Analog noise on Atari Falcon

[leonk]

Exxos suggested I open a new topic on this, so here goes. While, with his help, I managed to solve a lot of the clock SDMA issues I've been having, and things are working more as expected; I did discover that there's analog noise in the output of my Atari Falcon.

When I plug headphones into the headphone jack, I can clearly hear white noise. It's always there; when the system boots and during game play (that doesn't even touch DMA). This system was recapped using only Nichicon 105C caps (from Digikey) and has a brand new MeanWell PSU. It does seem to have some small patches in the analog side (mostly ceramic caps added in some spots)

I was wondering; Falcon owners,

- when you plug headphones into the headphone Jack, how clean is the audio output on your system?
- Is there any information online on these ceramic cap factory fixes? I'd like to verify them (make sure nothing is missed or shorted)

IMG_8045.JPG (445.95 KiB) Viewed 290 times

[leonk]

FWIW, one thing that makes the noise go to 0 is if I run the DMA2DSP.PRG test. It shows all green with white border and I'm able to run JOY demo perfectly fine. So I don't know if the noise is analog issue or something else. And given I have only 1 Falcon, I can't A/B compare it to anything.

[exxos]

Those mods are I believe c-lab filters.

Audio white noise, if you mean "hiss" then that's always going to be there with electrolytic caps in the audio path. Non-polerised caps are a lot worse. If types got mixed up during recapping or inserted backwards , they can all be a problem. The schematics and silkscreen can also be wrong on a lot of Atari machines. It all needs properly investigating and verifying.

Did it do that before the recap?

Also while going for branded caps is fine to a point, you need to study the data sheets for the lowest ESR ratings. Unfortunately just going with a brand isn't enough. You could even inadvertently replace caps with worse specs than you took out if your not careful. Some manufacturers, even well known ones, can produce total garbage. So you always need to check the data sheets from several top manufacturers to get the best ones... I've talked about all this in the past a lot already..viewtopic.php?p=56#p56

You might just be hitting on inherent flaws which are in most audio circuits. Bad grounding, bad caps etc. I've worked on quad amps which are supposed to be good amps and hiss like mad. The only cure for that is to replace all the electrolytic caps in the audio path with ceramic caps. Beyond that, you'll probably just have to put up with it.

[leonk]

Thanks. You are correct. This is the C-Lab audio mod on a rev K Atari Falcon. I found a utility called FALCON 030 VOLUME CONTROL by Simon Sunnyboy and when I set the volume level to media, all the background hiss goes away. It seems the Falcon boots with max volume on the headphone jack which is especially noticeable when you plug in headphones.

Unfortunately, with DSP audio (playing MP3 or demos) the audio level is reset and what's worst, theres cracking/compressed audio in the output. I compared to captures from other Falcons and they don't have it. Argg.. I might have to just live with it.

[Steve]

I don't know what to suggest other than experiences I've had. The falcon shouldn't output noise by default, but it really depends how hard your listening. If you really crank up the volume of course you will hear some soft white noise, but that's normal for most audio equipment in amplification. I don't know if it helps but here's a photo of a re-cap I did for a Clab Falcon a few years ago. I used Nichicon gold series audio caps.

audio_section_after.jpg (471.46 KiB) Viewed 233 times

[leonk]

@Steve thank you for your reply and image. I need to study the schematics a bit more, but I think I have a prime suspect. My clues are:

- The scratchy digital noise can be heard in the on board speaker, headphones and video output from the video port audio pins (the only OP amp they share is U21 but I'm not sure it's the cause)
- Another FB group user had a similar issue. He replaced every single OP amp on the PCB and the problem didn't go away at which point he considered my culprit
- 2 months ago I recapped a C-LAB branded MK1. I noticed a particular IC was replaced by C-LAB (put into a socket) It's the high amount of flux residue around this chip that hinted to me what was done
- Your image shows this as well!

Here's what this system currently has:

IMG_8050.JPG (369.79 KiB) Viewed 218 times

I suspect C-LAB was replacing ITT made CODEC's with a Crystal chip (CS) 4216-KL CODEC. Your system has a CS.

How's my logic thus far?

Here's the audio circuit. You can see only U21 is common.

Screenshot 2024-08-26 at 8.23.11 AM.png (452.11 KiB) Viewed 218 times

Screenshot 2024-08-26 at 8.23.50 AM.png (526.38 KiB) Viewed 218 times

[exxos]

My Falcon isn't a C-lab one, and this is what mine has. I can't say I ever heard anything odd in the audio, but I rarely have audio on anyway, and my cheap TV doesn't help either.

IMG_2640.JPG (292.25 KiB) Viewed 214 times

So are you suggesting the ITT one is "bad" then ?

[leonk]

I'm suggesting that the ITT CODEC is not as good as the CS branded one. In many posts comparing what C-Labs did to the Falcon in MK1/MK2/MKX, the replacement of the CODEC is always listed.

Time to get one + a new socket and do some testing.

[Steve]

I'm suggesting that the ITT CODEC is not as good as the CS branded one. In many posts comparing what C-Labs did to the Falcon in MK1/MK2/MKX, the replacement of the CODEC is always listed.

Time to get one + a new socket and do some testing.

Yes I'm very surprised that falcon doesn't have the crystal codec, I've never seen one with ITT before. Yeah that'd be worth installing a socket and trying the crystal codec.

[exxos]

I'm suggesting that the ITT CODEC is not as good as the CS branded one. In many posts comparing what C-Labs did to the Falcon in MK1/MK2/MKX, the replacement of the CODEC is always listed.

Interesting. None of the Falcon images we have on here seem to show the ITT part. Looking in the service manual its "C303003-001 / C303170 / CS4216 / ASCO 2300 ". So maybe originally they used the ITT and it changed to CS later on

Time to get one + a new socket and do some testing.

EDIT:

I found both datasheets if someone wants to give themselves a headache..

ASCO2300.pdf

(271.7 KiB) Downloaded 5 times

CS4216.PDF

(802.35 KiB) Downloaded 6 times