Ziggurat passive shunt for chipamp
Two tiers in order to achieve the lowest ESR - top is using 4700uF and bottom, 15,000uF. Total around 0.8F, 25V.
I plan to try this out first with my TDA8566 but that's limited to 18V so if that sounds promising I'll look for a new chip which can handle the full 25V supply capability.
Update1 : Wired the top tier only underneath my new TPA3116 class D amp board from Taobao. Seems here the dynamics aren't limited by the capacitance on the output stage supply, rather on the analog supply (pin17 of the TPA3116) which goes to an internal LDO regulator. No amount of capacitance on this pin (I went up to 100,000uF) gives the chip the LF dynamics of my class AB amp so I've been off in search of a class D chip which doesn't have an internal reg so might be amenable to a hexacap on its supply...
Update2 - NXP's TDA8932 is on order now - this is a chip which can be DIYed because its in a manageably small SO32 package, so no need to order up an eval board. Over on TI's site, a reviewer from Germany says NXP's class D chips sound better than the TPA3116.... I tried to add my own review there but the website has a bug (which I reported). I wonder if the German engineer filtered pin17 - an older chip from TI has the recommendation to do this which seems to have gotten lost in the DS for the 3116.
Update3 - I submitted my review of the TPA3116 to TI's website, they eventually fixed the bug. I'm now using a pair of the 4700uF/25V hexacaps to power my newest creation, a double TDA1521 amp (bridged, paralleled). Will write about that amp in my next blog post probably. Anyone else tried using TDA1521?
Update3 - added pic of my heavily tweaked TPA3116 board.
I plan to try this out first with my TDA8566 but that's limited to 18V so if that sounds promising I'll look for a new chip which can handle the full 25V supply capability.
Update1 : Wired the top tier only underneath my new TPA3116 class D amp board from Taobao. Seems here the dynamics aren't limited by the capacitance on the output stage supply, rather on the analog supply (pin17 of the TPA3116) which goes to an internal LDO regulator. No amount of capacitance on this pin (I went up to 100,000uF) gives the chip the LF dynamics of my class AB amp so I've been off in search of a class D chip which doesn't have an internal reg so might be amenable to a hexacap on its supply...
Update2 - NXP's TDA8932 is on order now - this is a chip which can be DIYed because its in a manageably small SO32 package, so no need to order up an eval board. Over on TI's site, a reviewer from Germany says NXP's class D chips sound better than the TPA3116.... I tried to add my own review there but the website has a bug (which I reported). I wonder if the German engineer filtered pin17 - an older chip from TI has the recommendation to do this which seems to have gotten lost in the DS for the 3116.
Update3 - I submitted my review of the TPA3116 to TI's website, they eventually fixed the bug. I'm now using a pair of the 4700uF/25V hexacaps to power my newest creation, a double TDA1521 amp (bridged, paralleled). Will write about that amp in my next blog post probably. Anyone else tried using TDA1521?
Update3 - added pic of my heavily tweaked TPA3116 board.
Total Comments 0
