The Birth of the 3M M56 16 track Audio Recorder

Assistant Ellen Burke with the 3M-M56  during a session at Wally Heider studios San Francisco in 1971

 The Birth of the M56by Dale Manquen

The 1960s saw a tremendous amount of tape recorder development. In addition to new Ampex products, which included the AG350, MR70, AG440, and MM1000, there were new companies that entered the market. Scully, 3M, MCI, Stephens and AutoTec were among the new entries. At the same time, individuals were building custom configurations with new track formats using heads built by Lipps, Nortronics, IEM, Norton and Applied Magnetics. The 3 and 4-track recorders using ½” tape were soon eclipsed by 8 and 12-track 1”, and then 16 and 24-track 2”. The jump from 4 to 24 tracks took place in the blink of an eye – about 5 years! During this time studios were also scrambling to update their consoles with more inputs, more busses, and more flexible monitoring to accommodate overdubbing, the new way of recording.

3M entered the 2” market in a haphazard manner. I was working on a new electronics package that would reduce the size of the 3M module from 5 ¼” to 3 ½” using an inexpensive package consisting of a monocoque shell and two plug-in circuit boards. This simplified package, which had no internal provisions for Dynatrack, was primarily intended for 8 and 16-track machines. Unfortunately, I had made a few significant wrong assumptions when I started the design.

The 3M machine that never was


The first error was to use horizontal printed circuit boards with the components on the underside. This arrangement traps the heat against the components, causing transistor overheating. The second error was not understanding what was needed in terms of quality of the input transformer. I had chosen a small UTC transformer for the input circuit so that I could mount the transformer directly to one of the circuit boards.

Being the young fool that I was, we built up a batch of chasses and showed the breadboard machine at an AES Convention. I even presented a talk on the new electronics package. The breadboard had mostly unpopulated PC boards, except for the pots and switches. At least one channel had enough components to give a believable ‘show and tell’ at the convention.

Mystery electronics front w strips off
Mystery electronics front with cal strips removed and printed circuit boards
Mystery electronics both pcb assys
Mystery electronics both pcb assys

Once I understood the shortcomings of the design, I wasn’t really enthused about working on this model. Our Technical Director would ask how things were coming, and I would give enough of an encouraging report to keep him off my back.  At about the same time, we began experimenting with a 2” transport. Fortunately, the basic M23 design was rigid enough to accommodate 2” tape without any major structural changes.

I found a pair of the next size larger Bodine spooling motors used on a Mincom instrumentation recorder, and I figured if those motors would handle a 14” roll of 1” tape, then it should be able to handle a 10 ½” roll of 2” tape. The problem was that the motor was bigger and wouldn’t fit between the webs of the transport casting. I took apart the motor and decided that we could shave the sides off the motor without affecting the inner working parts. The machinist who was assigned to the job wasn’t quite so confident. When he had everything set up on the mill and he had cut away most of the side, he came into our audio lab and declared that the motor was going to fall apart, and he didn’t want to be responsible. We followed him back to his mill and assured him that everything would be OK. After we absolved him of any blame, he proceeded to make the final cuts that broke through the case, exposing the interior of the motor. He was obviously disappointed that the shell didn’t fall apart. In fact, he took a tool and poked at the motor, trying to prove to us that it was going to crumble. Nothing happened. The motor was still solid. We turned and walked away, and he finished that motor and its mate.

Figure 3 Cut-down Bodine spooling motor and Mincom instrumentation 1” reel spindle
Cut-down Bodine spooling motor and Mincom instrumentation 1” reel spindle

A few months later, when we were ready to take the prototype transport to the AES convention, those same motors were still in the deck. If you lifted up the transport, you could actually look inside the motors! Since this wasn’t a great confidence builder for prospective customers, we needed to do something to improve the appearance. Don Kahn, our mechanical designer, put strips of masking tape over the gaping holes and sprayed the tape with black Krylon paint. The holes were gone and everything looked quite normal.

