Edman wrote:

Hi there, I would like to ask for little help regarding my old Blaster. I know it's not the best one but it's all I'v got. I can hear quite a lot of noise. It subsides while lowering volume but it never entirely stops. I read here on many occassions that people were replacing caps to get better sound and also - that getting rid of all the noise might not be possible (depends on the model). I would like to try to recap my card. I also read that Creative engineers originally designed the card with different caps on some spots. Can someone please advice me where to get the schematics or datasheets for this kind of card so I know what caps were ment to be there? Thank you in advance.

Can you point to some info which talk about these supposedly different caps?
You won't get schematics of the card, but of course the card has also standard non-Creative ICs that have examples in their datasheets.

Usually recapping means to change only the electrolytic capacitors. There are also other kind of capacitors (ceramic, tantalum, polymer..) that usually don't go bad and these hardly ever need replacing except when tantalums explode. Electrolytic caps are easy, because you can read their capacitance, voltage and temperature rating, and if you are interested how good/bad those capacitors are compared to what you are going to buy as replacement, they also read the manufacturer and capacitor model series information. Ceramic surface mount caps rarely have any markings on them, so their capacitance and other ratings are only a guess.

I also recall some SB16 cards were pretty noisy (I think my noisiest was a Vibra 16C) so if the noise comes from the ICs, just changing the electrolytic caps on sound card won't help much with the noise. Also standard electrolytic caps are not so good at stabilizing high frequency noise, so they are mainly used as bulk energy storage and the higher frequency noise is smoothed with local ceramic caps near every IC.

The noise also depends on power supply, motherboard, other cards, hard drives etc that generate noise on supply voltages.
Just by changing the sound card to a different ISA slot further away from other cards could help with noise emitted by nearby cards.
Also, just a thought that I have been thinking; how does it really affect the system, if you put the best capacitors you can get to a sound card, and the sound card into a system with just regular capacitors, do the caps on the sound card have the largest ripple current over them, as they have the lowest impedance and they want to stabilize the voltage better than the other caps? I don't know, but if this happens, it surely does not sound good either.

There's about fifty electrolytic caps there, and if I had to estimate, about half of them stabilize the supply voltages, and the other half are there for passing audio between ICs where DC bias must be blocked.

I personally would not blindly do a full recap, but of course if a cap is broken or leaky it should be changed.

This is what I found:

+Sound Blaster 16+ (CT2770): For the sake of a refresh all necessary capacitors will be replaced with Panasonic FM or FC line ca […]
Show full quote

+Sound Blaster 16+ (CT2770):
For the sake of a refresh all necessary capacitors will be replaced with Panasonic FM or FC line capacitors.

1) C18, C132, C86 and C95 replaced. Digi-Key P# P11212. These are the four initial filter caps for the Voltage supply rails of the ISA bus. No obvious change.

2) C68, C79, C104 and C105 replaced. Digi-Key P# P11212. These four filter caps are tied to the rails of the two on-board voltage regulators. In my opinion, bass was improved.

3) C87 replaced. Digi-Key P# P12924. This is the Supply Voltage Rejection capacitor and reduces ripple. Some of the hiss in the background seems to have subsided and the music sounds overall smoother.

4) C89 and C99 replaced. Digi-Key P# P12923. These are the Feedback capacitors. (also referred to as Inverting Input DC Decoupling capacitors) I'm uncertain whether there was a change. The music sounds as if it may have a bit more power behind it. According to the datasheet, this capacitors can affect the low frequency cut-off. It calls for a 100uF capacitor, but Creative Labs used a 47uF capacitor. I may rectify this, later.

5) C83 and C? replaced. Digi-Key P# P12924. These are the Bootstrap capacitors. excerpt from datasheet: "The bootstrap connection allows to increase the output swing. The suggested value for the bootstrap capacitors (100uF) avoids a reduction of the output signal at low frequencies and low supply voltages." I cannot say whether I hear a difference or not, but this is still an important place to apply good quality capacitors. Logically, it would seem that this becomes more of a factor as you increase volume and more power is required. This would likely be a real issue if you were stressing the circuit.

6) C93 and C103 replaced. Digi-Key P# P12376. These are the Output DC Decoupling Capacitors. Their capacitance affects the low frequency cut-off. At 470uF and 24-Ohms the cut-off would be 14Hz. (It's hard to say, exactly since Creative used smaller Feedback capacitors than the Datasheet calls for, but much larger Input capacitors than the Datasheet calls for.) I did not change values, but very low-frequency performance HAS improved. Also, I noticed some details in the various areas of the frequency range in the tracks I have been listening to that I did not notice before. Again, they are much better caps than the originals and they are fresh. These capacitors are also important if you are using headphones as they block DC which would otherwise damage your headphones. Ideally, for best quality they should be non-polarized, but for the moment, I wanted to keep the capacitors the same diameter. Maybe I should have done this first; maybe it would have allowed me to more easily noticed improvements from replacing other caps. doh!

