I have some old tapes (the compact cassette variant) and an old Technics deck. I want to digitally archive these tapes as best as I can.

The original deck is from the late 80s and contains a standard playback-only 180 ohm DC resistance stereo playback head with one common wire (so 3 wires total). Since the original deck sounds pretty bad by today's standards (lots of noise and bad dynamics) I want to replace the analog signal path with something more modern. There is not much information available on the subject, most of it deals with stupid things like tube tape preamps or other absolutely inaccurate forms of amplification. I know that tapes themselves don't sound great, but I want to preserve basically everything you can reasonably get off of them to throw them out afterwards.

I explicitly don't care about IEC equalization or Dolby decoding, or any hardware-level noise reduction (other than noise that comes from the signal path itself obviously) since I will do that in software when I actually need the tapes in playable form.

My tentative design currently consists of two LT1115 differential amplifiers (one for each channel) feeding a single AD1871 two-channel 24bit Audio ADC. The LT1115 are going to run off of regulated and filtered +15V and -15V rails, the ADC has its own regulated 5V rail (and a 3.3V digital rail). I'm open to swapping these components for better ones, this is just what I came up with.

I'm not sure about what layout I need on the analog side. Can a coil with three wires (one common one) even reasonably be treated as two differential inputs (since they shouldn't really share a common wire)? Do I need any additional components in the analog path like a bleeder resistor for the coil charge or some form of matching or clamping?

I'd really appreciate it if somebody explain the analog behavior of tape head / inductor amplification.

  • \$\begingroup\$ I personally would keep the LT1115s, and ditch the ADCs, just running the output to Line In on any PC sound card. Why build a high speed digital interface to get data off the ADCs when it's all done for you in the sound card anyway. Especially as this is a one-off exercise, and you're not trying to make a commercial product. \$\endgroup\$
    – Neil_UK
    Apr 30, 2018 at 5:27
  • \$\begingroup\$ @Neil_UK I can feed the I2S output directly into a SoC board, I have some RK3328-based ones and Pi3s lying around. But for testing I could definitely take the audio and feed it into a normal line-in. \$\endgroup\$
    – Lorenz
    Apr 30, 2018 at 5:38
  • \$\begingroup\$ I think assuming that you can make a better preamp than you can buy (in a reasonable amount of time) is a good strategy to sink lots of time. The alternative is to use what you have and see how bad the results are. There may be noise already on the recording. You could benefit from a good post-processing suite depending on the value of your data. \$\endgroup\$
    – loudnoises
    Apr 30, 2018 at 8:41

1 Answer 1


Don't imagine "modern" is necessarily any way "better". I'm not about to look up the LT1115's noise performance but unless it's below 1 nV/rtHz it can easily be beaten by a discrete component tape amp.

Tape head amplification is not merely gain, thanks to (a) the characteristics of inductive pickups (tape heads) but also the (different choices of) frequency shaping aka "equalisation" which differ for ferric and chrome tapes.

(And you're on your own with Dolby B decompression, you'll need some serious DSP to mimic that on the digitised signal)

Assuming you mean cassette rather that open reel, the only serious published design for tape playback I know of was by John Linsley-Hood, in the British "Wireless World" magazine in the mid-70s. May be findable by online searches if you're lucky. Either copy that directly (JLH designs were regarded as seriously high quality) or read, understand, and adapt/modernise as you see fit.

EDIT : found here (scroll down to May 1976) - the BC214 input stage provides pretty low noise, but you might want to upgrade a bit from the 741 in a later stage of the replay amp!

  • \$\begingroup\$ The LT1115 actually has 0.9nV/rtHz typical noise performance and the deck has a single-chip preamp, not one based on discrete components. I don't care about the equalization (not hard to do as DSP) and I have a Dolby B/C decoder that performs pretty much on par with high-end analog ones. The main issue that remains is dealing with inductive signal pickup and potential mitigations needed in the analog signal path. I'll see if I can find the article you mentioned. \$\endgroup\$
    – Lorenz
    Apr 30, 2018 at 14:40

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.