Here you go Ricardo
Brent
An externally hosted image should be here but it was not working when we last tested it.
Brent
Thanks Ricardo
My opinion.....
It's not the same as ours or Audiocom reg (you can see that in the picture). It's more like a Raygulator but should be better has it does not (I think) use a LM type device. A circuit diagram of it would allow for a full assessment.
The v adjust seems a good idea, this is something we thought about but I was worried if a faulty pot went open circuit and made the reg output full voltage into customers delicate DACs etc so I stayed away from that idea.
If it's $16 then it should be good value + it is nice and neat.
Brent
My opinion.....
It's not the same as ours or Audiocom reg (you can see that in the picture). It's more like a Raygulator but should be better has it does not (I think) use a LM type device. A circuit diagram of it would allow for a full assessment.
The v adjust seems a good idea, this is something we thought about but I was worried if a faulty pot went open circuit and made the reg output full voltage into customers delicate DACs etc so I stayed away from that idea.
If it's $16 then it should be good value + it is nice and neat.
Brent
Today I received some silver solid core wire.... I did not know it was naked wire (looking at the photo in the site it looked covered with PTFE).
Can I use thermo sleeve or is it an issue ?
I am preparing the DOS and want to use the wire for PWM input and analog out.
Can I use 0.4mm diameter on the input or should I use 0.8mm ?
Ricardo
Can I use thermo sleeve or is it an issue ?
I am preparing the DOS and want to use the wire for PWM input and analog out.
Can I use 0.4mm diameter on the input or should I use 0.8mm ?
Ricardo
6h5c said:
You haven't really read what I wrote? The PWM analog supply is not being fed off the +12V, it's shunt regulator is being fed of +12V analog, and the shunt regulator is set to 5.05V actually.
In the CD43/53/63/1010/1020 feeding the DAC off the 12V analog supply (with appropriate local regulation, of course), is not the best solution (if you were building a SM5872 based stand-alone DAC you would certainly do it differently), but it is the best possible solution, save for building a new power supply. It is certainly far more 'correct' than feeding it unfiltered digital +5V the way it is done originally, or deriving it's supply from unregulated servo supplies. The simple fact is, the +12V analog output supply is regulated (so, pre-regulated if there is local regulation for the DAC) and while it is far from noiseless, it has the least noise of all the available supplies. So, it makes it a good candidate for PREREGULATION. Never attempt to feed the DAC with 12V, it will fry - the datasheet clearly states the limits.
So, on the basis that the clock and PWM sections should connect to the same ground, it should be the digital ground under the DAC.
This would mean disconnecting the PWM grounds from the analogue grounds (starred to the o/p socket) and connecting them to the same digital ground under the DAC that the clock and digital sections use.
Presumably this is where ideally the analogue and digital grounds should connect together, but I don't think this is possible without creating a ground loop.
I hope I've understood this situation...
This would mean disconnecting the PWM grounds from the analogue grounds (starred to the o/p socket) and connecting them to the same digital ground under the DAC that the clock and digital sections use.
Presumably this is where ideally the analogue and digital grounds should connect together, but I don't think this is possible without creating a ground loop.
I hope I've understood this situation...
6h5c said:I think it's obvious the DAC is not fed with +12V...
Still, I wouldn't connect the DAC's PWM section to the analog PSU, not even with a shunt regulator.
Just my
Ray
Well, yes, great - now if only you would give a reason for your opinion?
Glenn2 said:So, on the basis that the clock and PWM sections should connect to the same ground, it should be the digital ground under the DAC.
This would mean disconnecting the PWM grounds from the analogue grounds (starred to the o/p socket) and connecting them to the same digital ground under the DAC that the clock and digital sections use.
Presumably this is where ideally the analogue and digital grounds should connect together, but I don't think this is possible without creating a ground loop.
I hope I've understood this situation...
A ground loop will be created if the return ground to the analog regulators (+-12V) is not disconnected from the common ground star on the +10V smoothing capacitor, where it actually goes by 3 parallel runs. This is one of the routing errors on the board.
Yes, normally the grounds (clock, analog, digital) would all meet under the DAC in a single point, while the analog, digital and clock power supplies all have completely independent (floating) grounds. This would prevent any ground loops.
