LME49830 with 2SK1058/2SJ162 output stage

[sam xt]

Hello,

I recently started to build an amp based on the LME49830 with a MOSFET output stage using 2SK1058 and 2SJ162. Everything seems to be working except I get high frequency oscillations (around 400kHz with a magnitude of maybe 200mV) that appear on the the positive output peaks of a sin wave test signal. The amp is running off of a +-45V supply. Of note is that when I disconnect the mains and let the amp run just off of the supply caps the oscillations go away.

I have tried several things to try and make these oscillations go away. The schematic I am using is pretty much identical to the National Semi ref design. I will try and post some photos of the amp board, power supply etc. but in the mean time I was wondering if anyone else has encountered a similar problem?

Any help would be greatly appreciated.

Thanks,
Sam

[abraxalito]

As your oscillations disappear when the mains is switched off, you might be seeing noise from your rectifiers. Try fitting series RC snubbers across each diode. The optimum values can be determined by experiment but start off with 100R and 10nF - see what happens.

[AndrewT]

Originally Posted by abraxalito

As your oscillations disappear when the mains is switched off, you might be seeing noise from your rectifiers.

monitor the rectifier/smoothing cap voltages for ripple and oscillation, when you know the amp is reproducing properly and when it is misbehaving.

[sam xt]

Thanks,
I will try adding RC snubbers to the rectifiers. Currently I have 0.1uF caps across them but I will try RC as well. I have monitored the rails and do observe oscillations on both the + and - rails, however oscillations only show up on the positive peak of the output. Recently I did add a RC snubber to the output of the power supply and did notice the oscillations reduced.

I will post my results later today after I make the changes.

[AndrewT]

Originally Posted by sam xt

I have monitored the rails and do observe oscillations on both the + and - rails, however oscillations only show up on the positive peak of the output. Recently I did add a RC snubber to the output of the power supply and did notice the oscillations reduced.

does the rail behaviour change with the output misbehaviour?

[sam xt]

Andrew,
I went back to look how the rails behave. They show the same pattern of oscillation, that is, they only oscillate when the amp output is at its most positive point.

When the output from the amp is increased the oscillations on both the amp output and supply rails are reduced both in duration and magnitude. This has me very puzzled. It appears that as more load is placed on the power supply it tends to have less oscillations.


Later today or tomorrow I will try to add RC snubbers to the rectifiers and post back.

[AndrewT]

I think what you are confirming is that the power amp kickstarts a potentially unstable PSU system when the PSU does not like a fast transient demand for current.

It's that potentially unstable PSU system that needs sorting.
However, this may be made worse by applying amp decoupling in the wrong places/values.

[Mooly]

It could be the amp that's unstable reflecting back into the PSU.

Is your layout good, with correct grounding and feedback returns used ? Short low impedance paths to the FET's ? Any zobel network correctly returned ? Gate stopper resistors mounted as close to FET's as possible ?

Might be worth adding a small (0.1uf) from each FET supply (drain or source depending on configuration) to PSU ground.

[Sebastiaan]

Originally Posted by sam xt

Hello,

I recently started to build an amp based on the LME49830 with a MOSFET output stage using 2SK1058 and 2SJ162. Everything seems to be working except I get high frequency oscillations (around 400kHz with a magnitude of maybe 200mV) that appear on the the positive output peaks of a sin wave test signal. The amp is running off of a +-45V supply. Of note is that when I disconnect the mains and let the amp run just off of the supply caps the oscillations go away.

I have tried several things to try and make these oscillations go away. The schematic I am using is pretty much identical to the National Semi ref design. I will try and post some photos of the amp board, power supply etc. but in the mean time I was wondering if anyone else has encountered a similar problem?

Any help would be greatly appreciated.

Thanks,
Sam

Just a thought,

Can't it the amplifier get unstable above a certain rail voltage? If you disconnect the mains, the voltage get lower and slide down.

The best way to find this out is, use a variac. And start at a low voltage operating point while you montor the sine wave in a load. Raise the voltage and see if the sine wave change. Big chance above a certain threshold you see the oscillation kicks in.

I had it too in one of my designs. Everything worked great till +/- 35 VDC. But going higher I got oscillations. A change of miller capacitor value and some capacitance from Collector to base on the driver transistors solved my problem.

With kind regards,
Bas

[jackinnj]

Originally Posted by sam xt

I have tried several things to try and make these oscillations go away. The schematic I am using is pretty much identical to the National Semi ref design. I will try and post some photos of the amp board, power supply etc. but in the mean time I was wondering if anyone else has encountered a similar problem?

Try using a somewhat larger Ccomp -- I don't think that Troy never got around to writing the application note describing the care and feeding of this cap. The demo board design uses IRFP240/9240 which is a lot different from the lateral 2SK1058/J162.

Take a look at the other piece he wrote: Application Note AN-1645 "LM4702 driving a MOSFET output stage"

As mentioned, the gate stopper resistor should be mounted as close as possible to the gate of the output devices. Sometimes the N and P channels will benefit from different values of this resistor.