Features for function generator

I missed following features for the function generator:

• Phase - moving the outputs relatively to each other
  
• Connection between "duty" cycle and triangle signal - being able to make everything between an inverse saw and a saw signal ( from /| to /\ to |\)
  
• Two connectors - It seems to be possible to add signals together by putting multiple signals onto parallel one line. But this kinda hurts in the eyes of an hardware electronics guy. Should be connected in series
  
• File input - would be great, when there would be a way to input a (e.g. raw)file, to make individual waveforms. Might need something like frequency.
  
• Higher quantisation / Achieving multiples of other frequencies - The quantisation seems to get visible, when the frequency is set to 120 or 300. It adds some non zero decimal places. It's in general not a problem. But - just for fun - I superimposed some func-gens in order to get my individual waveform. Due to the quantisation, the resulting waveform was wobbling. (Well, I know - it's not the main goal of simulide to be a SPICE..)
  

There are some nice features for the function generator.
I will have a look.

About quantisation:
Have you tried "Quality" property?
By deafult is set to 4, which is a compromise betwen quality and speed.
If you want the best possible waveform, set it to 5.
Some more info here: https://www.simulide.com/p/blog-page_1.html

Yes, I tried quality=5, too. But this did not change the wobble. To correct my first message: It's not really a wobble, it is a beat signal (I had to look for the translation).

At first I was confused, since I tried to add some sine waves for a simple sawtooth emulation, for a check (V1 = 4V @ 100Hz, V2 = 2V @ 200Hz, V3 = 1V @ 300Hz, V4 = 0,5V @ 400Hz, V5 = 0,25V @ 500Hz). It looked nice in the beginning. However, while running and adding some other stuff to the circuit the signal was slowly changing into a shaky line and then back to a kind of sawtooth again..

Then i don't understand exacly the problem.
Can you provide some more info? specially circuits or pictures.

I just made some images of it:

The first image shows the valid signal (a sawtooth-like signal) directly after starting the circuit:






The second image shows the signal of the same circuit after about 10 seconds (no values changed, only time was running):



The cause might be the frequency of the middle waveGen of 300,0300x Hz, which shows a little offset.

I will try your circuit and have a look.

One thing i forgot to mention:
Limit for "quantization" is always simulation step, that is why certain frequencies can't be achieved exactly. This also limit the maximum frequency of wave generator.
You can get more accurate frequencies using smaller simulations steps.

Anyway i think it is posible to get exact frequencies, but that could cause other inacuracies, for example square waves not exactly simetrical or triangle wave tips cutted.
But maybe in some cases exact frequency is preferable.

The issue is definetly because 300 Hz wave is not exact frequency.

For sine wave is quite easy to get the exact frequency without distortions, so i tried it and it works perfectly.

Hera after a few minutes running the resulting saw wave is still intact:




It is probably better switching to exact frequencies for all waveforms even whith it's downsides.

For example this is the effect in a triangle wave at 6000 Hz with simulation speed 1 MHz:



There is always a prize...
What do you think?

I'm working to solve this limitations, those problems whith oscope resolution and some other timing issues.

I'm testing a new simulation model with time resolution of 1 ps that will allow a few things:
- Exact frequencies in wavegen without loosing quality with frequency limit in the GHz order.
- Oscope resolution of 1 picosedond, solving the "zooming" and other limitations.
- Logic Components propagation delay could be configured (with resolution upto 1 ps).
- Some others that don't come to my mind now.
- All this increasing simulation performance (in same conditions).

This doesn't mean that with this change you will be able to simulate a wave at 1 GHz in real time, your cpu limits are still there, but it will be possible to simulate something in the KHz order with 1 ps precision in real time. Or higher freqs. at lower simulation speed.
And i hope that performance will improbe in most cases (in same conditions) and at the same time have 1 ps accuracy.

For example it will be possible to generate a wave at exactly 1.014281 KHz and do it with less cpu usage.

Not sure yet if I will include this in next RC, there is still lots to change.

Related to this:
https://simulide.forumotion.com/t792-new-signal-generator-component#5433