squareguy wrote:

I had a chance to try out my new Sound Blaster Pro 2 (CT1600) for a little while this morning and I like it.

First, I would like to know how to identify which revision it is and how many revisions are there known to exist?

SB 2.0 CT1350 CMS chips dumped

always interested in this kind of stuff, but what are the plans here? sell kits? sell parts? testers?

The plan is I will make one up and take good pics with information and details. Then if someone could replicate it from the instructions and have test equipment to measure noise before and after, that would be great. All I can do is subjective listening without measuring equipment. This is not a radical makeover and be easily made back to original if wanted. A picture is worth a thousand words so hopefully I can show some within a few weeks.

Ah great, can't wait to see what you come up with.
I have a CT-1600 here that I can mod, I'll certainly do that if your ears say it's worth it 😀
No measuring equipment here either...

squareguy wrote:

The voltage regulators are fixed, linear type. I would use adjustable, linear type instead. This would be built on tiny PCBs above the card, it would only need the regulator, two resistors and a small capacitor. The reason? Well, variable voltage regulators allow you to bypass the adjustment pin with a capacitor which can reduce the output noise of the regulator by about 20dB. You cannot do this with a fixed regulator since all of that is within the chip itself. So, I would use a LM317 circuit in place of the 7809, a LM317L circuit in place of the 78L05 and a LM337L circuit in place of the 79L05. I would stay away from trying to use modern LDO regulators and stick with linear.

Yes, somewhat interested. I've been thinking something similar as well.

Depends on the specific manufacturer and type, but some 78XX types from some manufacturer have better noise properties than some types of LM317 from some manufacturer. The capacitor may or may not affect noise at higher frequencies that much, it mainly affects ripple rejection which is measured at lower frequencies. Ripple rejection may be 15dB better at 120Hz, but not much difference at 5kHz, and PC power supplies are switching type so they don't have 120Hz ripple but they do have ripple that is over 5kHz. The noise should be smaller as well with the cap, but the datasheets rarely quote how the noise is affected. So in general, the cap on adjustment pin may not be any better than just getting a good regulator to begin with. Plus you will have some extra effort to electrically isolate the LM317 tab as it is not ground, while 7809 tab is ground and it is bolted tight to a copper plane which is most likely connected to ground. The isolation makes it conduct less heat so the LM317 regulator would run a bit hotter.

I think you really should take a look at modern LDO regulators; while they may suck at ripple rejection but they will excel when low noise is required, and they are perfectly normal linear regulators (not switching regulators which I think you mean to stay away from). Just take a look at http://www.ti.com/lit/ds/symlink/tps79501.pdf. LDOs almost always require low-ESR ceramic capacitors instead of electrolytics, so if you mod ceramic capacitors to filter sound card power, they might be the best caps in the system so they might start to filter the ripple of the whole PC, unless you isolate them with inductors so they only filter the sound card power.

You could also just add some inductors and/or ferrite beads to filter the noise even with noisy regulators.

And there are many other places the noise can couple into the output. PCM DAC (I believe integrated into the mixer) might get coupled noise from the DSP chip digital output, mixer chip might have mic/line input volume set very high so it amplifies noise to output, or just some games leave few FM channels running but at very low amplitude instead of stopping the FM channels dead silent. So try some things with a mixer controlling utility first to see where the largest noise comes from, before you start modding the board. The noise could even come from a board in the next slot.

Circuit off the board on small PCBs, no hot lm317 tab touching ground and shorting out
No LDO because of stability
Bypassing adjustment pin effects wide band noise as here http://www.tnt-audio.com/clinica/regulators_noise2_e.html

squareguy wrote:

Circuit off the board on small PCBs, no hot lm317 tab touching ground and shorting out

OK, you got that covered.

squareguy wrote:

No LDO because of stability

In what way they are not stable?

I have been working with commercial audio and video circuits with LDO regulators and never had any stability issues.
LDO regulators just need the correct type/ESR of capacitors on input and output and they are pretty happy. If you don't follow the rules in the datasheet, then of course it won't be stable.

squareguy wrote:

Bypassing adjustment pin effects wide band noise as here http://www.tnt-audio.com/clinica/regulators_noise2_e.html

Yes, I won't argue with that, cap on bypass lowers noise. But the LM317 circuit used on that web page is not according to datasheet suggestions anyway so it might have larger noise to begin with. There is no 100nF ceramic on input nor 1uF ceramic/tantalum on output, only the bulk electrolytics. Ceramic and tantalum capacitors have much smaller ESR so they are better than electrolytics to suppress high frequency ripple, noise and transients.

Fine, I won't do it. It was a bad idea. The data sheet does talk about bypassing the adjustment pin for better performance. peace out