sangram said:
Yep, the mods are a result of reading head-fi threads
My system is DC coupled too and I did not have the slightest form of DC on the card output.
Local supply bypass for OPAs would be great if you can find a ground point around the OPAs (try measuring with a multimeter). I usually use a 100uf cap on the opamp itself, (between the + and - rails).
Ryssen said:
It's not a schematic, it's a power supply schematic. By the way, the actual schematic is a near-to-identical copy of the reference board designs from Cirrus and AK for the analog board. You can dig those by searching the corresponding parts for the DAC and ADC on google. DAC is CS4398 and ADC don't remember, it's easy to read on the board though:
Output stage: I use 100k input on the next stage (directly poweramps) and there is no problem, when using 22R resistors in order to isolate from interconnect capacitance. It won't drive long balanced lines or headphones, but it's more than enough for load impedances greater than say 10k.
Well, in my book the CLC approach has most of the sonical benefits of an expensive regulator. I took the out of the reg, desoldered it from the board and put a cap followed by an inductor right there to form a CLC filter with the existing cap on the board.
Another thoughts are - in the digital board: change the 3.3 Vd cap to OSCON and the 5V (from PCI) cap to OSCON too (C80 & C39). You can use 6.3V caps in here.
Yes one could use better caps of the same values.
AND REMOVE THE DAMN MUTING TRANSISTORS TOO
lucpes said:
Yep, the mods are a result of reading head-fi threads
My system is DC coupled too and I did not have the slightest form of DC on the card output.
Local supply bypass for OPAs would be great if you can find a ground point around the OPAs (try measuring with a multimeter). I usually use a 100uf cap on the opamp itself, (between the + and - rails).
Good to know about the DC offset.
No there isn't a ground point less than an inch away from the opamps, I already checked.
I was thinking of connecting a cap from pin 4 to 8, should work just as well as the opamp is operating on a floating ground anyway - just the way you've done, but with a small 1.5uF Tantalum cap or .47uF MKP instead of a big cap.
It seems the card already has pretty solid power and ground planes so the extra bypassing may be superfluous, but should strike while the iron's hot!
I'm not planning on any changes on the digital board except the caps that feed the analog section supply. Plus a foil shield for the cable between the boards, and a few other mechanical tricks.
I thought the 49720 and LM4562 were the same opamp, except the name - I could be wrong though, but there was a discussion on these boards about this...
Edit: yup here it is:http://www.diyaudio.com/forums/showthread.php?s=&threadid=107344&highlight=
I already finished two mods, and warped one from the heat and tore off two tracks on the other. Both work again with some patchwork, so I'm a little more confident of chips. My trouble is the cap holes actually, there's never a way to get all the solder out clean enough.
Should be OK this time - third time lucky?
Should be OK this time - third time lucky?
Solder wick with enough flux usually gets it clean. http://www.seattlerobotics.org/encoder/200002/mvc-010f.jpg
Cut the old opamp legs prior to desoldering or just heat 4 on one side at a time using a decent amount of solder to do it...
Cut the old opamp legs prior to desoldering or just heat 4 on one side at a time using a decent amount of solder to do it...
Thanks guys. I do use a desolder wick, sometimes wetting the joint again to get a good and melted solder flow.
Desolder wick leaves behind this gummy residue which even acetone has difficulty taking out.
Sometimes I'll be lucky and it leaves a neat hole.
@ryssen: I wouldn't dare using a drill, no matter what. I'm just too clumsy and my eyes aren't that great any more. I use a small length of guitar string - that doesn't cling to the solder and is just the same width as a component lead.
Desolder wick leaves behind this gummy residue which even acetone has difficulty taking out.
Sometimes I'll be lucky and it leaves a neat hole.
@ryssen: I wouldn't dare using a drill, no matter what. I'm just too clumsy and my eyes aren't that great any more. I use a small length of guitar string - that doesn't cling to the solder and is just the same width as a component lead.
Actualy,this is the mod that made most difference..and I have opamps in class A-> removed muting transistors -> LM4562 opams -> changed nearly all e-lyts -> next to try is output direct from DAC,if that works better that opamps ----> No ned to do a separate PSU for them....
Edit: I have the analog card in box outside PC.
Edit: I have the analog card in box outside PC.
Hi Sangram,
You can remove the " ..leaves behind this gummy residue .." with NC paint thinner which we get in our hardware stores.
Great solvent to clean up a pcb after soldering.
However you might have to buy from a good store to ensure it doesn't leave a white deposit. The deposit can be rubbed off in any case.
You can remove the " ..leaves behind this gummy residue .." with NC paint thinner which we get in our hardware stores.
Great solvent to clean up a pcb after soldering.
However you might have to buy from a good store to ensure it doesn't leave a white deposit. The deposit can be rubbed off in any case.
Dc?
Hi guys
I did the recapping of the analog board and changed out the opamps. Will report back in some days about the sonic changes.
Because the system is totally DC-coupled, I changed the coupling to 47uF FCs but soldered to be easily removable (on reverse of board) so I can switch for BG NX soon.
One Observation: Each capacitor +2.5 DC on its input leg. The DC would pass through the output stage as is. I thought about it and realised that the opamps inside the CS4398 must be powered off the +5 analog voltage rail and the output is exactly at half-bias.
So if your system is DC-coupled and single ended (balanced systems will AFAIK ignore the DC as it is very similar on both outputs) you will have to deal with a few volts of DC at the input if you short the coupling caps.
Hi guys
I did the recapping of the analog board and changed out the opamps. Will report back in some days about the sonic changes.
Because the system is totally DC-coupled, I changed the coupling to 47uF FCs but soldered to be easily removable (on reverse of board) so I can switch for BG NX soon.
One Observation: Each capacitor +2.5 DC on its input leg. The DC would pass through the output stage as is. I thought about it and realised that the opamps inside the CS4398 must be powered off the +5 analog voltage rail and the output is exactly at half-bias.
So if your system is DC-coupled and single ended (balanced systems will AFAIK ignore the DC as it is very similar on both outputs) you will have to deal with a few volts of DC at the input if you short the coupling caps.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.