ARP Avatar Monophonic Analog Synthesizer
Last Update 12-28-2019
The ARP Avatar is a 1977 guitar synthesizer designed to use the guitar
to play the synthesizer. All you had to do was install the special
pickup on your guitar. Unfortunately the guitar interface was
hopelessly unreliable because the Avatar was an unfinished product and
was rushed to market. The Avatar was the symbol of intransigent
infighting at the top echelons of ARP management - ARP founder Al
Pearlman was not a supporter of the project and the development party
headed by ARP president David Friend deliberately shut him out of the
project and brushed off his prophecy of failure.
The Avatar team bet the farm on the principle that there were four
times more guitar players than keyboard players. Back then there was a
great market interest of coupling the guitar to the synthesizer.
Intriguing market opportunity, but market research failed to answer
whether guitar players used to buying $50 pedals and $300 amps were
willing to dole out $3000 (in 1977 dollars!) for a guitar
synthesizer. Guitar players can be a notoriously conservative
lot, barely able to contemplate anything more sophisticated than the
tone controls on their guitar and their amp. I doubt that market
research considered that guitar players could grasp the concept of
analog synthesis with words like oscillators, filters, LFOs...
The Doomed
Development Project
ARP built two prototype polyphonic synthesizers; one was to be a
keyboard system while the other to be a guitar system, but the keyboard
concept was dropped early in the project in favor of the guitar project
which was renamed Centaur
VI. The products were a brute force
attempt at polyphony using the limited technology of the time, and they
were not programmable. “Brute force” meaning polyphony was
achieved without CPUs and with discrete circuitry, which amounted to
VERY HIGH component count (and high retail price). Due to
excessive heat from the 115 printed circuit boards inside a case that
was deeper than the legacy Oberheims, they ceased to function in a
short time. Al Pearlman did a failure analysis on the system and
determined that their mean time between breakdowns (MTBB) was about two
hours, which is very poor and not fit for the market. The product
would had retailed for about $20,000 (in 1977 dollars!). The
prototypes still exist but as of 2019 their whereabouts are unknown.
The Centaur was never finished and it was ultimately abandoned in favor
of the concept of re-using existing Odyssey synthesizer circuit boards,
coupled to a guitar interface. This product would be monophonic
and not polyphonic, and would be christened the Avatar. The
re-use concept was adopted because of the excessive R&D expenses
into the Centaur project. But the pitch-to-voltage conversion,
especially with the complex timbres and dynamics of an electric guitar,
was never finished to a reliable design and could not be achieved with
1970s technology.
Indeed, a thorough study of the Paul DeRocco’s patent on ARP’s
pitch-to-voltage conversion (US 4,300,431) revealed a very convoluted
design. At the time, guitar synthesizers were a very hot product
and ARP used every tool to conceal the working details from the
competition. The critical components were secretly built in a
room separate from the production line. The patent was not even
assigned to the ARP company,
probably an attempt to hide it from competitors. A related patent
US 3,851,265 predated DeRocco’s patent and was the work of a Leroy
Young, who was a consultant on the Centaur/Avatar products. This
invention was no less convoluted.
At the time, ARP’s convention was to “pot” their critical circuit
designs in modules to prevent competitors from stealing their designs
(although they actually hid infringement when Moog discovered their
patented lowpass VCF in one of ARP’s modules). ARP’s early
modules were potted entirely in rock hard epoxy; by the time of the
Avatar they
had changed the potting process to a thin epoxy top with silicone
filling to facilitate repair (still not easy!). The crucial
Avatar circuit boards (dubbed the “D Prime module” and “E Prime
module”) was encased in near-impenetrable and unrepairable rock hard
epoxy modules (on later Avatars the D Prime module was left
unpotted).
The patent reveals little information of the contents of the E Prime
module, only a partial abstract circuit. And it’s a big module
that is hardly representative of the circuit shown on the patent.
