#lm-1

An excellent view of a Lunar Module (LM-1) ascent stage under construction at Grumman's Bethpage, Long Island facility.

Date: 1967(?)

777 9311 Hours of Work – Restoring the First Digital Drum Machine

I’ve always been a huge fan of Prince. His trademark sound owes a lot to the Linn LM-1, the first digital drum machine ever, designed by Roger Linn in California in 1979. Only around 500 were made and Prince actually had three of them at Paisley Park studios. Here’s the story of my #315 made in 1981.


I never expected to find a LM-1 for sale ever. Despite its scarcity in early 2009 I tracked down a LM-1 on a second-hand music gear website in Germany. The seller confessed it had for years served as a cellar door stopper in a Swiss recording studio. It was serviced by Bruce Forat (a former Linn Electronics employee) in the US — who reportedly was not able not fix it 100 %. But we’ll come to that in a minute.

She Blinded Me With … Magahony


LM1 #315 looked amazing, quite impressive in size. Perfect layout, orange print on black powder coated steel, luxurious mahogany side panels. 12 glorious 8 bit samples at 28 khz plus analogue CEM 3320 filters for bassdrum and toms/congas.


The CEMs are adjustable internally via trim pots, which was only introduced on revision 2. As in revision 3 though, there was no external sync on this LM-1 originally and no clock rate switch in the back. But it had a low output jack which was supposably removed on revision 3. All that makes my LM-1 some kind of revision 2 and a half, technically. The clock input was retrofitted at some point in the past.


That’s it for the basics. The rest turned out to be pure esoterics. 


I never figured out when or when not it would work. Sometime I had to turn it up 20 times, before it would run. On some occasions, it would just go berserk in the middle of a programming session, preferably while messing with complex rythms and before I saved it that pattern of course. Faders where performing poorly, too. And only recently I found out it was also sensitive to room temperature … and maybe the axial tilt of the earth?

All that called for a checkup pretty early in my ownerhsip, but who’d do it if even Bruce Forat had turned it down? And why would I keep that four decades old piece of early digital gear with little to no documentation at all? 


The answer is simple: Nothing sounds like the LM-1!

Bring the Beat Back


I introduced Alex, a talented friend of mine with a masters degree in electronics, to the quirks and features of the LM-1, and then we started off restauration.

First, I got new faders from synthpatchers in Canada for a whopping 239 $. That’s a lot of money for a de facto 99 ¢ mouser.com fader. (The original faders are Mexico 112 1002 94, the same as used in ARP synths.) The only problem is to find out which of the thousands available fit into the LM-1, because the sockets are said to be different from the stock one, but I can’t verify that.


You either have to take the stock ARP ones and modify them, use other faders and built an adaptor yourself — or just go with the synthpatchers.com parts that will just work as designed (the original faders were performing very poorly so cleaning was no option either).

The faders arrived after four weeks and another whopping 80 € for shipping plus customs. Globalization doesn’t work for everyone …

According to the service sheets and schematics, the LM-1 power supply runs on 15 volts. Actually, it’s a 12 volt Power-One HTAA-16W power supply with a little “power boost” that did not provide equal current anymore. We were unsure how long the power supply would still work correctly after 40 years. Since it’s a crucial component and we wanted to avoid future consequences for the rest of the circuitry, instead of recapping it we simply replaced it with a modern switch power supply running stable on exactly 15 volts.

Those rechargeable batteries were also replaced along with a few rotten voice board capacitors (see white circle). Some display pins had also gone loose and were fitted back in.

It turned out the startup issue was a consequence of a glitch in the power supply and the display pins. When you switch on the LM-1, the OS boots within seconds and if that initial current is incorrect, it won’t boot at all.

More Light

This is a historic piece of gear, so any restoration is of course also a question of philosophy. Do you want a museum piece, maybe not working 100 %, but 100 % original? Or do you want to use it the way Roger Linn designed it — as a musical instrument, ready to perform? I decided on the latter.

Repairs aside, any non-reversable modification was not an option for me. But an easily removeable and musically justifiable mod – why not?

After Alex had figured everything out, the LM-1 worked perfectly. Since it’s a revision 2 with three unused holes covered by blank jacks in the back, we were going through possible reversable modifications to the circuitry and having them controlled with pots sitting in those unused holes.

We decided on a filter cut-off and resonance mod. One that allowed for both the bassdrum CEM and toms/congas CEM to be controlled from the back.

To achieve cut-off change one resistor had to be slightly decreased in value, so the internal trim pot would cover the entire filter range, not just half of it like in the factory setting (it was not intended to be changed as a “sound feature”). For resonance Alex soldered in two extra resistors. Then he took two pots with two dual caps as CV source and wired them to the filters (see photo below; new switch power supply also installed).