However, as the PCB is done such that it is a copper area rather than a point, and there are multiple ground returns because the analog and digital transformer windings share a common ground pin, the only thing you can do is try to break the superfluous loops, attempting to reduce the area of the ground to the copper plane under the DAC. As you can't completely break apart the grounds (due to the aforementioned transformer issue - of course, you could use a different trafo and solve this in it's root), you have to use the regulators to break the ground loops. I know this sounds odd, but what you actually do, is try to enclose the power supply current loops as close to only where they must go, by either local filtering, or regulation - in the case of regulation, you need to move the regulator ground reference points so that the regulators 'regulate out' ground bounce voltages due to ground loop currents, and/or the filters only alow a DC current component (or very low frequency) through the ground loops, while all the signal (AC) component stays in a loop locally to the DAC or output or whatever actual part we are talking about.
As I said, this is really not that difficult if you just take the PCB image from the service manual, and actually DRAW where the current loops of the various sections go. Sometimes it simply is not possible to understand exactly what is going on unless you do this.
6h5c said:
OK, this would be what is usually called a 'knee-jerk reaction'.
You are aware that the PWMs high frequency noise is the result of it's carrier frequency plus the noise shaping, and it goes directly into the analog stage as an input? Well then, you will have to explain to me how the PWM can pollute te analog supply 'in reverse' going past a heavily bypassed shunt regulator, and through a very high impedance current source, into an equally heavily bypassed output of the analog regulator, which BTW is separately and additionally filtered by local RC filters for each place where it is used. All in all this amounts to well over 120dB of attenuation of this noise, WORST CASE, and that is not counting the output stage PSRR. Are you certain your output filtering does this job even close to these figures, preventing the same noise going into the output stage as a signal (so no immunity to it at all)?
If my explanations are not enough, please DO draw the relevant current loops and impedances and see for yourself, or if in doubt, just measure.
I would not call this a knee-jerk reaction, on the contrary, the suggestion could almost be taken as an insult.
There are lots of years of experience regarding this subject in this thread, and that shouldn't be dismissed that easily.
If my output filtering can't reach these nice numbers, why should your shunt do better? Your theoretical approach is correct, but we all know the numbers do not make the sound.
Regards,
Ray
There are lots of years of experience regarding this subject in this thread, and that shouldn't be dismissed that easily.
If my output filtering can't reach these nice numbers, why should your shunt do better? Your theoretical approach is correct, but we all know the numbers do not make the sound.
Regards,
Ray
I have a question about regulating the clock part of the DAC.
Presumably the regulator (e.g., 7805, LM317) can do absolutely nothing to regulate the current drawn by a 16.9MHz oscillator, as its error amplifier has no gain at these frequencies. So if we let it take care of line regulation and accept that load regulation of this circuit is not within its capabilities, is it beneficial to RC filter the output of the regulator to further filter out line noise and regulator self noise?
I see that the stock player has 4.7R+220uF at the clock which is a corner frequency of about 150Hz, so it starts to bite roughly where the 7805 ripple rejection starts to decline.
So even when giving the clock its own regulator, is it better to leave the series resistor in to form a low pass filter with the bypass cap?
Presumably this would also dampen out any ringing caused by the output inductance of the regulator and the IC's bypass cap, which should be low-ESR types.
Presumably the regulator (e.g., 7805, LM317) can do absolutely nothing to regulate the current drawn by a 16.9MHz oscillator, as its error amplifier has no gain at these frequencies. So if we let it take care of line regulation and accept that load regulation of this circuit is not within its capabilities, is it beneficial to RC filter the output of the regulator to further filter out line noise and regulator self noise?
I see that the stock player has 4.7R+220uF at the clock which is a corner frequency of about 150Hz, so it starts to bite roughly where the 7805 ripple rejection starts to decline.
So even when giving the clock its own regulator, is it better to leave the series resistor in to form a low pass filter with the bypass cap?
Presumably this would also dampen out any ringing caused by the output inductance of the regulator and the IC's bypass cap, which should be low-ESR types.
You need one heck of a regulator to be able to handle such high frequencies. So even if you have one with a decent bandwidth (let's say, 1MHz?), some form of passive filtering is a good idea, to keep the HF noise from entering the regulator. I think a ferrite with low inductance capacitor will do a good job here.
Regards,
Ray
Regards,
Ray
Hi Ray,
Yes, I was thinking that normally following a reg with an RC filter would kill the load regulation (as load current developes a voltage across the series resistor) but at these frequencies there is no load regulation so the pros may outweigh the cons.
(I.e. the RC filter after the reg)
Yes, I was thinking that normally following a reg with an RC filter would kill the load regulation (as load current developes a voltage across the series resistor) but at these frequencies there is no load regulation so the pros may outweigh the cons.
(I.e. the RC filter after the reg)