INDEX OF ALL CASSETTES
Introduction
I recently purchased a good-condition Nakamichi Cassette Deck 1 (aka CD1), for no other reason that I lusted after such machines back when I could not (remotely) afford them. I also always wanted to learn the real performance of the cassettes I so lovingly used in the 80s and 90s (first with BPC, then with a CR-2), until CD-R and self-recorded DVD-A (yes!) wiped out everything.With the new Nak, and with the excellent audioTester software the time was finally there.
I use a laptop, with as bus-powered USB sound system an Alva Nanoface, running at 96kHz. I only use the Alva for measurements, it is handy and reasonably good. It has XLR microphone/line inputs, and RCA line inputs and outputs. There is one issue with audioTester (also with WFGUI), and that is that in these programs only inputs 1-2 can be used, which are the XLRs, and these have a low input impedance of 6 kOhms. However, with all the cassette decks I measured so far this has not been a problem (provided, of course, that the deck's output level control is at the maximum).
I calibrated the CD1 for the three tapes I intend to use most: XLII (1994), Metal XR (1995), and UR (1994). I did not use the service manual instructions (bring 400Hz and 20kHz to the same level), but rather graphed the entire 20-25kHz frequency response at a level of -20dB (ref. Dolby level = 200 nWb/m ANSI) and judged the resulting plots for flatness. The record level was aligned at -6dB (I found that often at 0dB slight compression already set in). Before all of this the CD1 playback was verified with a TEAC MTT-356 tape, and any channel imbalance was corrected.
In what follows frequency plots are made at -20 dB (shown on the graph at -30 dBFS), -10 dB (shown at -20 dBFS), 0 dB (shown at -10 dBFS), and +6 dB (shown at -4 dBFS). The -20 dB sweep was always made from 20Hz to 25kHz. The other sweeps have progressively less extension, to avoid the curves running into each other. Some sweeps have their low frequency part start at 200 Hz: I initially did so to save on time, but later I figured that having a look at bass compression at 0 dB and +6dB would be interesting, so all later plots start at 20 Hz at each level.
Each tape's sensitivity, relative to the XLII of 1994, can be read as the difference of its -20 dB plot and the -30dBFS line in the graphs. I also estimated Maximum Output Level (MOL) at 400 Hz, with the limit set at 3% distortion , third harmonic. (I referred the result to the input side, i.e. I noted how much the input signal for 3% output distortion was above the input signal that resulted in Dolby level on the tape. This method is actually wrong, because it does not account for compression. The industry practice is to measure MOL entirely at the output side. This is what I did in my later series, using a BX-300 deck.)
I sometimes also estimated Saturation Output Level (SOL) at 10 kHz, looking for the highest input level where the output level still was equal to the input. Again this was compensated for sensitivity at 400 Hz, but not for a tape's frequency non-flatness in the treble. As such this a crude estimate only. (In the later instalments of this series the MOL and SOL measurements got more refined, see the comments in those articles. I will update the older articles when I find the time to redo the measurements.)
Before each new batch of tapes I first verified the setup with a control sample of XLII (1994), ascertaining that -20 dB still plotted at -30 dBFS and that the tape path was clean. Then for each tape the treble frequency response was plotted, and the CD1's front panel bias knob, ranged -5 to +5, was tweaked for the flattest and/or most extended response. The plot titles reflect the actual bias setting in use.
The decks used have 'positive control' for the front panel bias knob: turning clockwise, towards '+' increases bias current and thus reduces treble. I write this because there exist decks with the opposite operation.
The Nakamichi literature does not tell us what range the front panel bias fine control operates over. At any rate the effect of the bias knob depends on the tape type: it does almost nothing for metal/type IV. The graphs below are the mid and treble frequency plots of the type II calibration tape (XLII), with the bias knob at 0, -2.5, -5 and +2.5. This gives you an idea.
The photographs are of the actual tapes used for the test, all purchased in Europe. Many of these cassettes are very old now, and have been used in cheap portables and car stereos. Some wear is evident, but generally not much. I tend to take care of my things.
Click on the graphs for a larger version.
Maxell XLII (1994) (what the deck was calibrated for)
Relative bias: (reference)
Relative sensitivity: (reference)
MOL400: +5.5 dB
SOL10k: -5 dB
Bias noise: -55.3dB, -59.6dB(A)
Not much to say. Like many modern tapes the lower treble has a peak that cannot be calibrated away with increased bias, unless a severe loss above 10 kHz is accepted. I assume that such humps were engineered-in to obtain a flatter response at 0 dB.
