Output level of TDA1545

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

Great, we can now say that "voltage at I ref"=1,67V is important to keep us at the sweetspot of Ifs=2mA.

In the case of Vref for IOL/IOR we just have to keep within the previously mentioned limits. From ca 1,9V up to ca 3,8V with center at ca 2,9V when using 100-150ohm Rload.
 
The reference voltage does get applied to some switched circuitry, based on internal design. Also, a shunt can be made to have very low noise and very good ripple/noise rejection.

So not the "standard" TL431 application, but the little enhanced one (with c3 to set HF gain to unity to improve noise)? But I never saw you implement it this way..: Simple Voltage Regulators Part 1: Noise - [English]
Or with a led instead of zenner in this one:
Simple Voltage Regulators Part 1: Noise

Thank you very much for the explanations about the 861! I will make a small PCB for it in a week or two (busy week now). We might get a nice new NOS dac project out of this ;-) I always felt the 1545 was not well represented between all these 41 and 43 projects...

Since the 1545 is CMOS, no I2S attenuators here but just some series-R's?

Thanks,
Berny
 
Hi,

So not the "standard" TL431 application, but the little enhanced one (with c3 to set HF gain to unity to improve noise)? But I never saw you implement it this way..

Well, the 431 has uses. But it is a bit noisy for really critical stuff. Sorry, really cannot say any more.

And BTW, where I use 431's and have a PCB I always add the cap to give unity gain at AC, usually way oversized too... I only omit it if I make a handwired "adjusted Zenner Diode" with a 431 and two SMD capacitors to fit instead of a small Cap on existing PCB's or for hardcore hardwiring.

Since the 1545 is CMOS, no I2S attenuators here but just some series-R's?

Yup.

Ciao T
 
Hello Thorsten,

I started working on a layout for the TDA1545, it will include the Salas shunt power supply to power the 1545 and probably a OPA2211 as two unity gain buffers to both feed the Iref-pin directly and generate Vref at 2/3Vcc for the OPA861. The OPA861 I/V will have its own shunt regulator (6.6V). The concept from Pedja you posted makes me doubt about how to continue after the OPA861...

The discrete I/V from Rbroer I used, was well improved by using a B1 (2SK170 source follower biased by a 2SK170 CS and all of this powered by its own symmetrical shunt reg) to buffer its output. What is your experience in using the OPA861 just as a buffer, compared to a well implemented source-follower?

The plan was to stay in SMD :) (except the power supply) and using BF862 JFETS that would be very similar to the 2SK170 as a buffer à la B1, instead of a second OPA861.

Thanks,
Berny
 
Hi,

I started working on a layout for the TDA1545, it will include the Salas shunt power supply to power the 1545 and probably a OPA2211 as two unity gain buffers to both feed the Iref-pin directly and generate Vref at 2/3Vcc for the OPA861. The OPA861 I/V will have its own shunt regulator (6.6V). The concept from Pedja you posted makes me doubt about how to continue after the OPA861...

Sounds like a plan. Why not breadboard this?

What is your experience in using the OPA861 just as a buffer, compared to a well implemented source-follower?

TBH, I never extensively compared this.

I am lazy. I get very good sound using Tubes the easy way, getting good sound from solid state is hard work.

Ciao T
 
Hi,
Sounds like a plan. Why not breadboard this?

I believe that for the 1545 I already have my test-platform, but you made me think... I will first try a couple of power-supply topologies. I just tried to have a feeling of noise on VCC with my prehistoric scope and was very surprised to see the 50Hz test signal I was using on Vcc with 5mVpp. At about 250Hz it is still 2mVpp and it further lowers with frequency. The decoupling cap is 470uF. There must be someting wrong somehow.. some additional work is needed. For the measurement I soldered a small coax directly to the power supply connection I made on the board. Beside that, I don't see a lot of noise.. (within my measure capability of maybe 1mV :) )

I will take out the shunt regulator to test it separately and I will also make a basic TL431 + CS to compare.

Regards,
Bern
 

Attachments

  • predac.jpg
    predac.jpg
    580.3 KB · Views: 488
  • dac.jpg
    dac.jpg
    487.9 KB · Views: 467
  • Scope 50Hz 1545 Vcc.jpg
    Scope 50Hz 1545 Vcc.jpg
    344.6 KB · Views: 443
  • Scope 250Hz 1545 Vcc.jpg
    Scope 250Hz 1545 Vcc.jpg
    394.2 KB · Views: 419
Last edited:
Thorsten,

What would be an acceptable noise/ripple on the VCC and Iref-pin?

I found the schematic on the site from Walt J. of a rather simple differentian probe using a SSM2017.. I guess it should be good enough as a start, what do you think? I believe it has a gain of 100, so I would see 50 -10 uV...

Berny
 
Hi,

What would be an acceptable noise/ripple on the VCC and Iref-pin?

Well, the Iref pin has NEGATIVE PSRR, so noise needs to as low as you can. I use discrete shunts there that are well below 0.2uV.

The Powersupply shows 30dB supply noise rejection, so with 2V out and aiming at ~ 100dB suppression we are talking at least -76dBV, round up to -80dBV supply noise, that is no more than 100uV.

A 431 has around 100nV|/Hz Ein at 1KHz and around 200nV |?Hz at 10 Hz. I'd estimate this as 18uV.

So 431 shunt with a ccs should deal with noise well enough. For reference you need a lot more filtering from that though.

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