Hotrodding the X-Fi: A Layman's Guide (No 56k) - Page 90

Awesome Mod, and a question

Hi all,
First off thanks for the info on this awesome mod. I did the basic version replacing the 4 (primary and surround) opams with LM4562s and the the power filter cap with the Blackgate 2200uF 16V. In removing the cap I think I may have messed it up a bit. I assume if the cap is not installed properly I would either get no sound or smell that sexy burning odor? I stuck a DIMM on the cap and got 1.2v. Is this correct?

Thanks in advance

My RMAA results are ugly especially the frequency response, but it sounds great, and that's all I care about

Looks like you are running Windows Vista & have mismatched sample rates between record & playback in the OS or the RMAA is set to a different sample rate than the OS. I have Windows Vista & thats what it looks like when mismatched. It will still sound fine as it seems to effect only the record part of the equation especially if you have bit matched playback enabled. Make sure all your areas that relate to sample rate are matched & you will see very flat frequency response.

Thanks!

What do you mean "to the internally regulated powersupply for the DACs"? You mean the 100uf and 47uf are both in parallel to the DACs power supply (+5v and ground) ?

Those 2 220uf ones, what do you mean "those are the +/- 12v rails for the opamps"?

I'm a little bit confused.

Thanks! If change the 24.576mHz crystal oscillator, do I need to pay attention to the CL of 30pf (there is nowhere indicating it's 30pf)? I looked through some Web pages, when they're selling the good oscillators, they do not say what CL the clock is.

Also, if I change it to a 4-pin TCXO one, do I need to change/remove/add other components on the board for it? Where to get the best +5v power supply on the card?

Can you suggest some better TCXO oscillators? I remembered some one said, there are 1ppm or even 0.5ppm ones.

Thanks!

Na I'm still on XP, tried various sample rates and tried to normalize it to the best of my abilities but it's still a bit high. The music and games sound great so if it's recording related it won't bother me.

Try going to creatives mixer, switch modes to Audio Creation then go to settings in the lower left hand corner & enable bit matched playback. You can also go to the creatives audio console & enable both bit matched playback & record. That should cure your problem. Otherwise all sound goes through the Kmixer which adds distortion especially if the sample rate is anything other than 48KHz.

Woot! Didn't have both bit matched playback and recording checked. Now the results look much better.

Just wanted to say thanks to all you guys, especially Germanium, Coldt and Bichi for your hard work and patience.

Glad that fixed it. Happy listening.

Germanium, I bypassed the main powersupply (16v 3300uf BG) and the 2 opamp powersupply (50v 1000uf ELNA) with a 22uf SOLEN each.

I want also to bypass the 3 powersupply caps with 2uf, 1uf, 1uf SOLEN in addition.

1.Can there be any improvements if I did the additional bypass? No ill impact to the circuit?

2. Since my RMAA graph has lots of spikes after 30kHz of noise level, can the newly added caps reduce these high-end spikes?

3. If do so, will it break the balance of low/mid/high?


Waiting for your guidance, thanks much!

There really isn't any that effect the output like the ones I mentioned. However I found by direct coupling the Line-In then turning on the microphone from the same jack as the Line-In there is about 5volts D.C. put on one of the channels & the sound will not work for the mike from the flexijack on the back. This may be compromising performance som as some of these circuits are still hanging out there when not in use. This will possably effect RMAA some so I'm trying some different routes.

This may not have been much help but will report back later, ok?

Yes, CL is what the manufacturer uses to tune the crystal. I got the CL from the DAC tech sheets. You can actually use a slower oscillator if you want less internal oversampling. But I wouldn't bother trying that unless you really want to.

I'm not sure how you would intergrate a 4 pin, but I wouldn't try, since the crystal is sensitive.


To find the CL of the oscillators you will need to look at their tech sheets.


EDIT:

On a side note, I replaced the coupling caps with 16V 100uF UCC PSA series polymers. Improvements across the measuring board, and more detailed than my FM's i was previously using.

Thanks, you shorted only 2 caps for line-in, but 4 for front-out, were you missed 2 others for line-in/mic? That may allow 5v DC for one channel.


Waiting for your update.

Also, do you mean that there will be no use if I bypass additional 3 2uf, 1uf, 1uf caps? Thanks!

The 5 volts for the mic is injected between the flexijack & the line in coupling caps & if you don't use the flexijack for a mike then the 5 volts is not injected.This voltage is injected on the right channel, in other words the ring on the tip ring sleeve 1/8th inch plug. In Windows Vista you can actually disable individual inputs so they don't show & can't therefore be selected. For those with Windows XP I'm not sure if that can be done so I no longer recommend shorting the line in coupling caps on XP machines. If you short them on a Windows Vista machine make sure you disable the mike input for the flexijack first.

Update, tried to find where to disable inividual inputs on XP but was a no go.


For those with Windows Vista to disable the mike inpute for the flexijack right click on the speaker icon in the taskbar & select recording devices then right click on micrphone & select disable.

There are two other coupling caps in the line in circuit after the line in opamp. These are 47uf caps on the other side of the ADC & very very close to those 4 large caps above the ADC.

I'm waiting on a new board as I was probing around on the board when it was live & accidently shorted the power supply to the relays to the left channel line-in. When there is no plug in the flexijack the left channel is shorted to ground. When I accidently shorted the relay power to the left channel line-in ther was no plug in the flexijack & I ended up blowing the left channel trace & seems like all of the opamps that lead to the ADC may have been fried as well, even the ones behind the second set of coupling caps. I will be getting new Elite pro card on tuesday if all goes well. The output section on the old card still works so I'm still able to listen to music on old card, just not able to do anymore RMAA tests on this card.

The other caps other than the ones I mentioned to bypass with metalyzed film caps are for the line-in & have no bearing on the sound you hear at the speakers though they may improve the specs as read by RMAA & help if you plan on doing any recording with the line-in.

SOIC8 to DIP8 Carrier
- made from spare surplus parts
- still needs epoxy "potting" for added mechanical strength
- SOIC socket installed on SB0460
- allows DIP8 opamps to be tested with SB0460 and SB0550

Intial test of adapter/carrier, plugged in an old Texas Instruments TL082, DIP8, circa 1980, JFET frontend.
- surprised by audio quality and did RMAA tests, out of curiosity.
- seems to be a nice match with Line-Out caps, Panasonic S, SMT, 33uf @ 25vdc
- short audition with "Put Your Records On," Corinne Bailey Rae, source 48/24
- firm bottom, accurate image, clean highs, very pleasant overall...

SOIC to DIP8 Carrier:
http://www.esnips.com/doc/4e49edc3-b.../X-FI-MOD-075b

SOIC to DIP8 installed on SB0460:
http://www.esnips.com/doc/4c65f579-f.../X-FI-MOD-076b

RMAA Results: 48/24/48 and 96/24/96
SUMMARY: http://www.esnips.com/doc/5eda7602-0.../X-FI-MOD-077b
DETAILED: http://www.esnips.com/doc/b413cc2d-7.../X-FI-MOD-078b

Texas Instruments TL082-CP:
http://www.ti.com/lit/gpn/tl082

EDIT: reloaded eSnips stuff with new links

Deleted