Torsten in reply to your comments about the pre on my site..
Lol at the pots again.. The noise is about -110dB down. It's not a layout issue, and given the ppm distortion performance it's definitelynot..
Secondly, as I state clearly, the volume pot was st to the mechanical mid point, and the inputs not shorted which often how these things are measured.
No need to replace the opamps I used. The 4562 is a fine opamp, as are some of the ADI offerings. I am sure both would measure up extremely well and on a DBT you would be had pressed to tell which was which.
Lol at the pots again.. The noise is about -110dB down. It's not a layout issue, and given the ppm distortion performance it's definitelynot..
Secondly, as I state clearly, the volume pot was st to the mechanical mid point, and the inputs not shorted which often how these things are measured.
No need to replace the opamps I used. The 4562 is a fine opamp, as are some of the ADI offerings. I am sure both would measure up extremely well and on a DBT you would be had pressed to tell which was which.
It seems they are just a brand owned by Harmon now? The point is it gets class A in Stereophile, the same thing you guys harp on constantly, with an op-amp front end.
Torsten
Non linear wiper behavior?
Look at the distortion plots. I think they will confirm this is not an issue. But, you have stated on this forum that ultra low distortion is no big deal anyway.
Crappy cables? Yes, most likely where the noise came in.
Noise from the PSU? Unlikely, but then again it is >100db down
Non linear wiper behavior?
Look at the distortion plots. I think they will confirm this is not an issue. But, you have stated on this forum that ultra low distortion is no big deal anyway.
Crappy cables? Yes, most likely where the noise came in.
Noise from the PSU? Unlikely, but then again it is >100db down
Last edited:
Hi,
Hmm, I would not be so sure, but suit yourself. So what is the problem then? Bad cables?
I think you may wish to evaluate this nevertheless. It may be worthwhile to understand the non-ideal properties of components. Once you do, you may find that using a J-Fet input device is preferable in the position where you have a placed a bipolar Op-Amp.
I have no beef that for an audio Op-Amp the LM4562 is not bad. But it has a bipolar input stage, which has inescapable consequences in your application.
Just as using looped feedback systems with low open loop bandwidth has consequences for the CMRR and PSRR, which IMNSHO are visible in your tests (and underscore the need for tests and the ability to interpret them - just measuring low THD is meaningless)...
But as said, as long as you like the results, that's all that matters.
Ciao T
Hmm, I would not be so sure, but suit yourself. So what is the problem then? Bad cables?
I think you may wish to evaluate this nevertheless. It may be worthwhile to understand the non-ideal properties of components. Once you do, you may find that using a J-Fet input device is preferable in the position where you have a placed a bipolar Op-Amp.
I have no beef that for an audio Op-Amp the LM4562 is not bad. But it has a bipolar input stage, which has inescapable consequences in your application.
Just as using looped feedback systems with low open loop bandwidth has consequences for the CMRR and PSRR, which IMNSHO are visible in your tests (and underscore the need for tests and the ability to interpret them - just measuring low THD is meaningless)...
But as said, as long as you like the results, that's all that matters.
Ciao T
Hi,
So, where are the 30 - 40dB excess distortion in your measurements (compared to what teh LM4562 should exhibit) coming from?
It is not, however different sources of distortion give rise to different audibility.
OH NO. You meant to tell me CABLES MAKE A DIFFERENCE!?
Well, the noise is 50Hz, your local mains frequency and you do use an unshielded torroid, with mechanical arrangements that will give significant coupling to the magnetic fieldlines it emits (basically, bad layout).
Ciao T
So, where are the 30 - 40dB excess distortion in your measurements (compared to what teh LM4562 should exhibit) coming from?
It is not, however different sources of distortion give rise to different audibility.
OH NO. You meant to tell me CABLES MAKE A DIFFERENCE!?
Well, the noise is 50Hz, your local mains frequency and you do use an unshielded torroid, with mechanical arrangements that will give significant coupling to the magnetic fieldlines it emits (basically, bad layout).
Ciao T
Last edited:
Maybe this is interesting for the high open loop bandwidth camp:
Website of Wayne Stegall - Musical Feedback Amplifiers
Website of Wayne Stegall - Musical Feedback Amplifiers
Torsten
At 2v out this simple pre is delivering single digit ppm distortion into a 600 ohm load as measured in what was not an ideal environment.
