My Dream Cassette Deck

[braxus]

Realizing I dont really need another tape deck (I already have too many- yes I have a problem), and the fact the market today will probably never get back into cassettes (not like Reel to reel decks that are making a comeback) and make such a deck, I would trade the majority of them in for a dream deck like this one. I was milling around remembering the days when cassettes were popular. Had CD-R not killed the format for mainstream use, I wondered what kind of ultimate cassette deck could be made with what technology we have today. I came up with a Sony/Aiwa/Tandberg/Nakamichi/Revox hybrid. I'd love to hire the techs that worked on the best Tandberg (3014A), Nak (ZX-9), Aiwa (S9000), and Revox decks and get them to design this new deck. If I had the chance to design a high end deck for professional and consumer use today, here's what I'd ask for:

-High quality build with pro grade parts inside.
-3 choices of output stages- A: Solid state. B: Tubes in output stage. C. Digital output via firewire with choice of 24 bit 384/ 192khz or 1 bit DSD
-4 motor dual capstan with pressure pad lifter (capstans able to handle faster speeds)
-3 choices of speeds- A. 1 7/8 IPS. B. 3.75 IPS. C. 7.5 IPS.
-fast forward and rewind speed brakes to slow down cassette before stop
-real time tape counter with memory point programable
-4 choices of noise reduction with HX-Pro (all usable at all speeds) plus OFF (no noise reduction)- A. Dolby B. B. Dolby C. C. Dolby S. D. DBX type 2
-seperated sections for power, input, and output with honeycomb copper plated chassis (aka Pioneer CT93 and Sony TC-K555ESJ), center cassette position (Sony TC-K909ES), and vibration reducing wood base (Aiwa XK-S9000).
-best material for heads for sound quality (aka Naks), but also capable of very high input/output levels
-6N pure copper wiring inside
-3 head system with seperated record and playback heads (aka Naks)
-cassette shell damping clamp (aka Sony and Aiwa decks)
-digital display VU meters from -40 to +20 db with *post-equalized* metering and 0 dB corresponds to a 250 nanoweber/meter (nwb/m) flux level
-input direct mode bypassing balance control and other sections
-seperate fine bias control for left and right channels (aka Naks and Teacs)
-record EQ setting plus record level for Dolby/DBX
-auto tape bias/level control/head alignment for recording with 4 test tone generator (aka Pioneer CT-93)
-record mode that turns off erase head for better signal to noise ratio on virgin tapes (aka Aiwa XK-S9000)
-balanced signal paths (aka Rotel), common ground paths (aka Luxman), and dual mono audio board (Sony TC-KA7ES)
-auto azimuth adjust(aka Naks)
-computer controlled transport system for very high head contact capability and accuracy, plus reduced head wear (aka Naks)
-speed adjust trim for slower or faster playback (pitch like control) with on-off bypass switch
-shielded sections between stages
-external isolated twin R core power transformers (aka Aiwa XK-S9000/ Excelia XK-009 and Sony TC-KA7ES)
-quartz locked direct drive with low wow and flutter
-transport safe enough to use with 120 minute length cassettes
-15hz to 27khz +-3DB at -10db level normal speed, 20hz to 27khz +-0db at -10db level 7 IPS mode
-flat frequency response from 20-20khz +1/-2db at all speeds at 0db level
-65 to 70db SNR without noise reduction
-XLR mike inputs with phanton power
-switch for mike or line level inputs via RCA connections
-choice of RCA or XLR output jacks (connected to the two optional analog output stages)
-external battery pack attachment for use on location
-internal power conditioner
-less IC circuits and more discrete circuitry

Did I miss anything? This deck would have to be on par and beat the performance and sound quality of the best decks including the Tandberg 3014A and the Nak ZX-9. Also to be released with this deck are the return of a modified Metal cassette designed to be used in this deck at high speeds. This deck would be for people who want reel to reel sound quality on cassettes that are still available, plus to kill by offering better then CD sound quality.

Target sales: To prosumers to convert their tape collections to digital, people who need a high end analog recording system with still available cassettes, or people who just want the best freakin' deck possible to get.

Price-$5000 or so

I'd suspect at that price point they could sell enough to merit it, especially since there are still die hard analog users out there with decks that you cannot get parts for anymore. And with the fact many burned CDs dying within a few years of being made, and that cassettes have better longevity, then it might be a good time for a return of one good deck out there.