7) C88 and C98 + C107 and C105 replaced. These are the Input DC Decoupling capacitors. The first pair seem to be tied to the second; what is going on here, I am not certain. Anyway, sound was overall more detailed and even the bass seemed clearer. Actually, I have to admit that I had read that my headphones are bass-heavy, but on the SbLive! that was absolutely NOT true. After replacing these signal caps, the bass does seem a bit heavy; enough to shroud some of the upper range that I used to notice, while the mid-range is beautiful. Applying film bypass caps to the filter caps of the ISA voltage rails did not seem to yield any real improvement as it did on the Live! And it may never make a noticable difference since I can't really do anything about running electrolytics on the outputs. Changing 1uF signal caps on the SB16 to films will likely make a large difference.

That pretty much finishes up a basic refresh for use with hard-drive audio. There are many other traces, such as the Line-In and the several CD-A inputs which could be recapped, but I have no interest in these. Actually, I may refresh the Line-In for use with a Playstation 1 I am refreshing, mainly to be used as a simple CD-Player, but also to improve sound while gaming.

I do plan to replace many of the signal capacitors between the DSP and OpAmp as well as those between the Line-In jack and DSP with Metallized Polypropylene Films. I am still considering options for replacing the 470uF output capacitors as films would be too costly and bulky. Likely, non-polarized electrolytics bypassed with films will be the end result.

I am considering one or two very simple modifications for the card as well. For example: The original schematic for the TEA2025B uses a 100uF filter capacitor at the Voltage Supply pin. Creative Labs saw fit to simply place a 10uF electrolytic before and after the voltage regulator. I tried adding a 100uF, swapping it from in front of to behind the regulator to see what happens. I didn't notice any real difference.

And also this is an article on SB16 from Wiki:

As many Sound Blaster 16s are now well over 20 years old, many cards suffer from symptoms related to aging capacitors, ranging f […]
Show full quote

As many Sound Blaster 16s are now well over 20 years old, many cards suffer from symptoms related to aging capacitors, ranging from muffled or distorted output to the cards working improperly. In addition, with regard to the amplifier design on most boards Creative did not strictly follow the datasheets' recommendations on capacitor uF values, negatively impacting the amplified output's sound quality. Replacing the capacitors with fresh ones of the recommended values can noticeably improve both amplified and line-level audio quality, in addition to restoring proper operation.

On many TEA2025-based boards, Creative used 47uF capacitors for the amplifier's inverting input DC decoupling (connecting the Feedback pins on the amplifier's pin-out), whereas the datasheet recommended 100uF units against an increase in low-pass cutoff.[7] In addition, Creative installed 1 uF input decoupling capacitors on the TEA2025's inputs, which according to the datasheet should have been 0.22uF units and were only necessary when a volume control slider or knob was present, but were installed regardless if the volume knob was present. (These capacitors may be replaced with short jumper wires if no volume control slider or knob is present.) Finally, Creative used polarized capacitors where non-polarized capacitors should have been used.

On boards that use the TDA1517 amplifier, Creative used 470uF capacitors for the outputs where the TDA1517 datasheet schematic suggested 1000uF units.[8] Depending on the board, an undersized capacitor for supply voltage rejection (connecting pin 3 of the TDA1517 to ground) may also have been used; the datasheet recommends a 100uF unit for this application. The two output capacitors have a DC voltage present, so polarized units may be used.

Sound Blaster 16 sound cards with the CT1747 chipset frequently have the internal preamplifier gain set too high, causing clipping and crackling in the output that wasn't present on sound cards built using the larger and less integrated CT1746B chip. Changing the mixer levels has no effect on the clipping; the only way to fix this would be to decrease the preamplifier's gain level.

Some Sound Blaster 16 cards made after 1992 may have grounding loops and overall less effective filtering on the +5 volt and +12 volt DC traces coming from the ISA bus. A symptom of these cost-reduced designs is an audible hiss or buzz which is present even when the card is not playing any sounds, which may affect both amplified and line-level outputs. Fixing these problems may not be trivial depending on the card's design.

Please note that I won't be the one doing the actual replacement since I do not trust myself with such a complicated operation. As you pointed out, the card is filled with capacitors - some of which are pretty tiny and I have very shaky hands. More over - I had no idea there were types of caps that do not need replacement, hell I don't even know what the Digi-Key P# stands for and so on. I have just only begun to scrape the surface of this stuff - not knowing much about electronics. The refresh will be done in a professional hardware service I trust from past experience. I'm just trying to save them time by looking for the data they can build upon. Thank you for all the info so far anyway!

