Next time you'll mention rise and fall times?frequency is about 200 KHz.
Hi againWith using your idea i can't adjust rise and fall in real-time manner .
I must change rise and fall from software in real-time.
If you can accept exponential first order response as adjustable rise and fall time, yes. The electronical variable resistance point is still not so easy.It's to simple. you can use an RC circuit with an buffered opamp at the output of your Fast TTL Circuit. If you want to control from software you can use an adjustable resistance or may be pump circuit.
Depends on the intended rise time. You need at least 5 samples per ramp time and a good filter (e.g. 3rd order gaussian) to get an acceptable ramp waveform.Another approch is an simple DAC with an FPGA triangle wave generator. It is very easy to control but a little difficult to implement in HDL language. If you want this approach i can help you to implement on a small cheap CPLD!!!
Next time you'll mention rise and fall times?
Classical analog pulse generators with variable rise and fall times are using switched programmable current sources, an integrating capacitor and a clamp circuit to define the high and low level.
---------- Post added at 17:58 ---------- Previous post was at 17:51 ----------
A principle schematic can look like below:
I plan at lest 64 (6bit) sample with cheap resistance DAC network and a small buffered opamp. It doesn't need the filter or just a cap. You can simulate the output using HDL. I have made a resistance DAC (for only DAC not like your case) and it is working fine. We can try it. But base on my experience it will works fine.
Another approach is a pump circuit and a programmable current source. It again needs the buffered opamp. but the circuit is difficult.
Depends on the intended rise time. You need at least 5 samples per ramp time and a good filter (e.g. 3rd order gaussian) to get an acceptable ramp waveform.
A reconstruction filter is essentially a low pass. Because you want a time domain response without overshoot, you need to select a respective low Q filter, Bessel or Gaussian.but how would this work. So you would send 5 or so digital sample points to the DAC and then use a digital or analog gaussian filter to "fill in" the rest of the area between the sample points, like an interpolation/reconstruction filter or some some kind of up-sampling (is this correct)? Also, why a Gaussian filter?
A reconstruction filter is essentially a low pass. Because you want a time domain response without overshoot, you need to select a respective low Q filter, Bessel or Gaussian.
Ultimately, the generator will produce a monotonic output signal even for a single input step.
Not necessarily, it's a matter of low pass characteristics. Most today's universal function generators are using DAC with reconstruction filter.a low pass would be a poor choice because it could produce overshoot and ringing between samples.. correct?
I think a low pass 10 nf cap is enough.
We use cookies and similar technologies for the following purposes:
Do you accept cookies and these technologies?
We use cookies and similar technologies for the following purposes:
Do you accept cookies and these technologies?