Comparator and oscillationS

[zanzeoo]

Hi every !

I designed this circuit below , the purpose is to "convert" a signal from +/- 10Vac , 400Hz  , into a 0 - 5 Vdc .

I used an AD790 powered by a +/- 15 Vdc supply ,  the particularity of this comparator is the output which can be TTL ( see the datasheet) so Vlogic is 5 VDC .

I add a +/- 100 mV hysteresys with external resistors  , but it still oscillate a bit .

So wonder if i add a feedback capacitor , will it be better ?

How do you determine the value of the feed back capacitor ?

Thx

Sorry for my english

 

[zanzeoo]

up

[danadak]

Looks like you only have a couple mV of hysterisis


http://sim.okawa-denshi.jp/en/compkeisan.htm


Regards, Dana.

[Andy Watson]

Have you applied all the recommendations given in the datasheet, specifically with regard to layout and "Bypassing and Grounding". What/how have you connected the latch input?
How do you know it is oscillating? - If you have applied an oscilloscope probe directly to the output of the AD970 the additional capacitance could be making the AD970 unstable - isolate the probe with 100-200\$\Omega\$ at the probe tip.

[mash107]

Hi every !

I designed this circuit below , the purpose is to "convert" a signal from +/- 10Vac , 400Hz  , into a 0 - 5 Vdc .

I used an AD790 powered by a +/- 15 Vdc supply ,  the particularity of this comparator is the output which can be TTL ( see the datasheet) so Vlogic is 5 VDC .

I add a +/- 100 mV hysteresys with external resistors  , but it still oscillate a bit .

So wonder if i add a feedback capacitor , will it be better ?

How do you determine the value of the feed back capacitor ?

Thx

Sorry for my english

As a rule of thumb, you don't want to limit the bandwidth of the positive feedback. It's not depicted in your schematic.. but say you had a noisy power supply that you're dividing down, and you had to add a large filter capacitor (in parallel with R4). Then you would need a cap like C1 to increase the bandwidth of the hysteresis. Otherwise it would be a large R (R1) charging a large C, and that might be too slow for the AC signal you want to convert. Let me know if that helps out.

[PChi]

My experience of the AD790 is that it likes to oscillate at the transition even when care is taken with the layout and decoupling.
Are the decoupling capacitors right by the AD790 and are the connections to the inputs as short as possible to minimise the capacitive feedback?
Even with the hysteresis there might be a very short oscillation burst at the transistion.