I assume you resolved this by issuing a mute command to the DAC before switching modes?
Last edited:
There have been reports of problems with playback of dsd files with the Topping D50 ( dual sabre 9038's) which has just started shipping in quantity. Suggests that this dac chip is not easy to implement properly.
Review and Measurements of Topping D50 DAC | Page 19 | Audio Science Review (ASR) Forum
Review and Measurements of Topping D50 DAC | Page 19 | Audio Science Review (ASR) Forum
The way it's been described it sounds like a thermal problem with the DAC. Given the pics I've seen of the D50 I wondered whether that might be a problem given that there's apparently no thermal pad to conduct heat from the chip to the case. If that's what it is it'll be an easy fix.
From the pics I saw ( Review and Measurements of Topping D50 DAC | Page 15 | Audio Science Review (ASR) Forum post 293) the board&components sits very close to an internal plate in the casing , which is claimed to be milled from a solid billet of Aluminum . So it would allow for thermal pad placement or cutouts for heat generating components. It is a very tight casing, probably too tight.
Exactly. I think it was designed with both thermal conduction and EMI reduction in mind but unless there's an appropriate conduction pad (and properly sized metal stack) it won't work well and will actually make localized thermal problems much worse due to lack of airflow in the case. It's also possible that even if the ES9038 has an exposed thermal pad on the bottom that Topping didn't connect it to an appropriate plane.
I know that the thermally conductive silicone pads come in a wide range of thickness, but people will be better off using a thin one and filling the gap with a chunk of aluminum and using JB Weld or something to adhere that to the case (or just have a thin pad on both ends and use some other technique to keep it in place). Seems like there was some miscommunication between the PCB guy and the machinist on the case design.
I'm going to post a suggestion in that other forum if I can find it....I think it was on Amir's review of the D50, right?
Given the small size of the ES9038, if there's enough of a gap between the chip and the case the simplest fix would be putting a flat head machine screw or hex bolt in the case directly above the chip. That way one could easily adjust the gap and the pressure on the silicone pad. If there's not enough room as is then one could drill or mill out part of the case in that region since the case looked pretty thick there.
Of course the odds are high that assembly simply screwed up and forgot to put in the thermal pads to begin with, but given this pic it would have to be a pretty darn thick pad to allow good contact for both DACs while allowing clearance between the case and other components....
An externally hosted image should be here but it was not working when we last tested it.
news..
Katana is fully working with our new isolator. (it will be cheap and designed to work with Katana/Boss)
THD+N was measured today at -105 A weighted 1khz odbfs
H2 is at -118 , h3 at -125 , -129 , -128, 128 , 133 , 139..
Testing phase is over (over 7 weeks of testing and improving), last changes to PCB are finalized (based on above testing...mmm we found a few improvements) and PCBs will be sent to PCB house next week (end of next week) ...for a limited run.
I would estimate max 8 weeks for mass production.
We are debating one thing... how to implement THD+N compensation (a really cool feature of the new gen ess IC). All manufacturers take one board , measure and program the parameters of that board for all boards. Of course this is an approximation and harmonics will suffer for most boards.
The right approach , is to take each board and measure the harmonics while programming the right register..so one by one( Benchmark does it individually). However that will introduce a lot of engineering time and will increase pricing. (30$ minimum)
Suggestions ?
Katana is fully working with our new isolator. (it will be cheap and designed to work with Katana/Boss)
THD+N was measured today at -105 A weighted 1khz odbfs
H2 is at -118 , h3 at -125 , -129 , -128, 128 , 133 , 139..
Testing phase is over (over 7 weeks of testing and improving), last changes to PCB are finalized (based on above testing...mmm we found a few improvements) and PCBs will be sent to PCB house next week (end of next week) ...for a limited run.
I would estimate max 8 weeks for mass production.
We are debating one thing... how to implement THD+N compensation (a really cool feature of the new gen ess IC). All manufacturers take one board , measure and program the parameters of that board for all boards. Of course this is an approximation and harmonics will suffer for most boards.
The right approach , is to take each board and measure the harmonics while programming the right register..so one by one( Benchmark does it individually). However that will introduce a lot of engineering time and will increase pricing. (30$ minimum)
Suggestions ?
Last edited:
Simply make programming of a specific board optional for additional cost. I would personally spend those $30 for additional tuning.
Hi Tim
first let me publicly say thx you for your great work. Its appreciated by the community and by myself . Katana will be sent for your testing from the first units.
Driver is almost ready.Will be submitted next week to mainline linux . You can change filters , volume etc. I will publish a doc on Monday from the main engineer.
first let me publicly say thx you for your great work. Its appreciated by the community and by myself . Katana will be sent for your testing from the first units.
Driver is almost ready.Will be submitted next week to mainline linux . You can change filters , volume etc. I will publish a doc on Monday from the main engineer.
"Less is more" >> "Less can be more"
...if you havn't done your homework. That's the Allo message.
We'll see.
Soundcheck , I was not so blunt , but you understood exactly what I meant.
Ok lets see..
I expect Katana to blow out of the water ALL dacs under 1k .
Thanks for the nice compliment! Much appreciated. moOde users have been asking about Katana
What about power supply(s)?
We are working very hard on the Nirvana
. testing is starting tomorrow ( we had a 24h delay). I will give you more details in a week...but basically its the " best 5V smps in the world" (testing pending
. I keep my fingers crossed, but to explain briefly , we designed the Nirvana like its the analog stage of a DAC...each component was placed like it was an analog component. Nirvana is not a SMPS its a analog stage that happens to be a power supply.Ioan,
I think it is great that you are even considering this level of final setup to make each unit shipped sound as good as it can. Def Kudos!!!!
I suggest doing some in-house checks first... compare tweaked boards to un-tweaked both via measurement and listening and publish your results here. Then consider doing each board for the 1st run, but also send out some un-tweaked boards to a few for further listening comparisons. The difference MAY be so slight that it won't be worth it long-term. AND it may be revelatory. I suspect the former, as you have said repeatedly, implementation is everything and you are really focusing on getting that as right as you can within your design and cost space.
But only one way to tell.
My 2 cents.
Greg in Mississippi
Hi Greg
We did test tweaked vs non. Fom machine (audio precision) point of view the difference is clear. 112 h2 on non vs 118-119 h2 on tweaked.
We will try doing some blind tests and also talk internally about what we can do to keep Katana in a sensible price range , while delivering great SQ .