Reel to reel machine, new build or Hack?

Google "tape amplifier schematic" and you'll see plenty of solid state (low voltage) amplifier circuits to get you started.

 

It would be a lot less work and higher quality to use digital recording techniques.

Bob

 

The playback speed (as well as the record speed) of a "reel-to-reel player" is governed by the rotation speed of its capstan tape drive, in conjunction with a pinch roller that presses the tape against the capstan without slippage between tape and capstan. The only purpose of the reel motors is to provide take-up and supply for the magnetic tape passing over the playback and recording heads, providing only sufficient tension in the magnetic tape between the capstan drive and the tape reels to allow tape to feed from the supply reel and be wound onto the take-up reel. Depending on how fast the tape must be transported across the tape heads to accomplish your purpose, the tape tension method will become the overriding concern in how you will hack an existing reel-to-reel player or design a purpose-built one from scratch. The best performance is achieved with vacuum columns that hold a goodly length of tape that is taken up and replenished from each tape reel as the capstan moves the tape at high speed over the magnetic tape heads.

All the above requires some sort of mechanism to sense tape tension and drive the reel motors in such a manner that (more or less) constant tape tension between the drive capstan and the take-up reel is achieved while the tape is moving in intimate contact across the magnetic tape heads. A typical way for inexpensive reel-to-reel players to sense tape tension is the use of a rotating, spring loaded, lever with an attached freely spinning roller over which the tape passes. Increasing tape tension rotates the lever against the restraining spring until an equilibrium position is reached. If the lever position is electronically sensed, this information can be used to control the torque applied by the take-up and supply reel motors. Other ingenious schemes, not involving electronic feedback, can also be used to provide pseudo-constant tape tension.

So, your mission should you choose to accept it, Darren, is to devise a means to provide a continuously variable capstan rotation speed, as determined by an analog signal from your synthesizer, speed range to be determined by you.

Some means to dis-engage the tape from the tension mechanism is also desirable to allow you to implement fast-forward and fast-reverse tape positioning, ideally while still maintaining intimate tape contact with the highly polished magnetic tape heads to allow audible monitoring of the signal recorded on the tape. A precision tape position counter would also be invaluable as a means to locate sections of tape for over-dubbing or editing. Considerable help with this project would be obtained if you could channel the shades of Les Paul and Mary Ford, who were pioneers in the 1950s creation and use of this type of now-obsolete technology, later taken up by other recording and performing musical artists. In fact, over-dubbing became so common a practice that "live" performances were impossible to perform without "lip-synching" the artists to their pre-recorded work. See Dick Clark's American Bandstand for numerous examples.

If you are serious about this, you will abandon the "old school" methods and embrace modern digital technology. Although I must admit a certain fascination with "steam punk" technology, for example as demonstrated on the TV series Warehouse 13, I also realize the totally impractical nature of this genre of "design" much less its actual construction. Trying to resurrect technology from the 1940s and 1950s is a waste of your time unless you just want to consider the effort art... as everyone knows, art knows no dimensions or boundaries... remember to have fun and not expect too much in the way of results.

 

A digital solution would involve an ADC to digitize your audio, memory to store it, a DAC to play it back and a microcontroller or microprocessor to orchestrate it all. A Raspberry Pi or similar small micro has all of the hardware you need. Your job would be the programming.

Bob

 

Not quite. The RasPi doesn't have the ADC, or even any analogue inputs.

 

I could have sworn it had audio inputs, but I guess I am wrong.

Bob

 

An Arduino is not going to have the performance and storage needed. You probably would need a ARM based micro like Rasp Pi, but hopefully with analog or audio inputs and outputs. Or, you could add an ADC externally to rasp pi.

Bob

 

I think it is admirable that you would consider both an analog as well as a digital approach. Certainly, given that you want play-back only, and just a variable tape speed for that, the analog approach is simplest. Using a C90 1/4-inch tape cassette may be a viable "reel-to-reel" implementation. I have some experience with that, both in analog as well as digital formats, from my first "real" job as a graduate electrical engineer in the late 1970s. I cannot recommend the cassette tape approach however because most cassette players are not precision audio equipment but are mass-produced junk, to be sold to the uneducated consumer and later discarded.

