High-End Regulated Buffered Inverted Chipamp

[Franz G]

Boaaahhh

Absolutely exiting, the sound out of my breadboard:



Based on the circuit from Carlos, but actually with NE5534 as buffer.

As thank, I offer this thread to Carlos, because his old thread is still closed.

Soon, I will compare my three different chipamps with my diy 300B amp and then write about.

But first, there is a small problem to solve, with this new experimental amp: with a 1kHz sine I remark some distortions between 8-10 watt output.

The amp begins clipping at 40 Watt, about 18V RMS 8 Ohm, what I think is O.K. for this design.

Franz

[carlmart]

Originally posted by Franz G

Based on the circuit from Carlos, but actually with NE5534 as buffer.

Soon, I will compare my three different chipamps with my diy 300B amp and then write about.

But first, there is a small problem to solve, with this new experimental amp: with a 1kHz sine I remark some distortions between 8-10 watt output.

The amp begins clipping at 40 Watt, about 18V RMS 8 Ohm, what I think is O.K. for this design.

The NE5534 may not be the best to use as a buffer and it does have a large DC offset, which demands big electrolytic capacitors at output to work well. Other fast single chips might do better. Hopefully you used sockets for the buffers.

Have you fed a signal directly into the amp, bypassing the buffer, to see what might be causing the distortion?


Carlos Martinez

[Nuuk]

The NE5534 may not be the best to use as a buffer and it does have a large DC offset, which demands big electrolytic capacitors at output to work well.

Carlos, in defence of the NE5534, I should point out that this is only partly true!

You don't have to use 'big electrolytic capacitors' on the the output, a 4.7 uF polypropylene works just fine!

You can even get the NE5534 to have zero DC offset although at a cost of a lower input impedance. Or you can use the 22 pF caps (for stability) and configure it in unity gain with zero DC offset and remove the cap altogether.

Recent listening tests with several opamps including the OPA627, have left me in no doubt that the NE5534 is more to my (musical) taste!

[Franz G]

I think it was you who recommended me to try as first step NE5534 from Philips, Nuuk.

They work in unity gain just as buffer, regulated with 7818/7918 and with 22uF psu-bypass, bridged with 0.1 uF.

I used sockets, fortunately: yesterday, the first time powering up the amp, the green overdimensioned resistors started to smoke (power supply for the buffers). Quickly I realized, that I just inserted the two NE5534 the wrong way

No damage, everything works fine now.

Today afternoon I will have a look at the mentioned distortion, by feeding the testsignal bypassing the buffers.

I will check the output (zobel, 0.4R resistor or LR) and I will change the bypassing (wima's under the board).

Maybe it is just this testboard: the outputcables direct over the heathsink could cause capacitive load?

We'll see.

Franz

[Nuuk]

Quickly I realized, that I just inserted the two NE5534 the wrong way

Steady Franz. I have seen a thread on which way to orientate resistors; but opamps!

[Franz G]

Now, I need some consulting:

The measured distortion results from the power stage, not the buffer.

First, I could see the problem only between 8-10 V RMS output (8 Ohm).

After removing all small foil bypass caps under the board (Wima MKP, 0,1 between +/- and GND, 0,15 betwen the PSU-connections), the distortion starts at 10V and changes with more power into clipping (still 18 V).

It looks like this at 1kHz:



This is the right channel, on the left channel I have the same distortion at the negative wave.

With some correction on the output side (zobel, rc network, lr network) I could not correct the problem.

What could be the reason?

- Board layout (DC paths, RF problems)
- Some unsymmetry with the PSU-supply
- other?

BTW: the amp is absolutely great (below 10V RMS, of course)!

Yesterday evening I had some session, comparing it to the 300B single end tube amp (with my Coral Beta 8 in FE206E horn enclosures). The first result: Different amps, different result, but in fact, comparable in a way!

My chipamp needs some aging, some time to get warm, to be under tension and to play music, before I tell more about the sound, compared to completely different designs, like tube amps.

Any help is welcome.

Franz

[analog_sa]

What could be the reason?

I would think oscillation. Do you get the same wave form if you bypass the buffers?

[Franz G]

Yes, this foto from the old Heathkit scope is bypassing the buffers, just the powerstage.

Franz

[sek]

Hi Franz,

This is the right channel, on the left channel I have the same distortion at the negative wave.

I find it odd that one channel shows this behaviour at the upper half-waves while the other does it with the lower ones!

Could it have to do with your power supply (wiring)?

Other than that, it reminds me of the evil SPiKe(tm)! Check the LM's datasheet to compare with the limiter's characteristic influence on the signal shape.

Hope this helps,
Sebastian.

[Franz G]

I fixed it, with the help from the excellent german forum www.jogis-roehrenbude.de (see link below).

A rude oversight from me: the NE5534 as buffer (unity gain), needs some 47pF between pin 5 and 8 as compensation.

I did not really bypass the buffer, when I fed in the signal at the LM3875: the feedback from the buffer was still active and disturbing, because the NE5534 was the oscillator...

Franz

Thread (german) at Jogis Roehrenbude