carlosfm said:Of course it was.
And I have two independent RCA outs on the player, I can even A/B.
There's no comparison between a bitstream Dac, the better that it may be, and a good multibit Dac.
Yes, the TDA1543 is good.
Forget the specs, well done it's very good.
Bitstream Dacs always sound undynamic to me.
Thanks Carlos - it's good to hear from someone who has given both types of DAC a fair go. I will probably end up doing something similar to you - ie. take the SAA7321 in the CD624 to it's max, and then do a good multibit dac for comparisson. The idea of being able to A/B two DACs in one CDP is great. Do you run them off the same Tent Clock at the same time?
Thanks for all the replies...
They have given me some more things to try.
I will try some of the KYW mods in the not too distant future.
I note with interest his comment about the C Found mods as I also had some doubts e.g. CF recommends changing the -11V PSU cap from 3,300u to 4,700u to match the +11V, (quote) "this balances the loading of the mains transformer windings and allows for a slightly higher current dissipation". Impressive stuff!!!!
The KC-7 is built on Elso's pcb and fitted <5cms from the signal processor - maybe I should connect it to BCK at the DAC as well (do I need to disconnect the original BCK i.e pin 48 of the SAA7378 from BCK input, pin 1 of the TDA1545?).
Elso - thanks for the OP275 recommendation. I replaced the OPA627s I/V in my Arcam with a pair of OP176s and was surprised to get a worthwhile improvement. The OPA627 is better than the OP176 in the o/p filter but it seems to be the other way round in the I/V. I will try an OP275 next time I open up the CD723.
My approach with the CD723 is in line with the points made by Fin and Jean-Paul - it's a bit of cheap, clean entertainment to see what can be achieved. Better to peel a few tracks off this pcb than a Wadia! Expectations are low so if it turns out good then that's just fine with me.
On another point - discrete I/V stages - I'm sure many will recommend this. However I had a bad experience with the LC audio ZAPfilter. The minimal instructions are wrong and eventually LC more or less agreed that the thing cannot accomodate unipolar current output without an external bias current.
Anyway, I ran the thing as a line stage to see how it sounds and was unimpressed, it's some way behind an OPA604 run off even the most basic of power supplies!
Might be tempted by KYW's OPA660 I/V circuit if I can find a decent pcb to accommodate it (haven't got time to design one 'cos I need to finish quite a number of projects).
Cheers
Dave
They have given me some more things to try.
I will try some of the KYW mods in the not too distant future.
I note with interest his comment about the C Found mods as I also had some doubts e.g. CF recommends changing the -11V PSU cap from 3,300u to 4,700u to match the +11V, (quote) "this balances the loading of the mains transformer windings and allows for a slightly higher current dissipation". Impressive stuff!!!!
The KC-7 is built on Elso's pcb and fitted <5cms from the signal processor - maybe I should connect it to BCK at the DAC as well (do I need to disconnect the original BCK i.e pin 48 of the SAA7378 from BCK input, pin 1 of the TDA1545?).
Elso - thanks for the OP275 recommendation. I replaced the OPA627s I/V in my Arcam with a pair of OP176s and was surprised to get a worthwhile improvement. The OPA627 is better than the OP176 in the o/p filter but it seems to be the other way round in the I/V. I will try an OP275 next time I open up the CD723.
My approach with the CD723 is in line with the points made by Fin and Jean-Paul - it's a bit of cheap, clean entertainment to see what can be achieved. Better to peel a few tracks off this pcb than a Wadia! Expectations are low so if it turns out good then that's just fine with me.
On another point - discrete I/V stages - I'm sure many will recommend this. However I had a bad experience with the LC audio ZAPfilter. The minimal instructions are wrong and eventually LC more or less agreed that the thing cannot accomodate unipolar current output without an external bias current.
Anyway, I ran the thing as a line stage to see how it sounds and was unimpressed, it's some way behind an OPA604 run off even the most basic of power supplies!
Might be tempted by KYW's OPA660 I/V circuit if I can find a decent pcb to accommodate it (haven't got time to design one 'cos I need to finish quite a number of projects).
Cheers
Dave
Re: Thanks for all the replies...
Hi,
I happen to know the gentleman, nuff said.
Seperate (galvanically) PSU with low earth leakage for that?
No, don't.
Try an LM6182 please. No reactive components in the feedback loop, just one resistor and adjust teh values for 2mA 0dbf on the TDA1545 (see datasheet).
Sayonara
Hi,
Dave S said:
I happen to know the gentleman, nuff said.
Dave S said:
Seperate (galvanically) PSU with low earth leakage for that?
Dave S said:
No, don't.
Dave S said:
Try an LM6182 please. No reactive components in the feedback loop, just one resistor and adjust teh values for 2mA 0dbf on the TDA1545 (see datasheet).
