"Super-Pair" I/V for TDA1543

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Jump to Latest
Bricolo

Bricolo

Member
Joined 2002
Terry Demol said:






Hi Rudolf,

Sorry to diverge slightly.

WRT comparisons of the "super pair" folded cascode I-V and
the single rail version, I noticed you are running about 16mA
through super pair and about 3.5mA through the single rail
version.

I think the differences you are hearing between these two
topologies is more due to current than topology type :)

I suggest running equal currents and re-compare :cool:

Regards,

Terry
Click to expand...


Terry, you seem to know how a super pair works (at least better than I do)
Do you know if there is a solution to set Vbe to zero?

BTW, where can I learn more about the super pair?
 
Terry Demol

Terry Demol

Member
Joined 2002
Bricolo said:



Terry, you seem to know how a super pair works (at least better than I do)
Do you know if there is a solution to set Vbe to zero?

BTW, where can I learn more about the super pair?
Click to expand...

Bricolo,

If there is 0V between base and emmiter (Vbe), the transistor
is turned OFF. I think what you mean is how to get DAC
IP point to GND, is that what you mean?

Terry
 
Bricolo

Bricolo

Member
Joined 2002
Terry: yes

Andrea: it's not as simple as this. that's why I asked this question to Terry.

The problem is that the voltage between Q2P's base and Q2N's emitter (this is what I called the "Vbe of the super-pair" in my previous post) isn't exactly zero (my sims show -50mV)

And 50mV is too much, according to the TDA1541A's datasheet.



Grounding Q2P's base directly would pose no ptoblem itself, I think (but I'm really not sure) that it would even be better, since the impedance of the voltage source is exactly zero (in fact, the voltage of the voltage source is also zero, if you compare it to the leds that have something loke 3V and 10R)

But you'll have those 50mV on the input :(


I see no solution yet. Using a pot (the wiper on Q2P's base, and the 2 other legs on V+ and V-) we would be able to set the input to zero. But the impedance will be high between the base and ground. A small value pot would draw too much current...

Maybe Terry or someone else has a solution
 
Terry Demol

Terry Demol

Member
Joined 2002
Bricolo said:
Terry: yes

Andrea: it's not as simple as this. that's why I asked this question to Terry.

The problem is that the voltage between Q2P's base and Q2N's emitter (this is what I called the "Vbe of the super-pair" in my previous post) isn't exactly zero (my sims show -50mV)

And 50mV is too much, according to the TDA1541A's datasheet.

Maybe Terry or someone else has a solution
Click to expand...

OK, you can not use a single supply for TDA1541. You need
a neg supply to set up a current source to bias N channel
GBS BJT on.
Use a diode or diode connected BJT with cathode to gnd to lift the
GBS BJT's base 0.7V above ground. It's emmiter will then be at 0V.
Put a current trim from diode to + supply and this can be fine
tuned if desired to get virtual GND within 10mV of GND.

OK now a simple "dumb" question for you, how do I post
a schematic here?

Cheers,

Terry
 
Bricolo

Bricolo

Member
Joined 2002
Terry Demol said:


OK, you can not use a single supply for TDA1541. You need
a neg supply to set up a current source to bias N channel
GBS BJT on.
Use a diode or diode connected BJT with cathode to gnd to lift the
GBS BJT's base 0.7V above ground. It's emmiter will then be at 0V.
Put a current trim from diode to + supply and this can be fine
tuned if desired to get virtual GND within 10mV of GND.

OK now a simple "dumb" question for you, how do I post
a schematic here?

Cheers,

Terry
Click to expand...

I will use a neg supply! I'm talking about the circuit Rudolf posted on this thread (not the single supply one, the folded cascode one!)

To attach a file, when you're typing your message there's a "Browse" button, just over the "submit Reply" one. Attach a file with it. As peter said, not all file formats are accepted. You have to limit the graphic and file size too, I think it's 800*600 and 100kb.



I'm not sure I understand your answear. The reference schematic is this one, what modifications should be done to put I-dac on 0V?
 

Attachments

  • superpair-iv.gif
    superpair-iv.gif
    9.2 KB · Views: 1,096
Terry Demol

Terry Demol

Member
Joined 2002
Bricolo said:


I will use a neg supply! I'm talking about the circuit Rudolf posted on this thread (not the single supply one, the folded cascode one!)

To attach a file, when you're typing your message there's a "Browse" button, just over the "submit Reply" one. Attach a file with it. As peter said, not all file formats are accepted. You have to limit the graphic and file size too, I think it's 800*600 and 100kb.

I'm not sure I understand your answear. The reference schematic is this one, what modifications should be done to put I-dac on 0V?
Click to expand...

Hi Bricolo,

OK, all is clear.

To move virtual gnd to 0V just omit D2, D3 and R1. Base of Q2P
will then be grounded. Emmiter Q2P = +0.7V so Emmiter Q2N
goes back to 0V.

Note however that the VBE of Q2P and Q2N will not be
exactly the same due to different voltages across them.
However this will be cancelled out to a degree because
Q2P is run at much less current than Q2N.

My guess is that it will be within 20mV of gnd.

Cheers,

Terry
 
mikelm

mikelm

Member
Joined 2003
Bricolo said:
In my simulations, it's not within 20mV of ground, but rather 50mV
And the TDA1541A has a voltage compliance of 25mV :(

That's why I wanted to be able to ajust it precisely.
Click to expand...

Would it not be OK to have resistor chain between the rails, two fixed values to the rails with a variable one in the middle that covers the critical area with a cap from the wiper of the trimmer to earth to deal with any noise and a connection base of Q2P also to the wiper of the trimmer.

This is a very simple possible solution - are there any drawbacks ?

Of course it will need to be carefully set initially but I think it should be OK if the cct's are left switched on.
 
rbroer

rbroer

Member
Joined 2002
Terry Demol said:
WRT comparisons of the "super pair" folded cascode I-V and
the single rail version, I noticed you are running about 16mA
through super pair and about 3.5mA through the single rail
version.
I think the differences you are hearing between these two
topologies is more due to current than topology type :)
I suggest running equal currents and re-compare :cool:
Click to expand...

The one posted in the thread is using initial values from the simulator.
The one in my cd player is actually 10mA Terry.
Didn't observe much improvement though...:cannotbe:
 
rbroer

rbroer

Member
Joined 2002
mikelm said:
Would it not be OK to have resistor chain between the rails, two fixed values to the rails with a variable one in the middle that covers the critical area with a cap from the wiper of the trimmer to earth to deal with any noise and a connection base of Q2P also to the wiper of the trimmer.
This is a very simple possible solution - are there any drawbacks ?
Of course it will need to be carefully set initially but I think it should be OK if the cct's are left switched on.
Click to expand...

Finally a solution :D

I would do something like this (capacitor from wiper to ground can be added):
 

Attachments

  • 0mv.gif
    0mv.gif
    2.9 KB · Views: 1,019
Bricolo

Bricolo

Member
Joined 2002
Terry Demol said:


Bricolo,

Stop simulating, stop worrying, start soldering :smash:

What is the problem with 47R at base of Q2P?

Cheers,

Terry
Click to expand...


I don't know what's wrong with it.

My question was if it is better to have a low impedance voltage reference from Q2P's base to ground.
I think it is (but maybe you have another opinion), so the voltage divider isn't optimal :(
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Top Bottom