Over the past months, I have been tweaking my Rotel RCD991 extensively. Doing all the usual stuffs like, SMD caps on the power pins, OSCON caps, inductors on the power lines, passive I/V conversion, chassis damping etc ...
However, when I replace the two 7805s in the CD player with LM1085s, the sound become worst. In summary:
7805: more body, especially in the mid range and overall warmer.
LM1085: More details but sounds a bit harsh, the mid range is quite thin and overall doesn't sound very good.
The output of the regulators were loaded with a Panasonic FA cap (22uF) and a ceramic cap.
So, what's happening here? I am lost.
However, when I replace the two 7805s in the CD player with LM1085s, the sound become worst. In summary:
7805: more body, especially in the mid range and overall warmer.
LM1085: More details but sounds a bit harsh, the mid range is quite thin and overall doesn't sound very good.
The output of the regulators were loaded with a Panasonic FA cap (22uF) and a ceramic cap.
So, what's happening here? I am lost.
Well, DC is DC. If all you see with a scope is DC, that's all you've got. If you're talking about 2 mV of noise on the power rails ( and I assume here you actually mean random, uncorrelated noise), and you've got a circuit with even minimal PSR, I'd sweat the more important stuff and stop trying to convince myself that there's something magical and mysterious going on.
Doing a quick and dirty calculation, let's assume 2 V max output of your CD. Let's assume 40dB PSRR at the top of the audio band(which is pretty mediocre). That means that the noise at your output would be 20 microvolts. That's about 100dB S/N at 20 kHz. As you go lower into the audio range, the PSRR will increase substantially- you'll probably have 100 dB or more PSRR at 1 kHz, even with old opamps. With 160 dB S/N at midband due to power supply noise, you're doing pretty well.
Doing a quick and dirty calculation, let's assume 2 V max output of your CD. Let's assume 40dB PSRR at the top of the audio band(which is pretty mediocre). That means that the noise at your output would be 20 microvolts. That's about 100dB S/N at 20 kHz. As you go lower into the audio range, the PSRR will increase substantially- you'll probably have 100 dB or more PSRR at 1 kHz, even with old opamps. With 160 dB S/N at midband due to power supply noise, you're doing pretty well.
Regulators
Hi Patwen,
I had a similar experience when trying the TDA1543. The Philips DAC did not like the Jung-like regulators I am using for analog supply in my DAC. Audionote is also using these in the Audionote1.1.& 1.2.
When connected to a LT1086-5 (fixed +5V regulator from Linear Technology) sound was fine but when connected to the Jung like regulator sound was "uneasy" not relaxed, fatigueing.
I wonder how the AD1865N-K reacts to this. Something not tried yet. It seems a matter of not harmonizing those two elements i.e. regulator and chip powered from it. I made a post on AA on this issue. Will see if I can find it back.
http://db.audioasylum.com/cgi/m.pl?forum=tweaks&n=32609&highlight=elso+regulator&r=&session=
Well that may be interesting but I meant this post:
http://db.audioasylum.com/cgi/m.pl?forum=tweaks&n=36904&highlight=elso+regulator&r=&session=
Hi Patwen,
I had a similar experience when trying the TDA1543. The Philips DAC did not like the Jung-like regulators I am using for analog supply in my DAC. Audionote is also using these in the Audionote1.1.& 1.2.
When connected to a LT1086-5 (fixed +5V regulator from Linear Technology) sound was fine but when connected to the Jung like regulator sound was "uneasy" not relaxed, fatigueing.
I wonder how the AD1865N-K reacts to this. Something not tried yet. It seems a matter of not harmonizing those two elements i.e. regulator and chip powered from it. I made a post on AA on this issue. Will see if I can find it back.
http://db.audioasylum.com/cgi/m.pl?forum=tweaks&n=32609&highlight=elso+regulator&r=&session=
Well that may be interesting but I meant this post:
http://db.audioasylum.com/cgi/m.pl?forum=tweaks&n=36904&highlight=elso+regulator&r=&session=
You don't always see just DC on the lines. That's why I like to test my regulators with test tones and with music playing. Along, of course, with the usual standard tests of banging the output rails, feeding signals to the output to check dynamic source Z, feeding signals to the input to check dynamic line rejection, etc, etc.
