[lan]

Quote:

Originally Posted by lucpes - add 1R5 2W resistor on the 12V line from the floppy connector. Cut the unused 5V and ground wires. Should help keep noise from entering the board ground. Seems like less grain after some time - 1 day, but it may be that the whole card is 'burning-in'.

The floppy connector is for 1820/m users only along with the 35V caps. No need to fool with those if you are using 1212m.

 

[zhoufang]

lucpes,

Thanks for your pictures.
They are amount the most useful informations that I have found for 1212M mod. If fact I was about to make something similiar, until I found yours.

There's one problem I don't understand.
for the 0202 analog card, the top left cap on your photo, i.e. C17 10/16(opa), I don think it belongs to the line out (DAC) circuit.

If you look at the back to the 0202 PCB, you can see that the DAC and ADC sections are separated to their own masking plane (the dark shade), that's a common praction for isolation of noise and prevent interference, and the C17 falls under the plan of ADC.

So I believe C17 is for the ADC circuit and not the DAC circuit, which to us Head-fiers are not in our interest of modding (since we only care about the line-out).

from the post "Fully modded E-Mu 1212M pics and impressions", Iron_Dreamer also mod C17 with a blackgate, which make me wonder that could I be wrong?

 

[lucpes]

Quote:

Originally Posted by zhoufang lucpes, There's one problem I don't understand. for the 0202 analog card, the top left cap on your photo, i.e. C16 10/16(opa), I don think it belongs to the line out (DAC) circuit. If you look at the back to the 0202 PCB, you can see that the DAC and ADC sections are separated to their own masking plane (the dark shade), that's a common praction for isolation of noise and prevent interference, and the C16 falls under the plan of ADC.

If you measure from the output OPA supply pins (don't remember whether it's the + or the -) there's a direct connection to this cap and it's on one of the +/- 12V supplies with a 4R7 resistor right in front. DAC buffer uses +/-12V and ADC buffer is +/- 5V as far as I have figured it out, so it makes sense to change it.

 

[lucpes]

Quote:

Originally Posted by lan The floppy connector is for 1820/m users only along with the 35V caps. No need to fool with those if you are using 1212m.

Oops

Right. That's what happens when you plug in the soldering iron before reading the manual. Then again, the capacitors were obviously poor quality and I happened to have the 470/25 FCs at hand

Updated pics and the following post to reflect that. Thanks for the input!

 

[lucpes]

If you're using the digital output from the 1010 card here's what I think a part of the $500 Empirical Audio mod ("removing some filtering elements") consists of:

Basicly, just bridge the two inductors and remove the filtering capacitors. Downside is that your card won't pass FCC standards anymore. Who cares?

Did not try this as I have only a 16 bit capable DAC around. Anyone willing to do it?

 

[zhoufang]

I just mearsured it again
the C17's neg pin is shorted to the ground.
the C17's +ve pin is shorted to the V+ pin of the 5 op-amps in the line-in circuit, about 14ohm from the V- pins of the DAC buffering OPAs and 9 ohms (the values may not be very accurate since my Analog multimeter is the junk type) from the V+ pin of the DAC buffering OPAs

Using pictures of 1820M PCB as reference I'm certain that each DAC uses two OPA and each ADC uses 5 OPA.

The overall I'm quite certain that C17 is there for the 5 OPAs in the ADC circuit.

the electrolyte caps for the two DAC buffering OPAs' power supply are C74 and C77 in the bottom right of your 0202 photo, which are for V+ and V- respectively. They are link to the OPAs power supplies pin by two 4R7 resistors, which are R41 and R72 for V+ and V- respectively.

C74 not only links to R41, it links to R21 (the 4R7 resistor right in front of C17 ) as well, which links to the V+ pin of the 5 op-amps in the line-in circuit.

Quote:

Originally Posted by lucpes DAC buffer uses +/-12V and ADC buffer is +/- 5V as far as I have figured it out.

I believe that the V+ of both group of OPAs are at the same level since they are both linked to C74 by a 4R7 resistor (R41 for the DAC's OPAs and R21 for the ADC's OPAs). I'm not sure of the V- for the ADC's OPAs yet though.

 

[lucpes]

I'll leave the card in until next week when the capacitors arrive, but if you're right there are three reasons why that cap should be changed too:

1) A capacitor not only filters DC power for a certain device. It also acts as a reservoir and keeps the rest of the supply circuit (which also powers our precious output buffers) free from fluctuations/noise when current draw suddenly increases from that specific device (be it the 5 ADC opa's). Thus, better sound on output.

2) If you're not going to use Black Gate, Panasonic FCs are great for analog decoupling and in this value are dirt cheap. So one/two more cap(s) changed won't hurt anyone's wallet too much.

3) Better RMAA numbers

 

[zhoufang]

lucpes,
your points are valid.
the 3rd point is rather interesting.


There's people against the idea of moving the 0202 card outside the CPU casing because althogh the cable carries only digital signal but the clock signal is being carried as well. And if the cable is too long the quality of the clock signal will be compromised.

http://www.audio100.net/6/96.htm
but when I look at the photo of 1820M's external dock, I don't find any clock generating crystals there.
Does that means 1820M also carry the clock signal by cable?
If that's the case if we use a sheilded cable for mounting the 0202 card externally it will not no worst than the 1820M in the clock signal regard?

I will imaging that a moded 1212 with external 0202 fedding on its own regulated linear power supply and away from interference, that will be pretty close to ideal.

