Help with power amplifier current gain stage

[qa9b]

Hi all,

I'm using an LME49830 audio amplifier chip which serves as the input stage and voltage amplification stage for a power amplifier. It has outputs Nout and Pout which connect to the gates of a mosfet current gain stage. (see attached schematic).

I'm having a bit of difficulty constructing this output stage, as things have a tendency to blow up. I've gotten it functioning just great with a single output stage (one nmos, one pmos), but when I added a second stage (my wiring was correct to the best of my knowledge) the thing went up in smoke. This is a little frustrating, as I now have to replace all the power mosfets, the LME49830, and hope this doesn't happen the next time I power it on. On a side note, does anyone know some good remedies for this? I was considering placing a temporary fuse of very small value (maybe .5A) on each of the power supply rails to limit short circuit current.

Anyway--my main question. In the schematic that is attached, I left out the source resistors. Their typical value is from .1 to ~.2 ohms. Is this a no-no when using multiple output stages? Also, I left out the zener diodes that are shown -- these are purely protection devices, correct? That is, if they're 10V zeners, they limit the Vgs to +- 10V?

Thanks for any help or general advice!

 

[alec_t]

Definitely a no-no!. The resistors are there to balance the FET currents. Without them most of the current passes through one FET, which fails and diverts current to another FET which also fails, and so on. Assuming the speaker has significant inductance it can generate a high back-emf which can destroy the FETS too. So the remedy is install the source resistors and zeners.

 

[qa9b]

Ah. Thank you. That actually makes quite a bit of sense. And also explains why my four mosfets quite hilariously began emitting fire in quick succession.

How should I determine the optimal value of the resistor to use there? Like I mentioned, I think I've seen both .1 and .22 ones used depending on the design.

 

[alec_t]

depending on the design

Precisely so. As your schematic has no component values (e.g. supply voltage, speaker impedance, FET type, output power required) we can't specify the source resistor values. But the values quoted are probably in the right ball park.

 

[qa9b]

Thanks for your help!

 

[Mr RB]

In amp testing I remove the + and - rail fuses and replace with 10 ohm fusible resistors, or any 10 ohms resistors will do. They will smoke up before the output drivers. They also allow easy current measuring or scopeing.

Also I use a resistive dummy load, and usually a higher value like 50 ohms, 27 ohms etc, whatever 25w resistors I have handy.

If the output stage produces a good high amplitude waveform into 50 ohms and the 10 ohm resistors stay uncooked then it is usually safe enough to reduce the dummy load resistance and put some fuses in for the next set of tests.

I usually leave a large resistor (27 ohms?) in series with the speaker too for the bulk of the testing, as most amp fails will go to full rail voltage and it will save your speaker from instant destruction.