Sayonara
Quote: "Seperate (galvanically) PSU with low earth leakage for that?"
1) Elso recommends using the digital +/-10V rails, not a separate PSU.
2) Given that Elso may be fallible, how do I get low earth leakage? Do I need an E-I core transformer rather than a toroid?
I added an IEC mains input, which means I have the possibility to earth the chassis - is this worthwhile?
Dave
1) Elso recommends using the digital +/-10V rails, not a separate PSU.
2) Given that Elso may be fallible, how do I get low earth leakage? Do I need an E-I core transformer rather than a toroid?
I added an IEC mains input, which means I have the possibility to earth the chassis - is this worthwhile?
Dave
LM6182 is discontinued, not sure if I can find NOS. Can I use AD812 in the same way?
Other duals in my bitbox: OP275, RC4227, NE5532, AD8620 (SO8) - any of these worth a try?
LM6172 looks viable if I can find one (not listed in Farnell, RS or Digikey), OTOH the feedback resistor, 3K3, seems a bit high for such a fast amp.
Should I change the o/p circuit from CF's, which is:
3K3 feedback resistor with 470pF polystyrene in parallel
47uF pana FC // with10nF Wima on opamp output (you aready told me to change this)
1K series resistor and 470uH and 820pF in shunt (after the coupling caps).
Thanks
Other duals in my bitbox: OP275, RC4227, NE5532, AD8620 (SO8) - any of these worth a try?
LM6172 looks viable if I can find one (not listed in Farnell, RS or Digikey), OTOH the feedback resistor, 3K3, seems a bit high for such a fast amp.
Should I change the o/p circuit from CF's, which is:
3K3 feedback resistor with 470pF polystyrene in parallel
47uF pana FC // with10nF Wima on opamp output (you aready told me to change this)
1K series resistor and 470uH and 820pF in shunt (after the coupling caps).
Thanks
What about Peufeu's claim that Philips CD723 is not bit accurate -
http://peufeu.free.fr/audio/extremist_dac/bit_accuracy/
can the player ever get beyond mediocre?
John
http://peufeu.free.fr/audio/extremist_dac/bit_accuracy/
can the player ever get beyond mediocre?
John
Re: Re: Thanks for all the replies...
Please do so
The BCK "happens" to be 8.X MHz.
There's enough time margin to prevent metastability problems (BCK output of decoder is lagging a bit with respect to clock input).
To be really sure, with the KC7 you can connect "normal" output to TDA1545A, and the inverted output to the servo/decoder chip.
IME the performance gain is higher than just connecting to the SAA737X servo/decoder only.
Kuei Yang Wang said:
Please do so
The BCK "happens" to be 8.X MHz.
There's enough time margin to prevent metastability problems (BCK output of decoder is lagging a bit with respect to clock input).
To be really sure, with the KC7 you can connect "normal" output to TDA1545A, and the inverted output to the servo/decoder chip.
IME the performance gain is higher than just connecting to the SAA737X servo/decoder only.
jkeny said:
I tested the TDA1387, scavenged from an old SB AWE64 Gold soundcard (much better than any Live IMHO).
This Dac is much alike the TDA1545A, and sounds very good indeed.
I could never find where to buy the TDA1545A, but listening to the 1387 I imagine this is a very good Dac.
The 723 has this dac, so it MUST be possible to make it sound VERY good.
Even if you have to take the chip out of the original PCB and make a new one.
Konnichiwa,
If that is done ground loops and ground bounce seems unavoidable. It is one of things that cause me problems with the claims that external clocks actually lower jitter. If you supply it locally the implementation must be handeled very carefully, with minimal loop areas to not cause problems.
If you supply it from a seperate supply connecting this supplies ground via a AC milliamp Meter (or even better a small resistor with a differential 'scope probe across it) will show just how much current "leaks" into the (clock) signal ground line.
Among those I have tested the lowest leakage items had 2-Chamber safety construction with a dairly thick plastic in the thing and in general small size. I find that you need to actually try a few locally available brands and you should find a good one. Often doubling up the transformers and wiring winding out of phase can get you a drastic improvement as you now have capacitances et all better balanced and leakage in opposite polarity between the two transformers. It's a bit of "black art.
Measure, with the Player in system using a good fully floating 'scope.
The AD812 appears to be the cmpatible AD part. Go for it.
EVERYTHING is worth a try, if only to confirm a negative fact.
The 6182/812 will be a better choice.
Of course. First change the resistors on the DAC's reference for 2mA full scale current. Then you need an I/V resistor of 5.656V/2mA = 2k8, no parallel capacitors or RLC circuits.
Then feed a suitable (positive) current into the inverting input to bring the output down to zero Volts from the around 3.3V it is at (this is well implemented is stable and fit & forget, no servo needed). IIRC I used a seperate 317 Reg (15V) from the raw supply some extra RC filtering and then simple resistors to each inverting input, with a +/-10% trim in series (you need to experiment to find the right fixed resistor value first) to trim out sample variation in offset.