 

[lucpes]

Quote:

Originally Posted by zhoufang Does 1820M carry the clock signal by cable? If that's the case if we use a sheilded cable for mounting the 0202 card externally it will not no worst than the 1820M in the clock signal regard?

Yes, it carries the digital data & clock by cable, but it must use a proprietary protocol for balanced transmission between the 1010 card and the AudioDock (that means no real clock signal on the cable, it's embeded in some other form of data transmission similar to what network cards use - check the transformer & some ICs present on both the 1010 card and the AudioDock). This protocol is sure to have error correction capabilities as the maximum cable length in this case is something like 10 meters as per EMU's recommandations.

Quote:

Originally Posted by zhoufang 2) I imagine that a moded 1212 with external 0202 fedding on its own regulated linear power supply and away from interference, that will be pretty close to ideal.

You will have to figure out which one of the 34 pins carries the power lines on the connectors, their coresponding grounds, cut the traces on the board and also reference the external linear power supply ground to the computer ground (better to the 1010 card itself), otherwise you're gonna get in big trouble.

Also the fact that there's real clock signal in the cable means that the cable must be short (shielded or not), otherwise you'll get into jitter problems - the stock cable is a tad bit too long as it is.

A bit too complicated IMO, you'll gain sound quality wise by using a shorter cable inside the PC and using something grounded to the PC's case to shield both the cable and the analog card against RFI. A quality PC power supply will also help.

 

[zhoufang]

here are some photos that I just took.
most silkscreen markings should be readable.
I hope this will help in the discusion.

1010 digital card

1010 digital card part 1

1010 digital card part 2

0202 analog card

0202 analog card
DAC section, most mods will be here

Thanks Floodo1 for providing the webspace!

 

[zhoufang]

regarding the 4 10/16(opa) caps on lucpes' 0202 photo, in order of left to right,

C17 is for V+ of ADC's 5 opa
C16 is for V- of ADC's 5 opa
C43 is for V+ of DAC's 2 opa
C65 is for V- of DAC's 2 opa
So the last two need very good e-caps since they are for DAC buffering OPAs, while the first two only affect ADC performance.

-------------------------------------------------
Regarding the Giant 1000uF/16V C72 on the 0202, I have found that it is for muting the lineout while the power is being turned on (so that you don't hear a loud pop when you turn on your computer)
The +12V supply from analog expansion male socket pass a limited current through D12.
Both C72 and the "muting while power on" circuit (R76, R71 and Q7, leading to R64, R52, R43 and R43 which leads to the muting transistor Q2-Q5) are at the other end of D12.
While power on, large C72 will draw whatever current that pass through D12, leaving the "muting while power on" circuit in a low state (the state that mute the lineout).
A second later C72 is charged up, so the "muting while power on" circuit goes to a high state which open up the line out.

After D12 there's also another circuit (R76, R79 Q8 C75, R71, D1).
I don't know exactly the function of it, but at least one of the link after R82 will link back all the way to the 1010 main card's smaller E-DSP.
Since this Circuit is affected by C72, I believe what it does is still turning off something while the power is being turned on to prevent damages from unstable voltage.

Thus I think there's absolutely no need to mod C72.

 

[lucpes]

Information about sound signature and performance of various Low-ESR capacitors:

http://www.diyaudio.com/forums/showt...436#post604436

You should check the entire thread, it's a good read.

 

[zhoufang]

I was looked at the Sanyo OS-con series, it says:

SP Large capacitance and Low ESR Optimum for Audio etc.

SEP Standard radial lead type Guaranteed at 105 degC for 3 000h Environmental product

for the same capacitance the SP series is physically smaller than SEP series, have the same ESR, but with significantly lower Tangent of loss angle (which I think is more important than ESR, you can get low ESR by using bigger caps, but not for tangent of loss angle).

So I think SP series in better in all aspect, and should be the first choice amount OS-con series.

 

[zhoufang]

Just found that the 1212 gets its -12V from the PCI socket.
I thought many low end PSU no longer produce that voltage, or it is generated by motherboard?

regarding all the electrolytic caps that need to be changed, it's right that the highest they will get is 12V. In fact more than half of them only need 5V or 3.3V, one of them only need 1.8V. So if we get two lines, 16V and 6.3V (4V cap typically aren't much smaller than 6.3V), then all the 6.3V caps will be physically much smaller/better grade. Not to mention saving a few penny, esp if you are using expensive caps.

I will post a picture of which cap need which voltage later.

 

[zhoufang]

regarding removal of muting transistors, I believe it can be potentially hazardous.

Unless I'm wrong, the giant C72 and the circuit at the left of it will mute the lineout when the power is been turned on.

And for 1212M the only DC blocking cap is found betweeen the DAC and the buffering opamp, which means whatever DC offset the op-amp produce will be passed to lineout.

Most opamps only work properly above a certain voltage (for the original JRC2068 it's +-4V). When the rail lines goes below that voltage the op-amp will not be stable, and may produce large DC offset at the output.

because of the capacitors and the 4R7 resistors in the +-12V regulation circuit found on the 1212 cards, the +-12V supply to opamp will take sometime to rise to the intended level. And this time will be longer if you mod your card with bigger caps.

So the overall result is that when you turn on your computer, not only you hear a pop sound but there may be a instantaneous large DC offset.
If your headphone/speaker is connected to be unprotected amp and the amp is on while you on your computer, there maybe chance for some damage.

Since the duration of such event will be short (a few ms at most), but the risk of damage might still be possible.