Thank you for the appreciation of our efforts. I think we designed a very good DAC. I hope you will enjoy it. I haven't received my final PCB's yet so my fingers are itching...
Keep to the BOM and compare it with the best DACs you can compare it with but only do so after the ES9023 DAC has been used for some time. It will improve in time. Call it running in (or whatever the reason might be): it just will perform better after some time.
Keep to the BOM and compare it with the best DACs you can compare it with but only do so after the ES9023 DAC has been used for some time. It will improve in time. Call it running in (or whatever the reason might be): it just will perform better after some time.
more build notes
Today I built a better 5v power supply for the DAC. I used an 8V 0.8A transformer, rectified with soft recovery diodes, and fed this to an LM317 to get ~ 8.5v. I then fed the 8.5v to a cap multiplier circuit. Current source feeding a resistor to generate ~7.5v which drives the gate of an FET with a Vgs of 2.5v. This puts out 5v with very low noise.
I listened to this for a couple of hours and I'm still letting it burn in. I did make one quick change from the BOM after this initial listen. I replaced the 4.7nf caps at the output, which are Vishay polyester caps, with some Wima 4.7nf polypro caps - Mouser part no. 505-MKP24700/630/5. I think this got rid of a trace of harshness in the highs.
Next, I may try building an FET buffer and LC filter as suggested by J. Gerhard in another thread. I'm curious to know how that will affect the sound. But I'll let the current version play for a while before any further experimentation.
---Gary
Today I built a better 5v power supply for the DAC. I used an 8V 0.8A transformer, rectified with soft recovery diodes, and fed this to an LM317 to get ~ 8.5v. I then fed the 8.5v to a cap multiplier circuit. Current source feeding a resistor to generate ~7.5v which drives the gate of an FET with a Vgs of 2.5v. This puts out 5v with very low noise.
I listened to this for a couple of hours and I'm still letting it burn in. I did make one quick change from the BOM after this initial listen. I replaced the 4.7nf caps at the output, which are Vishay polyester caps, with some Wima 4.7nf polypro caps - Mouser part no. 505-MKP24700/630/5. I think this got rid of a trace of harshness in the highs.
Next, I may try building an FET buffer and LC filter as suggested by J. Gerhard in another thread. I'm curious to know how that will affect the sound. But I'll let the current version play for a while before any further experimentation.
---Gary
Hi Gary, did you change the caps isolated from other mods ? I ask this because the Vishay caps are of known good quality. With modern film caps some excellent types are made (be it polyester/MKT or not). It might very well be the better power supply that improves the DAC (just like it should). I think you assume the power supply with LM317 is low noise or did you measure it ?
Of course experiment as much as you like but to be sure what is better or not really 2 DACs or more should be made. I only experienced "traces of harshness" when running the DAC in or when it is first powered on. I made quite a few and to my surprise I had very good results with OSCON for decoupling (the small 22/16 ones with blue sleeve and copper wires). Of course it was not a surprise BG NX HiQ 47/6.3 is a winner in this DAC ! Apart from this I would encourage builders to keep to the BOM. That is if you like a guaranteed known very good outcome. The power supply is coming to production and I think it is very well suited for our DAC.
Of course experiment as much as you like but to be sure what is better or not really 2 DACs or more should be made. I only experienced "traces of harshness" when running the DAC in or when it is first powered on. I made quite a few and to my surprise I had very good results with OSCON for decoupling (the small 22/16 ones with blue sleeve and copper wires). Of course it was not a surprise BG NX HiQ 47/6.3 is a winner in this DAC ! Apart from this I would encourage builders to keep to the BOM. That is if you like a guaranteed known very good outcome. The power supply is coming to production and I think it is very well suited for our DAC.
Last edited:
Hi Patrick. You use ES9022, AFAIK different regs, USB/I2S input and a completely different PCB so comparing situations is comparing pears with apples. However the FKP are better than the types we use in the BOM but they have a smaller footprint. I did build a few with 5 mm Wima polypropylene caps but the Vishays were slightly better IMO. This was by coincidence as Subbu sent me those.
Please let's not mix up 2 designs in this thread. Our design is a relatively low budget "open" group buy effort with a BOM for a large audience of mixed experience and we would like to keep this thread clean from having to answer every question on parts. Just suppose a few hundred people all using different parts asking us for the outcome....Probably the reason your design and PCB are less easy to obtain. No pun intended.
Please let's not mix up 2 designs in this thread. Our design is a relatively low budget "open" group buy effort with a BOM for a large audience of mixed experience and we would like to keep this thread clean from having to answer every question on parts. Just suppose a few hundred people all using different parts asking us for the outcome....Probably the reason your design and PCB are less easy to obtain. No pun intended.
Last edited:
You may haver tested all stuff in the world with your group but this is pollution of this thread. You very well know what our DAC is and what is yours. Our PCB's have been sent to the participants and some of us have received them and some are still awaiting their package. We should wait till the DACs are built with the BOM as a leading point. That way many participants will have a very good product. Modding/changing stuff in this stadium is confusing and there is no time or wish to advise on individual basis. I think this is quite logic.
We experimented with caps otherwise we would have used the ceramic ones at the output
We use a BOM to give the many participants a known good outcome. In fact we provide a low priced PCB design of good quality, a schematic and a BOM to build a very good DIY product for a good price. No mystic properties here, no open ends, just a very good product.
If you like to advise on our design you are a bit late. I suggest a new thread: "improving the Subbu/Jean-Paul ES9023 WM8804 DAC" I can sent you one if you like as I value your opinion. In fact it would be interesting to compare DACs.
