Analog Output Stage Mod for Marantz CD-67SE

October 23, 2002

Welcome to my Analog Output Stage Mod for the Marantz CD-67SE page. I should first refer you to my Simple Mods for CD Players modpage for general mechancal enhancements that I applied to the CD-67SE described here prior to incorporating the following mods, both so that you don't miss anything and I don't have to repeat them (!). Then come back here and read about the mods to the actual circuit that will like as not have a much more noticeable effect on the sound of the player.

Stage One--Disassembling the CD-67SE

Unlike the mechanical mods, many--if not most--of which can be applied without removing the PC board, these mods require that you remove it from the case. So first remove the cover by unscrewing the two rear top-lip case screws and the four side screws, which have captivated washers. I like the copper screws they used on this player, but that's me. The only downside is that if you bung them up, you can't touch them up with a black magic Marker--oh well.

Then remove the top brace by unscrewing the four side mounting screws and lifting it up out of the player chassis proper. Then, depending on your preference, gently disconnect, one by one, all of the cable assemblies. There are several, and they vary in type, but are different enough so that you can easily identify them when it's time to put the thing back together. I found that with the ribbon cables, especially with the wider one, it helps to pull them up into a loop so that you can get a good grip and pull them evenly out of their sockets. With the Molex-style plugs, gently lift them by pulling evenly on each end at the same time until they pop out of their sockets. There is one locking socket that grips the 3-wire cable that comes from the front panel. Lift the locking lever and release the cable so that you can gently lift it from its socket. Finally, one harness that emanates from the laser assembly is held in place with a nylon retainer that requires that you spread the sides so you can remove the very delicate wires, so don't force it or you'll be a sorry formerly working CD player owner.

Once the cables have been disconnected, gently unscrew the rear-panel analog and digital (and remote control) jack retaining screws by holding them from the inside with your fingers and gently unscrewing them from the outside. Watch that you don't strip the plastic blocks that hold them, especially when you reassemble the unit. Then unscrew and remove all of the board retaining screws, noting that only one has a captivated washer--the one closest to the front panel center. The last ones you remove should be the ones that hold the power transformer down to the chassis, but first gently spring the power switch activating bar from the board-mounted push-pull switch by spreading the collar with a small screwdriver and using another screwdriver, push it out and off of the switchbar. Then, after the power switch has been disconnected and all of the board retaining screws removed, move the cables out of the way and lift the board up from the front so that the rear-located connectors can be pulled inside and free of the back panel and then lift the board up and out of the player. Whew.

Stage Two--Removing the JRC 2114D Opamps

OK, look at the board. A lot of nice components have been used in the SE version of the CD-67, incuding now hard-to-get Elna Cerafines, Elna Silmics, and so forth. Much has been written about replacing all 47 mfd caps with 1000 mfd caps, which sounds ludicrous to me, and I'd be somewhat suspicious--and am--of replacing what are basically top-shelf caps with top-shelf caps of much larger values when the incidental cost of using them during production was not that much higher. Why try to out-think the engineers? Are you an engineer? Just because you play one on the Internet doesn't make you one. Anyway, find the two opamps that are surrounded by Elna Silmic electrolytic bypass caps, near the output and just prior to the copper-covered HDAM modules. By the way, don't bother removing those covers--there isn't much in there you can change to improve anything, and just because I did it doesn't mean you should--let me be the idiot. Anyway, now you have to suck it up, turn the board over, and desolder and remove the JRC 2114 opamps. They're 8-pin DIP style chips, and their orientation is clear from both above and below the board, so don't worry about getting it right when you put in the new ones later. This is no-turning-back time, by the way, so don't screw it up. Use good equipment, specifically a thermally-controlled, non-electrostatic soldering station (no, you don't need a $400 DE-soldering station), and some thin solder wick such as they sell at Radio Smack-O. Once you have the opamps out, install and solder the sockets, for which I used silver solder, or whatever it is that Radio Shack purports to be silver solder, mainly because it's really thin and tends to be safer to use around dense circuitry. You can use commercial 8-pin DIP sockets, or do what I did, impulsive guy that I am--make your own. I cut up and ground down two 16-pin DIP sockets and then glued little black plastic end braces I rescued from the cut pieces onto the free ends. Nuts, I know, but it worked fine. In they went and that was that.

At this point, if you're the paranoid type, you COULD plug the original opamps back in and put the whole thing back together, try it, and be relieved that it still works. But it'd require that you reassemble the player--and take it back apart again when you want to do the actual mod--and chances are greater that you'd screw that up than if you did the mod itself now. Why? Because you'd be working from panic. But if you're feeling nauseous after this part of the process, shut everything down, go upstairs and have a Martini or a Manhattan and take a break. Tomorrow is always another day (if you're lucky).

Stage Three--Replacing the Opamps With Better Ones

OK, it's tomorrow. Get a pair of Burr-Brown OPA2604's, which are dual opamps with internal compensation, and are pin-for-pin replacements for 2114's, LM833's, and a host of other more-or-less standard dual opamps. While they may not be the best--AD825's are purported to be better, but are SMD-type and must be obtained from LC Clock on litle boards (Type I, as I recall), cost a lot of money and have to be shipped from Europe, so forget that--they work fine. At this point you could--and I did--reassemble the player and see what you've done. You'll like it but you'll still feel like there is more to be done. There is.

Stage Four--Biasing the OPA2604's Into Class A

OK, I got this tip from Bob Fitzgerald, from a submission dated 9/29/02 on mods he made to the analog output stage of his his Kenwood KT-8300 tuner that I found on the Yahoo FM Tuner eGroup "sister site" to the controversial Tuner Information Center--a self-righteous and needlessly damaging venture if I ever saw one--and give him full credit for it. In sharing his information, I hope to help others in the position I was in, namely liking some aspects of the CD-67SE but hating others--like the sound, to be specific. Here we go:

On the underside of the board, you will have to solder four small 3.6K-ohm resistors--one from pin 4 to pin 1 and another from pin 4 to pin 7 on EACH of the OPA2604's (take them out of the sockets first, if you're smart). I used--gasp!--quarter-watt carbon composition 5% hand-selected resistors in mine. OK, I had them, so sue me. Here's your Roderstein, Vishay, whatever--I would have used quarter-watt carbon-films if I'd had them. All you're doing is biasing the opamps closer to Class A, so I'm not too worried about it. I hit them quickly with the soldering iron at its lowest temperature (since carbon comps are notorious for shifting value when soldered), and got through it OK. The player sounds terrific, if that's any indication.

There is one consideration one should be aware of when putting wide-bandwidth opamps into equipment that formerly had narrower bandwidth, slower, and less sophisticated opamps in it previously, and that's high-frequency oscillation. As Bob explained, it was necessary to use 5-7pf ceramic (or silver mica) caps across the feedback resistors in his KT-8300, but with the opamps in the CD-67SE removed, I measured the capacitance across the 27K (as I read it) feedback resistors, and saw what I think was slightly over 1000pf (.001mfd). So not hearing any sizzle or experiencing any deleterious effects from the switch, I left that part of the circuit alone. If I had a Spectrum Analyzer I'd look for problems, but I don't, so I can't. But if I get information to the contrary, I'll post it here.

Stage Five--Put a Choke On It

You'll notice that where the AC linecord enters the cabinet there is a lot of unused PC board space where a proper AC line (mains) filter ought to go, but doesn't. While I am researching an easy way to punch a clean .82 x 1.14 inch rectangular hole in the rear panel in which to mount a Corcom model EEA or EEB EMI/RFI filter (any ideas please email me), I decided to simply place an over-the-cord clamp-on low-Q choke (see The Shack for yours). It beats a blank, as we used to say in the music business.

Stage Six--Curse Me or Love Me

OK, you're done. Put the Burr-Brown opamps in their sockets, reassemble the player, and put your favorite CD in it and see what you think. I should say that while I was in the throws of all this change, I replaced four rather rudimentary Elna electrolytic caps with Nichicon MUSE polarized KZ-series replacements of the same value, and it made zero difference. It was at that point that I decided to go all the way, so to speak. And I should also say that I first used OPA2604's in a 1985 Magnavox model 1051D 14-bit CD player at the urging of Chuck DiGio--last name shortened to save space--with good results. I didn't bias them into Class A, but if I had four more 3.6K resistors on hand and some extra time I would, now that I've seen the improvement it makes. So thanks to Chuck for his encouragement, not to mention the fact that he gave me the four OPA2604's I ended up using.

After performing these changes on the CD-67SE, I listened to Bonnie Raitt's "Nick Of Time" and Quincy Jones' "Back On The Block" and almost had a religious experience. Seriously. This isn't one of those subtle changes you think you hear--it's major. I've often considered changing out the clock, but the nature of the sound made me more suspicious of the analog section than the digital part of the player, and I don't care what the "experts" say, how much of a role can jitter play when the data stream and decoding section of the player is self-correcting by design? What's up with that? What kind of jitter do you think you get in your car's player, your portable player, or when you turn up the bass? How about data retrieved from your PC's CDROM drive? A quartz crystal is used as the basis of an oscillator that drives circuitry specifically designed to handle errors from all kinds of sources. Sounds like snake oil to me, until I actually hear differently. Besides, $200 plus shipping for an LC Clock model X O2? Puh-lease! Wait until Shady O'Rack sells 'em, I say.

Anyway, that's it for now. Try these changes and see what you think--they may well improve the sound of your player at very little cost.
 

To return to WA3FLE homepage, click here.