According to the ES9023 datasheet, pin 11 is described as "Negative Supply (Internally Generated)". I think you are right that this pin should be connected to the -ve of the C32 instead of +ve as in the pcb.
In a few posts back in the EU GB thread, JP mentioned that he liked the sound of replacing C32 with a BG 0.47uf NX. Since NX is bipolar, that is certainly better than the reversed cap.
I replaced C8 with a 4.7uF Tantalum but didn't hear much of a difference. I think it is time to reverse C32 and listen hear again.
In a few posts back in the EU GB thread, JP mentioned that he liked the sound of replacing C32 with a BG 0.47uf NX. Since NX is bipolar, that is certainly better than the reversed cap.
I replaced C8 with a 4.7uF Tantalum but didn't hear much of a difference. I think it is time to reverse C32 and listen hear again.
GaryB, congrats on finding an error. You are absolutely right about polarization of C32.
This was a non data sheet extra cap that I found to improve things but I used hardwired BG NX non polar for C32 at that time on the rev. 1 PCB's. I still do use BG NX for C32 on most of my DACs so I never found this error as these are non polar electrolytic caps. These are hard to find so they were changed to polar Pana FC caps in the BOM as many wanted a PCB. Just checked and I only have one DAC board myself with Panasonic FC 4.7 µF indeed soldered in reverse. Although it is still early in the morning I already replaced it for a new one correctly soldered.
In my experience the value can be lower than 10 µF for optimal results. 1 to 4.7 µF is slightly better subjectively as I don't have an explanation why it is like that. I use 0.47 µF BG NX now for my own use but I am running out of NX. I will contact Subbu to have documentation changed.
This was a non data sheet extra cap that I found to improve things but I used hardwired BG NX non polar for C32 at that time on the rev. 1 PCB's. I still do use BG NX for C32 on most of my DACs so I never found this error as these are non polar electrolytic caps. These are hard to find so they were changed to polar Pana FC caps in the BOM as many wanted a PCB. Just checked and I only have one DAC board myself with Panasonic FC 4.7 µF indeed soldered in reverse. Although it is still early in the morning I already replaced it for a new one correctly soldered.
In my experience the value can be lower than 10 µF for optimal results. 1 to 4.7 µF is slightly better subjectively as I don't have an explanation why it is like that. I use 0.47 µF BG NX now for my own use but I am running out of NX. I will contact Subbu to have documentation changed.
Last edited:
Power Supply PCB group buy update:
Please visit thread http://www.diyaudio.com/forums/grou...y-pcb-es9023-wm8804-s-pdif-dac-group-buy.html for latest information
Regards,
Subbu
Please visit thread http://www.diyaudio.com/forums/grou...y-pcb-es9023-wm8804-s-pdif-dac-group-buy.html for latest information
Regards,
Subbu
With a decent layout, 4 onboard regulators, beads and good decoupling it is already a good design from the start. Some of us desperately want to improve things that are already good. Maybe 5 to 10 % can be won by tweaking caps. There is not much left to tweak anyway 
Admittedly BG NX make a bigger difference. To the tweakers modding is a natural habit which may seem strange to others. This comes from seeing many factory made devices that are full of design errors which must be changed in order to have a good performing device. If you're not that type: don't change it and stick to the BOM. Please mind that the layout of the DAC you are building already has known at least 30 versions.... Maybe Subbu has counted them, I only kept the final versions. Anyway it were many versions.
If you stick to the BOM you will have a fine sounding en performing DAC. With hindsight I would advise on FKP/MKP/KS caps for the 4.7 nF output filter caps as they're better. If you're adventurous you could try out different caps but again: keep em small.
You know how it works in audio: 1% improvement = "it seems like a veil is lifted", "background is more black" etc etc..
It is a pity that C32 was printed wrong on the PCB. Totally my fault.
Admittedly BG NX make a bigger difference. To the tweakers modding is a natural habit which may seem strange to others. This comes from seeing many factory made devices that are full of design errors which must be changed in order to have a good performing device. If you're not that type: don't change it and stick to the BOM. Please mind that the layout of the DAC you are building already has known at least 30 versions.... Maybe Subbu has counted them, I only kept the final versions. Anyway it were many versions.If you stick to the BOM you will have a fine sounding en performing DAC. With hindsight I would advise on FKP/MKP/KS caps for the 4.7 nF output filter caps as they're better. If you're adventurous you could try out different caps but again: keep em small.
You know how it works in audio: 1% improvement = "it seems like a veil is lifted", "background is more black" etc etc..
It is a pity that C32 was printed wrong on the PCB. Totally my fault.
Last edited:
Jean-Paul,
I agree that it is a good sounding DAC and again thank you and Subbu for your efforts. Regarding C32, don't feel too bad about it. Everyone makes errors from time to time. But I do think it is important to let all the builders know about this, since it makes a bigger difference in sound quality - much bigger than anything else I tried. Perhaps it would be helpful to add a note on the BOM or an update to the build guide so that people don't miss this as it gets buried in the thead? Or maybe even an email to the buyers if you've got a distribution list already set up.