My second option would be to produce a 1/4 inch reel to reel deck with cassette deck like features. Many reel to reel decks never used technology put into cassettes and this would be a good place to do this. A reel to reel deck with features I'd like is: 7.5 and 15 IPS, Dolby S, 1/4 track playback, 1/2 track play and record, bias fine with calibration mode, metal tape feature to use digital reel to reel tapes on, high headroom electronics, XLR mike inputs with phantom power, 3 choices of outputs similar to above list (tube stage, solid state, digital output (PCM and DSD), amd pitch control.

 

[goldear]

flat frequency response from 20-20khz at all speeds at -10db level

Real world machines can do better than that spec A couple of the better Naks could do 20-20K +1,-2 dB at 0dB! Since your going crazy here, why not just say DC to light? :lmao:

 

[braxus]

Ok I modified it.

And what is DC to light? DC power to a light socket? Or to the speed of light?

 

[wil816]

It's always fun to dream.

I think your tube version would have to forgo the option to use battery-powered for portable mode. The portability however is why I spent $140 to get my Marantz Superscope CD-320 pro stereo portable cassette machine refurbished just over a year ago. The equivalent new ones were several times that much, and now mine is practically in new condition. With the shoulder strap, you can record with it while walking, like a reporter. It's kind of a poor-man's Nagra. My open-reel machine with its outstanding performance sits in a closet for now, collecting dust.

The higher speeds would be nice, although I must say Nakamichi sure achieved great performance even at 1-7/8ips. The Portastudios I worked on when I worked at TEAC in the early 80's had really good performance at 3-3/4ips. They cut the high end off at about 13kHz and used a bias frequency of about 60kHz, which brought some impressive performance benefits. Even though the frequency response number itself may not look that good, they could maintain it into the high recording levels, unlike the other cassettes whose high end went to pot at 0VU.

I like your idea about the wider-range meters. Standard mechanical ones are the pits. If they have to be mechanical, at least we could use a circuit that would linearize the range a bit so everything from -7 or -10 and below wouldn't be all scrunched up at the bottom end with no resolution left. If you want to record a test tone at -10VU for example, you can't get it exact at all with common analog meters. You need external equipment.

Also to be released with this deck are the return of a modified metal cassette designed to be used in this deck at high speeds.

The metal tape didn't seem to give any real benefit at the higher speeds, which is why it never got used in the professional studio equipment with open reel machines that could go 7.5/15/30ips.

-shielded sections between stages

Not really necessary. When I worked at TEAC, few of the printed circuit boards were even double-sided, let alone multi-layer; but I've designed a lot of very complex boards in my work with up to 12 layers, and effectively kept switching-power-supply noise and digital noise out of the audio without shielding one section from another-- just using the techniques available with multi-layer boards. The crosstalk performance bottleneck will mainly be between sections of a head.

I definitely like the variable speed, as it's important for practicing an instrument with a pre-recorded tape. If everything is not exact, it is much easier to turn the tape speed up or down a bit instead of re-tune the instrument. Some instruments have a very narrow tuning range anyway.

-65 to 70db without noise reduction

If you mean a SNR of 65 to 70db, that's mostly a function of the tape, tape speed, and track width. There's not much else the machine can do in this area. However, to the NR options, I would add dynamic noise limiting. A good DNL circuit cuts the tape hiss by 10dB on tapes recorded with no NR, without hurting other areas of performance.

Did I miss anything?

I would add auto-alignment (bias, eq, levels for NR, meters, etc.), especially for audiophiles who people who realize the importance but don't have the expertise to do it by hand. For the rest of us, it would speed up the process. I would also include several presets for different kinds of tape; so for example, you could align for several different brands of tape, and have it remember all the settings so it would not be necessary to go through the process each time you change. There could be a feature to lay down brief alignment tones at the beginning of a tape, and automatically check and adjust PB alignment when you go back to the tape at a later date.

It would be nice to have the tape counter count in actual minutes and seconds, and be able to punch in a target location and have it go right to it.

For portable recording as mentioned in my first paragraph, it is often useful to have at least a record-level limiter (although I almost never use my automatic record level feature.) For portable use, it must also have a built-in speaker. It's nice to have as good a speaker as will fit in the small space, and I think the one on my Marantz may be better than the little one that came on the $4000 Nagras. There's the choice to turn the speaker on or off while recording, although of course the other outputs (earphone, line, and speaker-amp) are always online.

Since the Philips patent ran out years ago and we've gotten away from their standard by having faster tape speeds, four tracks in one direction, dbx, etc., how about an option for two wider tracks in one direction only (or two directions when mono is ok)? The wider tracks and higher speed (3-3/4ips anyway) would improve performance considerably. Should there even be an option for full track?

