Btw Honinbou, what clock do you use?
If its the standard cristal from CDP it would differ a lot when you try a separate clock to 7220.
regard:
If its the standard cristal from CDP it would differ a lot when you try a separate clock to 7220.
regard:
Sorry to resurrect an old thread, but I just implemented directly clocking the TDA1541a in my Rotel RCD855 and am having problems.
I am using a Tent XO on a PinkFlea board (http://www.acoustica.org.uk/t/naim/35clockreg.html) that I laid out and etched myself with an additional 74HC74 to do the reclocking. Q/ is tied back to the data pin on the 1st FF and the 11.23xxx MHz clock signal is going to the clock pin on the FF, so output is /2 or 5.6xxx MHz as intended (verified on a scope).
Previously I had installed the board and was only replacing the players clock into the SAA7220... this worked fine.
I cut the trace (2 spots) from pin 16 on the 7220 to pin 2 on the TDA1541a and soldered minicoax center to pin 2 and then out to the clock board /2 (47ohm resistor between FF and output). I tried both Q and Q/... both result in staticky noise but can faintly hear music from the left channel. On this player, pins 2 and 4 of the DAC chip are tied together... could this be the problem? Any other suggestions?
I am using a Tent XO on a PinkFlea board (http://www.acoustica.org.uk/t/naim/35clockreg.html) that I laid out and etched myself with an additional 74HC74 to do the reclocking. Q/ is tied back to the data pin on the 1st FF and the 11.23xxx MHz clock signal is going to the clock pin on the FF, so output is /2 or 5.6xxx MHz as intended (verified on a scope).
Previously I had installed the board and was only replacing the players clock into the SAA7220... this worked fine.
I cut the trace (2 spots) from pin 16 on the 7220 to pin 2 on the TDA1541a and soldered minicoax center to pin 2 and then out to the clock board /2 (47ohm resistor between FF and output). I tried both Q and Q/... both result in staticky noise but can faintly hear music from the left channel. On this player, pins 2 and 4 of the DAC chip are tied together... could this be the problem? Any other suggestions?
The player stock from Rotel comes with a TDA1541a but pins 2 and 4 are tied together on the board... it worked fine with the TDA being clocked thru the SAA7220 😕 Service manual schematics also show 2 and 4 tied together. I'll take a look again at everything and see if I spot anything with my implementation.
Chris
Chris
Also, comparing the 1/2 clock coming out of the 7220 vs. the /2 clock from the FF, the 1/2 clock from the 7220 appears to be delayed more (0.4 divisions at 0.05uS/div).
I did not try disconnecting pin 4 from pin 2 on the TDA as I looked at my Pedja Rogic DAC that I am building and it too has pins 2 and 4 tied together. I think my problem is a timing issue and I need to add some more delay to my /2.
I was hoping Guido Tent might catch this thread and comment as well.
Thanks,
Chris
I did not try disconnecting pin 4 from pin 2 on the TDA as I looked at my Pedja Rogic DAC that I am building and it too has pins 2 and 4 tied together. I think my problem is a timing issue and I need to add some more delay to my /2.
I was hoping Guido Tent might catch this thread and comment as well.
Thanks,
Chris
You can send a PM to Guido, and wait some for answer, he is a busy man.
The "A' 1541 works with and without pin 4 connected to pin 2, experienced that myself.
The "A' 1541 works with and without pin 4 connected to pin 2, experienced that myself.
Rereading this entire thread, it seems that my problem is similar to what honinbou was seeing. It would also be nice to know if Fin ever got his working, and if so, what he wound up doing.
honinbou seemed to be saying that he connected the new /2 clock to pin 2, but that the original clock was still connected to pin 4
unless he just meant that pins 2 and 4 were tied together?
honinbou seemed to be saying that he connected the new /2 clock to pin 2, but that the original clock was still connected to pin 4
unless he just meant that pins 2 and 4 were tied together?
Pars said:
You may have a problem because the DAC now receives a bit clock that is too delayed compared to the other signals.
Try to feed your 1541 with an inverted version of that clock you're using.
Try using the Q/ (or the Q if you were already using Q/) of your DFF.
Bricolo said:
Understood. I think that you meant to say that the BCK going to the DAC from the XO->FF is not delayed as much as the original BCK from the 7220?
I tried both Q and Q/ from the FF... neither worked. I will try the Q/ -> (1) 74HC04 gate and play around with it some more if that doesn't work. It pretty much has to be a timing issue, and although I could not find propogation specs for XSYS (IIRC.... the raw crystal clock in) -> pin 16 in the 7220 datasheet, common sense would seem to indicate that it could not possibly be faster than a single HC FF.
Thanks for the replies, and Guido, thanks for taking the time to reply via email... much appreciated.