Thanks,
---Gary
p.s. I've got a few remaining Black Gate NX caps, so I'll be giving them a try. Have I understood you correctly in saying your preferred location and values are 4.7uf for C32/Vneg and 100uf for C10b/Vreg?
Hi after trying I use 47 µF 6.3 V NX HiQ for C10B and 0.47 µF 50 V NX HiQ for C32. I have some N series 100/6.3 that I want to try for C10B but then I want a N series 1/50 for C32. I have one (!) N series 1/50 coming my way as a gift from a member so I hope I will be listening to that combination soon. For C9 I have BG PK series 10/50 as it is not so critical at that spot it seems otherwise I would have used NX HiQ too. Both the 10 µF caps for the WM8804 are PK 10/50 as well in my DACs.
In the few DACs I built with Panasonic caps I used Pana FC 4.7/50 for C32 and Pana FM 100/25 for C10B and Pana FC 10/50 for C9.
In the few DACs I built with Panasonic caps I used Pana FC 4.7/50 for C32 and Pana FM 100/25 for C10B and Pana FC 10/50 for C9.
Last edited:
Just came home from a listening session. We were supposed to listen line conditioners and mains filters (PS Audio Duet/Quintet) but I decided to bring a newly finished "all BG NX" ES9023 DAC. For the output filters I used the standard MKT Vishays as I don't have ERO KS caps anymore. It was finished today so I left it on from noon and hoped for the best. The NX need much longer to be performing optimally. In the hurry I reverse connected the PS which is a very big "no no" in my book but it happened and luckily I realized it when I switched it on (and very fast off again). No damage, pffff
To make a long story short: our DAC had to compete with an older Krell SBP32X connected to NAT tube amps and Usher speakers.. I won't brag but the owner of the Krell gave our DAC a 8 to 9 on a scale of 10....This was a analog top quality set so there was a very nice gigantic turntable with a 40 kg spindle and analog was the winner of all (as expected).
To make a long story short: our DAC had to compete with an older Krell SBP32X connected to NAT tube amps and Usher speakers.. I won't brag but the owner of the Krell gave our DAC a 8 to 9 on a scale of 10....This was a analog top quality set so there was a very nice gigantic turntable with a 40 kg spindle and analog was the winner of all (as expected).
Last edited:
Promised not to brag but it certainly was better with the highs, lows and stereo imaging/depth. The Krell was slightly better with the mids IMHO. The Krell owner wanted me to build our ES9023 DAC for him. Others said both DACs were different but not one of the two was really much better or worse. We agreed that our hobby project was successful
Just think what happens when the NX have broken in and 2 x 4.7 nF MKP/FKP/KS will be used.
BTW I get annoyed by the "WM8804 ES9023 SPDIF DAC" as we call it. Let's call it the "Subbu DAC" from now on. I hope Subbu don't mind but I like the name much better. If you are around Subbu please let us know if you have objections of lending your name to our DAC
Just think what happens when the NX have broken in and 2 x 4.7 nF MKP/FKP/KS will be used.
BTW I get annoyed by the "WM8804 ES9023 SPDIF DAC" as we call it. Let's call it the "Subbu DAC" from now on. I hope Subbu don't mind but I like the name much better. If you are around Subbu please let us know if you have objections of lending your name to our DAC
Last edited:
No they are sold out. We get a lot of PMs and email regarding this subject and I can tell that it is tiring to have to answer all that mail. From today I will not answer such emails/PMs anymore.
This was risky business to get involved in as there was no guarantee to success ("who needs a PCB for ES9023 ?" style) and it was a one time only affair. Those that did not have fear went along and listen to the DAC now.
Those who played safe and waited played safe indeed as there will be no DAC to build Those that came too late are just too late. Life goes on, better DACs will be around.
This was risky business to get involved in as there was no guarantee to success ("who needs a PCB for ES9023 ?" style) and it was a one time only affair. Those that did not have fear went along and listen to the DAC now.
Those who played safe and waited played safe indeed as there will be no DAC to build
Those that came too late are just too late. Life goes on, better DACs will be around.
Last edited:
Hey Subbu is the COO of our risky business and normally other items are named after the COO...
Anyway we did a lot of wrongs and some rights and I suppose renaming a device when it is going EOL is not too clever either but "Subbu DAC" it is when the COO himself does not object.
One of my prototypes was called "spaghetti DAC" because it was wired... eh because I like spaghetti
Anyway we did a lot of wrongs and some rights and I suppose renaming a device when it is going EOL is not too clever either but "Subbu DAC" it is when the COO himself does not object.
One of my prototypes was called "spaghetti DAC" because it was wired... eh because I like spaghetti
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.