There are many different approaches to controlling tape tension, but it is generally not a problem with cassettes because the reel diameters and the amount of rotating tape mass are both quite small. Generally, the capstan and pressure roller approach will suffice to pull the tape past the magnetic tape heads at a constant linear velocity, the tape being tensioned by the supply motor driven slightly in reverse and operating as a torque motor, thus pulling on the tape in tension between the supply reel and the capstan with the tape heads located in between. Very little tension on the take-up side of the capstan is required, so the take-up reel needs to be driven in the forward direction, but again with very little torque.

The sound cards on most personal computers, combined with appropriate software, will produce any of the functionality that you might be able to obtain with analog reel-to-reel equipment. There are abundant sources of free, open-source, software as well as professionally developed software, available on the Internet. You might want to visit some of the links on this page to get an idea of what is available and the capabilities of the software. Or... climb the learning curve and write your own personal computer, sound card based software. Yeah, I know, it's a real PITA to learn software programming when you already know how to wield a screwdriver and a soldering iron, but its the 21st Century way of doing things. To get you started, check out this web page on the free Audacity software for Windows 10. It is also available for Mac and Linux.

Back in the day, my first engineering job after graduation from college was to replace a Kansas City Standard audio cassette "digital" recorder with a real digital cassette tape recorder. The so-called Kansas City Standard was just an ordinary telephone modem whose audio tone-encoded output was recorded on ordinary audio tape recorders, typically similar to cheap Radio Shack cassette recorders. The resulting "digital" audio tone-coded data was then played back into the modem to recover the original digital data. It sort of worked, but was not a reliable means of digital data storage for the exchange of data between machines.

Completing the "upgrade" task required over a year, and ten field-portable data acquisition systems were converted to Intel 8085 microprocessor-controlled digital cassette tape recorders. The irony of this task was the need to continue to use acoustically-coupled telephone modems, in motel rooms located in remote locations, to send the digital tape data back to the Mother Ship in Dayton, OH, for subsequent data processing. The original tapes were also mailed to Dayton, but the modem transmission was necessary because the data collected was time-sensitive.

As an historical note, sometimes tape cassettes were specially constructed to allow a continuous loop of tape to be installed as a Mobius loop. This doubled the playing time for a given length of tape and allowed recording to occur on both sides of the tape. Not very practical IMO but apparently it did work. More elaborate schemes allowed a conventionally spliced loop of tape to continuously cycle through the recorder and were used for various effects, including reverberation delays. Fortunately, such contraptions pretty much vanished from the recording industry with the introduction of fast, high-resolution, A/D and D/A converters, inexpensive random-access-memory (RAM), and microprocessor data processing along with field-programmable gate arrays (FPGAs) and digital signal processors (DSPs). Going digital opened up many new and exciting pathways to product development that were simply impractical with analog signal processing techniques. Video cassette recorders (VCRs) were just the first of many mass market applications, although VCRs continued to mainly use analog recording techniques until their fairly recent replacement by compact disk recordings, which in turn have been replaced with solid-state FLASH memory recordings.

 

I'm not clear what effect you are trying to achieve by varying motor speed, but tape transport mechanisms usually involve a capstan motor with a substantial flywheel to stabilise the speed to eliminate wow and flutter. This would rule out any intended rapid speed variations.

 

True. So remove as much moment-of-inertia as practical from the capstan to allow it to be quickly accelerated and decelerated. Add tachometer feedback to "tighten" the capstan speed control loop. Brute force the damn thing if you have to in order to achieve whatever speed profile you deem necessary. And welcome to the world of power electronics. Adding inertia anywhere in an electro-mechanical system often introduces more problems than it solves, so a direct capstan drive from a low-inertia pancake servo motor will generally yield a frisky system. This was all worked out by analog engineers from the beginning to about the middle of the 20th Century to everyone's satisfaction, so no need to re-invent that particular wheel... unless you just want to discover it as part of a personal learning process. Fascinating stuff, analog electronics and their servo control mechanisms.

Modern personal computers have the power to model analog circuits with excellent accuracy, replacing PITA manual computation methods that were previously necessary. Why model analog servo mechanisms? Because tossing around a few tons of steel can be highly dangerous to people and property if your servo control "gets away from you," and this can and does happen because you didn't model and test the design before building it. However, you will probably be safe enough playing around with reel-to-reel tape recorders.