i'm trying to build a discrete class D amp, i've implemented dead time and added a gate driver, but my outputs a little distorted. Could someone point out where i can improve my design. This is the first time i'm working with these circuits so please excuse any dumb mistakes i've made.

 

I don't know the different classes of amplifiers but if you tell me what you want I can help you with that.DC gain(open loop,close loop adjustment),AC gain(open loop,close loop adjustment,bandwidth)

 

I don't know the different classes of amplifiers but if you tell me what you want I can help you with that.DC gain(open loop,close loop adjustment),AC gain(open loop,close loop adjustment,bandwidth)

The output from the h bridge should ideally be smoother i guess. I have to pass this through a low pass filter to get the audio signal.

 

View attachment 365823

The output from the h bridge should ideally be smoother i guess. I have to pass this through a low pass filter to get the audio signal.

Spikes usually mean high frequency or transient response components so passing it through a low pass will absorb the spikes.

 

View attachment 365823

The output from the h bridge should ideally be smoother i guess. I have to pass this through a low pass filter to get the audio signal.

What if u added a low pass filter before the input of the MOSFET

 

hi dee,
Check out this link
You require a suitable load for the output drive
E
https://www.powerelectronicsnews.com/class-d-power-amplifier/

 

What if u added a low pass filter before the input of the MOSFET

You would slow down the switching speed, which would keep the MOSFET in its linear region for longer, which would cause it to dissipate more heat, and probably fail.

 

View attachment 365823

The output from the h bridge should ideally be smoother i guess. I have to pass this through a low pass filter to get the audio signal.

Why does it only get down to 2.6V? With no load it should get to zero. Is the upper MOSFET switching off properly?

 

i've implemented dead time and added a gate driver,

That looks good.

my outputs a little distorted.

The output looks better than most people's try.

M1 Gate is being driven incorrectly. Try using a "Half bridge gate driver". Like the LTC444. See below. This IC is built for this job.


If M1 was a P-MOSFET, and the signal was inverted, then it probably will work.

 

OF COURSE a class "D" amplifier will haave a distorted output. That is what the filters are for.

 

You really should show all of the relevant circuit - then we can help with the gate drive.

 

then we can help with the gate drive.

M1 Gate Drive is wrong.
Red trace is G to S on M1. Gate to Source voltage is 1V, 7.4V and 9.8V. (not Gate to ground)
Green trace is the output.

 

As said just earlier - if OP would post complete clear schematics ( some of us do not want to download .asc files ) then the matter would be quickly cleared up - as usual ( always ) it is the gate drive.

 

An .ASC file means nothig to me because I don't have that simulator to use itwith. Even a screen shot of the circuit woud be useful.

 

Even a screen shot of the circuit woud be useful.

I have the SPICE loaded. This part is supplies and on op-amp.

Here is the Gate Drivers and the two MOSFETs. I think if you click on the picture it might open up bigger.

 

AWild guess is that not all of the transistors switch at the same speed. THAT would be the cause of the over-shooting distortion. AND, any resistance in the power feed to the gate driver would slow the drive voltage rise and fall times.

There are definite reasons as to why very low distortion amplifiers are expensive: They AREmore complex, often quite a bit more complex.

 

I think 500kHz would be pushing it without antisaturation clamps on Q7/Q11

 

Schematic.

Waveforms

See attached SPICE file. There are some problems with it but that is all the time I have right now.
The first one of these I built was in about 1980. It worked but the switching frequency was too low for good audio. Back then we fought to switch above 20khz. Now we are switching at 1mhz.
RonS.

 

Schematic.
View attachment 365872
Waveforms
View attachment 365873
See attached SPICE file. There are some problems with it but that is all the time I have right now.
The first one of these I built was in about 1980. It worked but the switching frequency was too low for good audio. Back then we fought to switch above 20khz. Now we are switching at 1mhz.
RonS.

Is M3 upside down?

 

Is M3 upside down?

Yes, and there are other problems.