If all you see is stable DC under those conditions, then any perceived differences are illusory. There ain't no such thing as magic. DC is DC.
If all you see is stable DC under those conditions, then any perceived differences are illusory. There ain't no such thing as magic. DC is DC.
2200 uF is too large behind the regulator. What is the purpose of such values ?
I once modified a Sony SACD player by changing the too big caps with smaller ones ( 100 uF ) and it sounded better.
I also remind a Marantz CD 60 SE that had too large caps behind the regulators. The result was when switched on it took almost 15 minutes before the supply was stable. It was "motorboating" ( Low frequency oscillation ). Changed the caps to 100 uF and the problem was solved.
I once modified a Sony SACD player by changing the too big caps with smaller ones ( 100 uF ) and it sounded better.
I also remind a Marantz CD 60 SE that had too large caps behind the regulators. The result was when switched on it took almost 15 minutes before the supply was stable. It was "motorboating" ( Low frequency oscillation ). Changed the caps to 100 uF and the problem was solved.
DC is DC
This is an interesting topic and 60 pages of post could still not exploit all the factors involved in power supply design for audio and digital circuits. I will throw out a few observations and stand back and wait for the backlash.
1. Power supplies have an enormous influence on the sound of circuit and factors all the way to the AC outlet have an influence.
2. PSRR of the regulator (particularly at RF frequencies) has even more of an impact than output impedance on many regulators circuits.
3. Transient response of a reglulator has as much to do with sonics as absolute value of the output impedance. This may explain the preference of simple followers over feedback circuits by many designers. Large amounts of negative feedback are ofter detrimental to audio circuits and there is no reason that the power supply circuits should be free of these conserderations.
4. Most regulated supplies have an inductive output impedence that is parallelled with a capacitor for high frequency stability. The more feedback in the regulator, the lower this inductance is. This makes it even harder to get good stability at high frequencies and makes the particular choice of output capactor critical.
5. At the frequency the the output inductance and output capacitor have the same impedance resonant circuit is formed the freuency and Q of this resonant circuit can be very important. Add the self resonant frequency of the output capacitor and resonances of small value bypass caps to this and you can have quite a mess to considerd in the design.
6. The Jung regulators can be probamatic with digital and D to A circuits because this resonances are more easily excited by the large high frequency currents. The new designs are even more difficult to optimize due to the stabilty compromises made in the quest for low impedance numbers.
7. I hope this will dispell the notion that we are dealling with DC here as a voltage regulator is really a high speed amplifier with a capacitive load and a big DC offset. These factors are also important in three termimal regulators and that devices from different manufactures are audibly different is no suprise. Try the JRC regulators by the way. Also experiment with many different values and types of output and input bypass caps.
Just a regulator guy,
Fred
This is an interesting topic and 60 pages of post could still not exploit all the factors involved in power supply design for audio and digital circuits. I will throw out a few observations and stand back and wait for the backlash.
1. Power supplies have an enormous influence on the sound of circuit and factors all the way to the AC outlet have an influence.
2. PSRR of the regulator (particularly at RF frequencies) has even more of an impact than output impedance on many regulators circuits.
3. Transient response of a reglulator has as much to do with sonics as absolute value of the output impedance. This may explain the preference of simple followers over feedback circuits by many designers. Large amounts of negative feedback are ofter detrimental to audio circuits and there is no reason that the power supply circuits should be free of these conserderations.
4. Most regulated supplies have an inductive output impedence that is parallelled with a capacitor for high frequency stability. The more feedback in the regulator, the lower this inductance is. This makes it even harder to get good stability at high frequencies and makes the particular choice of output capactor critical.