 

[goldear]

Ok I modified it.

And what is DC to light? DC power to a light socket? Or to the speed of light?

Neither: It would be the entire electromagnetic spectrum up to visble light.
DC = 0Hz. Light = An insanely high frequecy (somewhere > 1,000GHz, I'm guessing).

 

[braxus]

I guess I didn't add 2 track instead of the standard 4 track, to keep compatibility with older tapes. Part of the reason for the deck is to transfer old cassettes into digital for archiving in that format. Connection to the computer would be an obvious choice.

As for portability- I guess I was thinking of this deck as a standard home style deck with the option to use battery power. I often wanted to bring a deck on location without having to worry about finding power. This is not a deck I'd wanna carry around. I guess the tube outputs could be disabled when run off battery power. Somehow a deck with all these features I'd suspect would be a rather big deck. At least as big as the good Naks were. Mostly for indoor use.

As per Metal tapes and high speed- wouldn't you be able to extend frequency response way into the late 20khz area (like 27khz) and have the ability to crank more input to the tape (meaning higher saturation)? With those specs getting a flat 20 to 20 at 0db would be easier and give better sound.

As for SNR I mentioned, some decks get around 58-60 db SNR. The best decks I've seen go as high as 62 db SNR with Dolby Off. I'd suspect they could go better then this now, especially at higher speeds.

 

[goldear]

As per Metal tapes and high speed- wouldn't you be able to extend frequency response way into the late 20khz area (like 27khz) and have the ability to crank more input to the tape (meaning higher saturation)? With those specs getting a flat 20 to 20 at 0db would be easier and give better sound.

Yes, in theory you would get more extended FR, and more HF headroom. But you might actually loose stability by running such a narrow tape so fast, and inadvertently add scrape flutter by going so fast. The cassette mechanism was simply not designed to run that fast. But this is just a wild guess.

A genuine downside of faster speeds is slightly less extended Low Frequency response. Beleive it or not, most R2R mastering machines will have flatter, and slightly more extended low frequency response when running at 15ips, than they will when running at 30ips (even though the other aspects of the performance at 30 ips are clearly superior).

 

[tracker-x]

I can attest to that myself- I've heard slower speed tapes that I actually preferred overall, to faster speed tapes of the same recording with obvious better high end response in them.

It's the overall mix that counts.

I've also heard 1/4" tape machines that I preferred, over similar machines with 2k-4k high top end response.

That high end response seems to come at a price.

Now, since we're getting techical here, can one of your gurus tell me, why this is true ?

bandwidth at 0 dB level, is not as wide as bandwidth at -10 dB level

and bandwidth at -20 dB looks wider yet

why is that ? (in reference to printed specs)

for instance on this BIC machine freq response graph- the FR goes past 20K at -20 dB, but it falls off at 15K at 0 dB

what's the standard then ?

and how does dB relate/convert to readings on a tape deck VU meter ? what is the conversion factor ?

thanks ahead of time- but this question nags me- how can anyone hear anything at -20 dB ? it makes no sense to me

http://i11.tinypic.com/4dgxj7p.jpg

 

[Arkay]

Neither: It would be the entire electromagnetic spectrum up to visble light.
DC = 0Hz. Light = An insanely high frequecy (somewhere > 1,000GHz, I'm guessing).

Actually, I wouldn't want a cassette deck in my home that would put out significant microwaves, infrared, UV, X-rays....

 

[EchoWars]

Now, since we're getting technical here, can one of your gurus tell me, why this is true ?

bandwidth at 0 dB level, is not as wide as bandwidth at -10 dB level

and bandwidth at -20 dB looks wider yet

why is that ? (in reference to printed specs)

for instance on this BIC machine freq response graph- the FR goes past 20K at -20 dB, but it falls off at 15K at 0 dB

Tape saturation, pure and simple. The EQ curve to squish all that high-frequency material onto skinny, slow-moving cassette tape 'saturates' the tape (tape has reached the point at which it can no longer accept any more signal). Cassettes typically had frequency response measured at the -20db input level. I'm extremely sceptical of the ability of ANY cassette to record any frequency tone above 12KHz or so at 0db without serious tape saturation. Dolby HX will help by reducing bias with high-level high-frequencies, but 0db at 20KHz seems a tad unrealistic.