A former ARP employee revealed that the details of the module were kept
under lock and key at the factory, schematics were distributed
internally only to people with the “need to know”, and the construction
and calibration of the converter module was done in a closed secure
room separate from the production line. Even today, no one knows
the contents of the E Prime module - hardly a loss as it was a failed
design that never worked.
The ‘E’ prime module
converted the pitch of the guitar to a digital
pulse train; the ’D’
prime module was simply a discrete phase lock loop
that converted the pulse train from the E module to the CV for
controlling the synthesizer pitch. There is a lot of
pre-processing and detection of the guitar signal prior to the ‘E’
prime module, and
trying to follow the convoluted operation of the conversion will make
your head spin. One of the design goals was to reduce the latency
of the pitch conversion of low notes - on previous designs, the delay
between picking a low pitched string and the converted signal was very
noticeable. But because the Avatar only played one note at a
time, the pre-processing had to detect the strongest signal of the six
strings of the guitar. That meant the guitar player had to use
very clean picking technique on a single string, and even then the
conversion process was not always accurate.
The ARP guitar synthesizers were a concept enthusiastically hatched by
ARP president David Friend; Al Pearlman had no involvement on them and
did not share Friend’s enthusiasm as he foresaw the complications with
pitch-to-voltage conversion. In addition to the already excessive
R&D expenses, the company began purchasing a large inventory of
parts in anticipation of a successful product.
ARP rushed the Avatar to market before it was finished. They ran
out ot time (and dangerously depleted the company revenues) before the
conversion process could be perfected. Besides
the high $3000 retail price, guitar players were put off by the
unreliability of the pitch-to-voltage conversion which seldom resulted
in any consistent results. During the two year production only $1
million dollars worth of Avatar sales was produced. After the
sobering results became known, the company had an operating loss of
$700,000 and an inventory write-off of $300,000 from unsold
Avatars. The Avatar was a colossal market failure. The poor
sales and losses were far from adequate to recover the $4 million
dollar R&D expenses that went into development. The company
never emerged from the deep financial hole created by the
Centaur/Avatar debacle and ARP only survived a few more years.
While other products like the ill-timed Quadra and the hopelessly noisy
16-Voice Electronic Piano did not do the company any favors, many
former ARP employees point to the Avatar as the primary impetus of the
ARP bankruptcy and ultimate liquidation.
After the failure of the Avatar, Quadra, and Electronic Piano the board
of directors asked David Friend to resign. Paul DeRocco redeemed
himself as chief engineer of arguably ARP’s greatest (and last) hurrah
- the polyphonic Chroma, highly advanced for its time. top
The Avatar
Rediscovered
The Avatar was the unwanted orphan child for decades until synthesizer
enthusiasts in the vintage synth renaissance realized it had the same
synthesizer guts of an ARP Odyssey. The Avatar included
CV/Trigger input and output jacks so it can be readily utilized as a
standalone synth. Many owners use them as an “Odyssey expander”
module without keyboard, thus it was only about 2015 that the Avatar
begun to be appreciated. One peek at the picture with Odyssey
next to Avatar at the top of this page reveals the close similarities.
For a long time I pretty much avoided ARP products due to their
achilles heel: slidepots. The slidepots seize up with age because
the lubricant
dries out, and the carbon resistive elements wear out with use.
Frankly they were not long life parts. Re-lubing them required a
labor-intensive disassembly, and I found that they could be broken by
the disassembly process too easily. For a long time, brand new
replacement slidepots were nowhere to be found. You could buy
“NOS” parts, but the risk of wear was still present. CTS still
makes the slidepots but they could only be purchased in large
quantities and were not available through low quantity outlets like
Allied,
Digikey, Mouser, Newark, et al. There is no industry standard PC
board footprint for slidepots, which complicates locating a direct
replacement. Only recently were new replacements available which
solved the footprint problem by using an adapter PC board between the
pot and the circuit board.
