ESS ES9008/9018 Sabre dac

[Paul Hynes]

Hi Leo,

Thanks for the welcome. Pleased to help the project.

The 3v3 and +-15v shunt regs will be the same type I fitted to the Altmann.

You will need to increase the input voltage of the 3v3 reg from 6v5 to 8v5 to allow for the regulator current source dropout voltage. I will set the available current to 250ma to give a little headroom.

The +-15v regs will need to see an input voltage of 20 volts and I can set the available current to 100ma.

Can you let me know the VA ratings of your mains transformers so I can assess their current capability?

The 1v2 series reg (three terminal drop-in replacement) would run pretty hot (too close to ratings for comfort) in this position so I will have to devise a different solution here. I do have a high performance discrete component series reg circuit that will that will operate down to 1v2, but will need to update the circuit board layout before you can have one of these. This will probably take a week or two to get ready.

I will need to know your postal address which you can e-mail to me at paul@paulhynesdesign.com

Regards
Paul

[leo]

Hi Paul,

Well I hope you can stick around when you get time

Altering the pre regs is no problem

The transformers specs are

2 x 0-18v 2 x 2.22A
2 x 0-9v 2 x 555mA

Before the pre regs the voltage drop supplying the 3.3v (160mA ) LDO's is 8.9v (regulated down to 6.5v) and the voltage drop before the pre regs supplying the 1.2v 56mA LDO's is 11v
These voltages are measured after a 10R resistor on the end of CLRC filtering before the pre regs

If the transformers aren't suitable I can see what others I have, also I presume using pre-regs before your shunts are ok and won't effect their performance ?

No problem regarding the 1.2v, I'd have thought the 3.3v is as important as any seeing as though it supplys the XO, comparator and dacs VD

Mail sent

Cheers,
Leo

[Marco]

Hi Paul,

Welcome to AOS

I have upgraded your account to trade member (Manufacturer) status to reflect your business position. Would you like your website http://www.paulhynesdesign.com/ featured in our Links Box shown below?

http://theartofsound.net/forum/showt...?p=123#post123

Marco.

[Paul Hynes]

Hi Leo,

Your power supply will be fine powering the shunt regs.

Pre regulation is more likely to be of benefit as less interference will be heading for the post regulation.

Hi Marco,

Thanks for the welcome. I would be happy if you include my web address in the Links Box. I do not have a logo at this time.

Regards
Paul

[Marco]

No problem, Paul. I'll organise it later

Marco.

[leo]

Thanks Paul, I'll adjust the pre regs output when needed

Leo

[leo]

I've re-drawn out the schematic for part of this dac module to give a better idea, it seems one socket supplies the 3.3v LDO reg, the LDO reg then supplies the XO,comparator, the 1.2v LDO reg for the core and the non inverted input of a LM4562, the 3.3v outputs of the LM4562 supply the dacs VREF left and right

It seems the VA socket supplying the onboard LM4562 supply rail has quite an influence , I tried a simple zener,cap multiplier set for 7v output to feed the VA socket for the onboard LM4562, this sounded worse than the standard LM317 based reg.
I'd imagine the output impedance of the cap multiplier was too high

C8 1uf SMD cap is right next to the +v LM4562 pin

[leo]

The LM4562 is quite a decent op-amp in analogue circuits , its crap at driving capacitive loads so needs a resistor on the outputs and seems to require a very good supply to stop it from sounding nasty

The question is, is it the best type of op-amp to be used in the above circuit for a voltage supply?

[Paul Hynes]

Hi Leo,

I think that the LM4562 could be improved upon however, as this part is essentially providing a buffered 3v3 for the Sabre left and right reference pins, it may be worth trying a pair of mini shunt regs in place of the LM4562. I will send you two to try with the other regs later this week. I will fit bendy wire leads so you can manipulate them into place. The noise performance should be slightly better than the LM4562 but the more important parameters in this application, like slew rate, settling time and low output impedance over a much wider bandwidth will greatly improve on the LM4562 specs.

Regards
Paul

[leo]

Hi Paul,

Thanks mate, it seems the VA socket supplys the onboard LM4562 only, I was surprised how much effect that supply has.

Rather than remove the actual SMD LM4562 I was thinking it would be ok to remove R4 going to the non inverted inputs of the LM4562 and also not connecting any supply to the VA socket?
I could then run the shunts outputs to the 1uf SMD ceramics going to the Vref L & R

Cheers,
Leo

BTW, theres an update for the IVY input filter which Twisted pair posted, I tried it the other day and its slightly improved the sound mainly the upper midrange.
Heres a link to the updated circuit, obviously the 100k resistor across the differential op-amps is for the sim so those are not fitted to the actual circuit http://www.diyaudio.com/forums/attac...amp=1224629138