A little info:
http://arts.ucsc.edu/ems/music/equipment/analog_recorders/Analog_Recorders.html#Cassettes

 

[madpioneer]

So when are the first ones off the assembly line?
I want one! I have my checkbook ready......:yes:

 

[goldear]

Tape saturation, pure and simple. The EQ curve to squish all that high-frequency material onto skinny, slow-moving cassette tape 'saturates' the tape (tape has reached the point at which it can no longer accept any more signal). Cassettes typically had frequency response measured at the -20db input level. I'm extremely sceptical of the ability of ANY cassette to record any frequency tone above 12KHz or so at 0db without serious tape saturation. Dolby HX will help by reducing bias with high-level high-frequencies, but 0db at 20KHz seems a tad unrealistic.

A little info:
http://arts.ucsc.edu/ems/music/equipment/analog_recorders/Analog_Recorders.html#Cassettes

I agree with much of your statment; but not quite all of it. Head saturation is just as responsible here as is tape saturation. The graphs posted above are very typical of cassettes in general. And were I not to have seen flat FR curves for a couple of Nak models out to 20k at 0dB record level myself, I would not have beleived it either. I will attempt to post some scans of some reviews out of Audio Magazine of some old classic Naks if I have time tonight to prove my point to EW that Nakamichi was actually able to do this with a few of their better models, even without the aid of HXpro! :yes:

 

[wil816]

And were I not to have seen flat FR curves for a couple of Nak models out to 20k at 0dB record level myself, I would not have believed it either.

The May 1980 issue of Modern Recording had a review of all the major high-end cassettes, tested on a Nak. At 0VU record level, none of the type-II tapes even reached 10kHz, even though the TDK SA made 23Hz-25.6kHz at -20VU. The maximum output level at 10kHz was -2.5VU for TDK SA and -4.0VU for Maxell UD-XLII. In the graphs above, even the metal tapes' response at 0VU record level has a -3dB point at about 11kHz, and the Fuji FX-II's -3dB point is at about 7kHz at 1-7/8 ips.

 

[goldear]

The May 1980 issue of Modern Recording had a review of all the major high-end cassettes, tested on a Nak. At 0VU record level, none of the type-II tapes even reached 10kHz, even though the TDK SA made 23Hz-25.6kHz at -20VU. The maximum output level at 10kHz was -2.5VU for TDK SA and -4.0VU for Maxell UD-XLII. In the graphs above, even the metal tapes' response at 0VU record level has a -3dB point at about 11kHz, and the Fuji FX-II's -3dB point is at about 7kHz at 1-7/8 ips.

You will get no disagreement from me on this. The fact is that none of the decks that were made in 1980 could perform this remarkable feat (not a one)! That includes the legiondary Nak 1000ZXL, and the 700ZXL!

However, I have reviews from 1984 (I believe) in Audio Magazing for both the Nak Dragon, and for the Nak BX-300 which clearly show this level of performance. I know that it is hard to believe, but look them up yourself if you don't believe me (roger russle dot com has copies that you can purchase)! I will try to post a couple of scans from the reviews in question very soon, so that I can prove my point. OK?

 

[EchoWars]

Lol!

Anxiously awaiting. :yes:

 

[goldear]

Lol!

Anxiously awaiting. :yes:

Okay Mr. Skeptic, here you go...

I am attaching two pfd format scans. One was from a Nakamichi BX300 review in Audio magazine. Shown in this scan (among other things) is the Frequency Response of this machine at both -20dB and at the Dolby Level (= 0 dB on a Nakamichi) with Dolby C engaged. There is also a table listing the cut-off frequencies at both -20 dB and at 0dB with and without Dolby. Notice how dolby C actually increases the high frequency headroom of the deck. This has absolutely nothing to do with HX pro (which all nakamichi's lack). Dolby C has this so-called "spectral skewing" feature built-in, which increases HF headroom, although very few machines, even with the aid of Dolby C, can match this amazing level of performance.

The second scan shows a page out of a review of a Nakamichi Dragon. This also shows graphs at -20dB as well as 0 dB (the Dolby level). Like the BX-300, the frequency reponse with all tape types closely approached 20-20kHz, even at 0dB! Also like the BX-300, there is a significant improvement in HF headroom when dolby C is engaged.

As an aside: another interesting graph that is shown here demonstrates the Dragon's auto Azimuth correction system in action. It shows how the Dragon is able to completely correct for azimuth errors on a recorded tape during playback, and just how dramatically this fantastic feature can effect the frequency response of the tape that is being played back.

 

[braxus]

Man if only reviews in magazines could be as good today as they were years ago. They used to test everything then.

