Hi Mooly!
Interesting experimental result. It leads to the conclusion, that two LM833's discoursed versions have not only different output stages, but also different input stages too, that is, they are completely different!
Well, in this case, it seems, the last word has to say their sound quality, which, as we know, not always depend on how perfect technical characteristics are.
With respect and best wishes
Interesting experimental result. It leads to the conclusion, that two LM833's discoursed versions have not only different output stages, but also different input stages too, that is, they are completely different!
Well, in this case, it seems, the last word has to say their sound quality, which, as we know, not always depend on how perfect technical characteristics are.
With respect and best wishes
Thanks. I'm very pleased with the subjective performance in my Sony MiniDisc recorder... who knows... MiniDisc is an 'imperfect' source, perhaps the characteristics (quasi output) of these new 833's lend themselves to giving a really satisfying subjective result.
I'm hoping to test these a bit more, maybe with a new listening test for you all. So watch this space !
I'm hoping to test these a bit more, maybe with a new listening test for you all. So watch this space !
I cannot agree with using 6800uF and you don't have noise. Noise is mostly from the input LTP that depends on input resistance and current......yada yada.......We all know that!!! That has nothing to do with power supply filtering.
You need to filter the power supply, BUT if you look at any data sheet, power supply rejection ratio is very very high at audio frequency particular at bass frequency. I cannot see any advantage to anything over 100uF. You worry more, put a small resistor in series with the power rail before the cap. These are low power devices, you can use quite high resistance without worrying too much. So if it sag slightly, so what, there is so much power supply rejection ratio it doesn't matter. I think that best thing one can do for lowering distortion is to use a pull up or pull down resistor that draw enough current from the output and keep the output stage in class A 100% of the time to avoil cross over distortion all together. Save the money on the caps and pay a little more for a better opamp like the LM4562.
There are enough sneak oil around already.
Last edited:
+1 to Alan0354
I came onto this late, was then considering a comment but found DF96, JCX and others covered my points pretty well. Now, at the end, Alan 0354 summed all nicely.
One thing hopefully learnt is that some designers apply r.f. techniques to a.f. and 'intuitively' come to the wrong conclusions. Where is the data when praising the (naturally ...) improved bass, 'speed', this or that before glibly making claims? 'Better, worse, improved' and such depends on subjective judgment - and not diminishing such (not re-opening that debate here!). But as Doug Self said, in audio it pays to be sceptical.
Threads like this should repeatedly appear if only to make us conscious again of what the basics are. The OP needs to be thanked for that. He should keep on asking questions. It is not stupid, it is the right way to learn.
I came onto this late, was then considering a comment but found DF96, JCX and others covered my points pretty well. Now, at the end, Alan 0354 summed all nicely.
One thing hopefully learnt is that some designers apply r.f. techniques to a.f. and 'intuitively' come to the wrong conclusions. Where is the data when praising the (naturally ...) improved bass, 'speed', this or that before glibly making claims? 'Better, worse, improved' and such depends on subjective judgment - and not diminishing such (not re-opening that debate here!). But as Doug Self said, in audio it pays to be sceptical.
Threads like this should repeatedly appear if only to make us conscious again of what the basics are. The OP needs to be thanked for that. He should keep on asking questions. It is not stupid, it is the right way to learn.
Alan is completely right.
What many people don't understand is that a cap without an impedance in front of it does not work as a filter. Of course, there is no such thing as no impedance in a real wire, so a large capacitor will always filter somewhat. But it is much more economical and effective to put 20 Ohms or so in front of a smaller cap. Philips used to do this all over their products.
Where the pssr of the active device slopes downward with frequency, the filtering effect of the RC combination becomes larger, ensuring a combined large PSSR across the entire frequency band.
What many people don't understand is that a cap without an impedance in front of it does not work as a filter. Of course, there is no such thing as no impedance in a real wire, so a large capacitor will always filter somewhat. But it is much more economical and effective to put 20 Ohms or so in front of a smaller cap. Philips used to do this all over their products.
Where the pssr of the active device slopes downward with frequency, the filtering effect of the RC combination becomes larger, ensuring a combined large PSSR across the entire frequency band.
Yes,
using an rC style PSU works but can show up limitations in HF PSRR.
converting to rCRC or rC(L+R)C helps significantly with HF PSRR.
using rCCCCCC costs a lot for little improvement at HF.
Yes, from the point of view of classical radio techniques these approaches are quite right for choosing bypassing caps for Op Amps.
Nevertheless, there are confirmed experimental evidences of improving the sound due to increasing the bypassing caps.
In the acknowledged Hi-End Amp ASR Emitter these caps more then two orders are larger, then are recommended here (85680..513680mkf vs ..6800mkf).
http://www.asraudio.de/Epdf/IEEmitter1.pdf
One man, who lifted up it's lid and looked inside, said, that PS of the AD843, the "heart" of the Amp, is based on chip voltage regulators. Ever since, there is a great question for me, how did they distribute that giant cap's store on parts before and after voltage regulators?🙂
Nevertheless, there are confirmed experimental evidences of improving the sound due to increasing the bypassing caps.
