Yes, aiming for the 6.2 - 6.3 range. With 21 turns for L2 it was a little high, peaking in the 6.4 region so I tried 23 turns which brought it down to 5.8, so I've settled on 22 turns for the moment which seems about right with the ability to squeeze or open the coil a little.
OK.
I did see that you had used the Wondom board, they seem to be pure unobtainium in the UK, I have accounts with Farnell and their sibling CPC and they don't have them.
Forgive me, I was wrong and it appears that Farnell does
not carry them nor does AliExpress any longer ("no longer available"). Well that's a bummer. When I started this thing in 2021 they were quite available but of course 2 years is an eternity in this industry, especially with the disruption of COVID-19.
I'll trawl AliExpress and Banggood to see what they have.
I have tried a clone 100w mono board from eBay also based on the TPA3116 but no joy at all. I'll give it another go this weekend when I have some more time.
I have a DROK (or Bangood?) module or two and they are problematic in this application. The audio cuts out intermittently, especially at high output and I assume it to be the over-current protection circuitry doing this. The protection appears to be a bit more over-zealous than a real TPA3116-based module. (The U-LULU is something like a 3-4 Ohm load according to simulation.)
Whatever you end up going with, it needs to drive at least 5 Amps peak at 4 Ohms (100 Watts into 4 Ohms differential = 5 Amps RMS). Of course, we are using the amplifier single-ended with the U-LULU. Four Ohms differential (such as a loudspeaker) is equivalent to 2 x (2 Ohms single-ended to ground). If it can handle a 4-Ohm speaker then it should be good to go. Of course the Bangood and DROK are supposed to drive 3 or 4 Ohms too, so.....
For this weakness and a few other reasons (that I am too lazy to describe right now) I suspect Banggood and DROK modules are built around counterfeit TPA3116 chips that can't drive a 2-Ohm load like the real thing from TI.
Caveat emptor.
Show Posts