We experimented with caps otherwise we would have used the ceramic ones at the output
We use a BOM to give the many participants a known good outcome. In fact we provide a low priced PCB design of good quality, a schematic and a BOM to build a very good DIY product for a good price. No mystic properties here, no open ends, just a very good product.If you like to advise on our design you are a bit late. I suggest a new thread: "improving the Subbu/Jean-Paul ES9023 WM8804 DAC"
I can sent you one if you like as I value your opinion. In fact it would be interesting to compare DACs.
Last edited:
Jean-Paul,
I did make the changes one at a time. First I upgraded the power supply and listened for a couple of hours. Then I switched the output caps and listened again for many hours. In my system I still like the Wimas better but it's a matter of personal preference.
I agree that the best way of evaluating changes is to build 2 DACs and change one at a time. I actually do have 2 working boards, but I didn't use the 2 boards for this comparison. But I will use it to compare the sound of adding a buffer.
Regarding the noise, I checked it out with my scope, but it's not sensitive enough to see down to the uV level that you're aiming for with your power supply. But there are many articles on the web showing that a regulator such as the LM317 followed by a gyrator/cap multiplier has very low noise. So I am trusting in the results of others and haven't measured that directly on mine.
One final question - is it OK to have discussions like this in this thread? You came down pretty hard on EUVL for offering up parts suggestions. Should we be moving discussions related to anything other than the "official BOM" elsewhere?
Thanks,
---Gary
Again we must have a kind of "standard" to be able to guarantee (sort of) a good result. That does not mean one can not experiment !
I was too hard on Patrick (EUVL) and I regret that but we put a lot of time in this and IMO it is way more handy to await the builders having their DACs finished following the "standard" before changing things. If I could still change my text I would do so. Sorry Patrick I guess I am a bit overreactive regarding our design. Just think of disappointment from our side when chaos kicks in and people start posting the DAC sounds like garbage....
Anyway, do anything to make your DAC sing even better but I think it is more practical to open a new thread. I will contribute as well. Here is a new thread:
http://www.diyaudio.com/forums/digital-line-level/216223-improving-es9023-spdif-dac.html#post3093127
I was too hard on Patrick (EUVL) and I regret that but we put a lot of time in this and IMO it is way more handy to await the builders having their DACs finished following the "standard" before changing things. If I could still change my text I would do so. Sorry Patrick I guess I am a bit overreactive regarding our design. Just think of disappointment from our side when chaos kicks in and people start posting the DAC sounds like garbage....
Anyway, do anything to make your DAC sing even better but I think it is more practical to open a new thread. I will contribute as well. Here is a new thread:
http://www.diyaudio.com/forums/digital-line-level/216223-improving-es9023-spdif-dac.html#post3093127
Last edited:
It is OK.
I was not upset by any means. But there is no point for any more comments till things have calmed down.
If you go back to my previous posts, all I suggested was ONE cap for you to try.
It is size and pin compatible, and costs less than the postage you have to pay.
If some of you might like it, then keep it.
If not, desolder it and change back to Vishay MKT.
No real harm done.
I can only recommend again that you give it a try.
I like it.
And yes, I like polystyrene too.
But they are a bit bulky and not pin compatible.
It's only hobby. We should all have fun.
Cheers,
Patrick
I was not upset by any means. But there is no point for any more comments till things have calmed down.
If you go back to my previous posts, all I suggested was ONE cap for you to try.
It is size and pin compatible, and costs less than the postage you have to pay.
If some of you might like it, then keep it.
If not, desolder it and change back to Vishay MKT.
No real harm done.
I can only recommend again that you give it a try.
I like it.
And yes, I like polystyrene too.
But they are a bit bulky and not pin compatible.
It's only hobby. We should all have fun.
Cheers,
Patrick
It is TWO caps You still don't get my point but it is not important.
I like polystyrene/styrol/KS caps too but they're large, fragile and they're not pin compatible as Patrick pointed out. I would advise not to use types rated higher than 63 V. The smallest you can get your hands on would be the best bet. They probably will pick up noise from the charge pump circuit and/or the 50 MHz clock. I have a lot of them but will not try them in this application. The smaller caps the better.
The Siemens/EMZ orange colored molded 1 % KS types would have been a good choice as the outer foil is marked (so could be connected to low impedance point) which makes them less susceptible for stray signals from the charge pump or the XO. They are out of production AFAIK. Still have a lot of those but no 4.7 nF ....
You still don't get my point but it is not important. I like polystyrene/styrol/KS caps too but they're large, fragile and they're not pin compatible as Patrick pointed out. I would advise not to use types rated higher than 63 V. The smallest you can get your hands on would be the best bet. They probably will pick up noise from the charge pump circuit and/or the 50 MHz clock. I have a lot of them but will not try them in this application. The smaller caps the better.
The Siemens/EMZ orange colored molded 1 % KS types would have been a good choice as the outer foil is marked (so could be connected to low impedance point) which makes them less susceptible for stray signals from the charge pump or the XO. They are out of production AFAIK. Still have a lot of those but no 4.7 nF ....
Last edited:
Hi Jean Paul & Subbu
Please clarify the point concerning Balanced DAC version.
According to me you need 2 DAC PCB populated as described in Subbu's guide. True or not ?
Any chance for PSU BOM ? GB is actually freezed ...
I know your're doing your best
You need 2 x DAC PCB for balanced operation and only 1 power supply PCB.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.