#linnlm1

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: