|
Post by jeffc on Aug 19, 2012 8:27:56 GMT
Comments/advice anyone? A mate of mine has a stack of coaxial cables, some quite exotic, and a while ago made me one with left over Oyaide silver coax cable terminated with silver RCA bullet plugs. Not sure which level cable, but silver and expensive. www.oyaide.com/ENGLISH/AUDIO/products_category/digital_cable/index.htmlAnyhow, use of the Oyaide cable improved sound quality overall and quite markedly compared to some Kordz silver coated copper coax I had been using, and even so after re-terminating the Kordz with the same RCA plugs to ensure these weren't contributing. And of course there are heaps of exotic coax cables out there touting all manner of reasons for why they offer SQ benefits over lesser cables. In some Sunday arvo web browsing, I came across what seems to be an issue with SPDIF implementations in most CD players that would accentuate SQ differences between cables, and what looks to be some fixes for these problems by using either a tube or a jfet buffer. lampizator.eu/digilampizator/CD%20Transport%20Digi-Lampizator.htmlTubes are out for me, but the jfet solution looks easy enough to try, but it would be nice to be able to tap into a +5V DC digital supply close to the transmission IC rather than look for a +9/12V DC supply as suggested in this schematic. Anyone have experience with SPDIF signal transmission buffers like these and be able to give me some advice on what might be worth a try. I’ve got an old Rotel RCD-965BX that’s still serviceable and uses the SAA7220 transmission IC mentioned, and I’d be up to lifting the lid to try a jfet buffer. cheers.. jeffc
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 19, 2012 8:53:33 GMT
Hi Jeff It's now 2012 , not the dawn of the CD era >30 years ago ! My Pioneer DVD-A/SACD, CD player isn't exactly a whippersnapper either! ;D Regards Alex Attachments:
|
|
|
Post by jeffc on Aug 19, 2012 9:26:25 GMT
Yep Alex... it is 2012..... but I can still remember vividly picking up that Rotel CDP at Sydney airport in about 1985 in transit from Brisbane to HK or somewhere, box left for me at the information desk, to save shipping costs. No way would this happen today due to security risks that weren't about back in those good-ol’ days. Anyhow, back on track, based on the schematic in the service manual, this old CDP seems to be amongst those that seem to have not implemented SPDIF transmission quite as well as could be to avoid cable influences, as supported by my listening experiences, and as its still in use it in my grotto system for playing CDs, I thought a jfet buffer might be worth a try, seeing as though it looks quite simple needing only a few parts, and doable P-to-P without needing to remove the PCB. What I was hoping for was advice on whether the BF245C jfet could be usable at +5V or not, whether another type might perform better, and whether the resistor values as specified are ideal. I like fixing things, or at least trying to, you know that. cheers.. jeffc
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 19, 2012 10:10:49 GMT
|
|
pagan
Been here a while!
Posts: 512
|
Post by pagan on Aug 19, 2012 11:52:52 GMT
Jeff You need to look at the circuit for the digital out of the cdpayer first then work out if it could be improved. Does the cdplayer use a transformer on the digital out?
Just putting a buffer after the damage is done is a waste
|
|
|
Post by gommer on Aug 19, 2012 12:38:45 GMT
I'd like to help, but i'm unfamiliar with what seems to be a symmetrical JFET (interchangeable D and S). Furthermore, i've got no clue as what levels come out of the SAA7220 (5V TTL, 3,3V TTL?).
I'd rather use a picogate inverter with schmitt trigger input, followed by a normal logic gate n-channel FET. Not the kind of thing you'll find in a commercial design, but not because it's worse. Commercial design usually try to reuse common components as much as possible because of the economy of scale. Therefore you'll find discrete components more easily used than dedicated IC's.
|
|
|
Post by jeffc on Aug 19, 2012 21:41:19 GMT
Thanks guys, Alan, from the CDP schematic, the digital signal from the SAA7220 goes to a 74HCT08 IC then via a transformer. The jfet cicuit shown would be fitted direct to pin14 of the SAA7220, so no messing. gommer, I was really attracted to the jfet solution for its simplicity, and probably now need to try it out anyway just to find out first hand whether it works OK. I know the sound of the digital out from this CDP well, and as its been in a stable system for a while, I should pick any changes/improvements in SQ if any should eventuate. If this helps, from the SAA7220 datasheet. Pin 14 = DOBM peak-to-peak voltage across 75 ohm load via attenuator (ref Fig 10), VL(p-p) = 0.4V to 0.6V A pic of Fig. 10 and RCD-965BX schematic from pin14 of the SAA7220. cheers.. jeffc
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 19, 2012 21:53:15 GMT
Thanks guys, Alan, from the CDP schematic, the digital signal from the SAA7220 goes to a 74HCT08 IC then via a transformer. The jfet cicuit shown would be fitted direct to pin14 of the SAA7220, so no messing. gommer, I was really attracted to the jfet solution for its simplicity, and probably now need to try it out anyway just to find out first hand whether it works OK. I know the sound of the digital out from this CDP well, and as its been in a stable system for a while, I should pick any changes/improvements in SQ if any should eventuate. If this helps, from the SAA7220 datasheet. Pin 14 = DOBM peak-to-peak voltage across 75 ohm load via attenuator (ref Fig 10), VL(p-p) = 0.4V to 0.6V A pic of Fig. 10 and RCD-965BX schematic from pin14 of the SAA7220. cheers.. jeffc Hi Jeff You could also try taking the output from the 74HCT08 where it goes into the transformer at present .My understanding,(although possibly in error) is that the parallel buffers not only give added current drive capabilities for driving the output transformer, but they they also "square up" the waveform from the previous stage and give consistent output voltage levels. Regards Alex
|
|
pagan
Been here a while!
