fallow ~ of course when putting your ear right next to the speaker the noise is clearly audible even in mute condition - or in our systems something goes wrong 
… but it is nothing to say about – the noise level is really low, audible only from the listening distance shorter than 50cm from the speaker. When I sit 2 m from the speakers I can hear nothing But the thing is, that even with this very low noise level PGA2310 works not so good with additional output buffer The optimal configuration I found is simple the PGA2310 without any input output buffers for unbalanced signals and PGA2310 + OPA1632 for balanced ones. I still looking for something better than PGA310 or for the better configuration with this chip. I send the question to Texas Instruments technicians and waiting for an answer.
… but it is nothing to say about – the noise level is really low, audible only from the listening distance shorter than 50cm from the speaker. When I sit 2 m from the speakers I can hear nothing But the thing is, that even with this very low noise level PGA2310 works not so good with additional output buffer The optimal configuration I found is simple the PGA2310 without any input output buffers for unbalanced signals and PGA2310 + OPA1632 for balanced ones. I still looking for something better than PGA310 or for the better configuration with this chip. I send the question to Texas Instruments technicians and waiting for an answer.some time ago i tested some volume chips:
pga2310, pga2311, cs3310.
one of the pga was noisy, maybe (not shure) very sensitive to power supply noise! the other was ok, but sound was a little veiled.
the winner was cs3310, low noise and clear sound.
(and no, i dont work for crystal)
but: you will always (!!!!) hear some noise at the tweeters:
calc: chip has (pga..) 9uv noise out, cs.. <4uv noise,
now with power amp gain x30 = 120uv (or 270uv with pga..),
if speaker has about 90db/1w (=2,8v) --> noise about 3dba at 1m --> nothing to hear....but if speaker 100db/1w --> 13dba, possible to hear under very quiet conditions; now going close to speaker, noise is up to 20db more audible, so you will always have it there!
(at least , if you have more than 0,5uv at your preamp out, but thats not easy to get...)
pga2310, pga2311, cs3310.
one of the pga was noisy, maybe (not shure) very sensitive to power supply noise! the other was ok, but sound was a little veiled.
the winner was cs3310, low noise and clear sound.
(and no, i dont work for crystal)
but: you will always (!!!!) hear some noise at the tweeters:
calc: chip has (pga..) 9uv noise out, cs.. <4uv noise,
now with power amp gain x30 = 120uv (or 270uv with pga..),
if speaker has about 90db/1w (=2,8v) --> noise about 3dba at 1m --> nothing to hear....but if speaker 100db/1w --> 13dba, possible to hear under very quiet conditions; now going close to speaker, noise is up to 20db more audible, so you will always have it there!
(at least , if you have more than 0,5uv at your preamp out, but thats not easy to get...)
3 years ago i tested the cs3310 versus the LM1972 (plus jfet buffer)
The ciruits was batteries powered and it was a differential configuration
The results was clearly in favor of the LM1972.
I haven't a try on the PGA family.
One of the advantage of the LM1972 is the absence of AOP at the output.
The ciruits was batteries powered and it was a differential configuration
The results was clearly in favor of the LM1972.
I haven't a try on the PGA family.
One of the advantage of the LM1972 is the absence of AOP at the output.
Alfach ~ You right – PGA2310 is really very sensitive to supply voltage quality analog as well as digital voltage – only low noise voltage regulators gives really good effect – I use REG104 or REG1117 (Burr-Brown) for +5V and LM2941/LM2991 pair (National Semiconductors) for +/- 15V with quite good results – better than for LM317/337 pair … I never tried the old best reference chip CS3310, probably I still have one or two saved in my chip archive. So, I have to try also them I have also other idea to build real low noise voltage regulator using REF102 -> AD797 -> BUF634 to supply PGA2310 (or CS3310 if it is even better) the work is ongoing but with the long breaks in between because of well known problems with free time …
analog as well as digital voltage – only low noise voltage regulators gives really good effect – I use REG104 or REG1117 (Burr-Brown) for +5V and LM2941/LM2991 pair (National Semiconductors) for +/- 15V with quite good results – better than for LM317/337 pair … I never tried the old best reference chip CS3310, probably I still have one or two saved in my chip archive. So, I have to try also them I have also other idea to build real low noise voltage regulator using REF102 -> AD797 -> BUF634 to supply PGA2310 (or CS3310 if it is even better) the work is ongoing but with the long breaks in between because of well known problems with free time …
alternately u could change PGA 2310 with 2311, all u ll have to do is to change the voltage regulators, but do seperate AC supply of digital and analogue. Later has a better noise performance.
For wm8816 u would need input and output buffers, better choice would be DS1802 and DS1666 in that case. Also soldering WM8816 and making suitable PCB wont be that easy.
For wm8816 u would need input and output buffers, better choice would be DS1802 and DS1666 in that case. Also soldering WM8816 and making suitable PCB wont be that easy.
Hi,
you are rigth, just wanted to make clear, when
people say the device is dead silence(at mute or idling....),
It is not possible, at least with any volume control chip, unless there is possibility to cut the signal wire by some short of relay etc.,
Have to try to use the chip battery operated to see, if there is something wrong with power supply. The only problem is that I have no equipment to measure signal strength below 2,5 uV(expect trying to hear the difference....), unless I place some amplifier to the signal, which may introduce some noise also.
Where should ground's be connected(digital and analog), if I use two different trafo's? near the chip?
If the battery operation does not recude the noise, are there some tricks to try, beside the voltage trick I'm already using. amplify the signal by 4(if voltage in is 2Vp-p as in standard the amplification could be higher) and after the PGA-chip the output is divided by 20, this means that the output voltage is 4/20 from maximum(at least is my use the volume is still high enough, usign about -20 dB setting for loud listening). With this kind of arrangement the noise is reduced. (There is also new noise component, the resistor network which is used to divide the output voltage, but it is lower than the reduced noise).
Could tis kind of arringement work, or is it just feeling than this is working. (at least in spectrum analyzer the noise floor dropped from 10 uV to below 2,5uV(the resolution of spectrum analyzer).
you are rigth, just wanted to make clear, when
people say the device is dead silence(at mute or idling....),
It is not possible, at least with any volume control chip, unless there is possibility to cut the signal wire by some short of relay etc.,
Have to try to use the chip battery operated to see, if there is something wrong with power supply. The only problem is that I have no equipment to measure signal strength below 2,5 uV(expect trying to hear the difference....), unless I place some amplifier to the signal, which may introduce some noise also.
Where should ground's be connected(digital and analog), if I use two different trafo's? near the chip?
If the battery operation does not recude the noise, are there some tricks to try, beside the voltage trick I'm already using. amplify the signal by 4(if voltage in is 2Vp-p as in standard the amplification could be higher) and after the PGA-chip the output is divided by 20, this means that the output voltage is 4/20 from maximum(at least is my use the volume is still high enough, usign about -20 dB setting for loud listening). With this kind of arrangement the noise is reduced. (There is also new noise component, the resistor network which is used to divide the output voltage, but it is lower than the reduced noise).
Could tis kind of arringement work, or is it just feeling than this is working. (at least in spectrum analyzer the noise floor dropped from 10 uV to below 2,5uV(the resolution of spectrum analyzer).
fallow said:
I would start by deleting the two output opamps. They can only contribute distortion and noise. The PGA's can easily handle anything you throw at them.
Deleting the output opamps also lets you run the PGA at higher gain which also helps the S/N ratio. And you save money as well. Have your cake and eat it too!
Jan Didden
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.