 

[wil816]

Man if only reviews in magazines could be as good today as they were years ago. They used to test everything then.

I don't think they did test everything here. It looks like what was done to get the high-frequency response was to back off the bias when there's a lot of high-frequency amplitude; but notice that they never tested the lows and highs at the same time. I expect that if you put in 400Hz and 20kHz at the same time, the lower bias used to get the 20kHz through would make the 400Hz more distorted. They only did one tone at a time, sweeping through the range. I can't say what the effect would be when you include psychoacoustics though, because the only time I can think of that you'd get so much hi-frequency amplitude would be in something like a cymbal crash, and for such a short moment, it might be more important to get the highs than to keep the distortion down on the rest of the orchestra.

 

[goldear]

I don't think they did test everything here. It looks like what was done to get the high-frequency response was to back off the bias when there's a lot of high-frequency amplitude; but notice that they never tested the lows and highs at the same time. I expect that if you put in 400Hz and 20kHz at the same time, the lower bias used to get the 20kHz through would make the 400Hz more distorted. They only did one tone at a time, sweeping through the range. I can't say what the effect would be when you include psychoacoustics though, because the only time I can think of that you'd get so much hi-frequency amplitude would be in something like a cymbal crash, and for such a short moment, it might be more important to get the highs than to keep the distortion down on the rest of the orchestra.

Welcome to AK wil816!!! I wish that I could provide a friendlier welcome to a NewBe the moment, but I'm afraid that I simply can't let this go unchallenged ... :sigh:

May I respectfully ask what in the heck you are talking about?!? :headscrat Did you happen to notice the overlaid Frequency Response graphs that show the traces for all three tape types at both -20db and at 0dB for the BX300? Those graphs were photographed directly off of a very high-quality real-time spectrum analyzer! That means that they tested all of frequencies simultaniously (using pink noise)! They clearly did not lower the bias simply to boost the highs. If they had, the distortion figures would be terrible, yet they are, infact, excellent. :yes: The reason that they were able to overlay these graphs is because this instrument featured memories which allowed for multiple traces to be displayed, and offset to show the all of those nearly flat lines that we see on that one screen.

The numbers quoted inside the table were only the band-width limits taken from the frequency response measurements. It does not mean that these were the only frequencies that they tested. Audio magazing (RIP) would always point-out when a piece of equipment did not meet its published specs. These particular decks actually exceeded them.

These guys were in the business of measuring equipment. They knew very well how to do it properly. If you can't believe your own eyes, then I'm not sure what else to tell you here. :sigh: Just because you may not happen to understand the concept of how the "spectral skewing" feature which is incorporated inside of Dolby C increases HF headroom, doesn't mean that it doesn't exist. FYI - It has nothing to do with bias reduction; that was a completely different technology called HXpro (which Nakamichi never used BTW).

 

[wil816]

Welcome to AK wil816! I wish that I could provide a friendlier welcome to a NewBe

Thanks. I've been here for a couple of months, with an interest mostly in tape recorders, but it looks like I've only posted 8 times so far.

I am familiar with spectrum analyzers. Using pink noise and a spectrum analyzer does not give distortion information. And I should have clarified that I did not mean the people doing the testing reduced the bias, but that apparently this one of the things Dolby C does, constantly adjusting the bias based on the program material. That's not all bad, but I'm very suspicious of anything that appears to give you something for nothing. But as I said earlier, the Dolby designers may have concluded that such extended frequency response at high record levels is seldom needed for more than a fraction of a second at a time for something like a cymbal crash, and that more distortion under those circumstances is more acceptable than a cymbal that sounds more like a toilet flushing because its sizzle is gone.

Another thing Dolby C apparently does is some companding, and not just around a lower threshold like Dolby B does. I used dbx for awhile on cassettes, and although the improved dynamic range is wonderful, the fact that it accentuates the cassettes' output inconsistency drove me nuts. Any little variation, even if not really enough to call it a dropout, became objectionable.

I wish I knew a lot more about tape recorders. Actually, I could learn by just taking the time to experiment. I have the equipment and know-how for the electronics portion. I'm an electronics engineer and did have to design tape R/P electronics for a project in the late 1980's, but it only required good voice quality on one track and digital on the other, not music. I am able to design digital solid-state recorders (and have), but something about tape recorders still attracts me. I also have an associate's degree in recording and worked as a repair tech at TEAC in the early 1980's and had an outstanding repair record, but I got out of recording before TEAC had any machines with Dolby C, and I am not familiar with it. All I know about it is what I got from the pdf'ed articles you posted the links to.