Another topic with question about filter caps

[QBAS]

Hi again

I read topics about filter caps, and I have decided to use 22n becouse some guys said that it is safe value. Also I noticed that on C64 orginally was 22n too.

So firstly I used canibalized ceramic 22n only for testing. They give me beauty sound.

Since I`m finish additional 7 sid modules I bought regular ceramic 22n, but they give me  similar sound, not the same like previously. Personally I like more sound this from older one caps.

What is your proposition for me:

a)looking for another 22n

b)changing to another value (what are you propose? I looking for brighter sound since I use external filter too)

c)waiting with this new caps, becouse after some time they will play better (it is my theory)

[QBAS]

What a caps are you use? It`s a regular from shop or a special one for audio?

And what value for 8580?

[TK.]

I'm using these 22nF caps:

http://www.midibox.org/forum/index.php?topic=9132.msg64907#msg64907

Best Regards, Thorsten.

[Sasha]

I am using 22nF caps too (block) as I couldn`t find better quality - styroflex with low tolerance. Never have chance to test different caps but seams they affect sound more than I think.  :-\ I really like the sound of my SIDs but worth some more effort if they can sound even better.

[QBAS]

Thank you for fast response.

Your arguments are good for me. Ok I will try to get this 100 superb caps. For SID I can do everything.

[Sasha]

For SID I can do everything

...same, but don`t let your girlfriend find out.  ;D 

[QBAS]

don`t let your girlfriend find out

Yes this is danger. I cannot be in obsesion.

Fortunately she loves SID sound too.

I would like to build delay into sid  (with precise time setting+possibility equalizing feedback). Do you know any DIY project with this kind delay?

[stryd_one]

You might be interested in the BBD delays, personally I like the PT2399. You can find lots of guitar stompbox designs that should do nicely around the net :)

[QBAS]

Thank you stryd_one, revelation!

[stryd_one]

Sorry to hijack...QBAS check out the robote 2.5 ;)

[QBAS]

Yes I see. Very simple and also can be (probubly) controlled from AOUT_LC by SID - inside patch.

I will do 2 stage delay: first one short (for modulating flangers) and second for longer delay, but first I will compare with your another great propositions at this forum :)

[QBAS]

I don`t wanna bother, but could somebody explain me (or put a link) why it is important to  select the most the same values of filter caps?

What happens when they are different?

Please be patient for my questions.

[Wilba]

The filter caps control the filter's cutoff frequency.

I don't know how different values of filter caps on one SID makes a difference to filter cutoff, it might just average between the two values, or maybe the cutoff frequency won't be stable...

But when using two SIDs for stereo, if the cap values are different, you will notice the difference in cutoff frequency between them, which may cause phasing/stereo effects when you didn't want them.

[stryd_one]

I don't know how different values of filter caps on one SID makes a difference to filter cutoff,

It changes the linearity of the cutoff curve... Something about the output of the filter being the product of the audio signal passed through each cap compared with each other... I've got a bookmark somewhere... Mike?  ;D LOL

[TheAncientOne]

From the original Commodore data sheet:

These pins are used to connect the two intergating capacitors required by the programmable filter. Both capacitors should be the same value. Normal operation of the filter over the audio tange (approximately 10Hz-12KHz) is accomplished with a value of 2200pf for C1 and C2. Polystyrene capacitors are preferred. In complex polyphonic systems where many SID chips must track each other, matched capacitors are recommended.. The frequency range of the Filter can be tailored to specicfic applications by the choice of capacitor values. For example a low cost game might not require full frequency response. In this case, larger values for C1 and C2 could be chosen to provide more control over the bass frequencies of the Filter. The approximate minimum Cutoff Frequency of the Filter is given by:

FCmax = (2.6E - 5)/C

Where C is the capacitor value. The cutoff frequency extends aproximately 9 octaves below the maximum Cutoff Frequency.

For the non-mathematical 2.6E-5 = 0.0000026 To save calculation, using 2200 pF, (2200E-12 Farads), plugging in that value in gives 11.8KHz.  NOTE: I've used a "." as a zero seperator. The English and the US use "." Europe uses ",", Which is why, if the story is true,  an engineer at a large French music institution  has a big polyester paper weight. It's probably the same mythical person said "Mikro Farad", to a German engineer, and ended up with a very small bicycle.

I can't find an 'opened out' schematic, showing the filter, but I guess it's some form of twin integrator loop, (clue in the text), so the capacitors must match or the thing might go unstable at one end of the range.

Mike

[QBAS]

Thank you for this. Ok I see that it was a good choice to order this superb http://www.midibox.org/forum/index.php?topic=9132.msg64907#msg64907 capacitors.

It seem that it is a time to buy better measure tool (with ac and dc ampers, and with capacity), ahhhh .... nice one could be also have sonometer. It`s my dream: to have sonometer for checking rms level of my speakers.

OK guys - Thank you one more time (not last of course)  ;)

[TheAncientOne]

Now I'm confused.....

The Commodore 6581  datasheet I have, says 2200pF = 0.0022uF, or 2n2

You're using 22n = 22000p or 0.022uF

Data sheet for 6582 says 6800pF = 0.0068uF or  6n8

(Same for 8580)

Have I got bad data sheets, Do we just prefer the sound of a lower filter cutoff?

I guess the real answer is to hurry up and finish my Basic MIDIbox SID and do some tests.

Edited to add - I've ordered a pack of the 22nF too, so if anyone in the UK needs a few, memo me.

Mike

[Wilba]

The values in the datasheets don't match what is in real C64s... the 22nF value was suggested to me originally by a HardSID user, confirmed as the actual value in C64 models with 8580, and backed up with audio tests and that formula, which  proves it's a good optimization of cutoff frequency range vs. resolution. In short, 22nF gives you smoother bottom range than 6nF, if you want a smoother bottom range, then increase it to 47nF or 100nF. However, those caps I suggested are a great deal, you won't be disappointed  ;D

[TheAncientOne]

Brilliant Wilba, and thanks for clearing this one up, a bit baffled for a moment. I've actually got some 22nF polystyrenes, but they're HUGE.

Near to this topic, I wonder if Smash might like to put a slightly bigger cap area, perhaps with a few extra holes, on his PCB, to facilitate experimentation? The present board is a bit of a tight fit for anything other than 2.5mm pitch caps.

Mike