I own a old Sony TA-FA3ES integrated amp which sounds decent with CD
as source but sucks via Phono MC (with a Dynavector DV23 RS MK.2) and
is far to noisy.
I made a little websearch and learned that the phonoeqpart consists
of an input stage with two 2SC2545 and a RC4560D Opamp. I also learned that
later models (for example the FA30ES) use the same circuit diagram but
with a NJM2068DD Opamp.
The whole phonostage is fed by +- 15V.
My question is would it be worth while to change the Opamp to NJM2068DD
or an probably better Opamp like LME 49720 in my amp? What other options do i have to better the sound of the give phonoamp ?
as source but sucks via Phono MC (with a Dynavector DV23 RS MK.2) and
is far to noisy.
I made a little websearch and learned that the phonoeqpart consists
of an input stage with two 2SC2545 and a RC4560D Opamp. I also learned that
later models (for example the FA30ES) use the same circuit diagram but
with a NJM2068DD Opamp.
The whole phonostage is fed by +- 15V.
My question is would it be worth while to change the Opamp to NJM2068DD
or an probably better Opamp like LME 49720 in my amp? What other options do i have to better the sound of the give phonoamp ?
Probably the schematics help a bit:
As you can see its a differential input stage with two transistors followed
by the opamp with riaa-eq in the feedback-loop. MM/MC switches the
input impedanz and gain.
What i want to know is if i can change the Opamp and probably - but not
necessarily - the transistors in the input stage to 2SK170 FETs without making
a new pcb. Is there any hope for that ? And what other components i have
to change to enhance the the soundquality and lower the noise.
I have checked this little phonostage with a couple carts. Its ok with MM
and even a AT OC-9 is ok,but with the DV23R it simply sucks big time.
Im not an expert in electronics,just a interested layman,and im standing
deep in the woods here.
Im thankfull for any input.
An externally hosted image should be here but it was not working when we last tested it.
As you can see its a differential input stage with two transistors followed
by the opamp with riaa-eq in the feedback-loop. MM/MC switches the
input impedanz and gain.
What i want to know is if i can change the Opamp and probably - but not
necessarily - the transistors in the input stage to 2SK170 FETs without making
a new pcb. Is there any hope for that ? And what other components i have
to change to enhance the the soundquality and lower the noise.
I have checked this little phonostage with a couple carts. Its ok with MM
and even a AT OC-9 is ok,but with the DV23R it simply sucks big time.
Im not an expert in electronics,just a interested layman,and im standing
deep in the woods here.
Im thankfull for any input.
some searching finds this link - might be of interest:
http://www.diyaudio.com/forums/solid-state/59986-good-phono-preamp-circuit-3.html#post675071
also, if you can find schematics of older (late 80's, early 90's ) yamaha and denon top line receivers and preamps to see phono preamps with discrete jfets feeding a "good" op amp.
it might be somewhat trivial to make a decent sounding moving magnet phono preamp, but doing the same for moving coil is no joke. as an example, check the search engine for the hps designs by syn08. some useful discussion to explain some of the challenges.
good luck,
mlloyd1
http://www.diyaudio.com/forums/solid-state/59986-good-phono-preamp-circuit-3.html#post675071
also, if you can find schematics of older (late 80's, early 90's ) yamaha and denon top line receivers and preamps to see phono preamps with discrete jfets feeding a "good" op amp.
it might be somewhat trivial to make a decent sounding moving magnet phono preamp, but doing the same for moving coil is no joke. as an example, check the search engine for the hps designs by syn08. some useful discussion to explain some of the challenges.
good luck,
mlloyd1
Last edited:
mlloyd1@: thanks for guiding me in the right direction.
I did read a lot here lately and also found some very interesting reads
in the HPS and Phonoclone threads ,respectivly in the linked websites there.
WOW! 😱
On the other side i did some research on old amps like you suggested and found
out the mentioned design above is all over the place. With various inputstages
and various opamps. Pretty much the industrial standart solution for a cheap
phonostage with decent quality i suppose. It is found in cheap amps and in
more costly ones too like the Marantz PM-8000 (2SK369BL FETs +NJM 2114) for
example.
