Glitch on the rising edge of the THS4021

Fausto Bartra

Prodigy 160 points

Other Parts Discussed in Thread: THS4021, OPA830

I have a very simple circuit.

Comparator, 9V single supply, Vinput = 0..3V square signal, Tr,Tf < 100ns.

see below

The output has an overshoot of 0.5V and a "drop" down to 1V on the rising edge. the load is only 1K.

if I remove the 1k, the "drop" disappears, but the overshoot stays.

Is this a phase reversal problem?

I have tried changing Rf to 1K , it did not help.

any suggestions?

1157.Fig2.pdf

over 14 years ago

0 Kris F over 14 years ago

TI__Genius 17905 points

Hello Fausto,

I will look into this in the lab and will get back to you tomorrow. I will test out your configurations to see if I can duplicate what you are seeing.

Ideally, it is not recommended to use op amps as comparators and operate them in an uncharacterized, non-linear fashion. Can you explain the desired function of the circuit?

I have a few initial debug questions and suggestions:

Where is the 2.5V coming from? Is it well decoupled to ground (0.1uF caps)?

Can you verify proper linear op amp operation? For example, remove Rb14 (leave it open to disconnect the 2.5V) and reduce Rf to 1kohm. By removing the 2.5V reference, we make sure that it does not contribute to this issue. The circuit should be a non-inverting amplifier with a gain of 2V/V and the output should be ~0 to 6.6V square wave. Do you see the output overshoot and drop to 1V?

Still with the 2.5V reference removed, increase Rf to 3kohm, so that the gain is 4V/V and the output high should hit the positive rail output. Do you see the output overshoot and drop to 1V as you mention above with all components on board?

I look forward to hearing from you.

Regards,

Kris

0 Kris F over 14 years ago in reply to Kris F

TI__Genius 17905 points

Fausto,

I'm in the lab with an EVM about to do some testing. My apologies, the THS4021 requires a minimum gain of 10V/V (or -9V/V) to be stable. Are you able to add attenuation in front of the THS4021? That is, instead of my recommended tests above, put the THS4021 in a gain of 10V/V by using Rf = 1kohm, Rg = 110, and apply a 0-0.5V square wave at the input. The output should be a 0 to 5V square wave, however, since you are running off of 9V single supply and the output of the THS4021 is limited to about 1.6V above the Vcc-, I would expect your output to saturate the amp on the low end and be a ~1.6V to 5V square wave.

Please e-mail me directly at k.flores@ti.com

Best regards,

Kris

0 Kris F over 14 years ago in reply to Kris F

TI__Genius 17905 points

Normal 0 false false false MicrosoftInternetExplorer4

This post apparently did not get through earlier:

Fausto,

I did see the same issue on my board. This is due to violating the common mode input range at the non-inverting pin of the THS4021. The datasheet lists the input common mode range for Vcc = +/-5V as +/-3.8V, or basically 1.2V above the negative rail and 1.2V below the positive rail. This is the range over which we test and specify the op amp performance, as this is the intended operation range of the op amp. With the 9V positive supply, the input range is 1.2V to 7.8V and is violated by the 0 to 3.3V square wave violates. I changed the negative supply to -2V supply and the output dip issue disappeared.

The figure you attached indicates a 100kHz 50% duty cycle square wave – is that the input signal in the real system? I suggest the OPA830 as a pin-compatible replacement. Its common mode input range includes the negative rail, which makes it more suitable for single-supply applications.

Again, please e-mail me directly at k.flores@ti.com. We can get into more detail about your design over e-mail and what modifications can be made if you are tied to using the THS4021.

Thanks,

Kris

Amplifiers

Amplifiers forum

Glitch on the rising edge of the THS4021