BC550 BC560 Very low noise RIAA

[lineup]

good work people!
we may endup with some nice PCB layouts!


Any ideas for a good quality dual regulator voltage supply?
Would LM317 + LM337 and some good decoupling caps be good enough?
I don't think it would be too bad, actually.

The supply current is low,
like 10-12 mA per channel + 5mA for the LM317, LM337 reference resistors
gives like 15-17 mA per channel, and like 30-35mA total, drawn from transformer.

The trafo needed is not very big.
Transformer 2x15 VAC, 100mA ( >= 3 VA )
would feed 2 channels
with each channel using separate LM317+LM337 regulators.


There are plenty of such dual regulated supply circuits
and layouts for supply PCB boards
to find in this forum.
Used for other preamplifier designs feeding chip Op-Amps.


Also peranders, one of our low noise RIAA amplifier specialists,
has got some extremely good DIY supplies and PCB at his website.
These 2 extreme low noise supplies
are specially designed for HiFi RIAA amps:
QSXPSR0 http://home.swipnet.se/~w-50719/hifi/qsxpsr0/index.html
QSXPSR1 http://home.swipnet.se/~w-50719/hifi/qsxps/index.html

I don't know what is the price for his boards and total cost including components for his supply,
but anybody can download his schematics for free
and have a look, how he designed these very good supplies.

I also doubt will be needed such extreme designs for this BC550C BC560C discrete RIAA.
A plain commonly used LM317 LM337 supply may perform very well.
And bothering for anything more advanced
may not improve just about anything to mention or notice at listening.


Regards
lineup

[peranders]

Originally posted by lineup
I also doubt will be needed such extreme designs for this BC550C BC560C discrete RIAA.

If you have an amp which is built by discrete parts you will probably need a better PS than with opamps since the PSRR is much better for an opamp. You can never have a too good PS. Compare the debate about the Jung Super Regulator.

The two boards mentioned above is Revision 0 (don't sell it anymore) and a developed version, Revision 1. I have added options for 10 mm snap-in caps, bigger resistors (more current) and some more things.

[lineup]

a couple of detail I have to add:

1)

the R17 resistor: 923.804 ohm
is an extra 'adjustment resistor'
and this value was what i needed in my simulator
to get zero offset at the output, between Collectors of U17-U18
Absolute zero is impossible,
but using all 6 decimals: 923.804 ohm for R17, was the closest to 0 Volt.
A multimeter in my Multisim then shows: +0.000000488 volt offset ( +488.2nV ).

Where does this offset come from?
Well, BC550C NPN and BC560C PNP does not have same gain.
PNP transistors are always slightly different from NPN.
In this case an average PNP has got higher gain, than NPN.

This means will be a bit more base current going into U1-U4
then comes out of U9-U10 bases.
This difference of bias current is called in datasheets for Op-Amp,
for input bias offset current.
It is very small, but it has to go somewhere.
And the only way it can go here
is take the road through input resistor R16 to ground, 0V.

And when we put a current into a input resistor
we get an input offset voltage.
Even if current is extremely small it creates a voltage of some mV in 56 kohm.
In this case a negative voltage.

This voltage is amplified to the output and so an offset is created at output.
To compensate for this negative offset voltage in output
I increase the current coming from U17
by adding an adjustment resistor in parallel with R11, 150 ohm.
In this case R17 = 923.804 ohm made practically 0 V at output.

On an eventually PCB, it is good if make an extra place
for parallel adjustment resistor at BOTH postive and negative side

Because somebody may need to from a POSITVE offset to zero.
now such an offset is not critical for good operation of this amplifier
as we used output capacitor C2.
But for some of us, it feels nice to have close to 0.0 volt at output.

I can tell without adjustment, I could read only an output offset of
-0.075 V ( -75mV )
It this is of no importance for good operation.


==============

2)
8-9 mA Class A idle current in output is a bit high for BC550/BC560.
And is way above what output will need to drive almost any practical load.

For example 2 V RMS output into 4.7 kohm load
takes only +-0.6 mA peak ....
... and a Class A push-pull output stage running only at idle 0.5 mA
has got the theoretical capacity to put out +-1.0mA peak
So even for 2 V RMS into into 4.7 kohm, 9mA class A
means an over-capacity of almost 20 times!