We needed a 2” capstan, pinch rollers and reversing idler. The M23 used 4 sets of ridges and grooves on the Isoloop capstan for a 1” tape width. This gave us the ability to run ¼”, ½”, ¾” and 1” tape on the same capstan. The equivalent Mincom instrumentation machines used much wider ridges since they were never used for ¼” or ¾” tape. With this in mind, we chose a 4-ridge pattern for the 2” machine, giving us two ridges for 1” and 4 ridges for 2” tape.

We had experienced significant manufacturing problems with the Neoprene rollers that we had been using on the M23. The M23 rollers are made by wrapping a thin sheet of rubber around the metal core, and then vulcanizing the wound core. The beginning edge of the first layer can produce enough of a discontinuity in the rubber cylinder to create a flutter once-around bump. We decided to use a homogenous cast urethane called Monothane for the 2” rollers to eliminate this problem. Nobody mentioned that urethane could turn to goo after several years, and in those days the machines saw steady heavy service which required roller replacement due to wear long before the rollers could soften.

We also needed a reversing idler that was 2” tall. In our haste, we built the first idler out of solid aluminum, and we didn’t include any air bleed grooves to bleed off the air blanket that formed between the tape and roller at high speed. The blanket was so effective in disconnecting the tape from the roller that you could run the machine up to a high rewind speed and then use your finger to stop the roller and spin it in the opposite direction while the tape continued to spool at hundred of inches per second! The slippage during fast wind wasn’t a real problem, but it meant that the idler might not stop when the tape stopped! All later idlers had bleed grooves.

Two inch reversing idler 002
Original 2” reversing idler without air bleed grooves on left, standard unit on right

We ordered 2” 16-track heads from Applied Magnetics in Goleta, CA. Their salesman, Trevor Boyer (who later worked at Saki Magnetics), served as our liaison. I had them build a set of erase, record and play heads in the 16-track format. I later had a pair of 2”, full-track heads built for making calibration tapes. These heads had 1000 2-mil laminations stacked up to create the 2” core.

We needed twice as many connector pins on the head assembly to handle the 16-track heads. Fortunately, the M23 had been designed to be adaptable to IRIG instrumentation recording, which requires 2 head on each side of the Isoloop. As a result, the casting was designed with connector wells on both sides of the loop. We designed a thicker head plate with connectors at each side, at the same time eliminating the mounting shoe that had been used on the base of the M23 heads.

M23 vs. M56 head assemblies 2
M23 head plate on left, M56 head plate on right. The M23 plate normally wouldn’t have 2” heads, but this assembly has the full-track 2” head mounted to it.

Now we had a functioning deck, but no electronics! With the popularity of overdubbing, we decided that nobody in his right mind would want or need to record all 16 tracks simultaneously. Our solution was a Selective Recorder, the history of which is covered in “The Selective Recorder” episode of this history. Since the Selective Recorder control panel was developed for 8 channels of playback and 2 channels of record, we needed two panels to cover 16 tracks. This gave us four channels of selective record.

We assembled the machine and Wally Heider signed up to buy the first unit. He was ready to make another killing by renting the first machines to major studios. We displayed this prototype at the Spring AES convention in 1968 with mixed reactions from the audience. When Larry Levine of A&M studios told Wally that he would never rent a selective recorder, Wally bailed out of his commitment to buy any machines. Needless to say, we were very disappointed by this turn of events.

About a month passed without any real solution coming to mind. Wally was using a Tom Hidley custom converted Ampex deck with 16 electronics modules. And then on the Saturday of the Memorial Day weekend of ’68, Wally called me at home. (Calls from Wally were common, even in the middle of the night.)

Wally (who stuttered badly when he was excited) said, “DDDale, sssixteen-track is here, and I nnneed a machine NOW!. Nnnot something with only a few channels of record, a fffull machine! Wwwhat can you do?” I reminded him that he had passed on the Selective Recorder, and that we weren’t ready with anything else. I then promised him to reconsider the problem and call him back.

I went to the Mincom lab (on the holiday weekend) and decided to look at the magnitude of the problem. I gathered up a bunch of audio cards and transformers and built a big pile on the workbench. I measured the pile. Certainly this wasn’t going to fit in a 19” rack! I measured the lower compartment our standard wooden console and noted that it was 24” wide inside and about 26” high. Could I cram 64 printed circuit boards (erase, record, repro and line amp per channel), all the control circuitry and 32 input and output transformers into this space?

