What's happening is the TPA is going into an overcurrent fault, likely due to the output inductors saturating because they're undersized. And the board likely just has SDZ/FAULTZ tied together and MUTE tied to ground. You're buying a cheap board - implementing proper inductors and a proper startup circuit costs money!
So overcurrent protection trips, SDZ gets yanked low, FAULTZ releases and SDZ starts right up again without any delay. This causes the GAINSLV divider voltage to be improperly sensed and the corresponding gain reduction that you're seeing.
So overcurrent protection trips, SDZ gets yanked low, FAULTZ releases and SDZ starts right up again without any delay. This causes the GAINSLV divider voltage to be improperly sensed and the corresponding gain reduction that you're seeing.
Switch back to 20dB gain could be restart in default gain as gmarsh points. Now what limits output to 5Vpp in 26dB gainsetting , increasing to 10Vpp in 36dB gainsetting, saturationpoint inductors seem irrelevant, that doesn't change with gainsetting, inputsignal signalgenerator was variable, not limited/fixed???
Here's thermal image of the board after 5 minutes 1khz sine wave into 4 ohm resistors. As you can see, inductor heat-up is almost unnoticeable, but chip itself is quite hot.
Signal generator has no fixed output levels, it can be adjusted freely, but output voltage stability across the whole range is great, I've already checked that.
I agree with xrk, in the fact that this is also the board I got and have had no issues.
I've had hooked up to various devices such as smart phone and tablet through headphone cable, and CD source through passive preamp...all 3 sources using both SE and balanced connections at some point.
Had the board powered with 20V 5A brick and drove 8" full range loud with no issues.
Just mention all this to highlight that this particular issues must be individuals board fault, leaving aside the actual design implementation.
The soldering quality on mine isn't great out of the box. Before I redid joints, the amp would snap, crackle and pop anytime I put finger anywhere near one of the input caps.
I've had hooked up to various devices such as smart phone and tablet through headphone cable, and CD source through passive preamp...all 3 sources using both SE and balanced connections at some point.
Had the board powered with 20V 5A brick and drove 8" full range loud with no issues.
Just mention all this to highlight that this particular issues must be individuals board fault, leaving aside the actual design implementation.
The soldering quality on mine isn't great out of the box. Before I redid joints, the amp would snap, crackle and pop anytime I put finger anywhere near one of the input caps.
Last edited:
😀 You can try.
I've a box of Junk Chinese amps. My First 3116 is 'in transit' In truth All mine are Ta2020 types.
A couple from YJ too. The quality consistency is non existant, the parts are typically ALL fake from Caps to Chip itself.
A couple were DOA most others Died (lost a channel) within the week.
One YJ is surviving, 2 years later... Good 🙂
It's sounds are OK. In Garage tunes service though.
Nothing like my Diy Ta2020 veroboard Tripath data sheet version.
Likely because the Ta chip I used was Genuine.
Few to None are these days. It was the last one I found and hoarded, several years earlier
Hopefully the 3116 chips are not yet altered cheater copies??
Dunno.. as of yet.
Well, I'm also quite happy with my DIY 2x TDA2030A BTL amp, just I wanted something with smaller heatsink and smaller transformer 🙂
Please check the bootstraps connection and voltage.
(Square wave input is just thermal overloading the module)
(Square wave input is just thermal overloading the module)
But can someone explain, is 9Vpp normal for such amp?
By my calculations 9vpp @ 4 ohms means 20W per channel, while this chip should be able to output 50W. Maybe this board has another chip, say, TPA3130D2 ?
By my calculations 9vpp @ 4 ohms means 20W per channel, while this chip should be able to output 50W. Maybe this board has another chip, say, TPA3130D2 ?
I'm still suspecting it's inductor saturation. The TPA current limit is 7.5A typical, there's no minimum number listed in the datasheet but let's say it's 6.5A.
9Vpp single sided on the output means you're putting 18Vpp BTL voltage into your 4 ohm load, or about 4.5A DC current into the load. A 10uH inductor at 400KHz adds +-1.2A of delta-IL current on top of that bringing us up to 5.7A assuming we've got a good inductor. But the inductor saturates and its inductance drops by half to 5uH, that delta-IL current increases further to 6.9A... that's getting awful close to the current limit of the chip.
Who knows what the saturation current of those cheap inductors is...
9Vpp single sided on the output means you're putting 18Vpp BTL voltage into your 4 ohm load, or about 4.5A DC current into the load. A 10uH inductor at 400KHz adds +-1.2A of delta-IL current on top of that bringing us up to 5.7A assuming we've got a good inductor. But the inductor saturates and its inductance drops by half to 5uH, that delta-IL current increases further to 6.9A... that's getting awful close to the current limit of the chip.
Who knows what the saturation current of those cheap inductors is...
I have 50 amp Isat inductors from coilcraft 🙂, use them?
but if this is a lower power IC, would be symptoms be the same?
but if this is a lower power IC, would be symptoms be the same?
Coilcraft MSS1210 should drop directly in place of the existing inductors. Not quite as overkill as the ones you've got, but much more practical 🙂
Anyway, pull the heatsink off the chip. If it's got a top side thermal pad, it's a TPA3116. Glue it back on with some Arctic Alumina or something when you're done.
Used them lots of times, they're great, never had a bad PCB.
I use them, dirtypcbs.com, elecrow.com and iteadstudio.com - I'm pretty certain all of these companies order from the same board house though, as the quality, manufacturing specifications, color options, etc... is consistent across all of them.
Saturation by itself doesn't cause inductors to heat up, most heat will come from the DC resistance of the inductor windings and the RMS current going through them. If you're amplifying music, the RMS current is fairly low.
Even with 2A RMS (say) and 20mohm inductors, that's 80mW copper loss. In a 12mm inductor package that's barely noticeable. There will be some hysteresis loss too, but it'll probably be a small fraction of that.
Even with 2A RMS (say) and 20mohm inductors, that's 80mW copper loss. In a 12mm inductor package that's barely noticeable. There will be some hysteresis loss too, but it'll probably be a small fraction of that.