning, lubricating and re-soldering. I considered removing the computer and possibly reduce noise, but Brandon assured me the computer was adequately isolated, not a noise problem and I didn’t want to give up my convenient mute groups.
Center and channel PCBs for the Soundcraft Ghost. Testing and Re-testing
After re-installing all the PCBs, blowing out dust, exercising the jacks and replacing the bottom cover (essential for proper grounding and improved noise performance), Brandon reran the bench tests to try and quantify the before/after differences. I’ll leave the finer points of analysis up to Mr. Wells but, suffice it to say, improvement was seen in all areas — some more than others. With improvements seen in crosstalk, phase accuracy, frequency response and THD, the question was whether or not I could hear such changes. Ready to get back to my production work, I headed back across I-40’s mountainous curves to Charlotte for some critical listening and mixing.
Back in the Studio
I had planned some exacting before/after mix tests, but dropped the ball and had mistakenly used some colorful compressors that were no longer in my possession for my “before” mix. Unable to use proper methodology, I luckily stumbled across a far more meaningful test: A client of mine had recorded a fine, 10-song, full-length release, but ran out of funding before final mixing. I mixed all the songs for my own pleasure prior to the Ghost’s mod then, upon my return home, my client was ready to finally do the “real” final mixes. I submitted my 10 mixes, and they liked the sound of all 10, but required some content edits on five of those songs.
With the now hot-rodded Ghost, we remixed those five songs as closely to the original mixes as we could. I could hear specific and significant positive differences: much improved stereo separation and imaging, more depth of field, greater punch (especially heard on snare drum), extended frequency response (both high and low) and just more liveliness in general. Respectful of my client’s slim budget, I convinced myself that the five old mixes were good enough for use. Within hours after remixing, my client called to schedule the remixing of the previous five songs! He explained that he would gladly spend the additional money because the new mixes were “so much more punchy, clear and exciting-sounding” when directly compared to the old ones. Money talks—and so does sound quality.
Now that I’ve experienced the benefits of my modded Ghost, I feel compelled to mod the remaining 20 channels, even though the master section mods appear to be significant in their own right.
I’m confident that my clients do hear the difference, and that can justify the additional expense. This whole endeavor has made me second-guess my ears and their judgments, so please read the sidebar testimonial on this topic from legendary producer/engineer Michael Wagener, who did his own summing/mixing tests with interesting conclusions.
I’m sure that many of you are wondering how your own analog console would fare from such treatment. Let me first recommend looking into power supply improvement, as any improvements to components will result in more current draw and some headroom here is pivotal. You will find an abundance of modern components (op amps, caps, etc.) that can improve performance, but the selection of those components should be left up to the “experts” — those well aware of longevity and performance expectations — to maximize your benefits and value in investment.
Technicians such as Brandon and the team at Creation Audio Labs have the experience and insight to evaluate your particular console and make specific recommendations. Who knows? You may be only a few thousand dollars away from having a powerhouse of an analog mixer that’s already in your
control room — or storage closet.
Rob Tavaglione has owned and operated Catalyst Recording in Charlotte, NC since 1995.
• • • • •
My Soundcraft Ghost Mod Shootout
by Michael Wagener.
When I started building the new WireWorld studio, I did a detailed shootout to find the best summing box for the new setup. With the exception of the Soundcraft Ghost, there were no other consoles involved at the time, just dedicated summing boxes. I had eight summing boxes from different manufacturers plus a couple of digital summing setups.
All analog summing solutions beat the digital summing solutions in depth of the overall picture. The Ghost sounded the best of all analog solutions by a big margin, but I don’t know if it was because it was adding something to the sound that wasn’t really there, or if it was the most natural sound of all of them. My personal feeling is that it added just a touch of high end to the overall sound and therefore stuck out from all the other analog boxes.
In the end, I decided to go with the Solid State Logic’s AWS 900 analog console/DAW controller based on the sound of the SSL X-Rack, which was in the test (the AWS 900 was not in the actual test, but it does sound the same as the X-Rack, which I compared at a later point). SSL just has that “neutral” tone without adding its own color but still creates that space around the separate instruments in the mix, which is so hard to get in the digital world.
There were other summing boxes that created an actual sound/color — which might be good for some digital mixing situations or even tracking to digital — but that was not what I was looking for.
• • • • •
On the Bench: Tavaglione’s Soundcraft Ghost — Before & After Mods
Before and after measurements of the modifications on Rob Tavaglione’s Ghost reveal easily demonstrable improvements resulting from the Creation Audio Labs modifications and refurb. Modest gains were measured in phase performance.
Interchannel crosstalk between modules was markedly improved, nominally 9 dB or better across the audio spectrum below 10 kHz. Significant improvements were made in noise performance.
Measuring the residual noise within narrow frequencies through a swept bandpass filter, noise components (particularly in the 50 Hz to 1 kHz range) were reduced as much as -20 dB, with at least a couple of dB of improvement across the audio range.
[Fig. 1 Noise performance-residual signal levels with no input measured through a swept bandpass filter and plotted against frequency-was significantly improved by the Creation Audio Labs modifications (Top, green trace, before mods. Bottom, red trace, after)]
THD+N (Total Harmonic Distortion plus Noise) measurements were also markedly improved by the modifications. The clipping point moved a dB higher and THD+N was lower, likely a combination of lower harmonic distortion and improved noise performance.
[Fig. 2 THD+N plotted against output level, while the input level was swept upward. As is typical, the performance improves with higher input levels (more signal in relation to the residual noise) until the signal clips, making square waves with high levels of harmonic content. The top, green trace is before modification. The bottom, red trace is after modification.]
You can view additional before and after plots on the Creation website that mirror the confirmation of the improvements measured on Rob’s Ghost.