I took these dimensions home and started sketching a single card cage for the 64 cards. I drew a half scale sketch of the front side on graph paper with ¼” squares – 3 rows of PC cards with 24 cards in the top and bottom rows (8 channels) and a center row of 16 bias/erase cards, with two brackets breaking up the center row with 8 erase trimmer capacitors on each bracket. Brackets separated the rows of cards, with 3 relays and the input level pot for each of 8 channels on a strip. Then I drew a back view, with 4 rows of transformers and 2 rows of XLR connectors with termination switches. Next I drew the side cutaway view showing the cards, transformers and all the innards. As a final step, I cut pieces of cardboard to form a box and affixed the drawings to the front, side and rear of the box. And there before me was a half scale model of the M56 electronics chassis.

M56 half-scale cardboard mockkup front 1
M56 half-scale cardboard mockup front
Figure 7 M56 half-scale cardboard mockup side
M56 half-scale cardboard mockup side
M56 half-scale cardboard mockkup rear 1
M56 half-scale cardboard mockup rear

In addition to the card cage, we would also need a meter panel and a remote control. We could build on the remote control techniques that had been developed for the M23, but in this case the remote would be the ONLY control, with no record/sync controls at the machine. Individual input/repro monitor switches would mount above each VU meter and a master A/B switch on the remote would switch all channels.

With the basic ideas for the new machine well in mind, I called Wally to describe the concept. I had hardly finished the description when Wally blurted out “III wwwant two machines. I will send you a deposit of $5,000 each the next (mail) day.” I reminded Wally that 3M knew nothing about this machine and that they couldn’t possibly take an order without a detailed design. I also mentioned that 3M might not even want to build the machine! His response was “IIIf they don’t want to build it, we will buy the parts and you can build it for us!”

Later that day I went over to Don Kahn’s (our mechanical designer) house to see his daughter (my girlfriend.) I showed Don the cardboard mockup and gave him the sales pitch. He immediately bought into the design. The next day (still a holiday) I was in the lab, and Jack Mullin just happened to be there, too. I showed him the model and he signed on to the concept.

We didn’t tell 3M management about the new machine; instead, we began to build a prototype. We had a priority project going through the metal shop for building some 150-mil tape QC playback machines for RCA Indianapolis. We made some quick drawings for the sheetmetal parts for a prototype card cage and submitted them on the RCA priority number. We didn’t even put all the holes in the metal, and we only made one door. Don hand built the prototype, clamping the sheetmetal together and drilling the necessary holes with a hand drill.

We made up a remote control prototype by cutting two 8-track Model 23 units down and Heliarc welding them together. Don made a box out of cardboard.

The only thing we didn’t have was a meter panel with 16 meters. Since Don was going on vacation and there wasn’t time to get metalwork for the meters, we just drew 16 meters and pasted the drawing over a standard 4-meter panel.

I wired the first chassis. I had included transistors to control the A/B and sync relays, and these transistors were hard wired to the terminals of the relay sockets. Well, very early in the testing I got the power supply leads reversed, and it wiped out every one of the 32 transistors. Needless to say, I wasn’t too happy. We changed the design to include sockets accessible from the front for the 32 transistors, AND we added a diode across the input line to catch reverse-polarity hookups harmlessly.

3M M-56 prototype
Original Prototype with only one door. Note small casters.

We put everything together with the 2” development transport and began testing. One day Dick Dubbe, the Mincom Technical Director, came into the lab for a routine social call. As he wandered around the lab, he discovered our new 16-track machine. He was a bit surprised, and he asked me what this was. I gave him a rundown on this new product that used many existing components. His eyes kept getting wider and wider. He said, “Wait here. I am going to get Bob Herr (the Plant Manager)” and off he went. I thought I would soon be dead meat! Dick came back in just a few minutes and informed me that Bob wasn’t available, but they would be in first thing in the morning.

