D
Deleted member 148505
D
Deleted member 148505
I actually managed to make the PFFB maxed to the limit, deeper than all other implementations out there. (TPA3251 almost running at unity gain)
So, H2-H3 distortion, and IMD is for sure the lowest for all TPA325x.
Will try to send to Amir before selling this (and before anyone copies my implementation lol)
Next on the horizon is -120dB class-AB amp, the moment my TINA-TI ran fast on all simulations, I knew I'm on to something 🤐🤐🤐.
Last edited by a moderator:
D
Deleted member 148505
Never had problems with modules on free air, except when the + and - DC power supply cables are not close together.
Here’s a photo that I found on the net. Running the + and - dc cables in parallel works as well.
If the supply wires form a large “O”, expect nasty interference
D
Deleted member 148505
SylphAudio FB100 PFFB voltage gain is 0.895x !!! (-0.96dB), so the preamp gain section needs to be adjusted to 15dB to get ~14dB system gain. The pcb design is flawless but many parts values will be changed so 1 more iteration to go.
Proof of stability: square wave output with open load, and hard clipping output on 4 ohms (38VDC Supply). Square wave output on 4ohms load does not shut down the module and as expected, produces perfect square wave compared to open load (wasn't able to take pics)
It means that this module will not shutdown on hard clipping or on hard to drive speakers. (some TPA325x designs do shutdown on high power / high volume scenarios)
D
Deleted member 148505
Thanks! FBPreamp gain can be adjusted to 16dB,
FB100 gain is 14dB, so 30dB combined if both are used together.
THD-wise, performance shouldn't be affected. Noise-wise, maybe 2 to 3dBs worse.
I added resistor pad settings on FBPreamp for increasing the gain.
RU and RG are 4pc 0805 SMT resistor pads.
Feedback resistor value on opamps is 1k ohms for ultra-low noise, so LM317/337 supplies are enough.
Using a non C0g/NP0 on opamp feedback cap largely affects the THD (will make harmonics around ~ -80dBc).
Gain RU* RG*
0dB 91Ω Open
6dB Open Open
10dB Open 820Ω
14dB Open 316Ω
16dB Open 220Ω
Last edited by a moderator:
D
Deleted member 148505
Started voicing the FB100 for input coupling caps.
Nichicon FG (default): all good but has a smearing sound on highs (some kind of echo or reverberation on highs).
Nichicon Muse UKZ: Most natural sound (almost perfect), highs are inferior and less refined than FG and both Elna's (Elna's have a hint of plasticky sound).
Elna Silmic II (RFS): Best vocals, rounded midrange sound, fine highs (but something is missing compared with UKZ)
Elna Bipolar RBD: more open midrange but less refined highs compared to Silmic II. Very good alternative to FG
Nichicon FG (default): all good but has a smearing sound on highs (some kind of echo or reverberation on highs).
Nichicon Muse UKZ: Most natural sound (almost perfect), highs are inferior and less refined than FG and both Elna's (Elna's have a hint of plasticky sound).
Elna Silmic II (RFS): Best vocals, rounded midrange sound, fine highs (but something is missing compared with UKZ)
Elna Bipolar RBD: more open midrange but less refined highs compared to Silmic II. Very good alternative to FG
D
Deleted member 148505
I actually removed the 5mm * 5mm WIMA 3.3uF PCB footprint and replaced it with 7mm dia, 2.5pitch, electro footprint. Because in terms of sound, the highs are about equal, and in overall tonality, the electro caps are slightly even better.
Yes, it will be available in January 2024.
SE800 2 weeks from now.
FB100 maybe Q1 2024. FB100 will also have a slightly increase in price since the cost of too many C0G caps aren't expected.
Others are still on vaporware status since I need to upgrade our THD measurement equipment first before proceeding.
- Status
- Not open for further replies.