Cables. Yes I could probably have used ones with better shielding for the tests. That is what I meant.
You keep raising the mains frequency. With volume control at mid point mechanically is at -110db. At volume max it's at -100db. Some of that may be coming in internally through some coupling into the cables between the pot and the board, but I doubt all of the 10dB delta is due to that.
As for the noise, well it goes up as the pot is turned towards max, and you can see that from the plots. A sure sign it was coming in from the outside environment.
No doubt a FET input opamp would /could sound very good. But, the point is would the 4562 sound 'crap' next to it? Of course not.
At 2v out this simple pre is delivering single digit ppm distortion into a 600 ohm load as measured in what was not an ideal environment.
Cables. Yes I could probably have used ones with better shielding for the tests. That is what I meant.
You keep raising the mains frequency. With volume control at mid point mechanically is at -110db. At volume max it's at -100db. Some of that may be coming in internally through some coupling into the cables between the pot and the board, but I doubt all of the 10dB delta is due to that.
As for the noise, well it goes up as the pot is turned towards max, and you can see that from the plots. A sure sign it was coming in from the outside environment.
No doubt a FET input opamp would /could sound very good. But, the point is would the 4562 sound 'crap' next to it? Of course not.
Last edited:
There is something 'wrong' here. We are not trying to make 'mid-fi' and that is exactly what Bonsai has built. There is nothing wrong with it, except that it will not perform as well as the best hi fi efforts. Also, you could buy its equivalent over the internet or a hi fi component store, and save yourself a lot of time and trouble, sourcing the parts and making an acceptable product. Kind of reminds me of a Chevy 6 cylinder pickup, and saying it is all that anyone needs.
John
I made no claims about by effort other than those supported by the AP measurements. I happen to like designing, building and then writing about this stuff. And yes, the X-Altra mini was never pitched against $10k professionally built pre-amps. It's a $100 buck system. That said, thanks to the opamp, it does a fine job.
I made no claims about by effort other than those supported by the AP measurements. I happen to like designing, building and then writing about this stuff. And yes, the X-Altra mini was never pitched against $10k professionally built pre-amps. It's a $100 buck system. That said, thanks to the opamp, it does a fine job.
Performing like a sample 'hotdog' in my parable. You can make something, yet it will not be received with much interest, and fast food outlets might be more economical, than a small individual business effort who thinks all 'hotdogs' are the same.
To argue the 'taste' of the public is pointless. Some know the difference, some want to spend the minimum amount, and others are indifferent.
To argue the 'taste' of the public is pointless. Some know the difference, some want to spend the minimum amount, and others are indifferent.
Open loop Scott! That makes it non-linear enough to get class A status!
jan didden
This was meant as a question for everybody. I´m looking for some advice on these silicon pots, can´t try them all unfortunately (TI, cirrus, etc.)
I realize this will not be as good as a stepped attenuator.
-Alex
Hi,
Correctly applied these can be quite good. There are a few things to account for.
1) They have the same problems with non-linear wiper impedance as normal pots, except 100 times or so worse.
2) They use power supplies to bias the switch circuitry, the isolation between power supplies and switch is not infinite, but usually not specified.
3) They have clear and present level limits set by the supplies. Get these wrong and overload is NASTY (and I mean NASTY), so input voltage handling is severely limited.
4) If you cannot guarantee no ejit will be trying to stuff 50V DAC into the circuit or some such, you need series resistors and clamping diodes for protection. These do not improve measured or subjective performance.
5) Some have internal Op-Amp's, in my experience these op-amp's generally do not improve measured or subjective performance.
Ciao T
Correctly applied these can be quite good. There are a few things to account for.
1) They have the same problems with non-linear wiper impedance as normal pots, except 100 times or so worse.
2) They use power supplies to bias the switch circuitry, the isolation between power supplies and switch is not infinite, but usually not specified.
3) They have clear and present level limits set by the supplies. Get these wrong and overload is NASTY (and I mean NASTY), so input voltage handling is severely limited.
4) If you cannot guarantee no ejit will be trying to stuff 50V DAC into the circuit or some such, you need series resistors and clamping diodes for protection. These do not improve measured or subjective performance.
5) Some have internal Op-Amp's, in my experience these op-amp's generally do not improve measured or subjective performance.
Ciao T
- Status
- Not open for further replies.