In the acknowledged Hi-End Amp ASR Emitter these caps more then two orders are larger, then are recommended here (85680..513680mkf vs ..6800mkf).
http://www.asraudio.de/Epdf/IEEmitter1.pdf
One man, who lifted up it's lid and looked inside, said, that PS of the AD843, the "heart" of the Amp, is based on chip voltage regulators. Ever since, there is a great question for me, how did they distribute that giant cap's store on parts before and after voltage regulators?🙂
You sure it's not sneak oil? Sure those people are not from the tube world? There is some truth that different tubes have different sound as they are really mechanical devices also.
Worry about ripple on the supply rails, after using 22ohm and 100uF, use a scope set in AC, crank to 2mV/div and to look at the rail. If there is no ripple, it's sneak oil.
Pay more attention to make sure you have bypass caps point to point at the supply pins of the opamp, have a good ground plane, make sure you have point to point connection at the -ve input pin, that's the most most sensitive pin of the opamp. If you drive coax, make sure you have a 200ohm( at least 50 to 100) between the output and the coax to avoid excess capacitance that kill the phase margin. Use a pull down resistor from output to -ve rail to keep the output in Class A.
All the talks, more than a few times when I looked at the layout of opamps on pcb, I laughed. It's the layout!!!! All the talks is going to be ruined by a bad layout. Then they turn around and blame on everything else.
Worry about ripple on the supply rails, after using 22ohm and 100uF, use a scope set in AC, crank to 2mV/div and to look at the rail. If there is no ripple, it's sneak oil.
Pay more attention to make sure you have bypass caps point to point at the supply pins of the opamp, have a good ground plane, make sure you have point to point connection at the -ve input pin, that's the most most sensitive pin of the opamp. If you drive coax, make sure you have a 200ohm( at least 50 to 100) between the output and the coax to avoid excess capacitance that kill the phase margin. Use a pull down resistor from output to -ve rail to keep the output in Class A.
All the talks, more than a few times when I looked at the layout of opamps on pcb, I laughed. It's the layout!!!! All the talks is going to be ruined by a bad layout. Then they turn around and blame on everything else.
Last edited:
Some of you might find this opening post interesting... more to follow 😉
http://www.diyaudio.com/forums/ever...rstand-what-we-think-we-hear.html#post4395937
http://www.diyaudio.com/forums/ever...rstand-what-we-think-we-hear.html#post4395937
Hi ! very interesting point.
May i ask which kind of measurement gives evidence of a bad layout ?
Very interesting point indeed.
I am trying to learn something.
It would be very important to me at this stage to conclude that op-amps can be completely fine ...
maybe the way in which they are used is the real issue/challenge.
Thanks and regards, gino
Last edited:
Can I just jump in on that...
Simple measurements you can easily do.
Crosstalk. It should ideally be equal between boh channels i.e. if you have -70db L to R and 50db R to L at say 10kHz then that could be an audible issue as it indicates the channel separation is different with frequency. That could lead to an unstable and 'wandering' image.
Load each channel individually to its max output current (same test applies to power amps as well as opamp and small signal stuff but in the case of power amps load to a sensible level, say 10wrms equivalent across 8 ohm) and with a shorting plug in the unused channels input, check that no output occurs in the unused channel as you attach and remove the load on the other channel. That can be a good guide to ground issues as the unused channel should remain at zero output whether the other channel is loaded or not. Use 400 or 1kHz as a test frequency for that.
Simple measurements you can easily do.
Crosstalk. It should ideally be equal between boh channels i.e. if you have -70db L to R and 50db R to L at say 10kHz then that could be an audible issue as it indicates the channel separation is different with frequency. That could lead to an unstable and 'wandering' image.
Load each channel individually to its max output current (same test applies to power amps as well as opamp and small signal stuff but in the case of power amps load to a sensible level, say 10wrms equivalent across 8 ohm) and with a shorting plug in the unused channels input, check that no output occurs in the unused channel as you attach and remove the load on the other channel. That can be a good guide to ground issues as the unused channel should remain at zero output whether the other channel is loaded or not. Use 400 or 1kHz as a test frequency for that.
Hi thanks a lot for the very valuable advice.
Usually if i connect more than one source to the amp i always get some leaking from a not selected channel to the playing one 🙁
I hear the music in the background
For this reason i switch off all the sources that i do not use.
I think that someone even use shorting plugs for the unused inputs.
Thanks again, gino
Usually if i connect more than one source to the amp i always get some leaking from a not selected channel to the playing one 🙁
I hear the music in the background
For this reason i switch off all the sources that i do not use.
I think that someone even use shorting plugs for the unused inputs.
An externally hosted image should be here but it was not working when we last tested it.
Thanks again, gino
Leakage from other inputs is another issue and typically occurs when a simple mechanical switch is used. That's a design aspect that needs addressing right in the early design phase.
Hi and thanks for the very helfpul advice.
Yes it has a mechanical switch actually.
I have to add that using mainly just one source (audio PC) the issue is less important.
I wonder if shorting the other unused inputs can provide some real benefits.
And in case which is the best way to short a rca input.
Yes it has a mechanical switch actually.
I have to add that using mainly just one source (audio PC) the issue is less important.
I wonder if shorting the other unused inputs can provide some real benefits.
And in case which is the best way to short a rca input.
- Status
- Not open for further replies.