well, this seems to be the simplist, most functional circuit so far..
simple question: if i were to build this using an OPA4134 would i be wasting my time?..
http://www.hagtech.com/images/bugleschem.gif
simple question: if i were to build this using an OPA4134 would i be wasting my time?..
http://www.hagtech.com/images/bugleschem.gif
RIAA...
One day, I had a count on all the "small" trafos I had lying around, and they were too small for PowerAmps...
So I started looking for RIAA & DAC-kits, and modified a few CD-Players....
No, really, when I got the Oracle Delphi in house, I tested some different cartridges, and some of them did not relate to the "advertised" sound-quality, so I suspected the Black Cube for beeing less than what I wanted...and I wanted to keep my hobby up.
Arne K
One day, I had a count on all the "small" trafos I had lying around, and they were too small for PowerAmps...
So I started looking for RIAA & DAC-kits, and modified a few CD-Players....
No, really, when I got the Oracle Delphi in house, I tested some different cartridges, and some of them did not relate to the "advertised" sound-quality, so I suspected the Black Cube for beeing less than what I wanted...and I wanted to keep my hobby up.
Arne K
Discrete input stage. Followed by OPamp filters/Output.
As mentioned earlier,
a low noise well matched discrete, transistor, input stage is a good solution.
-Either build a lownoise discrete OP, as first stage
or
-Use lownoise transistors to feed the input of first OP amp.
This is what
I would would use,
and this is a common solution for high quality RIAA amplifiers/MIC amplifiers.
The 0.1 dB accuracy of the filters is to overdo things.
A deviation in the region 1-2dB is well enough.
But use very high quality resistors and capacitors for setting the filter values.
As mentioned earlier,
a low noise well matched discrete, transistor, input stage is a good solution.
-Either build a lownoise discrete OP, as first stage
or
-Use lownoise transistors to feed the input of first OP amp.
This is what
I would would use,and this is a common solution for high quality RIAA amplifiers/MIC amplifiers.
The 0.1 dB accuracy of the filters is to overdo things.
A deviation in the region 1-2dB is well enough.
But use very high quality resistors and capacitors for setting the filter values.
Re: that is not my intention
The closer you are to the source the more important the little details are. As i understand it (i have not built my own RIAA) channel-to-channel accuarcy is more important than absoulutely matching the RIAA curve but 1-2% is kinda swinging wide -- real wide.
dave
halojoy said:
The closer you are to the source the more important the little details are. As i understand it (i have not built my own RIAA) channel-to-channel accuarcy is more important than absoulutely matching the RIAA curve but 1-2% is kinda swinging wide -- real wide.
dave
Much research remain to be done
when it comes how well our both ears falls in +-2 dB
from "linear hearing".
Our hearing curves should ideally be incorporated in the RIAA correction filter. Can be very different from ear to ear.
Luckily there is only 2 of them, so far in the evolution.
If we listen too much to 5.1 surround,
we might get 5.1 ears, as time goes by.
It it also very true that the room resonances have more influence
when listening far from speakers, than close to them.
And I have read somewhere that, like planet10 says,
the matching between the channels is more important.
If we build a Simple but good RIAA, it can be sometimes
hard to find the exact values for the components,
unless we use several in series or parallell.
But if we can get within some % from the timeconstants,
with a few components, We don't have to panic.
Of course Jocko has the point; the strive for perfection.
Perfection will never hurt!
when it comes how well our both ears falls in +-2 dB
from "linear hearing".
Our hearing curves should ideally be incorporated in the RIAA correction filter. Can be very different from ear to ear.
Luckily there is only 2 of them, so far in the evolution.
If we listen too much to 5.1 surround,
we might get 5.1 ears, as time goes by.
It it also very true that the room resonances have more influence
when listening far from speakers, than close to them.
And I have read somewhere that, like planet10 says,
the matching between the channels is more important.
If we build a Simple but good RIAA, it can be sometimes
hard to find the exact values for the components,
unless we use several in series or parallell.
But if we can get within some % from the timeconstants,
with a few components, We don't have to panic.
Of course Jocko has the point; the strive for perfection.
Perfection will never hurt!
Re: Discrete input stage. Followed by OPamp filters/Output.
0.1 dB is hard to hear but 2 dB?? RIAA curves creates a certain "klang" (sonic impression) so it's rather important that the curve fits to the reality. If you have one amp with -2 dB in here and there and an another one with + 2 dB.... The amps will sound different the SAME ears. I guess you don't change ears from time to time?
halojoy said:
0.1 dB is hard to hear but 2 dB?? RIAA curves creates a certain "klang" (sonic impression) so it's rather important that the curve fits to the reality. If you have one amp with -2 dB in here and there and an another one with + 2 dB.... The amps will sound different the SAME ears. I guess you don't change ears from time to time?
Re: Re: Bugle!
The bugle has 6 amp-stages (stereo), 3 for each channel, and that is what I'm trying to tell...
An OPA4134 is fine for MONO... add a OPA2134, (or another 4134, 2-unused), and we have STEREO!
BTW, I have good results running them i "Class A" , using a "J508" constant current diode (2,40mA)
Arne K
Optical said:
The bugle has 6 amp-stages (stereo), 3 for each channel, and that is what I'm trying to tell...
