My MOSFET amplifier designed for music

[Sadface]

Makes sense. Thanks

 

[Sadface]

Is a 5s delay at startup sufficient?

 

[prasi]

3of 4 holes lined up on the modu pre drilled heatsink….for the mosfets. Drill/tap the other 5.

Thats a very nice build... congrats..
Is there insulation on the bottom of mosfets?

 

[Mooly]

Is a 5s delay at startup sufficient?

Just about. It should be 99% silent after that time if you have retained the original component values that determine the time constants.

 

[Monk55]

Thx Prasi….there will be! Just test fitting.

 

[Preamp]

Just had a quick play with my newly acquired, pre-owned 4-channel Rigol late last night. Didn't have a USB stick handy, so I've made a crappy screenshot with the phone. It shows the startup behaviour of my build, as it is currently hooked up with only a 18V-0-18V transformer for testing. Channels 1 and 2 are the left and right outputs, channels 3 and 4 the power rails. Horizontal scale is 100ms/div, so the whole screen spans only 1.2 seconds. The pulse on the output only reaches about 12V (less than half the rail voltage) and settles in about half a second, which is in good agreement with what the simulation predicted.

 

[Mooly]

That all looks good and nice scope

Simulation is actually really good for looking at this kind of behaviour and its good to see theory and practice agree.

 

[poseidonsvoice]

Lasse,

I had a similar issue several months back and using the scope in Roll Mode as you have done to discover a ‘shut down’ behavior which was making a plump sound on my speakers. Turns out it wasn’t the main amplifier circuit. It was the preamp section and it happened within the first 100msec of shutdown.

So good sleuthing, and yes, it appears a speaker protect with 5 second delay would be MORE than adequate since your issue is resolved in under 500msec.

Best,
Anand.

 

[Preamp]

5 seconds more than adequate for my build, that is! Remember that I have used a different time constant for the servo circuit and that the settling time of Mooly's original is somewhat longer than mine.

 

[poseidonsvoice]

that the settling time of Mooly's original is somewhat longer than mine.

Do we have an idea of what that is? I will be building mine and my speaker output protect is set for 5 seconds on startup.

Best,
Anand.

 

[Preamp]

IIRC Mooly had posted some simulation results already, and after 5 seconds there'd be still some 10's of millivolts left, but that should not be audibly objectionable. I'm not in front of a computer right now, else I'd have plotted a simulation result on the quick.

Edit: Have a look here. Mooly beat me to it below.

 

[Mooly]

This is representative. The sim has a "real" PSU.

And from 2 seconds to 10 seconds.

 

[Mooly]

I timed my real amp and it set to about 6.5 seconds delay and the actual switch on noise is 99.9% non existent.

 

[poseidonsvoice]

Thanks. I’ll have to ask if the delay time of my protect board can be extended (if I needed it to be….best to test first). My voltage rails will probably be only +/- 40V max so that is also important to take into consideration.

Great data, both simulation and real world Mooly and Laase 💪 .

Best,
Anand.

 

[Chrisr3521]

Which type of inductor to use with this amp, air core or iron core? Does it matter?

 

[Preamp]

Yes. Always use air core (have a look at Mooly's photos for a size guide). Never use iron core, because they tend to introduce some unwanted nonlinearities.

 

[Mark Tillotson]

The jury is out on gadolinium cores though (!)

 

[Preamp]

Gadolinium below its Curie point of 20 °C (68 °F) is ferromagnetic

Better put the output inductor on the main heatsink then?

 

[Monk55]

I thought I read Mooly left out the inductor?

 

[Preamp]

You can see them in Post 1499. Looks like 11 turns with something like 1 inch inner diameter.