I suggest we lower the current in output to 5 mA
which is a current at which BC550/BC560 work better.
( BC550/560 should not run at more than max 6-7 mA )

By my pocket calculator, without I have have yet run Multisim of it,
lowering R9 and R10(4.7k) to 3.3 kohm
will give just about 5 mA in output transistors.

This will have some other good effects in quality of operation.
But I will not go into details here.

===============================

lineup


PS. I will post version 2, with these updates.
within a couple of days, when I have created new images of circuits.
It wont change anything that can effect PCB layout.
Only change in resistor value. DS.

[Nordic]

Not sure I understand... I have your circuit in eagle if you want..

[Nordic]

Think we can make that 2 series resitors then,so we can use say 825R and 100R

[lineup]

Originally posted by Nordic
Not sure I understand... I have your circuit in eagle if you want..

Nordic, my friend,
As you do not say what you do not understand, there is no way I can explain, sorry.

---------------------------------------------------------------

First one correction of my last post here above:
The actual offset, if nothing done to adjust output voltage
was actually more than -3.0 V
Too much!

So some offset adjust feature had to be added here!
Because the DC gain is very high!
Any small offset at input, will actually be amplified >500 times
The DC-gain is defined by
R19/R18 = 150.000/270 .... Gain ~ 556.

----------------------------------------------------------------

I chose to add offset adjust potentiometer at input.
A small compensation current is provided by putting
a voltage across R25, 1 Mohm
And this voltage is set by
a multiturn 100 kohm trimmer potentiometer, R24

As can be seen in my attached full circuit, version 2,
I had to turn this trimmer to 80% to get zero offset at output.
( well almost zero, +150uV, = +0.00015 volt )
-----------------------------

I also removed the previous offset adjustment resistor, R17.

-----------------------------------------------------------------------

R9 and R10 were changed to 3.3k.
This gives an idle Class A current of ~4.8 mA in output stage.
Which is much better value of current for the output transistors.
It is still high idle current, in relation to what is need for normal loads.

-----------------------------------------------------------------

The capacitor C1 at the input, has now got no value.
Because this depends of which MM pickup you use
and other factors, like length and type of signal cable
from pickup to this RIAA amplifier input.
Different pickups wants different pF capacitor
and different cables has got different capacitance per meter.

It is to point out
that this is a phono RIAA amplifier for MM pickups only.
Not for MC pickups, cartridges.


lineup

[Holger Barske]

Had a look at PA's supply circuits. Very nice, no doubt, but: Would cost 5 to 10 times the money of they preamp circuit itself.
What about a cheaper solution like a LM317 CCS feeding a TL431 shunt reg? Can be done at less than 1€ per regulator.

[lineup]

Originally posted by HBarske
Had a look at PA's supply circuits. Very nice, no doubt, but: Would cost 5 to 10 times the money of they preamp circuit itself.
What about a cheaper solution like a LM317 CCS feeding a TL431 shunt reg? Can be done at less than 1€ per regulator.

Yes, nothing and noone will be able to make use such advanced power supply.
Even how good it is.
Not for this simple but good RIAA. Or any other amplifier.
Because my own way is KISS - Keep It Simple.
Few but good components in as good harmony as possible.
Like often Nelson Pass has done it.


I have actually several discrete and fairly good regulators of my own.
One uses TL431 as reference. The one you mention.
working at a constant current, by a little trick I use some times.

The version of this great regulator is actually for high voltage low current apps.
Like Tube preamps Triodes.
So the output is of this very high precision regulator is:
150.001 VDC and gives currents out like max 50-100 mA

Using only 3-4 smaller transistors + one TL431.
--------------------------------


But for the easy build of this projects, for all new to DIY
I'd suggest we use LM317 + LM337 for the standard version.
These IC are very good, if setup correctly.
Not the ultimate, but good audio standard even for HIFi, for sure.

Then of course, anybody that is a bit more experienced in DIY audio
can use what he prefers personally.


lineup

[Holger Barske]

Simplicity is not an issue. Far better than 317/337 alone.

[Holger Barske]

Oops... forgot a resistor.