For a long time, the sole ARP product in my arsenal was the ProSoloist;
over time it suffered a worn slidepot that could not be salvaged (and
that control was hardly abused when I owned it!). Rotary pot was
the only option. Knowing that the other slidepots would not be
far from failure, I removed the remaining ones and replaced them with
rotary pots. There was ample room on the panel for rotary pots
and it was a satisfactory solution. Not so with most other ARP
products. The Odyssey - and the Avatar - possessed in excess of
thirty slidepots (the ProSoloist only had four), and the front panel
did not have sufficient room for rotary pots and it would had been
uglier than the ProSoloist. That left only refurbishing the
existing 30+ slidepots, which was a project I was not anxious to
undertake! top
Korg Reissues
the Odyssey
Korg introduced a respectable clone of the ARP Odyssey back in 2014,
and it sounded pretty good with a very vintage sound. One
addition that had great appeal to me was the
choice of multiple VCFs of ARP fame - 4023, 4035, or 4075. David
Friend was the consultant on the Korg reissue. They offered a
model with keyboard and a tabletop model without keyboard, and these
proved to be a popular seller.
To many other people, I recommend the Korg reissue over the vintage
deal because you won’t have to deal with bad slidepots. To date there
is one other company in Germany, which shall remain unnamed, offering
an Odyssey reissue but I refuse to give publicity to a chronic
plagiarist company with a history of stealing designs from other
companies and a history of bad quality products. top
Why I Chose
Vintage Over Reissue
I was playing in a 70s/80s classic rock band. The guitar player
owned a lot of music gear, and he was the original owner of an Avatar
that he didn’t know what to do with.
While I did recommend the reissue to users, there were many reasons why
an Avatar had greater appeal to me over a reissue. I wanted a
proper ARP synthesizer but I didn’t want another synth with a keyboard,
and the Avatar eliminated the trouble of replacing broken key contacts
and dried out bushings on the Pratt-Read keybed of the 1970s.
With the useless guitar
interface circuit
boards removed, this opens up additions such
as additional processors and a sequencer. The Avatar sports a
heftier power supply that can supply higher current, so with the
removal of the guitar interface this relieved the current demand on the
power supply which is available for other additions. There is
generous spare room on the front panel and inside the case to add new
controls, I/O patchpoint jacks, and circuit boards. Replacement
slidepots with LEDs in the control shaft were now available. It
is far easier to modify the 1970s through hole components of the
vintage than the SMT of the reissues. Korg chose to outfit a
professional synthesizer with “mini keys” which are cheap quality, and
I don’t like to modify my playing technique to a different key size
(Korg did offer a limited edition with full size keys). Korg even
replicated the ARP PPC pressure pads for pitch bend/vibrato, but I
never liked those controls. The MIDI implementation in the Korg
reissue was woefully inadequate for my needs (note on/off, no vibrato,
nothing else).
The potential of a keyboardless “patched out” Avatar with a potential
MIDI retrofit appealed to me, so I asked the guitar player to bring his
Avatar to practice. True to their reputation, the slidepots were
seized up. But he attempted to free them using WD40 which
actually makes them worse. I learned that he last powered up his
Avatar YEARS ago. Knowing that tantalum capacitors of this age
can easily fail, they did just that when I applied power and the unit
produced the familiar foul smell of a failing tantalum cap. No
sound at all, and the rear panel was hot which indicated a short
circuit on one of the power rails (tantalum caps often fail with a
direct short circuit). We now have a fixer-upper. I offered
to buy it, but he wouldn’t sell.
I liked the large case of the Avatar, and the modulation routings were
actually pretty decent (they weren’t obvious at first glance).
Avatars used later Odyssey circuit designs which are more
reliable. CV/trigger interfacing permitted control from a remote
controller (IE sequencer or the monophonic control from my
Memorymoog).
Retrofits for MIDI and adding I/O patchpoints were available, turning
the Avatar into a modular synthesizer - a “poor man’s ARP 2600”.