BTW, I also used a 317 (5V) with added RC for the DAC and 317/317 (+/-15V) without RC filtering for the OpAmp, of course low impedance bypass cap bypassing the Adj pin to ground in all cases, the negative voltage was actually an identical 15V positive (317) regulated voltage run via seperate wiring to the main PCB ground near the DAC where everything was "stared" for DAC & Analogue Stages.
Then a 100R build out resistor that forms the same LPF as previously in the feedback loop, so 15nF on the Output to ground, use a nice low inductance, audio grade type like Silver Mica or low inductance Polystyrene.
The actual Circuit remaind in my case completely on the CD-720 (I did this ages ago) Board and used original traces etc.
All that came from that board to the added supply board and transformers where two ribbon PSU cables one to carry DAC, Offset and Op-Amp +/- Supplies from one small EI Transformer, the other carying +/-12V (317/337 regged, classic symmetrical supply, again small EI transformer) to near the supply section where it fed the only the Servo Portion of the player.
Another transformer and regulator supplied the various Digital Decoder & servo IC's, I simply retained the original clcock arrangement de-noised on the supply lines and with a real X-tal instead of the ceramic resonator.
So in total I added 3 pcs small PCB mount transformers and a batch of regulators (8 or 9 pcs, would have to check my notes) on a Veroboard PCB to a CD720 and made a few small changes on the main PCB, plus I fixed the dreadful chassis. The result with a nice external mains conditioner sounded jolly decent, good enough to make a later bought Marant CD-67 SE to sound completely "declasse" in comparison.
Sayonara
Dave S said:
If that is done ground loops and ground bounce seems unavoidable. It is one of things that cause me problems with the claims that external clocks actually lower jitter. If you supply it locally the implementation must be handeled very carefully, with minimal loop areas to not cause problems.
If you supply it from a seperate supply connecting this supplies ground via a AC milliamp Meter (or even better a small resistor with a differential 'scope probe across it) will show just how much current "leaks" into the (clock) signal ground line.
Dave S said:
Among those I have tested the lowest leakage items had 2-Chamber safety construction with a dairly thick plastic in the thing and in general small size. I find that you need to actually try a few locally available brands and you should find a good one. Often doubling up the transformers and wiring winding out of phase can get you a drastic improvement as you now have capacitances et all better balanced and leakage in opposite polarity between the two transformers. It's a bit of "black art.
Dave S said:
Measure, with the Player in system using a good fully floating 'scope.
Dave S said:
The AD812 appears to be the cmpatible AD part. Go for it.
Dave S said:
EVERYTHING is worth a try, if only to confirm a negative fact.
Dave S said:
The 6182/812 will be a better choice.
Dave S said:
Of course. First change the resistors on the DAC's reference for 2mA full scale current. Then you need an I/V resistor of 5.656V/2mA = 2k8, no parallel capacitors or RLC circuits.
Then feed a suitable (positive) current into the inverting input to bring the output down to zero Volts from the around 3.3V it is at (this is well implemented is stable and fit & forget, no servo needed). IIRC I used a seperate 317 Reg (15V) from the raw supply some extra RC filtering and then simple resistors to each inverting input, with a +/-10% trim in series (you need to experiment to find the right fixed resistor value first) to trim out sample variation in offset.
BTW, I also used a 317 (5V) with added RC for the DAC and 317/317 (+/-15V) without RC filtering for the OpAmp, of course low impedance bypass cap bypassing the Adj pin to ground in all cases, the negative voltage was actually an identical 15V positive (317) regulated voltage run via seperate wiring to the main PCB ground near the DAC where everything was "stared" for DAC & Analogue Stages.
Then a 100R build out resistor that forms the same LPF as previously in the feedback loop, so 15nF on the Output to ground, use a nice low inductance, audio grade type like Silver Mica or low inductance Polystyrene.
The actual Circuit remaind in my case completely on the CD-720 (I did this ages ago) Board and used original traces etc.
All that came from that board to the added supply board and transformers where two ribbon PSU cables one to carry DAC, Offset and Op-Amp +/- Supplies from one small EI Transformer, the other carying +/-12V (317/337 regged, classic symmetrical supply, again small EI transformer) to near the supply section where it fed the only the Servo Portion of the player.
Another transformer and regulator supplied the various Digital Decoder & servo IC's, I simply retained the original clcock arrangement de-noised on the supply lines and with a real X-tal instead of the ceramic resonator.
So in total I added 3 pcs small PCB mount transformers and a batch of regulators (8 or 9 pcs, would have to check my notes) on a Veroboard PCB to a CD720 and made a few small changes on the main PCB, plus I fixed the dreadful chassis. The result with a nice external mains conditioner sounded jolly decent, good enough to make a later bought Marant CD-67 SE to sound completely "declasse" in comparison.