Finding a musically usefull range for resonance took a bit of fiddling. Due to a missing VCA, self oscillation did not make sense (you would just get a constanc hum) and also VCF envelope modification, though theoretically possible, was turned down due to irrelevant results (the samples are just too short).

It was important to me to have the factory VCF and resonance setting available at any time, because that’s the sound the LM-1 is famous for. The new dual pots are set in exactly that way: at 12 o’clock it’s the factory sound, to the left and right you can now either enhance or erase high frequencies and artefacts — for the VCF.

For resonance, the setting is slightly different, because resonance wasn’t used in the original design at all: for resonance amount zero (factory spec) the pot is turned far left. As you turn it up, resonance increases.

That’s it! A total nerd topic, but I hope someone will find this usefull because documentation, as I said in the beginning, is pretty rare on this machine.

I’ll leave you with the greatest LM-1 beat ever made:

Quantization is the process of correcting, or shifting, imprecise musical notes and beats to underlying musical representation or grid. To preserve more of natural human timing nuances, percentage of quantization can be applied to in many sequencers or DAWs.

While swing, in short, means a method of transforming straight grooves, by timing of notes, to shuffled patterns. And when it comes to swing, the MPC sampler series has an iconic status for its groovy musical timing. Its influence on electronic and hip hop music cannot be denied.

The MPC's creator, Roger Linn, has claimed that he stumbled upon note quantizing and swing by accident when developing the Linn LM-1 drum computer: by only permitting 16th notes using 1 byte per 16th note, the sequencer program was correcting played timing errors, hence quantization. And by delaying the playback of alternate 16th notes, and by varying the amount of delay, the swing/shuffle feature was invented.

Linn's implementation of swing applied to quantized 16th-note beats is merely delaying the second 16th note within each 8th note, or all the even-numbered 16th notes within the beat (2, 4, 6, 8, etc.)

The Douglas-built S-IVB upper stage intended for Apollo 1 SA-204 mission was erected at Launch Complex 37B, Cape Kennedy, Florida. It was repurposed for Apollo 5 to send the unmanned Lunar Module (LM-5) into Earth orbit later that year.

Date: April 10, 1967

The ascent and descent stage of the Apollo Lunar Module (possibly LM-1 or LM-2) under construction at Grumman's facility in Bethpage, Long Island, New York.

Date: March 14, 1967

Apollo 5 Saturn IB (LM-1/SA-204) at night on LC-37B.

Date: January 19, 1968

A schematic highlighting the major milestones of the Apollo 5 mission to test LM-1.

Diagram of the Saturn IB used for the unmanned LM-1 test flight.

"A nearly perfect performance by the Saturn IB placed the S-IVB-204 stage and its LM-1 payload into an initial 163 by 222 kilometer orbit with an inclination of 31.6° following 10 minutes and 3.3 seconds of powered flight. After 35 seconds in orbit, the nose cone was successfully jettisoned with the four panels of the SLA deployed 9 minutes and 15 seconds later. LM-1 used its RCS to separate from S-IVB-204 at 23:38:58 GMT about halfway through its first revolution and into a 167 by 224 kilometer orbit. After separation, LM-1 changed its attitude to cold soak its propulsion system for the next two orbits.

Diagram showing the configuration of LM-1 inside of its Spacecraft Launch Adapter (SLA).

With its primary duties concluded, S-IVB-204 performed a number of engineering tests including the dumping of residual cryogenic propellants and helium pressurant through the stage’s J-2 engine. This procedure would help lighten the stage for easier control in orbit and prepare future S-IVB stages for use as a 'wet' orbital workshop as proposed for the Apollo Application Program which was planned to follow the initial Apollo lunar landing missions (a program which later evolved into Skylab). After the propellant dump was successfully completed at 01:19:33 GMT on January 23, the stage was in a 155 by 223 kilometer orbit. Although it was not tracked, the orbit of S-IVB-204 was expected to decay ten revolutions after the separation of LM-1 about 15½ hours after launch.

An artist conception of LM-1 separating from its spent S-IVB stage.

Following the three-hour cold soak of LM-1, a pair of burns were planned for the descent propulsion system (DPS) followed by two burns of the ascent propulsion system (ASE). The first 39-second burn of the DPS would start at a throttle setting of 10% then ramp up to full thrust for the last 12 seconds to simulate the initial deorbit burn which would start the descent towards the lunar surface. The second firing of the DPS would last for 739 seconds and use a series of throttle settings representative of an actual descent to the lunar surface. Immediately afterwards, the abort staging would be tested with an initial five-second burn of the APS. A subsequent firing of the APS would continue until the stage’s propellants were depleted after about 445 seconds completing the primary mission about 6½ hours after launch. Because the LM ascent stage was expected to be left in a comparatively long-lived 315 by 815 kilometer orbit after the completion of the last APS burn, extended mission activities were planned until the ascent stage depleted its consumables about seven hours later.