I guess to make a serious effort in enhancing my phonostage qualitywise is
more necessary than changing a chip and a couple resitors. Silly me to think
that it would be that simple. 😱
I did read a lot here lately and also found some very interesting reads
in the HPS and Phonoclone threads ,respectivly in the linked websites there.
WOW! 😱
On the other side i did some research on old amps like you suggested and found
out the mentioned design above is all over the place. With various inputstages
and various opamps. Pretty much the industrial standart solution for a cheap
phonostage with decent quality i suppose. It is found in cheap amps and in
more costly ones too like the Marantz PM-8000 (2SK369BL FETs +NJM 2114) for
example.
I guess to make a serious effort in enhancing my phonostage qualitywise is
more necessary than changing a chip and a couple resitors. Silly me to think
that it would be that simple. 😱
Given the age of the unit, I'd be very interested in the ESR of the 2 'lytic caps inside the feedback loop. I'd strongly consider the LME49720 or LM4562 opamp as a replacement.
G²
G²
Hi,
this kind of circuit was quite popular in the 80s, when High Precision ultra-lownoise OPamps were very expensive or unobtainable.
It is the circuit that the well known Elektor Preamp and its later offsprings used (1986). The Preamp used 3 paralleled matched MATs and the OP27 industry standard OPamp (using the expensive MATs they gave up though on the original idea of reducing cost).
The 2SC2545 are very lownoise bipolar transistors made by Renesas/Hitachi but they are now obsolete. You can still source them from Reichelt Germany (ask for suffixes E or F).
They can´t be replaced simply with JFETS like the 2SK170 as drop-ins without serious redesign of the whole circuit. You needed to replace each single 2SC2545 with 2-3 paralleled 2SK170 and to run those on much higher Drain currents. Noise still wouldn´t improve, because JFETs are only superior at high source impedances, while bipolars are superior at low impedance values. It would be difficult to find better devices than the 2SC2545.
There are three points where the circuit might be tweaked noisewise.
1) slightly increasing collectur current of Q401/402 to 1-3mA each (R409-411 new, lower values)
2) paralleling input devices Q401/402 with one or even two additional pairs and running them on the higher collector current (roughly 3dB improvement in noise figure. Not needed for MMs but could help with MC cartidges).
3) replacing the RC4560 with a lownoise OPamp like the OP27/OP270 suffix A or E. The effect will be very small though.
I wouldn´t use the LMEs because they are imho vastly overrated parts.
When You look at the numbers that count, their performance is just mediocre.
Even the Grandpa OP27/270 shows lower noise figures both for voltage as well as current noise at low frequencies (where gain is very high) and it´s input offset voltage is specced just one tenth of the LME! The LME´s higher speed is of no positive use in this application, but could rather lead to oscillation problems.
From the use of the amplifying devices the circuit shouldn´t be noisy. It is actually a very lownoise design. I assume that the noise is coming from ´outside´of the circuit from the powersupply. The stabilizing circuitry of the positive supply voltage with just an simple zener-diode/capacitor leaves room for improvement. The input differential transistor pair is directly coupled via low valued resistors (how close might these be matched?) to the positive supply and just a resistor goes to negative side. So power supply rejection is not high.
Suggestion:
- replace R413 with 5k. This doubles the current through Q410/402 and reduces their noise contribution. In a later step R413 couuld be replaced by a decent current source.
- replace R414 with 1k and Zener diode D401 with an 10V type. This raises the supply voltage at R409-410´s junction. The headroom can now be used for noise filtering.
- cut the powersupply connection of R409 and 410. Insert a resistor of 1k between Zener D401 and the Resistors 409-410. Add a cap of 100µF to 270µF from the resistor´s junction to gnd.
This simple tweak reduces e.g. the positive powersupply´s ripple (100Hz) and noise by app. 30-40dB at the OPamps inputs. If for example there´s a ripple of 100Hz (rectified mains) of say 200mVpp on the +B line the OP amp inputs would still see app. 1.5mV of ripple which is much, especially when compared to the music signal´s voltage. As it is the circuit relies heavily on the OPamp´s commonmode rejection to filter it out. But any imbalance (tolerances of R´s) reduces the rejection ratio. So it would be good anyway to reduce the ripple and noise on the supply line, the more so the larger R409/410 differ in value.
jauu
Calvin
this kind of circuit was quite popular in the 80s, when High Precision ultra-lownoise OPamps were very expensive or unobtainable.