Maxell UDXLII (1980)
Relative bias (ref. XLII 1994): -5
Relative sensitivity (ref. XLII 1994): 0.1 dB
MOL400: +2.4 dB
Bias noise: -52.9dB, -56.3dB(A)
Beautiful, not? These were the first type IIs I bought (expensive!) and I used them in an ITT portable. I had to ask my friends to record them for me.
Bias had to be turned to the minimum, -5, yet sensitivity matches the more modern reference. MOL is low. There is 3.3dB more noise than the 1994 version.
Maxell XLII (1982)
Relative bias: < -5
Relative sensitivity: +1.1 dB
MOL400: +4.6 dB
These were the first tapes I bought in large quantities. By that time I had graduated to an ITT music centre (with metal, Dolby, and LED meters ;-).
Even at bias -5 a tad overbiased. Strange is the large jump in sensitivity compared to the 1980 version. MOL is much better, but that might be an artefact of the high bias used.
Maxell XLII (1985)
Relative bias: -4
Relative sensitivity: 0 dB
MOL400: +3 dB
One would expect this to be similar to the previous version, but no ...
Bias still much lower than 1994, at -4, but sensivity is back to normal. MOL is mediocre.
Maxell XLII (1986)
Relative bias: -1
Relative sensitivity: +0.2 dB
MOL400: +4.7 dB
Now it is getting interesting. This tape wanted bias at -1. From this version on XLIIs were fairly consistent in bias needs and sensitivity, and came with decent MOL.
Maxell XLII (1988)
Relative bias: 0
Relative sensitivity: +0.3 dB
MOL400: +5.3 dB
Bias noise: -52.8dB, -57.6dB(A)
Maxell XLII (1991)
Relative bias: 0
Relative sensitivity: +0.3 dB
MOL400: +5.8 dB
Bias noise: -54.3dB, -58.8dB(A)
Maxell XLII (1998)
Relative bias: -1
Relative sensitivity: +0.4 dB
MOL400: +6.1 dB
SOL10k: -6.6 dB
Bias noise: -55dB, -59.5dB(A)
The last real XLII? Compared to 1994 the MOL is improved, noise is the same, but SOL seems a bit disappointing.
Maxell XLII-S (1980)
Relative bias:-4
Relative sensitivity: 0.8 dB
MOL400: +3.8 dB
The very first XLII-S, and the prettiest. It does seem to better the contemporary UDXLII.
Maxell XLII-S (1982)
Relative bias:< -5
Relative sensitivity: +2.2 dB
MOL400: ???
Even at -5 it got overbiased, so much that I did not even bother with the higher-level sweeps and the MOL measurement. I never really liked these, but that was probably because the dual gold/transparent foil tended to delaminate on mine, turning these cassettes into an expensive, ugly, sticky mess. (Strangely my XLI-S tapes never did this.)
Maxell XLII-S (1986)
Relative bias:-2
Relative sensitivity: +1.2 dB
MOL400: +4.5 dB
SOL10k: -6.8 dB
One of my favourites, back then. That does not mean much, as I was using a TEAC V-455X BPC thing. Still too sensitive, but this is when the various XLII-S generations start becoming more consistent.
Maxell XLII-S (1988)
Relative bias:-1
Relative sensitivity: +0 dB
MOL400: +4.3 dB
SOL10k: -7 dB
I never purchased these. I only recently got a few samples, most of them heavily used. This is the best of the lot. Much like the 1986 version, but with reduced sensitivity.
Maxell XLII-S (1994)
Relative bias:+1
Relative sensitivity: +0.3 dB
MOL400: +7.3 dB
SOL10k: -3.7 dB
Bias noise: -54.2dB, -58.3dB(A)
This was my standard tape when I got my Nakamichi CR-2. Seriously improved MOL and SOL over the previous generations. As an aside, this Maxell measures very similar to the BASF TPII.
Maxell XLII-S (1998)
Relative bias:+2
Relative sensitivity: +0.1 dB
MOL400: +6.8 dB
SOL10k: -6.9 dB
Bias noise: -53.8dB, -58.1dB(A)
This was the final chapter. I purchased lots of these, they were dirt cheap. Cassette was clearly in its decline. I virtually stopped taping at this time, so my stocks are still sealed. From the looks one would expect similar performance to 1994, but while this is true for MOL, SOL has suffered a lot, and even noise is slightly worse.
BASF Chrome Super II (1991)
Relative sensitivity: -1.7 dB
MOL400: +0.8 dB
SOL10k: < -14 dB
Bias noise: -57.0dB, -61.3dB(A)
A true chrome tape. Low sensitivity. Low MOL. SOL was nearly impossible to measure: I ran out of patience.
Are these results normal and representative of the state of chrome tape in its heyday (the 1980s)?
