Non-Inverted Gainclone Clipping

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I have recently built a gainclone using the pcb from Peter Daniel, but when I apply a sine wave of amplitude 1 and frequency 1000Hz the amp shows a clipping of the lower portion of the output waveform. I have used the schematic that Peter Daniel supplied, but am unsure of why this is occurring.

When I power up the amp with a signal source from a Microsoft Zune, the amp sounds great (very low distortion that I can discern), but I am trying to do some performance calculations, and this is throwing me for a loop. Any help would be greatly appreciated. Thanks.
 
The clipping point is determined by:

1. the gain of the amplifier (determined by ratio of values of resistors in feedback network), and

2. Maximum voltage swing (determined by power supply voltage and amplifier's internal design).

For example, if feedback network resistors are Ra=22k and Rb=680 R, the gain is = Ra/Rb + 1=33.35 which means that for 1V input you should have 33.35 V on output, but it won't happen because the voltage swing is limited by power supply voltage (usually +-24V), so the amplifier will go into clipping.

All of this can be easily understood by carefully reading the datasheet of the chip (LM3875, LM3886...).
 
Hi,
clipping with 0.7V output or with 0.7V input.
Is this voltage AC waveform (RMS) or peak?

Peter's version runs from 1500uF on the chip amp and no smoothing at the rectifier.

This will have serious ripple on the supply lines. Can you measure the troughs in the ripple to ground voltage fed to the chipamp supply pins when you carry out your test?
Or measure the average voltage at the pins and measure the ripple voltage. Then, Vavg-[half ripple peak to peak]~=voltage ground to trough of ripple.

The output starts clipping when the troughs start to approach the peak output voltage.

Are the ripple and average voltages different on the +ve & -ve sides?
 
It is clipping with a sinwave of an amplitude of 1V. Your answer makes sense because I was testing the amplifier using a +/-25V lab power supply. So that would mean that a 1V input would need 33V of swing, which is not possible with a power supply of +/-25V.

Any suggestiong on what type of smoothing caps to use for the power supply? I have a 25-0-25 secondary transformer, and would like to use this for the power supply. Can the rails be brought up to +/-35V? Any help would be appreciated. Thanks a bunch.
 
Hi,
what is the regulation of the 25Vac transformer?

I'll assume 7% for this example.

Open circuit output voltage is [1+regulation]* Vac=1.07*25=26.75Vac
after passing through the rectifier the multiplier square root (2) must be applied and subtract 0.7V for the diode voltage drop giving [26.75*1.414]-0.7=37.8-0.7=37.1Vdc. Way above your target.

The smoothing caps must be rated for maximum mains input voltage to the transformer primary.
If your tolerance on mains is 6% then that 26.75*1.06 becomes the input for the rectifier and the maximum voltage across the smoothing caps is 39.4Vdc. This is a bit close to a 40V rating so you may feel safer using 50Vcaps.

Once you connect a load to the smoothing caps the voltage falls slightly at a low bias setting and more when the bias is higher.
Expect the chipamp to see a nominal 1V less than the open circuit voltage or about +-36Vdc from your 25V transformer.

Under load expect the PSU voltage to collapse. Maybe as little as 3V lost at maximum output power if your PSU is well specified or as much as 10V lost if instead you go for puny.

Now find out your mains tolerance and the transformer regulation and do the numbers for your installation.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.