It is the circuit that the well known Elektor Preamp and its later offsprings used (1986). The Preamp used 3 paralleled matched MATs and the OP27 industry standard OPamp (using the expensive MATs they gave up though on the original idea of reducing cost).
The 2SC2545 are very lownoise bipolar transistors made by Renesas/Hitachi but they are now obsolete. You can still source them from Reichelt Germany (ask for suffixes E or F).
They can´t be replaced simply with JFETS like the 2SK170 as drop-ins without serious redesign of the whole circuit. You needed to replace each single 2SC2545 with 2-3 paralleled 2SK170 and to run those on much higher Drain currents. Noise still wouldn´t improve, because JFETs are only superior at high source impedances, while bipolars are superior at low impedance values. It would be difficult to find better devices than the 2SC2545.
There are three points where the circuit might be tweaked noisewise.
1) slightly increasing collectur current of Q401/402 to 1-3mA each (R409-411 new, lower values)
2) paralleling input devices Q401/402 with one or even two additional pairs and running them on the higher collector current (roughly 3dB improvement in noise figure. Not needed for MMs but could help with MC cartidges).
3) replacing the RC4560 with a lownoise OPamp like the OP27/OP270 suffix A or E. The effect will be very small though.
I wouldn´t use the LMEs because they are imho vastly overrated parts.
When You look at the numbers that count, their performance is just mediocre.
Even the Grandpa OP27/270 shows lower noise figures both for voltage as well as current noise at low frequencies (where gain is very high) and it´s input offset voltage is specced just one tenth of the LME! The LME´s higher speed is of no positive use in this application, but could rather lead to oscillation problems.
From the use of the amplifying devices the circuit shouldn´t be noisy. It is actually a very lownoise design. I assume that the noise is coming from ´outside´of the circuit from the powersupply. The stabilizing circuitry of the positive supply voltage with just an simple zener-diode/capacitor leaves room for improvement. The input differential transistor pair is directly coupled via low valued resistors (how close might these be matched?) to the positive supply and just a resistor goes to negative side. So power supply rejection is not high.
Suggestion:
- replace R413 with 5k. This doubles the current through Q410/402 and reduces their noise contribution. In a later step R413 couuld be replaced by a decent current source.
- replace R414 with 1k and Zener diode D401 with an 10V type. This raises the supply voltage at R409-410´s junction. The headroom can now be used for noise filtering.
- cut the powersupply connection of R409 and 410. Insert a resistor of 1k between Zener D401 and the Resistors 409-410. Add a cap of 100µF to 270µF from the resistor´s junction to gnd.
This simple tweak reduces e.g. the positive powersupply´s ripple (100Hz) and noise by app. 30-40dB at the OPamps inputs. If for example there´s a ripple of 100Hz (rectified mains) of say 200mVpp on the +B line the OP amp inputs would still see app. 1.5mV of ripple which is much, especially when compared to the music signal´s voltage. As it is the circuit relies heavily on the OPamp´s commonmode rejection to filter it out. But any imbalance (tolerances of R´s) reduces the rejection ratio. So it would be good anyway to reduce the ripple and noise on the supply line, the more so the larger R409/410 differ in value.
jauu
Calvin
Last edited:
I would do first the follow:
1) an external, independent power supply (parallel regulated like NAD PP-2) and cutting the wires for the internal voltage supply - the "+" and "-" rail, but not the GND).
2) removing of the MM / MC switch and direct solder of connection wire either for "MM" or for "MC" (such switches causes extremely high loss of sonic quality, especially in the "MC" mode) For example have a look to attached scheme PP-2. Here you can use each value of capacity through loading about current source instead voltage regulator.
All other modifications will depend on whether you decide for MC or MM.
If you use moving coil cartridges with moderate input impedance and load impedance arround 1K (e. g. some Benz mid until high output moving coil models), then I would choice the "MC" mode and aditional you must do readjust the gain (feedback) through change of resistor R143, so that you get the same level than compact disc playback
If you use moving coil cartridges with very low input impedance and very low necessary load impedance like the mentioned above, then you must choice the MM mode and the additional use of a pre-pre in common base configuration like JC-1 or Hiragas "prepre"
look about
http://www.zenn.com.sg/ML_JC1.JPG
Hiraga MC Preamp
This is in this case the royal way
After do one of this both mentioned versions I would consider to change the OP-Amp respective the input BjT devices through jFETs.
