How low can you go?
The prices of DVD players just keep on going down here. Yesterday I bought my very first 99RMB [that's under $15] model in Carrefour. I couldn't wait to get home to see how it sounded.
First up I listened to one of my favourite BIS recordings - Mozart 40 in a version for piano, flute and strings. The sound really wasn't as bad as I'd expected - the cello sounded really 'reedy' rather than full-bodied and the flute was akin to a sopranino recorder in its higher register with very little bloom. But there was a fair approximation to a soundstage and really quite decent dynamics, not at all 'bitstream'.
Rebecca Pidgeon's 'Spanish Harlem' is the test disc I use to check for sibilance (it has lots of 's' sounds), and sure enough there was some but definitely not terrible considering the price.
This undoubtedly had potential for the price - I wondered if it could be improved. Inside almost all the functions are integrated on the MediaTek chip (although it has a heatsink, I know from previous incarnations this is an MT1389). So the DAC is onboard, no chance of souping that up. The output stage is based on a 4558 so I figured there was definite upgrade potential.
On tracing out the circuit (schematic to follow) I found that the output opamp provides just a bit of HF filtering, level shifting and gain. The DAC output is fed in through an 0603 cap which I measured as 1uF. Yuk! This must be a Z5U or something similar, totally non-linear and microphonic. So it needs to be substituted with an electrolytic and its also too small for decent phase performance in the bass. I chose a 4.7uF.
The rest of the changes I made are just based on the layout - a single groundplane is used over the whole board for 0V, the opamp is decoupled to this groundplane from both rails. The groundplane is also used as the reference voltage for the opamp - +ve input pin is directly connected to 0V.
Having removed the 4558 I bent up the two +input pins of the NE5532 I replaced it with and wired them direct to a ground on the MT1389. I also re-routed the grounding of the RC filter between the DAC and the opamp.
Lastly I found the feedback caps were a bit too large so they went down from 100pF to 50pF and the 1k series resistors on the output I decreased to 47R. The 5532 has more output oomph than the 4558 so can cope with a lower series isolation resistor.
...to be continued
First up I listened to one of my favourite BIS recordings - Mozart 40 in a version for piano, flute and strings. The sound really wasn't as bad as I'd expected - the cello sounded really 'reedy' rather than full-bodied and the flute was akin to a sopranino recorder in its higher register with very little bloom. But there was a fair approximation to a soundstage and really quite decent dynamics, not at all 'bitstream'.
Rebecca Pidgeon's 'Spanish Harlem' is the test disc I use to check for sibilance (it has lots of 's' sounds), and sure enough there was some but definitely not terrible considering the price.
This undoubtedly had potential for the price - I wondered if it could be improved. Inside almost all the functions are integrated on the MediaTek chip (although it has a heatsink, I know from previous incarnations this is an MT1389). So the DAC is onboard, no chance of souping that up. The output stage is based on a 4558 so I figured there was definite upgrade potential.
On tracing out the circuit (schematic to follow) I found that the output opamp provides just a bit of HF filtering, level shifting and gain. The DAC output is fed in through an 0603 cap which I measured as 1uF. Yuk! This must be a Z5U or something similar, totally non-linear and microphonic. So it needs to be substituted with an electrolytic and its also too small for decent phase performance in the bass. I chose a 4.7uF.
The rest of the changes I made are just based on the layout - a single groundplane is used over the whole board for 0V, the opamp is decoupled to this groundplane from both rails. The groundplane is also used as the reference voltage for the opamp - +ve input pin is directly connected to 0V.
Having removed the 4558 I bent up the two +input pins of the NE5532 I replaced it with and wired them direct to a ground on the MT1389. I also re-routed the grounding of the RC filter between the DAC and the opamp.
Lastly I found the feedback caps were a bit too large so they went down from 100pF to 50pF and the 1k series resistors on the output I decreased to 47R. The 5532 has more output oomph than the 4558 so can cope with a lower series isolation resistor.
...to be continued
Total Comments 1
Comments
-
Hello
I have two good dvd player, but for fun I buy most used cheap dvd players I can find for few dollars, just to look inside of them.
Like you I do look at the dac chip and output circuit and opamps to see what can be mods. And most of the dvd players do have on the main pcb some chips who heat a lot, so I put some heat sinks.
After fixing those cheap dvd players that I can fix, I alway look inside the power supply and all of them do have a very noisy switch mode power supply with very little filtering.
So, even with my good dvd player, I alway ad a good filtering on the switch mode power supply ouput and some bigger and better electro caps where it's need.
One other thing, most cheap dvd players don't have any oil in the brass bushing and shaft of the spindle motor, so I alway lubricate them.
Bye
GaetanPosted 6th May 2011 at 06:52 AM by gaetan8888