I was intrigued by the modular synthesizer renaissance of the 21st
century; I had already owned a modular synthesizer back in 1981 and did
not want the temptation of a money pit due to their perpetual
expansion. A modularized Avatar presented an optimal compromise so I
still wanted one. There weren’t many Avatars available for sale
(ARP only sold about 300 Avatars, far fewer than the Odyssey) and I was
on the lookout for a fixer-upper which would be cheaper. A few
months later an ideal specimen showed up on an auction site at the
right price and I snatched it up. The seller reported that he
stopped using it after he smelled smoke - despite the similar smoke
issue it is not the same unit from the guitar player.
While the Avatar uses identical synthesizer circuits of the Odyssey, a
few modulation routings were different. In order to make this a
true Odyssey “module”, some slideswitch routings had to be changed to
“S&H Mixer” which is crucial for certain effects. While the
rear panel CV/Trigger jacks are convenient for a MIDI converter, you
have to route the external CV to the portamento processor (stock
portamento only works with the guitar converter CV). Since I had
no intention of using this with a guitar, the guitar interface circuit
boards would be removed. That necessitated a change in power
distribution, because the power supply is a Kelvin system that uses
force/sense lines that terminate at a circuit board, which happened to
be one of the circuit boards for the guitar interface due to be
removed. You can’t just re-route the power busses to the Odyssey
circuit boards. So I had to add a prototyping circuit board as an
auxiliary board for power distribution and for adding new circuits.
The Odyssey power supply isn’t terrible, but the Avatar power supply is
a different design which is faulty. This is the infamous version
that puts out overshoot on power cycles, resulting in blown CMOS ICs
downstream and it is the same power supply used in ARP Omni/Omni2
string machines. ARP issued a field service bulletin that adds a
transient voltage suppressor to solve the overshoot problem and it will
be implemented in my specimen. The service bulletin changes a
resistor to lower the power rails from 15 to 13 volts, but this was
aimed solely at the string machine and will not be implemented on my
Avatar as the analog synth circuits - totally different from a string
machine - rely on 15 volt rails. On the plus side, the Avatar
power supply is capable of supplying more current than the stock
Odyssey supply. While the Omni string machines were stuffed with
many many first generation CMOS ICs that were vulnerable to damage from
the faulty power supply, the Avatar synth circuits (sans guitar
interface) only has a few CMOS ICs. In fact the audio path in the
Odyssey/Avatar voice is almost entirely discrete with the exception of
the LM3900 Norton opamps in the VCF module. top
Planning Out
The Project
These are the extensive plans for my Avatar:
- replace all tantalum capacitors with electrolytics
- new trimpots, new caps for power bypass
- fix the cutoff design error in the 4075 filter
- remove the guitar interface circuit boards, leaving many controls
available to repurpose
- restore some modulation routings (S&H Mixer) to revert it to a
“true” Odyssey voice
- add master volume (rotary pot) and headphone amp
- tap noise circuit for white noise, repurpose switch to select
pink/white noise
- install CMS Mega Lumina kit
- install Kenton MIDI retrofit, add MIDI clock, start/stop, note
velocity,mod wheel, aftertouch
- add vibrato LFO
- add Ian Fritz’s AD/AR circuit (stock Avatar AR is AR only)
- add several well known Odyssey modifications (LFO free run,
single/multiple trigger, etc)
- add some processors from the 2600
- add clone of 4023 VCF & switch to change filter modes
- add Juergen Haible Frequency Shifter (!)
- add I/O patchpoints
- clone ARP sequencer, expand to 16/2 or 32/1 (!!!)
This would give me a modular synthesizer complete with MIDI interface
and sequencer in a compact package. Restoring any ARP synthesizer
requires a lot of labor, this Avatar was no exception. On top of
them, I’m throwing in quite the gamut of modifications to make this a
fully patchable “poor man’s 2600”. But I was looking forward to
the effort as owners of “patched out” Avatars have reported that the
effort was well worth it. This project was extensive enough that
I had to plot out strategic sub-projects on a spreadsheet, which was
necessary to preserve my sanity! I didn’t know it at the time
when I purchased the Avatar, but it also fits the ATA case for my
Minimoog Voyager!
