Rate this Entry

TL431 output filter for TDA1543 current source elimination

Posted 6th September 2013 at 04:45 AM by abraxalito
Updated 6th September 2013 at 05:10 AM by abraxalito

Here's what I've ended up with in my SDcard DAC for eliminating the pull-up current sources within the TDA1543 - a heavily filtered 3.75V reference. For a single TDA1543 just connect the L and R 870R resistors to the output of this filter instead of the normal connection to 0V. Then the input's fed with the TL431 - I'm using 3k and 1k5 output voltage setting resistors. The first resistor and inductor on the left are my attempt to simulate the output impedance of the 431.

The electrolytics are Sancon 16V types with 0.1R ESR - the 440uF being made of two in parallel. The inductors need to have the right ESR (about 0,1ohm also) for which I had to wind my own 68uH - the ready made ones I have are too lossy

Varying these parameters might result in a lumpy frequency response.

Attenuation is -70dB by 20kHz but almost flat to 1.6kHz and output impedance remains low down to DC.

Eliminating the current sources means that pin7 of TDA1543 now goes unconnected. This arrangement maintains the inverting characteristic of the DAC,

Posted in DAC design and development

Views 1723 Comments 10

« Tweaking the TDA1543 SDcard player Main Introducing the Ozone DAC platform »

Total Comments 10

Comments

	Great idea. And at what voltage is the TDA1543 running, just asking before testing myself? 3,75 bias on the AOR/AOL seems to hint at 8 VDD. Just noting - I remember a schema that used an opamp to get the bias with the DC on AOR/L as source as an opamp-follower design.
	Posted 17th September 2013 at 11:20 AM by triode_al

	No, its running stock at 5V (7805) - the max compliance voltage at 5V supply is 3.8V, 3.75V just happens to be 2.5V * 1.5 so quite easy to set up with the TL431.
	Posted 17th September 2013 at 02:33 PM by abraxalito

	Me being a strong visual type, could you please point me to a schematic, or present one which have all that on? Thanks.
	Posted 19th November 2013 at 06:52 PM by miklos

Hello Abraxalito,

I have made a Dac using one TDA1543 (PSU 5V) and decided to try to unconnect Pin 7 and to feed the I/V resistors with a 3V2 virtual ground.

However, it doesn't work !
Well,it is working, but I get a very weak level.

When I connect back a 890 Ohms resistor between pin 7 and ground and the I/V resistors back to ground, it works fine.

What am I missing ????

Thank you for your help here.
Best regards,
Jean

Posted 7th May 2017 at 02:04 PM by Bidule Bidule is offline

	Hi Jean How did you create your 3.2V virtual ground? And you still use the same I/V resistors?
	Posted 8th May 2017 at 02:40 AM by abraxalito

	Hello Abraxalito, Thank you for taking care ! Yes I tried with the very same I/V resistors. Here is my Vref schematic, I use a TL431 In the mean time, I have read on a post that it might need to encrease Vref, right ? I don't understand how I calculate the current through the I/V resistors when using Vref. Thank you for your help. Best regards, Jean
	Posted 8th May 2017 at 09:20 AM by Bidule

	A quick look at the TDA1543A datasheet shows that with 890R from pin7 you'll get about 2.5mA out of pin7 through that resistor which translates to 5mA of bias. With 5mA of bias your I/V resistors should be 640R to get 3.2V as the maximum positive excursion. The negative going limit will be 1.7V. So you have 1.5V p-p or 500mVRMS.
	Posted 8th May 2017 at 01:32 PM by abraxalito

	Hello Abraxalito, So, it looks like it should work ok the way I did it and it isn't... Really don't understand why ! Best regards, Jean
	Posted 8th May 2017 at 02:00 PM by Bidule

Well, Indeed, what you explain is what I should get with 890 resistor connected between Pin 7 and ground.

When I do this exercice, with my R I/V of 690 Ohms, if I adjust Rref by hear (and scope) I get a resistor value between 820 and 1K05 Ohms. Ideal should be aroud 900 Ohms.

So let's compute it the way you did:
with R I/V = 690 Ohms, 3,2V Max gives 4,64 mA
That makes an I bias value of 2,32 MA (A=2)
Wich, in turn gives the Rref value (2,2/2,32)=948 Ohms

Ok, this is coherent with what I have found by ears. But all this is true with pin 7 connected to ground through a resistor Rref.
This translates into 4,64 - 2,3 = 1,14 mA min => Vmin=0,786 V
And this gives me an output of 2,4V pp ?

The data sheet also says that Rref is optionnal.
One can also use Vref to bias the I/V opamp for instance.

But here, what you did, and what I'm trying to do is to leave Pin 7 unconnected and to make a virtual ground.
So, my question is what happens when a voltage source is connected to R I/V ?
I understood that, in that case, the current goes from the virtual ground through the I/V resistors and flows into the DAC.

How can I calculate the I/V resistors then ?

Let say V VirtualGround is 3V4, how does it translates in bias current ?
Does it mean that across R I/V I have 3V2 - 3V4 = -0,2V max ? Which translates into a I bias of -0,145 mA

And with digital silence, I woud have x - 3V4 = - ?
And in fact the Vpp would be inverted and independant of R I/V ? The only effect of R I/V would be the bias current ?

Please, I would appreciate your help here.

Thank you.
Jean

Posted 8th May 2017 at 05:34 PM by Bidule Bidule is offline

Updated 8th May 2017 at 08:05 PM by Bidule

	Hi Jean, I haven't forgotten about this, it needs a chunk of my attention - working on a DAC design of my own right now.
	Posted 10th May 2017 at 01:51 AM by abraxalito

New To Site?	Need Help?
Register to Participate Search Privacy Statement Contact Us	Frequently Asked Questions Did you forget your password? Mark Forums Read