Indeed, before Jack Mullin even arrived the next morning after dropping his kids off at school, Dubbe and Herr had come to the lab, gotten the sales pitch on the new machine and given their thumbs up to proceed at all haste with the project.

Now with an official blessing, we retraced our steps and built a “real” prototype with meter panel, control box and fully-documented sheetmetal. I found some large Holtzer-Cabot motors that would fit into the openings of our web casting with only a minor cleanup of the taper on the webs. We assembled and tested the machine, then shipped it off to the ’68 AES Convention in New York City at the New Yorker, I believe. Our photo in the Convention Program showed our original model with only the one door!

Our display at the convention featured a white circular pedestal in the center of the booth for the M56. When the show opened, the pedestal was empty! Our machine, which had been shipped to the show cartage company in Philadelphia for transport to New York, was lost somewhere between Philly and New York. The machine finally arrived at the side entrance to the hotel about 5 p.m.

We had included some nifty features to make the machine more compact for shipping. The remote control had a mechanical “docking port” on top of the meter panel, and the control was held by a pair of thumbscrews for quick on/off mounting. The meter bridge could be removed and inverted for stowage inside the back of the machine. When the meter bridge was thus stowed, the machine was only as tall as the reel hubs.

M56 Proto Remote001cr
M56 Prototype Remote


When the machine finally arrived at AES, it was in the “tucked in” configuration. We wheeled it down the aisles to our booth and boosted it onto the display pedestal. Out came the screw drivers, and we quickly inverted the meter panel to operating position. By now a large crowd was gathering around to see what they had been waiting for all day. We connected the power cord and loaded tape reels onto the transport.


Ham Brosious
Ham Brosiou

Right at the front of the crowd was Ham Brosious, Sales Manager for Scully. He was eager to see his new competition. At this point the machine had traveled over 3,000 miles and not had power applied to see if it even worked. In what was probably not the smartest thing I have ever done, I invited Ham to step up and turn on the power for the first time. He pressed the Power button and the machine lit up. The tape reels tensioned up. Then he pressed the Play button. My fingers and toes were crossed, hoping that everything would work. Indeed, the machine was flawless! Ham exercised Fast Forward and Reverse, armed some tracks for recording and punched up Record. Then Ham just shook his head and walked away. He knew he had some real competition for his giant 16-track Scully.

Ray Dolby was also in he crowd. He asked me how we were able to package 64 audio printed circuit boards into a single chassis without any ground-loop problems. I was too naïve to know you shouldn’t be able to do this. I said that we had used multiple parallel bus wires (9 total on each row of cards) for grounding across the back of the PC board edge connectors. I believe that this conversation may have been the beginning of Dolby’s later M16 and M24 high-density packages for CAT22 cards.

M56 Inside Card Cage Rear001cr
M56 Prototype Inside Card Cage Rear

The machine was a big hit at the show. After the show, we packaged the machine up and flew to Nashville for a private reception at the Holiday Inn near Studio Row. We had very poor attendance. Nashville and New York were strong Ampex markets, and it took a lot of effort to get our foot in the door. Tom Welch, our 3M Tape salesman in the area, tried to get folks to come to the reception, but with little success. (We did get to go backstage at the Grand Ole Opry during a show, which was a big moment for me!)

Wally Heider was still waiting for his two machines, and he was getting nervous. The Hidley beast that he was using apparently had a very good signal-to-noise ratio, and folks had convinced Wally that the older 3M’s weren’t as quiet. We eventually took the M56 prototype down to Wally’s Studio 3 for a shootout. We aligned both the Hidley and the M56 using their test tape, making sure that we had apples and apples. The big test was to combine all the tracks and listen to the biased noise level. First we listened to the Hidley machine. Then we switched to the M56. Wally was convinced that the noise would go up, but instead it went    DOWN! Wally’s jaw dropped, and he quickly stuttered “Dddale, wwwhat did you do?”

