1984: Digital Signal Processing Increases Hardware Areal Density

PRML techniques enhance magnetic media storage capacity

Advances in data storage have been accomplished through innovations in recording media, read/write heads, mechanics, and, not least, in the electronics required to produce reliable binary data from the weak/noisy analog readback signals.

PRML (Partial Response Maximum Likelihood) detection provided significant gains in areal density (data stored in a given surface area). PRML operates at higher linear densities by allowing considerable interference between bits then applying relatively sophisticated digital signal processing to ‘unravel’ the original bit pattern. (“Partial Response” means that parts of the bit response appear in several read waveform samples. "Maximum Likelihood" refers to finding the most likely bit pattern) PRML and its successors have allowed 30-40% increase areal density.

PRML was first applied in Ampex’s 1984 magnetic tape recorder, the Digital Cartridge Recording System (DCRS). Development was led by Charles Coleman. For two decades, DCRS was the de-facto standard for high-data rate digital instrumentation recording for flight-test and military airborne applications. DCRS operated at 28 Mbits/sq.in. and 15 Mbytes/s. The largest cartridges offered capacities of 165 GBytes. The first application of PRML on a hard disk drive (HDD) was in the 1990 IBM 0681 developed at the Hursley UK laboratory. The PRML algorithms were developed by a team led by Francois Dolivo, IBM Zurich and implemented in hardware by IBM Rochester. The 0681 operated at 45 Mbits/sq in. and 3 MBytes/s and offered a maximum capacity of 471 Mbytes. All HDD suppliers adopted PRML by 2000.

Modern channels (2019) are about 10,000 times more complex (gate count). They operate by passing probability information iteratively back and forth between an NPML (Noise Predictive Maximum Likelihood) front-end and an LDPC (Low Density Parity check) back-end. Error-correction is implicit. The entire channel and HDD controller are integrated into a single VLSI chip. Other significant software developments that have improved the density, performance, and reliability of data storage equipment, include innovative compression, signal processing and encoding techniques such as RLL (Run-Length Limiting), signal dependent correlation-sensitive sequence detection, LDPC (Low-Density Parity Checking), and ECC (Error Correction) codes.

  • Kretzmer, E. R., "Generalization of a Technique for binary Data Communication" IEEE Trans. Commun. Tech., vol. COM-14, No. 1, pp 67-68, Feb. 1966.
  • H. Kobayashi, “Application of Probabilistic Decoding to Digital Magnetic Recording”, IBM J. of Res. & Dev., Vol. 15., No. 1, pp. 64-74, Jan. 1971
  • F. Dolivo et al., “Performance and Sensitivity Analysis of Maximum-Likelihood Sequence Detection on Magnetic Recording Channels”, IEEE Trans. Magn. Vol. 25, No. 5, pp. 4072-74, Oct. 1989
  • R. Wood et al., “An Experimental Eight-inch Disc Drive with One-Hundred MegaBytes per Surface”, IEEE Trans. Magn., Vol, 20, No, 5, pp. 698-702, Sept. 1984C.
  • H. Coleman et al, “High data rate magnetic recording in a single channel”, Journal of the Institution of Electronic and Radio Engineers, (Vol. 55, No. 6, June 1985) pp. 229-236
  • T. Wood, “Present and Future Directions for Ampex DCRSi and DD-2 Format Products”, THIC meeting, Ellicott City, MD, 16 October, 1996
  • “PRML Read Channels: Bringing Higher Densities and Performance to New-Generation Hard Drives” Quantum Technical Information Paper (1995)
  • Fisher, G.D., et.al. “PRML detection boosts hard-disk drive capacity” Spectrum, IEEE (Vol: 33 No: 11, November 1996) pp. 70 - 76
  • Kavcic, Alek and José M. F. Moura, "Method and Apparatus for Correlation-Sensitive Adaptive Sequence Detection” U. S. Patent # 6,201,839 (March 13 2001)
  • Gardner, Tom, “DEC AZTEC / RA80/81/82 (1982)” CHM Storage SIG Research Notes (June 21, 2011)
  • Lee, Jun & Kees A. Schouhamer Immink. “Advanced Signal Processing Technique for Storage Systems” IEEE Transactions on Consumer Electronics (Vol. 56, No. 4, November 2010) pp, 2373-2379
  • Cocker J.D., R.L. Galbraith, G.J. Kerwin, J.W. Rae and P.Ziperovich. "Implementation of PRML in a Rigid Disk Drive" IEEE Trans. Magnetics (Vol. 2, 6, 1991) pp. 4538-4543.
  • Hong J., R. Wood and D. Chan. "An experimental 180 Mb/sec PRML Channel for Magnetic Recording" IEEE Trans. Magnetics (Vol. 27, 6, 1991) pp.4532-4537
  • Howell, D.P. McCown et al. "Error Rate Performance of Experimental Gigabit per Square Inch Recording Components" IEEE Transactions on Magnetics (Vol . 26, 5, 1990) pp. 2298-2302
  • Taratorin, Alex. “Introduction to PRML” Characterization of Magnetic Recording Systems: A Practical Approach, Guzik Technical Enterprises (1996)
  • Thapar, H. and A.Patel "A Class of Partial Response Systems for Increasing Storage Density in Magnetic Recording" IEEE Trans. Magnetics, vol. 23, 5, (1987) pp.3666-3668 1987
  • Wang, Shan X., Alexander M. Tarartorin. “Channel Coding and Error Correction” Magnetic Information Storage Technology, Academic Press (1990) pp. 207 - 228
  • Wang, Shan X., Alexander M. Tarartorin. “PRML Channels” Magnetic Information Storage Technology, Academic Press (1990) pp. 361 - 421
  • Kavcic, Alek and José M. F. Moura, "Correlation-Sensitive Adaptive Sequence Detection IEEE Transactions on Magnetics (Vol: 34 No: 3, May 1998) ) pp. 763-771
  • Marvell 88i9422 Soleil SATA HDD Controller

1984_PRML_v6a Rev: 12.31.2019