Hi,
I am using TL3016 comparator in one of my applications for pulse shaping.
The comparator -ve i/p is maintained at a threshold and i/p pulse is fed to non inverting i/p. i have added a small amount of feedback to avoid ringing at threshold.
The problem is when the i/p is well above the thresh hold the i/p to o/p delay is about 15ns. If the i/p is just above the threshold (100 to 200mv above the threshold), the i/p to o/p delay is increasing to 85ns. this is causing timing issues in our design. is there any work around for this ?
thanks.
kpkar
Hello Prabhakar,
It is a normal comparator response for the propagation delay time to increase as the input overdrive voltage is decreased. The amount of increase you are seeing for the TL3016 is more than I would expect. Figure 6 in the data-sheet shows the typical propagation delay time vs. overdrive voltage for the TL3016. As you can see the time increases only a matter of nanoseconds as the overdrive is reduced a matter of tens-of-millivolts. Thus, the increase in time is inordinately large for your device when compared to the typical TL3016 behavior. This suggests that something else must be contributing to the increased time.
The TL3016 is an ultra-fast comparator and the circuit design, layout, drive and load conditions will affect the speed. Does the PC board layout follow good high-speed layout practices? Is the supply bypassing scheme appropriate for the speed of the device? Are the input drive and output terminations correct? Is the input drive signal clean and precise? Lastly, was the TL3016 obtained from a TI distributor?
Regards, Thomas
PA - Linear Applications Engineering
Dear Thomas,
Thanks for the reply.
I verified the points put forward by you and i have the following to share:-
1) I am using a multilayer board with separate ground and VCC planes. Decoupling of power supply near the chip is done with a 0.1mfd cap.
2) My i/p signal is a rectangular pulse of 1micro sec. width, amplitude varying from 1v to 5v(settable from external source), repeating every 61 micro sec.
3) the o/p drives a load of 47k with a 5 pf cap. in parallel
4 The pulse is reasonably clean generated by a Tektroniks function Generator
5) the comaparator ic is sourced by an authorized distributor of TI
when i studied the datasheet again, i found the common mode volatge spec for TL3016I as 1.25V to 3.5V for unipolar supply (5V). i am using industrial grade part with single 5v supply. The threshold set is about 1.1V. what is the effect of operating the comparator below / above its common mode voltage limit ?
prabhakar
It sounds from you description that your TL3016 PC board set-up is okay. The TL3016 wasn't characterized with the common-mode voltage set outside the specified range; therefore, I don't know how the device behaves in that region. The CMV range is usually established by the differential input pair, its current source(s) on one end, and a dynamic load on the other end. Basically, they share the supply voltage applied across the supply rails and the voltage applied to each must remain within their normal operating regions.
Usually, a device's performance degrades just outside the specification region and then may function poorly or cease to operate as it is driven further outside the range. It may be that one or more stages within the TL3016 are being driven into saturation when the comparator CMV range is violated. Operation in saturation increases transistor storage times and slows the device down. If it is possible for you to test the TL3016 with input levels within the specified CMV range you may find that its propagation delay times are within specification.