Sayonara
Kuei Yang Wang said:
Hi Thorsten,
I was thinking about this carefull implementation. I've been scratching my head for a while, and can't decide what's ok or not.
The PS pins of the clock will be connected directly after the diode bridge, that's sure. There's no other choice.
But where to connect the ground??
-at the transformer's ground
-at the nearest ground from the decoder (where the clock is connected)
-other?
Re:Crying in the Desert
Hi Dave,
I sometimes feel like crying in the desert here but a frog would not survive in such an environment.
Glad you like the OP176 though I am not sure the OP176 is exactly one half of a OP275. Maybe Scott Wurcer can shed light on this as the OP176 is obsolete. Where did you buy it?
I found the OP275 more harmonically right and subjectively having less distortion than the OPA2604. It sounds a bit like musical instruments are better in tune now. Also the OP275 has better and more firm bass.
As for the supply of the KWAK-CLOCK I tried both ways i.e. the raw digital supply of the player and a dedicated separate transformer and bridge rectifier and smoothing caps. The first approach sounded better and this was confirmed by several emails from guys who have built the clock. I second Rudolfs suggestion using both outputs of the comparator.
Where can I find the Chris Fountain mods?
I once "balanced" the raw supply in my Sony by providing equal 4700µF caps but ideally the current draw should also made equal with a ballast resistor.
Dave S said:
Hi Dave,
I sometimes feel like crying in the desert here but a frog would not survive in such an environment.
Glad you like the OP176 though I am not sure the OP176 is exactly one half of a OP275. Maybe Scott Wurcer can shed light on this as the OP176 is obsolete. Where did you buy it?
I found the OP275 more harmonically right and subjectively having less distortion than the OPA2604. It sounds a bit like musical instruments are better in tune now. Also the OP275 has better and more firm bass.
As for the supply of the KWAK-CLOCK I tried both ways i.e. the raw digital supply of the player and a dedicated separate transformer and bridge rectifier and smoothing caps. The first approach sounded better and this was confirmed by several emails from guys who have built the clock. I second Rudolfs suggestion using both outputs of the comparator.
Where can I find the Chris Fountain mods?
I once "balanced" the raw supply in my Sony by providing equal 4700µF caps but ideally the current draw should also made equal with a ballast resistor.
sreten said:
I simply don't care. My bog standard Cambridge Audio
CD4SE is miles better than any modded cheap CD player,
tubes added or not, and as far as I'm concerned thats it.
well, you should hear my modded Cambridge player
It's actually a pretty flimsy cheap player once you start poking around in the guts of it. Amazing that it does sound so good to begin with. Peter
Hi Elso,
I emailed you the CF article - it's no longer available on the web.
I am fairly sure that the OP176 is 1/2 of an OP275 - they have almost identical specs and both use the Butler front end. I got the 176s as samples 2 or 3 years ago.
Regarding balancing of the + and - loads - you are correct that it is not sufficient just to have the same value caps but the current drawn from the supply needs to be equalised by some means.
I do wonder what "higher current dissipation" means, maybe it's related to current woppling
I emailed you the CF article - it's no longer available on the web.
I am fairly sure that the OP176 is 1/2 of an OP275 - they have almost identical specs and both use the Butler front end. I got the 176s as samples 2 or 3 years ago.
Regarding balancing of the + and - loads - you are correct that it is not sufficient just to have the same value caps but the current drawn from the supply needs to be equalised by some means.
I do wonder what "higher current dissipation" means, maybe it's related to current woppling
Konnichiwa,
No CD Player utilising Digital Filters is "bit accurate" on the analogue output. As for your link, this is about the digital output, which on this particular Philips Chipset seems tapped of after a few stages of Digital filtering and volume control, If I gathered it right.
Sayonara
jkeny said:
No CD Player utilising Digital Filters is "bit accurate" on the analogue output. As for your link, this is about the digital output, which on this particular Philips Chipset seems tapped of after a few stages of Digital filtering and volume control, If I gathered it right.
Sayonara
Dave S said:Hi Elso,
I emailed you the CF article - it's no longer available on the web.
I am fairly sure that the OP176 is 1/2 of an OP275 - they have almost identical specs and both use the Butler front end. I got the 176s as samples 2 or 3 years ago.
Regarding balancing of the + and - loads - you are correct that it is not sufficient just to have the same value caps but the current drawn from the supply needs to be equalised by some means.
I do wonder what "higher current dissipation" means, maybe it's related to current woppling
Hi Dave,
I compared the photo micrographs of both chips and they are not exactly similar.
Thanks for the article. I already had seen it on a French forum.
I do not understand what is meant by "higher current dissipation". Did I write that?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.