-Animation of LM-1 in orbit

At 02:47:49 GMT on January 23 (just shy of four hours after liftoff), LM-1 was commanded to start the first of two planned burns of the DPS but the engine unexpectedly shutdown after firing for only four seconds leaving the spacecraft in a 170 by 222 kilometer orbit instead of the planned 215 by 330 kilometer orbit. After examining the telemetry, ground controllers quickly located the source of the problem. The LM’s guidance computer had been programmed to abort the maneuver and shutdown the DPS if it did not provide the expected acceleration level after four seconds – a situation which would normally indicate a problem with the DPS. Because the pressure-fed propulsion system was purposely running at lower than nominal pressure for these tests, it would now take six seconds to reach full thrust. It was this oversight which resulted in the premature shutdown of the DPS.

Cutaway diagram of LM-1 used for the first unmanned test flight of the Lunar Module (LM)

As a result of the problem, a preplanned alternate mission was adopted by ground controllers which would meet the minimum mission requirements while keeping LM-1 in touch with tracking stations for key maneuvers.

An artist conception of the firing of the LM descent propulsion system (DPS) during the Apollo 5 mission.

With the guidance system deactivated, the DPS was ignited by ground command for a 33-second burn at 04:58:49 GMT during the fourth revolution. The second burn of the DPS for the alternate mission sequence was commanded at 04:59:54 GMT for an abbreviated 28-second burn.

This was followed by the abort staging test and a 60-second burn of the APS. All systems worked as intended during this alternate mission’s three burns. The 228 meter per second total change in velocity from these three propulsive maneuvers boosted LM-1 into a 172 by 961 kilometer orbit.

-Animation of LM-1 Ascent Stage in orbit.

After these first three firings of the propulsion systems, the primary control system was reactivated for the balance of the mission. Unfortunately the guidance computer, which had been in a passive mode during the abort staging, had not taken into account the change in spacecraft mass and used excessively long burns of the RCS to control attitude as if it had a fully loaded descent stage still attached. This resulted in higher than expected RCS usage and eventual propellant depletion after only about an hour. Fortunately the RCS could be configured to draw from the APS propellant supply to provide attitude control during the mission’s final burn. Because of the timing and other requirements of the burns in the alternate mission plan, this second burn of the APS would be in the retrograde direction which would send the spacecraft into Earth’s atmosphere ending the Apollo 5 mission.

Flight Director Gene Kranz (left) and Dr. Gilruth (right) shown in the Mission Control Center at the conclusion of the Apollo 5 mission

With the ground track of LM-1 beginning to drift beyond the mission’s tracking stations due to the one-orbit delay to implement the alternate mission, the remainder of the mission had to be completed by the next revolution. The second burn of the APS started at 06:32:20 GMT during the fifth revolution. As planned, the sequencer automatically closed the valves supplying the RCS with propellant about 161 seconds later. Without attitude control, the ascent stage began to tumble as the APS continued to fire for another 190 seconds before its propellants were finally depleted. The last telemetry was received from LM-1 at 06:40:18 GMT on January 23 ending the Apollo 5 mission 7 hours, 52 minutes and 10 seconds after launch. The LM-1 ascent stage reentered the Earth’s atmosphere and was destroyed over the Pacific Ocean some 640 kilometers off the coast of Central America. The inactive descent stage of LM-1 fell from orbit on February 12.

"Map showing the ground track of the Apollo 5 mission as flown and the location of tracking stations supporting the mission.

Although the Apollo 5 mission had encountered problems forcing a switch to an alternate mission plan, the overall performance of LM-1 was good enough to satisfy the mission’s main objectives. And with the requirement to certify the LM for crewed test flights satisfied, a potential second unmanned test flight with LM-2 was cancelled allowing one more mission to be cut from the Apollo program’s increasingly tight schedule. With LM-2 being unsuitable for manned flight without significant reworking to meet new requirements in the wake of the Apollo 1 fire, it was set aside as work continued on LM-3 for the first manned LM test flight on Apollo 9."

"SA-204, the fourth Saturn IB launch vehicle, developed by the Marshall Space Flight Center (MSFC), awaits its January 22, 1968 liftoff from Cape Canaveral, Florida for the unmarned Apollo 5 mission. Primary mission objectives included the verification of the Apollo Lunar Module's (LM) ascent and descent propulsion systems and an evaluation of the S-IVB stage instrument unit performance."

Date: January 1968