what is the maximum ripple tolerable on class A amp? is this influenced by speaker sensitivity? the Rifa are a little bit expensive,the 47000uf ones so I`m wondering if 22000uf will also work instead of those? In simulations i get 90mV ripple
More capacitance is good in a single-ended amp because the single-ended topology doesn't have good power supply noise rejection.
I understand.Thank you for the moment.Will do more searching and maybe I will decide on each I will choose the money is a reason too...at least for the moment
32V rail Aleph J
I would like to try building an Aleph J with a small turbo. Having read the F5 turbo manual, I am thinking of using an "as close to standard as useful" Aleph J circuit but powering it with 32V rails.
For the power supply (1 PSU for both channels) I plan on using a 625VA, 2 x 24V AC transformer with a CRC PSU. Each rail would consist of:
I assume the JFets are OK (as in the 32V F5) if I use a heatsink. Luckily I have some Xen/EUVL dual TO-92 sinks. I knew they would come in handy!
What else would I need to change to get it runing properly?
Edit: I forgot to add that I plan to use a 4U 400mm Modushop case (0.20W/C per channel).
I would like to try building an Aleph J with a small turbo. Having read the F5 turbo manual, I am thinking of using an "as close to standard as useful" Aleph J circuit but powering it with 32V rails.
For the power supply (1 PSU for both channels) I plan on using a 625VA, 2 x 24V AC transformer with a CRC PSU. Each rail would consist of:
- C: 2 x 15000uF/63V Panasonic T-HA
- R: 4 x 0.47R/3W
- C: 2 x 15000uF Panasonic T-HA
+ 4 x 1000uF/35V Silmic
+ 10uF/160V MKP
I assume the JFets are OK (as in the 32V F5) if I use a heatsink. Luckily I have some Xen/EUVL dual TO-92 sinks. I knew they would come in handy!
What else would I need to change to get it runing properly?
Edit: I forgot to add that I plan to use a 4U 400mm Modushop case (0.20W/C per channel).
Last edited:
cascoding input jfets is recipe for better sleep
Ah, but then I would not be able to use the Peter Daniel boards.
How high can I go before I need to cascode? 2 x 22V AC be OK?
you can cut few traces and drill few holes
that's just 2 BC transstrs and 3 resstrs , plus one cap
that's just 2 BC transstrs and 3 resstrs , plus one cap
Thanks Zen Mod, but I am afraid it would be too difficult for me at this stage. I am a newbie. So far, I have only assembled an O2 (headphone amp) and a DAC (from Spencer/fetaudio). Both were complete kits, with instructions 🙂
Would I get away with a 2 x 22V AC transformer (29-30V rails) and keeping everyhing else standard?
An Alpeh J with a bit more power. Read the Firstwatt F5 turbo manual on the Firstwatt site. Or check out the F5 turbo discussions on the forum.
For the Z-th time : the numbers Modushop mentions are for Class AB amps !
At over 130W dissipation per channel, the heatsink temperature will be at least 60C.
You'll not get more powah merely by 32V rails, requires more action : either higher bias or larger turbo boost.
Be careful, as i am not 100% sure that the ACS will function properly with the diodes in place.
Thanks Jacco, but I don't understand why the Modushop enclosures would not work. I have a set of Aleph 2's (monoblocks) running in identical enclosures. A friend of mine made them for me 6 years ago, and they have been running fine ever since. If I remember correctly the Aleph 2 is rated at 300W dissipation (per channel), hence my choice for this enclosure.
Nelson explains that on the F5, upping the rails 32V will give about 50W.
Why would this not work on the Aleph J? I do not expect a big boost, but a bit more would be nice 🙂
It is about more than rails. I believe the AJ in standard setting is biased higher than F5. Pick your poison, just watch the temperatures. You could double up on the FE Jfets and increase number of output fets like i did in my monoblock build.
Aleph J is single ended, peak output current is 2.5A
2.5A peak is 25W continuous in 8 Ohm.
32Vdc rails means ~50W continuous in 8 Ohm, is 100W peak.
100W peak is ~3.5A peak current.
1A more current implies either more bias or fiddle with the dynamic current source.
2.5A peak is 25W continuous in 8 Ohm.
32Vdc rails means ~50W continuous in 8 Ohm, is 100W peak.
100W peak is ~3.5A peak current.
1A more current implies either more bias or fiddle with the dynamic current source.