Rev 1242 | Details | Compare with Previous | Last modification | View Log | RSS feed
Rev | Author | Line No. | Line |
---|---|---|---|
1229 | lvd | 1 | --== audio specs and reqs ==-- |
2 | |||
3 | |||
4 | 1. have sound from AY/YM and beeper/PWM-covox |
||
5 | 2. have some extra line-ins |
||
6 | 3. independent volume regulation for all of them |
||
7 | 4. line out/headset output also volume regulated |
||
8 | 5. be able to drive 32 Ohm headsets |
||
9 | 6. Use only +12v power supply for audio |
||
10 | 7. line-out also goes to extra video output pins |
||
11 | |||
12 | |||
13 | |||
14 | |||
15 | --== audio decisions ==-- |
||
16 | |||
1230 | lvd | 17 | (YM2149) |
1229 | lvd | 18 | |
1230 | lvd | 19 | voltage on outputs (loaded as per rev.C) |
1241 | lvd | 20 | 2.26v min, 3.40v max => 1.14 Vp-p [me] (max 1.35Vp-p by datasheet) |
21 | 2.22v min, 3.36v max => 1.14 Vp-p [dimka_m] |
||
1242 | lvd | 22 | 2.29v min, 3.42v max => 1.13 Vp-p [Nik X] |
1230 | lvd | 23 | |
24 | 1.5*1.14 Vp-p (A/C + half B) to 2.0 Vp-p is K=1.17; |
||
25 | |||
26 | |||
27 | (AY8910) |
||
28 | |||
29 | Per datasheet, max 2mA current per output, 2 Vp-p @ 1k |
||
30 | |||
1242 | lvd | 31 | (KC89C72) |
1230 | lvd | 32 | |
1242 | lvd | 33 | 0.5v min 1.27v max => 0.77 Vp-p [Nik X] |
34 | 0.5v min 1.29v max => 0.79 Vp-p [Nik X] |
||
35 | |||
36 | |||
37 | |||
38 | |||
1229 | lvd | 39 | (PT2258 related) |
40 | |||
41 | 1. Use PT2258 to control volume: |
||
42 | 1 channel for AY/YM |
||
43 | 1 channel for beeper/pwm-covox |
||
44 | 1 channel for output volume |
||
45 | 3 channels for line-ins (same number of line-ins as in rev.C) |
||
46 | total 2 ICs |
||
47 | |||
1241 | lvd | 48 | 2. max Vp-p for PT2258 is 2v (1v amplitude or 0.7Vrms, see PT2257 datasheet and THD vs amplitude there). |
49 | |||
1229 | lvd | 50 | |
51 | 3. Vcc for PT2258 is ≈9v, can't be 12v. So, there's AVCC9 regulated voltage |
||
52 | |||
53 | |||
54 | (headamp related) |
||
55 | |||
56 | 1. Use TPA6120 to drive line-one/headset: high power AND high quality |
||
57 | |||
58 | 2. Use AVCC12 (non regulated, directly from AT/ATX and passively filtered) |
||
59 | to get maximum output swing thus power |
||
60 | |||
61 | 3. To compensate for inevitable series resistors at the TPA6120 outputs, |
||
62 | use opamps. |
||
63 | |||
1257 | lvd | 64 | 4. Output resistors should be of high thermal power, preferably MELF or 1205 |
1229 | lvd | 65 | |
66 | |||
67 | (general) |
||
68 | |||
69 | 1. Use metal film feedback resistors (bigger than 0805, best if MELF), |
||
1241 | lvd | 70 | NP0/C0G capacitors for low pass filters. Footprint for 1206+0805 is ideal for that. |
1229 | lvd | 71 | |
72 | 2. DC blocking/high pass filtering can't be done with polypropylene/etc. caps (too big!), |
||
73 | so we have to use electrolytics. Try to have related cutoff freqs as low as possible (<<20Hz) |
||
1241 | lvd | 74 | to decrease distortion from them. Typical distortion disappearance happens at *20 cutoff frequency |
1229 | lvd | 75 | |
1241 | lvd | 76 | Gain (or loss) distribution. |
77 | 1. YM seems to have 1.14 Vp-p (or 0.57 ampl) outputs each. Not yet known what is with AY, but that's for future. |
||
78 | 2. for PT2258, supposedly optimal ampl. is 1V, more leads to distortion (guessed from datasheet for PT2257). |
||
79 | So we'll keep amplitude from each internal source like that. Afterwards it could be decreased before global |
||
80 | volume control and will be amplified by audio output circuit. |
||
1229 | lvd | 81 | |
1241 | lvd | 82 | |
1229 | lvd | 83 | (low-pass filtering) |
84 | |||
1241 | lvd | 85 | 3-rd order butterworth, 30kHz -3dB. |
1257 | lvd | 86 | First stage for AY is 2nd order Q=1, (multiple-feedback). 1st stage is 2nd order to keep input impedance more or less constant for AY/YM. |
87 | For beeper there's low impedance buffer, we use single 3rd-order opamp based stage. |
||
1241 | lvd | 88 | Second stage is just 1st-order RC |
1229 | lvd | 89 | |
90 | (high-pass filtering) |
||
91 | |||
1257 | lvd | 92 | "distributed" 2nd-order high-pass filters with electrolytics. |
1229 | lvd | 93 | <<20Hz cutoffs |
94 | |||
1257 | lvd | 95 | (line output vs. headphones output) |
1229 | lvd | 96 | |
1257 | lvd | 97 | for headphones there's a need for huge output capacitor. |
98 | For line output relaxation time would be then too big (like tens of seconds), so need different (smaller) cap. |
||
99 | Also maybe have line output without volume regulation, from summing opamps. |
||
100 |