Chris
If the new divided clock has the same frequency and amplitude as the old one (check twice), then it can only be a timing issue IMO. But that's strange that neither Q nor Q/ work
Cheers Guido 😉 How's life in Eindhoven?
Cheers Guido 😉 How's life in Eindhoven?
Got it working... I already had a spot on my test board for an inverter, so put one in. It works with either the Q or Q/ FF output, then thru 1 inverter gate. On a scope, the timing of the Q -> inverter looks closer to the 7220 pin 16 than the Q/, so that is probably what I will stick with. Thanks to all for all the suggestions.
Chris
Chris
Nice it worked out now.
I divided clock /2 also to TDA1541A (by 74HC163), and fed 11Mhz clock to SAA7210 and 7220, (and disconnected the CLAB between 7210-7220) Now has everything the same clock. Tried also dem reclock, took /16 from 1st HC163 and divided that again /4, this 174kHz is fed with coupling cap into pin 16 of 1541. For I/V Opamp/buffer OPA2132.
Dispite all these mods, a simple PCM56 nonos setup, direct from SAA7210 (with glue logic) and Pedja's AD844 I/V outperforms to my taste this TDA1541 experiment.
I divided clock /2 also to TDA1541A (by 74HC163), and fed 11Mhz clock to SAA7210 and 7220, (and disconnected the CLAB between 7210-7220) Now has everything the same clock. Tried also dem reclock, took /16 from 1st HC163 and divided that again /4, this 174kHz is fed with coupling cap into pin 16 of 1541. For I/V Opamp/buffer OPA2132.
Dispite all these mods, a simple PCM56 nonos setup, direct from SAA7210 (with glue logic) and Pedja's AD844 I/V outperforms to my taste this TDA1541 experiment.
Bricolo said:
It sounds very good from the listening I have done (not alot yet). My wife even commented that it sounded much "better" 😉 Currently I am running LM4562s in the I/V stage (stock otherwise, other than BG NH Hi-Q caps on the output and muting trasistors removed).
Next step is a discrete I/V, probably based on Jocko's simple I/V... any recommendations regarding where to find board layouts or at least a complete schematic? How does this compare with the Pass (I think) D1 I/V, such as the board that Promitheus has on these forums?
Thanks,
Chris
Hi Chris
My recommendation: Try Pedja's AD844 or some other of his good designed I/V's. It sounds better then Jocko's simple I/V imo. This transistor I/V has a typical transistor sound, with some lack of transparence. The TDA1541 has a transistor sound too (again imo) so it could be better to combine a fet or tube like I/V with the TDA1541.
Have build RBroertjes simple I/V wich was allmost similar of Jocko's, some sort of derivation a Jocko's I/V. Will take it apart soon to re-use some of its parts.
My recommendation: Try Pedja's AD844 or some other of his good designed I/V's. It sounds better then Jocko's simple I/V imo. This transistor I/V has a typical transistor sound, with some lack of transparence. The TDA1541 has a transistor sound too (again imo) so it could be better to combine a fet or tube like I/V with the TDA1541.
Have build RBroertjes simple I/V wich was allmost similar of Jocko's, some sort of derivation a Jocko's I/V. Will take it apart soon to re-use some of its parts.
Hi Pars
Sorry I wasn't able to help by saying how I got on with the direct /2 clocking.....but I still have not done it.
Anyway - it is good to hear that you got it working and that it sounds good. Well done (to you and the others who helped) for sorting out the timing issue.
I wonder if these timing issues are avoided if "DATA" and "WS" are reclocked using 74HC74 driven from the same master clock that does the direct /2 clocking of the Bitclock (also using 74HC74)? Does this mean that all signals entering the DAC would have the same delay and therefore no timing issues?
Sorry I wasn't able to help by saying how I got on with the direct /2 clocking.....but I still have not done it.
Anyway - it is good to hear that you got it working and that it sounds good. Well done (to you and the others who helped) for sorting out the timing issue.
I wonder if these timing issues are avoided if "DATA" and "WS" are reclocked using 74HC74 driven from the same master clock that does the direct /2 clocking of the Bitclock (also using 74HC74)? Does this mean that all signals entering the DAC would have the same delay and therefore no timing issues?
Interesting thought. IIRC, Data and WS are still going thru the 7220 on mine. My guess was that the 7220 would almost have to have more propogation delay than a single 74HC74, hence additional delay needed.Fin said:
Pars said:
The diagram for the Arcam shows an inverter in the BCK line between the FF and the DAC...but there is also an inverter in the clock line to the SAA7220. So - there is no compensation for timing changes caused by re-clocking BCK. I wonder why Pars required an extra inverter in the Rotel and one is not required in the Arcam?
tubee said:
Did you have a similar problem and have to use an inverter in BCK for direct /2 clocking?
- Status
- Not open for further replies.