ARP products have parts that do not age well - slidepots, trimpots, 1st
generation CMOS ICs, capacitors. ARP seemed to have favored RCA
for the majority of the active devices (CMOS, opamps, transistors)
which unfortunately have a history of failing with age. The
majority of parts can be replaced, something that wasn’t true until
about 2000 (mostly due to those damn slidepots). Since the
original CTS series 201 trimpots don’t seem to be available through the
usual low quantity outlets, I had to substitute them with a direct
replacement - Piher/Amphenol PT15LV18 series. Since the footprint
is symmetrical, I mounted the trimpots on the underside of the PC
boards so that I don’t have to disassemble the circuit boards to
calibrate it. top
Retrofits
There will be a bunch of orphaned controls and jacks for the
due-to-be-removed guitar interface, so they will be re-purposed.
First to go is the port for the guitar pickup, and I happened to have a
DIN jack for the MIDI retrofit that replaced it perfectly.
MIDI-to-CV conversion was supplied by the Kenton
ARP Odyssey MIDI Retrofit. This retrofit is for the
black/orange Odyssey but is suitable for the Avatar. Note that
this is a bare board retrofit which requires the services of a
competent synth service technician. This MIDI retrofit offers
more options than the MIDI implementation on the Korg clone - onboard
LFO configurable over MIDI, pitch bend, aftertouch, foot control, etc.
Renowned ARP specialist Phil
Cirocco of
CMS offered his Mega Lumina slidepot replacement kit for
Odysseys and Avatars. Some of the neat features of this kit is
the LFO rate slidepot LED blinks at the rate of the LFO, and the EG
slidepot LEDs show a visual representation of the envelope. At
the time Phil was not performing any installations so I grabbed one of
the last kits he had left in stock. The kit cautions that it
should be installed by a competent synth service technician, because
removing the old slidepots is a delicate task. While on the phone
with Phil, he inquired on my skill level. I assured him that I
had a BTEE degree and 30+ years of experience maintaining and modifying
my own analog synthesizers. Then I asked if he had heard of
“analoguediehard”, and I could hear the light bulb go off in his head
over the phone as all doubt was washed away. Nice to know he was
aware of my work from the internet.
Phil’s kit is designed for the Odyssey; it will work in the Avatar, but
it does not replace the three slidepots in the guitar control section
that were only on the Avatar. So I
ordered the extra parts from gmusynth.com.
These three controls will be re-purposed for the cloned
sequencer. top
Patchpoint I/O
Jacks
Patchpoint input jacks utilize a “normalled” architecture whereas a
default connection is made with nothing plugged in the jack, like the
ARP 2600. I always liked the concept of normalled jacks on the
2600 and it baffles me why no one has duplicated this in a commercial
product. Since I was adding a cloned ARP sequencer, this
architecture offered some exciting possibilities. There is a
patchpoint I/O kit offered for the Avatar on the net, but I passed on
it because I had reservations about the quality of the PC board mounted
1/8” jacks and the ribbon cable used for interconnects. The wire
gauge in ribbon cable is adequate for logic signals but not for power
rails or critical analog signals. I opted for the Switchcraft
1/8” jacks and plugs as they seem to be a solid interconnect, and
Switchcraft was one of the few OEMs that makes long life 1/8”
jacks. They did cost a little more than the kit, but I am a
strong subscriber of the adage “you get what you pay for”.
The Avatar has generous open space in the upper and lower margins of
the front panel where the sequencer and I/O patchpoints can be
added. But DON’T add anything above the AC wires in the power
supply, the high voltage lugs are exposed (!!!) and this is the most
restrictive space. I had seen patchpoint modifications done to
Odysseys and Avatars that included outputs of VCOs, LFO, Envelope
Generators, etc. One extensive job added ARP’s 4023 VCF - which
is a state variable multimode VCF - and added I/O jacks for the
multimode outputs. I wanted to do these for a full patchable
synth.