Actually, we really hadn’t done a lot. The playback preamp cards were our latest version, with a lower-noise input transistor and all metal-film 2% resistors, and the heads were the new style with 2-mil laminations, rather than the old IEMs that had 6-mil laminations. The head cabling was shorter, reducing the amount of resonant peaking. All in all the combined results was significantly better performance. Wally recovered from his shock and promptly started bragging about how good his new 3M machines would be!

We did an overall test at Studio 3. This was the first time for the M56 to be used in a typical overdubbing situation. We set up the machine and then placed the remote control box on the armrest of the console. The recording engineer proceeded to conduct the session without the aide of a second engineer at the machine. Everything was at his fingertips. Everyone was very please. Wally invited folks from Capitol and other studios over to observe the tests, and a few did show up.

M56 & Dale001cr
Dale Manquen and the production version of the M56.

We had great mobility because the M56 easily slid into the back of Don Kahn’s Ford station wagon. We could drive to a studio, unload and be ready to operate in just minutes. We also did a “station wagon” demo at RCA Hollywood one evening. Our test was with Henry Mancini, someone that I had idolized for years. Here we were going to be recording Henry Mancini on a machine that I designed! What a thrill — or maybe not. It turned out that Hank was recording some Dixieland for a film score. They stuck a mike on each instrument in the small combo and ran each mike to a separate track. We only used maybe 8 of the 16 tracks. Hank sat in the control room with his feet up on the producer’s table, smoking a cigar. He pushed the talkback mike button and counted the beat… and that was it! No big orchestra, no conductor, no nothing! “OK, try it again. One, two…” And that was our test!

We built a Pilot Run of 10 machines, and began delivering in the Spring of ’69. We had some remaining problems. The narrow female contacts in the sockets for the 6-pole Sync relays did not maintain good contact alignment with the relay pins (the relay type was changed to Parelco relays with good sockets.)   The illuminated Switchcraft mode switches in the remote control box got too hot because of heat produced by the 16 incandescent bulbs. The illuminated switches got replaced by conventional lever switches and separate colored indicators. (Too bad we didn’t have LEDs to illuminate the switches!)


Figure 10 Original Remote Control with Illuminated Switches
Original Remote Control with Illuminated Switches

The eventual design of the remote



The large reel motors caused erosion of the relay contacts, and these relays were later replaced by much larger relays. We increased the size of the casters to better cross cabling and curbs.

The 3M head department finally decided to build 16-track heads for us, ending our reliance on Applied Magnetics, and the 3M heads were better.

I had the inspiration that led to the M56, but the machine is the result of a team effort. Jack Mullin was my mentor, along with Wally Heider who taught me about the practical world of recording. Don Kahn was a very good mechanical designer, able to make things that were very straightforward. My technician was Jim Leatherman, and I drove him crazy because he was a “single track” processor and I kept him going on several different paths at the same time. (Jim transferred out of the lab and into the computer/accounting section where he could deal with one thing at a time.) We also owed a lot to Ken Clunis, on of the original design team for the Dynatrack machines that spawned the M23 and M56. Ken also solved the Mylar belt problems that allowed us to offer an improved capstan drive system.

3M M-56 design team
M56 design team 1968: Dale Manquen, Don Kahn, Jack Mullin and Jim Leatherman

I decided to go back to school for graduate work, starting in the Fall of ’69. Before I left, we hired a replacement (Clive Ross) and tried to plan ahead. One day the lab group went out to a local park and “blue skied” about what to do next. The main concept was to reduce the number of circuit boards in the M56. The next 3M model, the M-79 used only one circuit board per channel and greatly simplified the construction of the machine. Clive and our lab technician, Glen Phoenix, carried these ideas through to build another very successful product.

I went back to New Mexico State University and spent two years (not the originally-planned 3 years) in graduate school, ending up with a Masters in EE. I eventually concluded that getting an EE doctorate might make it hard for me to get a job in audio product development. I redirected some of my studies to Industrial Engineering and ended up only one course shy of a Masters in ME. I was able to find a job at Ampex (in spite of a hiring freeze there) that put me in charge of the Ampex MM1100, a Johnny-come-lately that embodied many of the M56 concepts. But that is a tale for another day….

Dale Manquen, August 2014 – all rights reserved

Leave a Reply