An OPA4134 is fine for MONO... add a OPA2134, (or another 4134, 2-unused), and we have STEREO!
BTW, I have good results running them i "Class A" , using a "J508" constant current diode (2,40mA)
Arne K
RIAA curves
The purpose of the RIAA corrections is, as we all know, to recreate
a flat frequency response. Hence, we are interested in the
flatness of the resulting frequency response. There seems to
be at least two things to consider here.
1) We don't know how perfect the inverse RIAA filtering of the
recording is. In the worst case the errors in the recording
and in our phono amp will add. If we have an almost perfect
RIAA curve in the phono amp, we need only worry about the
errors in the recording.
2) I read somewhere that the human ear is actually very
sensitive to deviations from a flat frequencey response. I don't
remember the exact figures but I think we can detect a dip in
the curve of only one or a few tenths of a dB, while we can
tolerate slightly larger peaks. (If anyone remembers the exact
figures or has a reference to this study, please post it). Of course
the human ear itself (by which I mean the combined physical
ear and the brain) has all but a flat frequency response, is
highly individual and even retrainable, but the brain seems to
know how to compensate for this with high accuracy.
The purpose of the RIAA corrections is, as we all know, to recreate
a flat frequency response. Hence, we are interested in the
flatness of the resulting frequency response. There seems to
be at least two things to consider here.
1) We don't know how perfect the inverse RIAA filtering of the
recording is. In the worst case the errors in the recording
and in our phono amp will add. If we have an almost perfect
RIAA curve in the phono amp, we need only worry about the
errors in the recording.
2) I read somewhere that the human ear is actually very
sensitive to deviations from a flat frequencey response. I don't
remember the exact figures but I think we can detect a dip in
the curve of only one or a few tenths of a dB, while we can
tolerate slightly larger peaks. (If anyone remembers the exact
figures or has a reference to this study, please post it). Of course
the human ear itself (by which I mean the combined physical
ear and the brain) has all but a flat frequency response, is
highly individual and even retrainable, but the brain seems to
know how to compensate for this with high accuracy.
Re: Re: RIAA curves
Yes, definitely and not the least when it comes to hearing.
However, the compensation I mentioned is perhaps not so
mysterious, except for the precision of the hearing. The brain
doesn't know what a flat frequency response is, it just knows
how things usually sound through the ears it happens to be
connected to and reacts if something sounds different from
what it usually does.
planet10 said:
Yes, definitely and not the least when it comes to hearing.
However, the compensation I mentioned is perhaps not so
mysterious, except for the precision of the hearing. The brain
doesn't know what a flat frequency response is, it just knows
how things usually sound through the ears it happens to be
connected to and reacts if something sounds different from
what it usually does.
Simple and very good OP
this OP-amp Schematic, posted by peranders
shows that it needn't be complicated
It uses LT1028, one of the most suited OP amps, for this application
and a few components - use 1% resistors
and try to find 2 capacitors that have close values
MKT or MKP
Of course a good quality Power Supply is important for
an amplifier with high gain.
peranders said:
this OP-amp Schematic, posted by peranders
shows that it needn't be complicated
It uses LT1028, one of the most suited OP amps, for this application
and a few components - use 1% resistors
and try to find 2 capacitors that have close values
MKT or MKP
Of course a good quality Power Supply is important for
an amplifier with high gain.
When we talk about deviation from flat response, not only the magnitude of deviation, but the frequency range (in log) spanned by the deviation also needs to be considered, as well as how sensitive the ear is to the frequencies affected.
For my own (commercial) RIAA amps, I design for a maximum 0.1 db deviation. Now if that 0.1dB error only affects one or two notes, it probably won't be noticed by too many listeners. But if that same 0.1 dB error spans an octave or more, chances are strong that it will be audible to many trained listeners. In fact, depending on where the affected octave(s) is located, a 0.1dB deviation can be enough to affect the overall perceived sonic personality of a design. Conversely, if the octave in question is located between, say, 10kHz~20kHz, perhaps a +/- 0.2dB deviation is in reality quite sufficient.
So you need to establish your allowable deviation target, then consider the range of the frequencies being affected, and their location. This is why it is a very good idea to listen critically to what you have built, and compare it to other designs that are regarded as being the best available.
BTW, the manufacturer of my speakers did match all of the drivers to 0.1dB for me.
hth, jonathan carr
For my own (commercial) RIAA amps, I design for a maximum 0.1 db deviation. Now if that 0.1dB error only affects one or two notes, it probably won't be noticed by too many listeners. But if that same 0.1 dB error spans an octave or more, chances are strong that it will be audible to many trained listeners. In fact, depending on where the affected octave(s) is located, a 0.1dB deviation can be enough to affect the overall perceived sonic personality of a design. Conversely, if the octave in question is located between, say, 10kHz~20kHz, perhaps a +/- 0.2dB deviation is in reality quite sufficient.
So you need to establish your allowable deviation target, then consider the range of the frequencies being affected, and their location. This is why it is a very good idea to listen critically to what you have built, and compare it to other designs that are regarded as being the best available.
BTW, the manufacturer of my speakers did match all of the drivers to 0.1dB for me.
hth, jonathan carr
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.