Hi,
So I'm new to this, and just before I powered everything up I came across a post suggesting that I need to load all the outputs before I power up or I could damage the amplifier.
Having had many commercial amps over the years, this is news to me. I've ordered some 100w 4 Ohm power resistors so I can load the outputs that aren't in use, but these will be dissipating quite a bit of heat.
What's the actual cause of self destruction of unconnected outputs? I saw some comments mentioning that the amp could go into self oscillation and shoot out high voltages killing the output circuitry. Can we not just put a high value resistor or something on the output to load it while not in use, or does it need to match the speaker resistance?
So I'm new to this, and just before I powered everything up I came across a post suggesting that I need to load all the outputs before I power up or I could damage the amplifier.
Having had many commercial amps over the years, this is news to me. I've ordered some 100w 4 Ohm power resistors so I can load the outputs that aren't in use, but these will be dissipating quite a bit of heat.
What's the actual cause of self destruction of unconnected outputs? I saw some comments mentioning that the amp could go into self oscillation and shoot out high voltages killing the output circuitry. Can we not just put a high value resistor or something on the output to load it while not in use, or does it need to match the speaker resistance?
It is highly unlikely, the output transistors, switch from rail to rail and no power is actually dissipated without a load.
I suggest respectfully, that the post you read is for valve amplifiers.
I suggest respectfully, that the post you read is for valve amplifiers.
Oh look, I'm an idiot when it comes to this stuff, it's all new to me!
Yes, it being a carry over from valve amplifiers makes sense.
I'll link the post I read, but here's the main comments talking about this.
First:
There's then a few people talking about how he damaged it by leaving outputs unconnected. This is the same/similar to the amp I'm working with, so I'm worried about it.
Here's the thread:
6 Channel TPA3116 + DSP
There is a small chance of damage in case the unloaded class-d-amp oscillates at the output filter frequency, about 50kHz. In that case substantial output current may occur creating overvoltage that may damage the output caps. Normally the chip is protected against overcurrent, but may fail if the output inductors saturate.
I suggest shorting the inputs of unused channels with outputs unconnected.
I suggest shorting the inputs of unused channels with outputs unconnected.
Khron and Bucks, basically saying the same thing? Once the channel input is grounded this is a non issue?
Certainly better than driving a 100W resistor and generating a ton of heat 🙂
Certainly better than driving a 100W resistor and generating a ton of heat 🙂
I think its the filter itself that's at risk - class D outputs are MOSFETs with body-diodes, so the outputs are clamped, but the filter itself could resonate until the capacitors fail if driven hard at the right frequency without load.
Hopefully this is unlikely with normal signals - and its entirely possible there's not enough Q in the inductors for sharp resonance anyway.
However there's also the self-heating of the inductors to worry about near resonance - ferrite can degrade over time if run hot, leading it to run even hotter.
Sticking a 'scope on the outputs would answer the question for any given setup of course.
Overcurrent protection requires a minimal output inductance of 1uH to shut-down the outputs in time. If circulating output current exceeds the inductor saturation current, this protection may fail.