There’s two challenges to these modifications: if adding I/O jacks
above the circuit board we must be cautious of components causing any
obstructions, and any switches or I/O jacks must clear the space
between the circuit board and the front panel. Barring any
obstructions, the Switchcraft jacks meet this restriction. A
rotary switch to manually switch filter modes would be too big and it
would have to be located in the margins outside the circuit boards -
assuming you can FIND the desired rotary switch. The better
approach would be to use a set of 405x CMOS switches with a single deck
rotary switch which could now fit between the circuit board and front
panel. top
Cloning the ARP
Sequencer
Of all the hardware sequencers I have seen on the market, the ARP
sequencer is the most intriguing design. Since I desired to own
one, I thought why not
shoehorn a clone of the ARP sequencer in the Avatar as long as I have
the surplus
current from the power supply and I was going to all this effort.
With the displaced circuit boards gone and ample empty space on the
front panel, there was plenty of room for it. It seemed an
obvious addition, and I hadn’t seen it done yet.
There were kits available for cloning the ARP sequencer. These
included footprints for slidepots, but I wanted to use rotary pots
which would be far easier to install on the Avatar front panel in the
upper margin. Every year I go to a local hamfest for gathering
surplus electronic parts, and I had collected a set of surplus low
profile knobs that were perfect for laying the controls out on the
front panel in minimum space yet easy to manipulate. From the
same hamfest I had collected a set of DPDT on-off-on bat handle
switches that were perfect for the buss assign switches. I also
wanted to expand the number of stages from 16 to 32, giving a 16/2 (two
independent 16 stage sequences) or 32/1 design. I also felt that
the quantizer circuit was far more complicated than it should be.
Since I was facing design changes, instead of
buying the kit I opted to clone the circuit, with design changes, on a
prototype circuit
board. The three orphaned slidepots in the guitar control section
were ideal for clock controls, I had to order the Bourns slidepots with
adapters separately from the CMS kit. The SYSTEM ON/OFF button
and LED were perfect for the start/stop function on the sequencer, as
well as the footswitch jack. The decommissioned ‘D’
prime module from the guitar interface had a matched NPN-PNP expo
converter that I scavenged for the VCO clock in the sequencer
clone. I had to overcome the challenge of sourcing the 4PDT
switch by using a SPDT switch controlling a set of 4066 CMOS switches.
Between the I/O jacks, the components for the sequencer (specifically
32 small profile rotary pots with metal shafts), the Kenton MIDI
retrofit, and the CMS Mega Lumina kit this upgrade did get expensive -
about $1500. These are the high cost parts, not including the ICs
and other components for the added circuits. Add that to the
purchase price of a “fixer upper” Avatar and it’s easy to approach the
original retail price of these things… but the stock unit never
had all these additions I added. top
Initial Repairs
IN PROGRESS - CHECK BACK LATER top
Epilogue
ARP president David Friend made the decision to rush the unfinished
Avatar into production, which set ARP on the path to bankruptcy and
liquidation. While Friend was ARP's greatest promoter, he was not
an engineer with any grasp on the technology and its limits. His
contagious enthusiasm would be his downfall. It's a darn shame
that ARP was lost, because they had some very talented designers on
their staff. While they seemed stuck in TOS land (used by string
machines & electronic pianos which were going out of fashion) and
bent on using discrete designs when ICs were readily available, their
last hurrah - the Chroma - would had saved the company.
You don't buy an ARP to emulate the sound of a classic Minimoog, but
the Odyssey/Avatar synthesizer voice has plenty of tricks up its sleeve
that a Minimoog can never pull off. It's a good sounding
synthesizer with decent modulation options that aren't immediately
obvious. David Friend did design the front panel of the Odyssey
so he doesn't deserve to be lionized for his failures at ARP. top
Home
