ESS Sabre Reference DAC (8-channel)

I'm looking a product brief of ES9039SPRO chip and finding that some pin assignments of the new chip do not match those of its predecessor ES9038PRO. I have a dual mono ES9038PRO DAC board and I want to replace two ES9038PRO DAC chips with new ES9039PRO chips. However, an special adaptor seems to be essential. Does anyone have any good idea for the chip replacement?
 
There may be other differences between the two chips, such as the I2C register assignments, new registers for added features, new register default settings, etc. Other than that, dac chips operating at high-ish RF frequencies require careful design of the ground plane and of bypassing around the chip. An adapter could add problematic inductance, to say the least. Might make a lot more sense to design a new board for the new chip.
 
  • Like
Reactions: 1 user
Member
Joined 2007
Paid Member
There may be other differences between the two chips, such as the I2C register assignments, new registers for added features, new register default settings, etc. Other than that, dac chips operating at high-ish RF frequencies require careful design of the ground plane and of bypassing around the chip. An adapter could add problematic inductance, to say the least. Might make a lot more sense to design a new board for the new chip.
Great answer. The control registers are completely different between the 9018 vs. the 9028/38, so if this is new silicon I'd expect that to be the case again. If you are using I2C, you could manage. But a mounting adapter is a poor approach, IMHO.
 
I'm looking a product brief of ES9039SPRO chip and finding that some pin assignments of the new chip do not match those of its predecessor ES9038PRO. I have a dual mono ES9038PRO DAC board and I want to replace two ES9038PRO DAC chips with new ES9039PRO chips. However, an special adaptor seems to be essential. Does anyone have any good idea for the chip replacement?
https://www.esstech.com/products-overview/digital-to-analog-converters/sabre-audiophile-dacs/
Looks like main specs hasn't changed from ES9038PRO.
 
Considering that the great gains in SQ between 9018 and 9028/9038 were mainly due to the new digital core (HyperStream II), it is very likely that the upgrade to HyperStream IV will also bring significant SQ improvements.

The reduction in power consumption is also welcome.. Even though the 9028/9038 also had heat pads and needed to be running at very high clocks to get hot (which was not necessary, in fact caused it to perform worse SQ-wise).
 
  • Like
Reactions: 1 users
Member
Joined 2007
Paid Member
I don't see how this news may influence that..
Agreed. One of the confounding factors behind the TPA audio 8-channel experiments seems to have been the cost and complexity of I/V conversion. That hasn’t changed for the latest ‘flagship’ 64 pin DACs. However, at a cost of some S/N, the ES9080 would eliminate that problem. I have no idea if SQ is up to TPAs high standards. It would need a new board since it is a 40-pin chip.
 
  • Like
Reactions: 1 user
Other than that, dac chips operating at high-ish RF frequencies require careful design of the ground plane and of bypassing around the chip. An adapter could add problematic inductance, to say the least.
Yes, very likely.
The most significant pin-assignment incompatibility is an inversion of polarity for some dc power input pins. I wonder why they have made such a change. This might be solved by adopting a tentative cross-wiring of DIY approach without any adapters.

By the way, does anyone have any idea or information how a new datasheet will be provided and when the new chip will become available for DIY users?
 
Agreed. One of the confounding factors behind the TPA audio 8-channel experiments seems to have been the cost and complexity of I/V conversion. That hasn’t changed for the latest ‘flagship’ 64 pin DACs. However, at a cost of some S/N, the ES9080 would eliminate that problem. I have no idea if SQ is up to TPAs high standards. It would need a new board since it is a 40-pin chip.
If that were the issue, they could just use an ES9028Pro and be done with it.. Then use the relatively affordable older I/V boards.

I think it's more likely that the market just wasn't there. That may be the reason that the multichannel Buffalo III was discontinued, in favor of the SE.
 
  • Like
Reactions: 1 user
Yes, very likely.

The most significant pin-assignment incompatibility is an inversion of polarity for some dc power input pins. I wonder why they have made such a change. This might be solved by adopting a tentative cross-wiring of DIY approach without any adapters.
It's not just a matter of taking out the old chips and soldering-on the new ones. They may have the same pin count but they use a different package. It is very likely that they are not physically compatible.
 
Agreed. One of the confounding factors behind the TPA audio 8-channel experiments seems to have been the cost and complexity of I/V conversion. That hasn’t changed for the latest ‘flagship’ 64 pin DACs. However, at a cost of some S/N, the ES9080 would eliminate that problem. I have no idea if SQ is up to TPAs high standards. It would need a new board since it is a 40-pin chip.
DAC can be used in V out mode without I/V but with a bit worse THD+N
 
Member
Joined 2007
Paid Member
If that were the issue, they could just use an ES9028Pro and be done with it.. Then use the relatively affordable older I/V boards.

I think it's more likely that the market just wasn't there.
Perhaps this is the other side of the same coin - a Buffalo 9028 plus 4 IVY boards comes (today) to $765 with no power supplies, which certainly would be a challenge to collect in our DIY marketplace. ...especially if the SQ was at all decreased by dividing the 9028 into single channels...
 
Come to think of it, even an 9038 in 8-channel mode would work with the older I/V boards. Each of its channels would put out as much current as an 9018 would, configured for stereo operation.

So that leads me to believe that there were no technical reasons for not releasing a multichannel Buffalo 9038 dac.
 
  • Like
Reactions: 1 user
Considering that the great gains in SQ between 9018 and 9028/9038 were mainly due to the new digital core (HyperStream II), it is very likely that the upgrade to HyperStream IV will also bring significant SQ improvements.

The reduction in power consumption is also welcome.. Even though the 9028/9038 also had heat pads and needed to be running at very high clocks to get hot (which was not necessary, in fact caused it to perform worse SQ-wise).
Not just running at high speeds. It depends. If you include the heat thrown off by the regulators, it can get quite toasty even on standard 45/49 XO. Yes, they had heat pads, bit small though. Think the QFN has larger vs size iirc. I prefer leadless packages these days too, s happy with that. Since i'm not hand soldering that much these days.
 
It's not just a matter of taking out the old chips and soldering-on the new ones. They may have the same pin count but they use a different package. It is very likely that they are not physically compatible.
Yes, you may be correct because we have no detailed datasheet of ES9039PRO yet.

However, as far as each document tells us,
ES9038PRO: https://www.esstech.com/wp-content/uploads/2021/03/ES9038PRO-Datasheet-v3.7.pdf
ES9039PRO: https://www.esstech.com/wp-content/uploads/2022/05/ES9039MPRO_ES9039PRO_Product_brief_v0.2.1.pdf
The two chips have the same size, 10 mm x 10 mm.

Anyway, we need more detailed information on ES9039PRO.