There's a long list of reasons. The main one is cost. They're $10 each * 4 supplies (I'm only using film on the analog supply). That's for 47uf, I can't get them any larger within my budget. Another reason is that due to their large size they force me to run large power islands which, I'm gravely warned, might resonate (if I don't use the power islands, there are worse problems). If the cells can stay within 300% of their rated impedance up to 200khz it'd solve a boatload of problems.
Yea, i dont think you are going to have any resonance problems. I dont know where you got that idea from. You wont get resonance in the dc supply thats providing 10ma to a preamp.
There is no need for film, good electrolytics will do the exact same job, and be tiny compared to the film.
Its a dc supply from batteries.
DC. From BATTERIES.
Look at the power supply section of a Cmoy amp. That will give you an idea of whats really required.
Here is the diagram from a stock cmoy amp.
Notice the ZERO caps across the rails, apart from those used to create the virtual ground, and standard electros are used.
http://www.mypage.tsn.cc/dakiller/cmoy.png
There is no need for film, good electrolytics will do the exact same job, and be tiny compared to the film.
Its a dc supply from batteries.
DC. From BATTERIES.
Look at the power supply section of a Cmoy amp. That will give you an idea of whats really required.
Here is the diagram from a stock cmoy amp.
Notice the ZERO caps across the rails, apart from those used to create the virtual ground, and standard electros are used.
http://www.mypage.tsn.cc/dakiller/cmoy.png
Last edited:
I'm designing the decoupling network using a 3d solver and taking into account parasitics of everything including the IC package. In my circuit there's a large impedance peak caused by the 47uf and parasitics its use introduces. I'm trying to keep the supply rails' impedance below 300mR from 20hz to 2.5mhz (1/(0.5*rise time)), so even if I dampen that anti-resonance, it still doesn't fix the other problems I mentioned. If the battery can stay below 50milliOhm/cell up to 100Khz it'd solve all those problems. Originally I was sceptical that it could but having done some more researrch, I think it might be possible.
Last edited:
Looks like the reservoir cap can go: http://www.dact.com/html/ct102_vs_batteries.html. Found some papers on IEEE that go only up to 10Khz but which tally with that website. Considering NiCD have 25mR/cell vs the alkaline's 250mR/cell, it looks like they're good up to 100Khz.
Try LiFePO4 3.3V batteries - internal impedance 8mOhm( a third of the NiCd cells) - leave out all the caps - problem solved 🙂
I'm actually using one of those to power the clock. They're amazing cells (120A pulse!). Not too expensive either. The DAC needs +/-5V unfortunately, so stuck with NiCD there.
any great ideas to get +/-2v out of 48V SLA(4x12V series)
currently I just connect through a centre tap, but it gets unbalance voltage after some play time
currently I just connect through a centre tap, but it gets unbalance voltage after some play time
- Status
- Not open for further replies.