Comparing LME49810, 49830 and 49811

[panson_hk]

Originally Posted by PauloPT

Is 0.22R the best value or should I use higher value than that?

Lower is better as long as the circuit is stable. I have used 0.22R with two pairs of IRFP240/9240 without any problem.

[PauloPT]

OK, 0.22R it is!

Thank you again Panson!

[Telstar]

Originally Posted by panson_hk

I implemented two-pole compensation to a LME49811 with ThermalTrak amp. The high-frequency distortion is reduced. Here is an example: 123W into 8 Ohm load. The result is very promising.

Yup, it is

[blueskynis]

Originally Posted by panson_hk

I implemented two-pole compensation to a LME49811 with ThermalTrak amp. The high-frequency distortion is reduced. Here is an example: 123W into 8 Ohm load. The result is very promising.

Cc1 = 22pF (single-pole cap)
Cc2 = 44 pF
Rc1 = 5.6 k

Yea, that's good result.

Now, if you have time, please try Cc2 100pF, unsolder the Rc1 resistor and solder a 1k8 resistor between capacitors and the output with a wire, just like I have shown in my previous posts. The THD results should be similar but the impulse response should be the same as with the standard miller compensation. With two-pole compensation you have overshoot.

[Sebastiaan]

Originally Posted by panson_hk

I implemented two-pole compensation to a LME49811 with ThermalTrak amp. The high-frequency distortion is reduced. Here is an example: 123W into 8 Ohm load. The result is very promising.

Cc1 = 22pF (single-pole cap)
Cc2 = 44 pF
Rc1 = 5.6 k

Dear Panson,

Such a distortion rise in the higher frequencies are sometimes suspective. Are you really sure that somewhere high in the MHz. range there isn't any oscillation?

I end up with this scheme, after advice from the National designer on this forum, and it killed the Mhz. range oscillation.

With kind regards,
Bas

[blueskynis]

Originally Posted by Sebastiaan

Dear Panson,

Such a distortion rise in the higher frequencies are sometimes suspective. Are you really sure that somewhere high in the MHz. range there isn't any oscillation?

[snip]

The THD diagram looks quite normal to me

And that is not a two pole compensation.

[Sebastiaan]

Originally Posted by blueskynis

The THD diagram looks quite normal to me

And that is not a two pole compensation.

It isn't a two pole. I wanted to point this out, because I experienced myself oscillation's far away, which wasn't visible or noticeable on first sight. This compensation scheme solved it.

Btw. My example shows one error, the cap should be connected to the sink pin instead of the source

With kind regards,
Bas

[klewis]

Originally Posted by Sebastiaan

Dear Panson,

Such a distortion rise in the higher frequencies are sometimes suspective. Are you really sure that somewhere high in the MHz. range there isn't any oscillation?

I end up with this scheme, after advice from the National designer on this forum, and it killed the Mhz. range oscillation.

With kind regards,
Bas

I'm doing the exac same thing as Bas. It's from AudioMan54 (Mark) from the National Op Amp Inflation thread. I haven't had any problem with oscillation.

As I recall this compensation scheme was created to help the amp come out of clipping smoothly.

I have had issues with overshoot in the square wave with the zobel connected. Disconnected it goes away.

Ken

[Sebastiaan]

Dear,

Here square wave measurements. In this case the LME49810 with a pair of STD03's. Compensation only 10pF miller cap without the extra added filter. No sign of oscillation or overshoot, still it was oscillating at 10Mhz. With normal squarewave test or THD measurements I would't have seen it.

Resp: 1Khz. and 10Khz. Square wave.

With kind regards,
Bas

[panson_hk]

Originally Posted by Sebastiaan

Dear Panson,

Such a distortion rise in the higher frequencies are sometimes suspective. Are you really sure that somewhere high in the MHz. range there isn't any oscillation?

I end up with this scheme, after advice from the National designer on this forum, and it killed the Mhz. range oscillation.

With kind regards,
Bas

Hi Bas,

No sign of oscillation when tested with 8Ohm and 4Ohm resistive load. The output was monitored by a 500 MHz scope. I will make more test including square wave after Chinese New Year.

Panson