I understand that it looks like I already have a list of caps that need to be replaced - and what should they be replaced with from the posted article. But - what I do not know is if his card is the same model to mine (even though the number is the same, I understand the PCB can look differently) and I also don't know if it's all that needs to be replaced since I don't understand the importance of all the mentioned caps.

Since my board only has two ISA slots, I can only move the card to the second one at the absolute end of the motherboard. The PC itself is a standard slot 1 P3 450 on SE440BX Intel board coupled with single 40GB IDE drive, GeForce 5500, PCI USB 2.0 hub, PCI Etherner card, a CD-ROM and a FDD. My PSU is 350W Chieftec. Do you suggest that I should also try to recap the PSU and motherboard for better result should the repositioning fail?

Personally I would only bother with such a project of I were doing it myself, simply to find out if replacing the caps actually improves the output on these old Creative cards.

I would be very very surprised if your repair guy charged less than it'd cost you to buy a different card. There are tons of people using vintage sound cards (often older than the ct2770) without these problems so I wouldn't think it was completely related to aged caps. It's certainly possible... but I would probably only invest this kind of effort into a rare or highly valuable part. The 2770 is pretty common on eBay and doesn't have any special features over other cards as far as I can tell (it's a value model).

You are probably right. But the refresh won't cost me that much and since I am not prepared to do it myself, I'm rather low on options since I have a connection to this card. As I reside in Czech Republic, much of the stuff from ebay usually costs half of what they charge for postage so this is no go for me at the moment. One day I would like to purchase a full fledged AWE32 but that is also little too much for me right now. I used to have one when I was young but gave it away many years ago. I loved that sound... But anyway, thank you for the opinion. I know it's hard to see reason behind my intentions.

Edman wrote:

I know it's hard to see reason behind my intentions.

Well, if it is a hobby, isn't that good enough reason 😀

About the supposedly wrong caps, yes, changing some of them with lower ESR can help with noise, and some don't, depending on what they do in the circuit. Changing some of them to larger capacitance than originally will help by allowing lower frequencies to pass the amplifier chip.

But, I would not directly say that the capacitors are of wrong value or of wrong type.

In the case of the TEA2025B amplifier chip, the manufacturer has suggested some example values for some application (like portable radio cassette players in this case), using it for something else (like home computer speaker amplifier) means it can use different component values. One thing is because their power supplies are different, portable stereos have batteries with some noticeable voltage sag during peak current consumption, while PCs have regulated power supply that can provide tens of watts without noticeable sag, so the amplifier chip just does not require that much bulk capacitance in a PC. Back in those days, if you connected passive speakers to a computer, they were typically very tiny and could not reproduce very low frequencies anyway, so why bother putting bigger and larger caps there to make the sound card bigger, heavier and more expensive. If you wanted more quality output, you would use connect the Line Out to home stereo system anyway and ignore the Speaker Out. Of course changing them to values suggested by amp datasheets will allow for lower frequencies to be heard better with headphones or large enough speakers connected to Speaker Out, that is a fact.

And the amplifier output DC decoupling capacitors that should be non-polarized electrolytics according to your source, well, he calls for low leakage current to protect earphones from DC, but I could easily find regular polarized electrolytics with lower leakage currents than what non-polarized electrolytics have. Non-polarized electrolytics are not so common to find in sizes you want easily, and they are more expensive. The thing is, since the amplifier chip is a single supply amplifier, the DC bias polarity over the capacitor never reverses, so the circuit does not even remotely seem to need non-polarized caps. Amplifier chip examples all have regular polarized electrolytics here, just to show that it does not need to be a special non-polarized electrolytic. And capacitor values required by the amplifier chip are so large the only reasonable standard option for these are electrolytic capacitors which are by default just standard polarized capacitors.

If the rant about polarized vs. non-polarized caps mean why there are 1uF polarized electrolytics instead of 0.22uF non-polarized capacitors as the TEA2025B input coupling capacitors, the answer is easy. If the design already had 1uF capacitors elsewhere, it will be OK use them there as well to reduce the amount of different components, instead of introducing new components. It could be cheaper to put for example ten 1uF capacitors per card, than to put eight 1uF capacitors and two 0.22uF capacitors per board. Increasing the coupling capacitance only makes the low frequency response better, and again, a polarized electrolytic is just fine there as long as the DC bias over it is positive or zero. The amplifier chip example puts a non-polarized capacitor there just to show it does not need to be polarized, and 0.22uF caps are so small in value that other types than electrolytics are usually used here and they are non-polarized by default. There is also talk about whether they should be removed because there is no volume slider, well, if there is DC bias voltage on the audio from mixer, then these capacitors must not be removed so that they remove the DC bias of course.

So yes, it is a typical consumer sound card for typical home use, not a high-end professional device for recording studios, that's why it is not perfect in every sense but just good enough, and it shows on the design and price of the device.

