Input ripple is only part of the issue, the other one is load-induced ripple.
The diyAudio gallery is easy to use and the photos stay internal to the forum. Check out the video. 
How to use the Gallery Video
How to use the Gallery Video
thanks for sharing.
if you don't mind my asking, what do you use to drive it and what do you use for the input cap, if any, that i'm guessing you may have mounted under the board.
No Input cap. My preamp is from Richard Murdey. There are two Black Gate caps at the output. The PSU is the snubber version. The Toroid ist 2x25 V 600 VA. All resistors are TKD from Thel Audio in Germany. The Feedback resistor is direktly soldered from pin 3 to Pin 9 at the LM 3886 with 9 pF Silver Mica cap in parallel.
The 0.18 Ohm are from Isabellahütte 3 Watt manganin.
https://www.hifituning24.de/passive_bauelemente/hochlast_widerstaende/isabellenhuette_pbh/index.php
The 0.18 Ohm are from Isabellahütte 3 Watt manganin.
https://www.hifituning24.de/passive_bauelemente/hochlast_widerstaende/isabellenhuette_pbh/index.php
Last edited:
sure there is drift.
with quality components drift during warmup is extremely low.
here's what i'm getting for each chip's output offset:
cold:
0mv 0mv 1mv / 1mv 0mv 0mv
hot:
1mv 1mv 2mv / 4mv 3mv 2mv
the worst case delta on the first board is 1mv
the worst case delta on the second board is 2mv
the worst case net delta side to side is under 3mv (with one side hot and one side cold).
3mv into either 8R or 4R is less than one milliwatt.
i'm reasonably confident the chip can handle that.)
with quality components drift during warmup is extremely low.
here's what i'm getting for each chip's output offset:
cold:
0mv 0mv 1mv / 1mv 0mv 0mv
hot:
1mv 1mv 2mv / 4mv 3mv 2mv
the worst case delta on the first board is 1mv
the worst case delta on the second board is 2mv
the worst case net delta side to side is under 3mv (with one side hot and one side cold).
3mv into either 8R or 4R is less than one milliwatt.
i'm reasonably confident the chip can handle that.
)Z in BPA 300
Hi.. all.
Sorry berofe about my english.
I am very interested in this excellent work.
How about Z input, and how to calculate it?. I am Newbie of this matter. My plan just build BPA300 half, does not build push pull and leave DRV. For my plan build a tube preamp that fits its for maching impedance for this Amplifier. tq.
Hi.. all.
Sorry berofe about my english.
I am very interested in this excellent work.
How about Z input, and how to calculate it?. I am Newbie of this matter. My plan just build BPA300 half, does not build push pull and leave DRV. For my plan build a tube preamp that fits its for maching impedance for this Amplifier. tq.
The simple answer is that you'll have to put a 4700uF or 10000uF capacitor in series with your speakers. There are other ways of doing it which don't have the LF cutoff dependent on the resistance of the speakers but they aren't catered for here. This is not an especially ambitious implementation of the 3886 and I didn't run mine this way. I'd also say that you are wasting most of what could be done with this by only running it as a PA 150. The advantage of running things differentially is huge. Apart from output impedance, which is again unambitious here with 0R2 resistors balancing the current, there is little to be gained by having 3 LM3886s in parallel except spreading the thermal burden - which is valuable and makes for some of the worst parts of the failings. But a single 3886 can deliver 11A, which is far more than any average speaker can manage.
If you can send the amp a proper differential signal, ie. one the inverse of the other (not that it matters from the point of CMRR) you will find it is a class above what you get from a single ended signal and all the trouble around finding a true ground. When I did that for the first time about 7 years ago, it was the biggest single improvement I had ever heard on any amplifier. It also allows you to drop the gain by half and thus extend the GBWP.
Now if you don't understand this then take Andrew's advice and build a single chip version. I would still advise you to drive it differentially. There is no solution to a non differential amplifier, and it just gets worse the more current you have at your disposal. There is a very good solution otherwise.
If you can send the amp a proper differential signal, ie. one the inverse of the other (not that it matters from the point of CMRR) you will find it is a class above what you get from a single ended signal and all the trouble around finding a true ground. When I did that for the first time about 7 years ago, it was the biggest single improvement I had ever heard on any amplifier. It also allows you to drop the gain by half and thus extend the GBWP.
Now if you don't understand this then take Andrew's advice and build a single chip version. I would still advise you to drive it differentially. There is no solution to a non differential amplifier, and it just gets worse the more current you have at your disposal. There is a very good solution otherwise.