mediatechnology wrote:"$70 Decoder for New CX Records," John Roberts, Popular Electronics, January 1982.
While looking for the Holman article I found this 1982 Popular Electronics Article by JR:
It made the front cover.
Even though records aren't CX encoded these days it still seems like it could be a useful gizmo.
Hirsch-Houck Labs liked it.
Deja VU man... Yup getting the cover was how Popular Electronics sold me on doing it. The Jan cover date is actually the on the newsstands at Christmas issue so is supposed to be the "money" cover. It wasn't my idea, I suspect CBS approached the magazine trying to gin up some interest and the editor of poptronics called me.
They made me a CX licensee for free and gave me the full licensee design package including some recommended decoder circuits and even a copy of the professional encoder schematic, designed by URIE.
I decided to use a better OTA than their recommendations, and added a few JR touches. One of the difficulties for accurate CX decoding is that it requires a calibrated playback level for the below threshold decoder to be accurate. I designed a slick level meter based on Q5, Q6 and a bi-color LED. The playback level was calibrated when both the red LED and Green LED were the same brightness. More red is too hot, more green is too low, and when playing music it makes a nice bi color light show.
A few subtle design points, the URIE encoder used a tantalum cap for the major time constant C, so I used a tantalum cap for my decoder (C5) so they would both exhibit similar dielectric absorption tracking errors. In prudent design you don't use tantalum for time constants (or sample and holds) but that's what URIE used for the encoder so I tried to track them with similar DA for playback.
I've told this story before, but it never gets old (to me). While looking at the recommended playback decoder schematic I noticed an error where their attack time was something like 10% off. IIRC they left the release resistor in parallel with the attack resistor shifting the math slightly. The encoder was correct and agreed with the published spec so i made my decoder agree with the spec and the encoder, and sent it along to the magazine to prep for publication.
I wrote a letter to CBS and advised them about the error I found in their documentation package recommended circuit, and then forgot about it. As I recall this article was cranked out on a pretty short time line to hit the jan cover date so i was busy.
That November I went to the AES show in NYC and while walking about I introduced myself to the guys in the URIE booth. Apparently they already knew who i was, I was that guy who found the CX mistake.
It turns out 2 or 3 other licensees did not catch the mistake and had already manufactured some tens of thousands of decoders, all wrong.
CBS decided it would be more expedient to just change the encoder and official time constant to agree with the decoder mistakes, but they didn't bother to tell me.
Now my CX decoder, the only correct one was the odd man out since they changed the spec.
It was then only weeks before the magazine printed, so i frantically called the editors and tweaked one resistor value to once again agree with the rest of the CBS world. I got it fixed in time for the magazine print date but it was pure chance that I happened to meet the right guy in the URIE booth to find out what was really going on.
In hindsight i didn't need to bother. I sold enough kits to justify my effort but it was not a winner, and CX records pretty much tanked in the hifi market. Based on my experience with CBS's lack of engineering rigor I am not very surprised.
I still have a bunch of CX albums that CBS gave me. I doubt i still have a working decoder around...
JR
PS That was near the end of my kit business days... by 1985 I was working at Peavey.