Please note, that extremely high values for the signal to noise ratio by phono stages completely absurd, because the limit in real life was set by the background noise of the LP records.
Last edited:
stratus46@: Right. My first thought was to spent C403,C408 and not to forget
C402 SILMICs and swap the RC4560D with an NJM2068D-D - as Sony did in later
Models. This is the cuircuit of the FA30ES phonostage,the direct follower of
the FA3ES:
Which shows some changes due to adapt the new opamp into this design and some
finetuning here and there.
Calvin@: Wow! Thanks for the insight and detailed tips.
The mentionend R409 & R410 are 5% carbon types. In fact all resistors in this
design are 5% carbon types,except R417 & R418 (1% carbon) and R416,which is a 1%
metalfilm type.
The powersupply. I guess that is another thing to improve in this amp. The amp
is supplied in a minimalistic way as you can see here:
But that probably comes later...
The phonostage is feed with this prior to the elements seen on the phonoboard:
Powerinput is comming direct from the powerampsupply. This cuircuit changes the +- 46V to +- 15V.
tiefbassuebertr@: Thanks. Another good advice. Probably its better to optimze
the phonoeq for MM. In fact Sony did exactly that in later Models like the
FB940R QS. Since the FA30ES - which must be a 1996-97 Model - they stopped
finetuning this phonostage but used it in any decent amp they sold - after they
decided to bury the bigger "ES" stereoamplines and go for surround effect acoustics instead.
And that was it than,the final version,MM only:
Even the pcb used here is still exactly the same like that used in FA30ES.
As for external headamps. I know,but no way in this case. Im on a tight budget.
If - and thats a big if - im convinced more fundamental changes must be undertaken than it must be placed inside the amp and in this case i think it
would be possible because the phonoeq in the FA3ES is on anice little rectangular pcb only connected by a socketrail to transport power and audiosignal.
Thats a top down nudi pic of a FA3ES. Top right corner is the phonoeq-pcb.
C402 SILMICs and swap the RC4560D with an NJM2068D-D - as Sony did in later
Models. This is the cuircuit of the FA30ES phonostage,the direct follower of
the FA3ES:
An externally hosted image should be here but it was not working when we last tested it.
Which shows some changes due to adapt the new opamp into this design and some
finetuning here and there.
Calvin@: Wow! Thanks for the insight and detailed tips.
The mentionend R409 & R410 are 5% carbon types. In fact all resistors in this
design are 5% carbon types,except R417 & R418 (1% carbon) and R416,which is a 1%
metalfilm type.
The powersupply. I guess that is another thing to improve in this amp. The amp
is supplied in a minimalistic way as you can see here:
An externally hosted image should be here but it was not working when we last tested it.
But that probably comes later...
The phonostage is feed with this prior to the elements seen on the phonoboard:
An externally hosted image should be here but it was not working when we last tested it.
Powerinput is comming direct from the powerampsupply. This cuircuit changes the +- 46V to +- 15V.
tiefbassuebertr@: Thanks. Another good advice. Probably its better to optimze
the phonoeq for MM. In fact Sony did exactly that in later Models like the
FB940R QS. Since the FA30ES - which must be a 1996-97 Model - they stopped
finetuning this phonostage but used it in any decent amp they sold - after they
decided to bury the bigger "ES" stereoamplines and go for surround effect acoustics instead.
And that was it than,the final version,MM only:
An externally hosted image should be here but it was not working when we last tested it.
Even the pcb used here is still exactly the same like that used in FA30ES.
As for external headamps. I know,but no way in this case. Im on a tight budget.
If - and thats a big if - im convinced more fundamental changes must be undertaken than it must be placed inside the amp and in this case i think it
would be possible because the phonoeq in the FA3ES is on anice little rectangular pcb only connected by a socketrail to transport power and audiosignal.
Thats a top down nudi pic of a FA3ES. Top right corner is the phonoeq-pcb.
An externally hosted image should be here but it was not working when we last tested it.
- Status
- Not open for further replies.