5. At the frequency the the output inductance and output capacitor have the same impedance resonant circuit is formed the freuency and Q of this resonant circuit can be very important. Add the self resonant frequency of the output capacitor and resonances of small value bypass caps to this and you can have quite a mess to considerd in the design.
6. The Jung regulators can be probamatic with digital and D to A circuits because this resonances are more easily excited by the large high frequency currents. The new designs are even more difficult to optimize due to the stabilty compromises made in the quest for low impedance numbers.
7. I hope this will dispell the notion that we are dealling with DC here as a voltage regulator is really a high speed amplifier with a capacitive load and a big DC offset. These factors are also important in three termimal regulators and that devices from different manufactures are audibly different is no suprise. Try the JRC regulators by the way. Also experiment with many different values and types of output and input bypass caps.
Just a regulator guy,
Fred
jean-paul said:
Large caps after the regualtor is OK if you have control over the regulation, no motorboating etc. You must have in mind though that it's not suitable for non-constant loads.
Some regualtor can't take too heavy capacitive loads, consult the datasheets, they can tell you a lot.
Printed out the 7 points and taped them on the wall
Hello Per,
We are talking about the 78xx series here and 2200 uF is overkill.
2200 uF is more likely to be used before the regulator.
Please see the datasheets for proof/ evidence or whatever it takes to believe it. In cdplayers the load is mostly non constant. I modified tens of cd players by changing the 1000 or 2200 uF caps behind 78xx series regulators and they all had better regulation afterwards. Especially the old Pioneer players have these outrageous values. Another thing is that in a lot of players the output cap is placed too far from the regulator. I place them directly on the pins of the 78xx. Example : in the very good Marantz CD80 caps are about 10 cm away from the regulator.
Change that and the player will have better regulation and less garbage on the supply lines. Check the 5 V supply to see what I mean.
The Marantz CD 60 SE is a good example. K.I. replaced all the original caps with the same value Cerafine ( 220 uF ). In this case 220 uF was too much. Sound was going from left to right in volume. I placed 100 or 10 uF and the problems were over and sound was better too.
Controversial thing: in digital OSCON behind the regulator is very good. In analog they sound worse than normal electrolytics.
Does somebody experience the same with OSCON ?
Hello Per,
We are talking about the 78xx series here and 2200 uF is overkill.
2200 uF is more likely to be used before the regulator.
Please see the datasheets for proof/ evidence or whatever it takes to believe it. In cdplayers the load is mostly non constant. I modified tens of cd players by changing the 1000 or 2200 uF caps behind 78xx series regulators and they all had better regulation afterwards. Especially the old Pioneer players have these outrageous values. Another thing is that in a lot of players the output cap is placed too far from the regulator. I place them directly on the pins of the 78xx. Example : in the very good Marantz CD80 caps are about 10 cm away from the regulator.
Change that and the player will have better regulation and less garbage on the supply lines. Check the 5 V supply to see what I mean.
The Marantz CD 60 SE is a good example. K.I. replaced all the original caps with the same value Cerafine ( 220 uF ). In this case 220 uF was too much. Sound was going from left to right in volume. I placed 100 or 10 uF and the problems were over and sound was better too.
Controversial thing: in digital OSCON behind the regulator is very good. In analog they sound worse than normal electrolytics.
Does somebody experience the same with OSCON ?
Regulators and Compensation
Hi All,
Very well said Fred!
As a illustration I found the following application note while cleaning up a pile of papers:
http://www.onsemi.com/pub/Collateral/SR003AN-D.PDF
(Compensation for Linear Regulators)
Hi All,
Very well said Fred!
As a illustration I found the following application note while cleaning up a pile of papers:
http://www.onsemi.com/pub/Collateral/SR003AN-D.PDF
(Compensation for Linear Regulators)
Fred, a couple of naive questions.
