Signal chain overload effects

First set of results for the Mixer/Amplifier combination showing a 'normal' input level and an overloaded input:

Firstly a two tone signal applied to the Mixer Input at a level so as not to overload the Input and monitored at the Amplifier output:


A spectrum analysis of this signal below:



Now with the input gain of the Cloud increased to maximum so as to overload the input stage; the Output waveform of the Amplifier is:


and the Spectrum of this composite 2 tone signal generates this:


It can be seen that there are a substantial number of artifacts compared with the non-overloaded input signal both as a result of the clipping and intermodulation products.
The test frequencies used are 960 and 1120Hz and the measured Amplifier output was of the order of 5V rms.

As to how much additional energy is passed to a 'speaker by the 'overload signal' compared with the base-line one, the Spectrum Analyser results show a substantial difference.

Mik

What is the normal and overloaded signal input level Mik?

Thanks for starting this thread- and the photos. I ran a few (quick) tests the other day with a Berry DCX2496 and must say that the output waveforms were a lot better than from the xover I used for my initial tests. I never had time to take any photos, but will. I´ve also got a couple of BSS FDS-360´s so I´ll test those and post some photos as well.




Edited by Earplug - 04 February 2012 at 1:48pm

The Berry DCX2496 uses the Analog Devices ADSP-21065L DSP IC for it's processing, the data sheet for this plus the source code for it's firmware is available here.

If you look into the source code instruction set for the firmware then your question relating to 'read ahead' will no doubt be answered...
Mik

Earplug
interesting results, will look at repeating your test method here.
What are your DSP Filter settings?

Think that Latency tests* are needed so have just switched on my bench to get the instruments 'warmed up' - takes a couple of hours for all to stabilise properly

Mik


*Think I can see some phase differences at the leading edges of your 'scope images which may be a result of either Latency in the DSP or Filter characteristics.

OK, initial tests using same setup as before with a 1kHz HP LR Filter setting for the LEM:

Squarewave at 1kHz input with the Amplifier output in the centre of the screen:

And the output spectrum of the filter:


Measurement Error: Next an expanded timebase of 0.1mS/div to look closely at the time difference of the zero crossing point of the output compared with the input signal to assess latency.


OK, - a fundamental error in measuring the Latency by using the 1kHz square wave input which obscured a 1mS difference from the real value of this timing difference as it appeared on the next cycle of the waveform.

Corrected in the 'scope image below, using a 100Hz square wave and 1mS/div timebase, which shows a more realistic figure of 1.2mS for the delay between input and Hi-Pass output of the test setup:


Mik

Edit 1: add timebase value
Edit 2: add corrected latency measurement due to timing error

Edited by audiomik - 06 February 2012 at 2:42am