By all means, do the recap, I just wanted to show that putting extremely high quality capacitors everywhere makes little point and bigger is not always better. Modern average capacitors are still better than average capacitors back then when the card was made. Just replace with equal capacitance and voltage rating and it will still be improved up to a certain point. To reduce noise even further, swapping better electrolytics is not enough, I am surprised nobody has yet modded their sound card with extra inductors on supply voltage wires for extra noise filtering.

It would be great if you could share your results! 😀 I've got CT1350B with terrible "thinking noise" here and after switching positions in slots and changing power supply I conclude that recap is all I have left. I'm not afraid to do it (I successfully recaped bunch of GameGears), but it's a lot of work for something I don't know will pay off 😀

Edman wrote:

I can hear quite a lot of noise. It subsides while lowering volume but it never entirely stops.

My first soundcard was an Aztech, wich died in 1995, so I had to buy the cheapest one and went for a CT-2770.
The first thing I noticed when I turned on the machine was a hissing noise and I could hear the HDD reading and writing.
That noise-experience are one of the clear things I remember from my 486 days. It's printed in my mind for some reason.
If there is no noise on the line out, then it is perfectly normal, what you are experiencing with you'r 2770.
Perhaps the card you have, have no bad caps. If the only issue is noise, then I think you don't need to worry.

Jepael: Thank you for a very detailed reply. This is certainly a lot to grasp 😀 I will hand it over to the service guy together with the original article - so he can put it to good use. This might make things easier I guess.

Sledge: I will post the result here. But as brostenen said, it might be that the card is in fact allright. The noise may be a natural thing to it. We'll see.

brostenen: As a matter of fact I have never tried the line out so I have no idea what level of noise it has. I'll check it when I reposition the card (presumably weekend). Thank you for the advice. I'll post the result.

Might be interesting to see if there's a potential performance upgrade in replacing the TEA2025 with something better.

If there is less to no noise on the line out, then plug in an external amplifier.
This should logically result in a better soundquality.
I remember the OPL part of the card as being good. Not bad, just good.
The games compatibility are also good with the 2770.
I never had any issues untill the card died of old age in 2009.
My card was working from 1995 to 2009.

@brostenen
I agree. I had a quick look at the SB hardware programming guide and the block diagrams make things quite plain. The analog signal path goes CT1703 DAC -> CT1745 mixer chip (on this card) -> TEA2025 (can be bypassed). The DAC and analog mixer chip heavily control the optimum audio quality of the card.

I wish that we had good data on the DAC and mixer chip, but we do know that later Vibra cards which replace the mixer chip definitely improve the noise floor, so I suspect the CT1745 is a big culprit as far as noise on the CT2770 cards.

gdjacobs: Thank you. Seems I won't be able to get rid of that noise afer all.

I'm the one who did the refresh on one of my SB16's, and it stated in the thread that it is in fact a CT2770. I don't remember having all these problems with HISS or background noise, but I also had a recapped PSU by a known, medium-quality brand (Antec).

I have the card, sitting right next to me, which I intend to install in this P6DBE-based tower, I have on the floor, here. Just trying to get around to it.

I noticed a lot of noise on my AWE32 CT2760 when I put it into my old P90. When I had it in my high-end Tyan motherboard, it was perfectly silent (I'm not exaggerating) even when the volume is cranked up and listening on Sony MDR-V6s headphones.

I'm not sure if it is the motherboard or the cheap power supply that's causing the noise. Maybe both, but I would start by experimenting with other power supplies to see if the noise subsides before trying to recap the SB16.

Thanks for the advice. I had some "bad" luck lately caused once again by my ignorance. Purchased not so cheap AWE64 Gold that was supposed to be in good condition, only to find out later that few connections to one of the chips were severed. So I had that repaired first place and recapped too. It works now but it's still not the sound I was looking for. Now I'm heading back to CT2770 troubleshooting 😀 I guess I'll buy a new PSU and try. The one I have in the machine is an old EuroCase, there would be no use of putting more money to that. Then, if nothing changes, I'll have the Intel Board recapped (only few caps there). Then - if nothing changes - I'll try to refresh the Blaster itself.

Also... I have some kind of VIA PCI card (USB HUB) installed to make use of the front USB ports on the case and of course to have the USB 2.0 ability. Do you guys think that the noise can be caused by that? I have to try without it.

Logistics: I understand there is more than one model of CT2770. Maybe my trouble are caused by different, perhaps cheaper build. I don't know. I'll refresh those parts around it and see. Thanks for the info though.

I got really excited about trying these on my audio cards until I saw the price tag. Geez!

http://sparkoslabs.com/product/ss78xx-discret … tage-regulator/

I get the impression they ask this price because they can. There's no way the component cost + R&D can equate to this price.