Posts: 512
|
Post by pagan on Aug 20, 2012 7:53:05 GMT
Thanks guys, Alan, from the CDP schematic, the digital signal from the SAA7220 goes to a 74HCT08 IC then via a transformer. The jfet cicuit shown would be fitted direct to pin14 of the SAA7220, so no messing. gommer, I was really attracted to the jfet solution for its simplicity, and probably now need to try it out anyway just to find out first hand whether it works OK. I know the sound of the digital out from this CDP well, and as its been in a stable system for a while, I should pick any changes/improvements in SQ if any should eventuate. If this helps, from the SAA7220 datasheet. Pin 14 = DOBM peak-to-peak voltage across 75 ohm load via attenuator (ref Fig 10), VL(p-p) = 0.4V to 0.6V A pic of Fig. 10 and RCD-965BX schematic from pin14 of the SAA7220. cheers.. jeffc Fom my meager understanding.... It's using a AND Gate to buffer and drive 5V TTL to the optical. There's 2 resistors going to the spdif transformer, 1 in series, 1 in parallel (and to earth) across the transformer, to drop the 5V to 0.6 spdif level. (The resistors are probably ~300 Ohm in series and ~75 Ohm in parallel. The AND Gate is also used to kill the Digital out. If there's a (spdif)digital out before it goes into the 7220 then use it, if not, there may be i2s going into the SAA7220, If you can get to these even better, you could always use a WM8804 to convert the i2s into and spdif out. Allan
|
|
|
Post by jeffc on Aug 20, 2012 10:35:05 GMT
Hi Alex/Alan, Apologies for the poor quality pics, done in a rush before work after getting up at 4 am to drop my son to the airport for the 1st flight out. FYI, the resistor values for SPDIF V out from the 74HC08 IC are 560R series 620R parallel with the transformer to GND. Way to complicated setup before the SAA7220 for me to figure what could be done there so……. in keeping with the KISS and probably naive principle…. I’ll blindly follow the simple SAA7220 pin14 jfet schematic as is, supplied by 12V-15V, maybe up the 120R to 130R, and see what eventuates. Only have some BF245A fets so will need to go shopping first…. Thanks for all the suggestions, although I seem to be ignoring them all , but will report what I learn. cheers.. jeffc
|
|
pagan
Been here a while!
Posts: 512
|
Post by pagan on Aug 23, 2012 13:04:53 GMT
|
|
|
Post by jeffc on Sept 2, 2012 8:09:38 GMT
Hi Alan, Thanks for this, interesting stuff. BF245C jfets arrived Friday so I've now buffered the SPDIF out from pin14 of the SAA7220 IC as in the schematic, except for omitting the 10nf cap as all DACs I have all have use a 100nf on their input. Does it work? ?? It works fantastic from what my ears tell me. No CRO so I have to take it at face value that what's been posted regarding square wave traces, is exactly what I have now. The SAA7220 has several pins (5, 8, 17, 19, 20, 21) not connected, so rather than using a small piece of vero board to make the circuit, I used these as solder lugs for P-to-P wiring to keep wiring as short as possible. +15V DC was stolen from the opamp output PSU, so probably not ideal, but the CDP PCB has a single gnd plain, and it was there so I used it, just added an extra 10uf oscon close to the BF245C. Also added a 100uf oscon/1000pf silver mica on the +5V Vdd pin 24 of the SAA7220. Noted approx 1V DC offset on the SPDIF output after the jfet, so if you're not sure whether you have an input capacitor in your DAC, best not to omit the cap from the schematic like me. If anyone has an old solid CDP (the ones before SMPS took over the world and made like a Sherman tank) still in service as a DAC source for playing CDs, and it uses the SAA7220 IC for SPDIF, this really is a very easy, useful tweak that adds heaps to listening pleasure. Loving it, it really does somehow make a substantial difference. cheers.. jeffc
|
|