*Would you agree that the job of a power supply is to provide stable DC under all operationg conditions of the powered circuit?
*And that the PS is doing its job if nothing but stable DC appears on the power rails when the powered device is in operation (whether playing music or test tones)?
TIA
*Would you agree that the job of a power supply is to provide stable DC under all operationg conditions of the powered circuit?
*And that the PS is doing its job if nothing but stable DC appears on the power rails when the powered device is in operation (whether playing music or test tones)?
TIA
the job of a power supply
Yep.... That's it's job all right. The problem is trying to do it in the presence of input signals from AC line variations, RF interference, voltages from the rectification of AC voltage to DC, and signals from the signal currents through the impedances on the input side of the regulator. At the same time the regulator is trying to maintain a very small AC error signal in the presence of the current demands of the circuit that it is regulating. Throw in the fact that the voltage reference for the regulator is being affected by all these factors as well as it's own zener or bandgap noise voltage contributions.
It sounded like a cushy job in the classifieds but it sounds like really having to work hard for a living to me. Power supplies are as difficult to get right as the audio circuit they are regulating. PSRR decreases with increasing frequency for feedback circuits and is zero for many no feedback circuits.
I'm not sure I answered your question, but I am not quite sure of what the point is that you are trying to make.
Fred
http://www.diyvideo.com/forums/showthread.php?s=&threadid=5336&highlight=voltage+regulators
Yep.... That's it's job all right. The problem is trying to do it in the presence of input signals from AC line variations, RF interference, voltages from the rectification of AC voltage to DC, and signals from the signal currents through the impedances on the input side of the regulator. At the same time the regulator is trying to maintain a very small AC error signal in the presence of the current demands of the circuit that it is regulating. Throw in the fact that the voltage reference for the regulator is being affected by all these factors as well as it's own zener or bandgap noise voltage contributions.
It sounded like a cushy job in the classifieds but it sounds like really having to work hard for a living to me. Power supplies are as difficult to get right as the audio circuit they are regulating. PSRR decreases with increasing frequency for feedback circuits and is zero for many no feedback circuits.
I'm not sure I answered your question, but I am not quite sure of what the point is that you are trying to make.
Fred
http://www.diyvideo.com/forums/showthread.php?s=&threadid=5336&highlight=voltage+regulators
No real point. Your post with a whole laundry list confused me a bit- it struck me that what a PS should do is actually rather simple and that once you have DC, it's DC.
Oh, yes, and that clearly the original questioner's power supply WAS doing its job, and if he was "hearing" differences in the absence of anything significant showing up on the rails with either regulator, the problem was more likely to be psychological than electrical.
Oh, yes, and that clearly the original questioner's power supply WAS doing its job, and if he was "hearing" differences in the absence of anything significant showing up on the rails with either regulator, the problem was more likely to be psychological than electrical.
DC is DC .. Not
Hi SY,
yes, I did see flat line on my "20MHz" scope, but I can sure you that life above 20MHz is much more "interesting".
And if I have a spectrum analyzer. wow, I can sure that power rail is not steady as rock, instead I expect to see firework displayed to them.
So DC is NOT DC... and power supply does make a huge difference.
Actually, we are not "amplifying" a signal, instead we are using the power on the power supply to "replicate" the input signal and get an output hopefully similar to the input one.
Good water make good wine, bad water make bad wine. It is that simple ... (of course not, but the basic is the same)
Hi SY,
yes, I did see flat line on my "20MHz" scope, but I can sure you that life above 20MHz is much more "interesting".
And if I have a spectrum analyzer. wow, I can sure that power rail is not steady as rock, instead I expect to see firework displayed to them.
So DC is NOT DC... and power supply does make a huge difference.
Actually, we are not "amplifying" a signal, instead we are using the power on the power supply to "replicate" the input signal and get an output hopefully similar to the input one.
Good water make good wine, bad water make bad wine. It is that simple ... (of course not, but the basic is the same)
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.