No, they aren't. It appears that chrome dioxide pigment does not age well, irretrievably losing sensitivity, MOL, and SOL, while shifting the bias needs downwards. Noise is luckily still very low, so this tape remains more or less useable, , but you will need a deck with a very low self-noise to exploit it.
You can find more details on chrome instability in this thread.
BASF Chrome Maxima II (1991)
Relative sensitivity: -1.5 dB
MOL400: +1.5 dB
SOL10k: < -13 dB
Bias noise: -57.3dB, -61.6dB(A).
Similar as CRSII, but at least the bias needs are closer to the Japanese ferro cobalt tapes. Noise is, again, very low.
BASF Reference Maxima TPII (1993)
Relative sensitivity: +0.4 dB
MOL400: +6.4 dB
SOL10k: -2.6 dB
Bias noise: -53.8dB, -57.7dB(A)
BASF's first ferro-cobalt tape. Allegedly there was some kind of cooperation with Maxell (according to some), or BASF just purchased and evaluated Maxell pigment (according to others). Excellent performance, aided by a great shell. And ... observe how similar this tape is to 1994's XLII-S.
Denon HD8 (1988)
Relative sensitivity: +2.2 dB
MOL400: +4.7 dB
Bias noise: -50.7dB, -53.7dB(A)
This is the second metal particle type II I ever encountered (the first was a promo freebie That's EM-X that I could not appreciate). I purchased it from a hifi dealer who was kind enough to dub the soundtrack of a concert VHS onto it (we did not have a HiFi VHS at home). The video machine tracked horribly, and the recording was pretty bad. That as an aside.
Pure or metal-enriched type II tapes as a breed should excel in SOL, giving a wider frequency response at -10 dB and 0 dB. They also tend to have a high sensitivity. MOL is decent, but below what ferro-cobalt type I and II can manage. Noise is high, of the same order of a decent ferric, but remember that this is with 70us equalisation, and that makes it a pretty bad performance in absolute terms. What surprised me is the low bias required.
Denon HD8 (1992)
Relative bias: -3
Relative sensitivity: +1.9 dB
MOL400: +4.4 dB
SOL10k: -0.6 dB
Bias noise: -51.5dB, -54.5dB(A)
That's CD/MH (1990)
Relative bias: -3
Relative sensitivity: +2.5 dB
MOL400: +5.5 dB
SOL10k: -0.5 dB
Bias noise: -51.1dB, -54.2d(A)
This tape's claim to fame is that now, in 2018, you can still find new old stock for reasonable prices. Much like the HD8 bias and sensitivity make it incompatible. Noise is high, MOL is average, and the excellent SOL can only be exploited with bright-sounding music. Perhaps best to be used in a auto-calibrating deck, with Dolby B, or for heavily-compressed music.
Sony Metal XR (1992) (what the deck was calibrated for)
Relative sensitivity: +0.2 dB
MOL400: +7.7 dB
SOL10k: +1.2 dB
This was a different sample than the one used for calibration, or the setting had drifted, hence the 0.2 dB offset at 400 Hz.
Sony Metal XR (1989)
Relative sensitivity: +0.5 dB
MOL400: +7.5 dB
SOL10k: +1.5 dB
If I remember correctly this was the first affordable metal to reach Europe, realised by marrying a more than decent tape to a welded budget shell. Magazine reviews were positive. I had my Nak CR-2E calibrated for this version, and purchased quite a few of them.
Sony Metal XR (1995)
Relative sensitivity: +0.5 dB
MOL400: +8.0 dB
SOL10k: +2.5 dB
Bias noise: -53.5dB, -57.6dB(A)
I only just found out my remaining stock is mostly the 1995 variant, and not 1992. So after this characterisation exercise I recalibrated my deck for this one.
Maxell MX (1980)
Relative sensitivity: +1 dB
MOL400: +9.4 dB
SOL10k: +3 dB
As a kid with no more than a small portable radiocassette I (foolishly) bought these, hoping that metal greatness would rub off on my pathetic gear. Or something like that.
Despite this, this second-generation MX is like new, and turns in a stellar performance.
Maxell MX (1982)
Relative sensitivity: +0.8 dB
MOL400: +7.8 dB
SOL10k: +2.8 dB
Maxell MX (1985)
Relative sensitivity: +1 dB
MOL400: +8.7 dB
SOL10k: -+3 dB
Maxell MX (1986)
Relative sensitivity: -0.2 dB
MOL400: +6.1 dB
SOL10k: +1.8 dB
This is the sole sample I have, and much like the 1986 XLII-Ss it hasn't worn well. So I don't know if the worse performance is characteristic of this generation.
INDEX OF ALL CASSETTES
