Slow Ramp Generator Design

[jeetz_mail]

I want to design a slow ramp generator with the following specifications:

Type of ramp – starting voltage 0V, ending voltage 1.22V, ramp time ~17.7 seconds.

I also want the ending voltage to be variable such as 1.744V, 2.44V, 3.0V etc.

Please give me suggestions on how to design this circuit.

Thanks
Jeet

 

[SherpaDoug]

While you could use an analog integrator, I would probably use a microcontroller and a DAC. Tell us what this will be used for. How accurate does it have to be? What power source do you have?

 

[jeetz_mail]

While you could use an analog integrator, I would probably use a microcontroller and a DAC. Tell us what this will be used for. How accurate does it have to be? What power source do you have?

Hi SherpaDoug,

The ramp generator circuit output will be connected to the input of an ADC (inside an ASIC chip). I intend to characterize the performace of the ADC. A slow ramp voltage signal at the ADC input will be started. ADC measurement will be initiated for a duration equal to the ramp time at its input.

I looked at using "The Howland Integrator" - Figure 7 at the following link:
**broken link removed**

But I wasn't sure if it would work as needed, i.e. to generate a slow ramp (~17.7sec) with variable ending voltages (1.22V, 1.744V, 2.44V, 3.0V)

Thanks,
Jeet

 

[SherpaDoug]

The Howland integrator is good if you must ramp both ways (up and down) but it will be hard to keep good linearity. For such a slow ramp you need either a very small current or a very large capacitor. Doing either precisely is difficult or costly. You will have to learn about capacitor leakage, soakage (dielectric absorption), nonlinearity, etc.

If you go with the uC and DAC the slow speed is not a problem. But you need a better DAC than the ADC you want to measure. You should also watch for systemic problems, where a fault in the DAC matches a fault in the ADC so you either don't see it or it gets distorted so you don't recongnize it.

 

[dick_freebird]

If you insist on a ramp then you have to create your own
most likely, but have you considered simply using a precision
DC source stepwise?

Building the ramp is one thing, but knowing with certainty
the voltage at the time of sample for each code is going to
be your test challenge. There, you might be better off to
"buy some accuracy" you can count on in-the-moment.