In this video, I show you how sub-sequencing for a larger system such that you can get multiple rails to power up and power down with a timing pattern of your choice. So what we have set up here is just an 8800 with two outputs. One for a core at 1V, and maybe an I/O rail set up at 1.8V. If we click on the sequencing tab, what you see is it comes up with a default timing parameters for start up. So let's walk through the sequencing panel and look at what every option does.
The first thing you can see is the graphical representation. If you highlight, or just hover your mouse over the two rails, you see these little drag boxes appear, at the beginning and at the end of the ramp. At the end of the ramp, if you just click on these and drag and drop, you can affect the soft start ramp time. The ones at the beginning set up with Ton delay. The Ton delay, what that means is, is the amount of delay that you get from the enable signal to when the device starts ramping. So, if you have multiple devices and you have one enable going through the entire system, you can then set up a timed based sequencing standpoint.
So let's walk through the sequencing box and let's look at all the other options above the graphical readout. So at the very top, what you see here on the left is a little eyeball. If you click on this, it's a visibility control. You're just turning on or off which rails you want to look at. The benefit is, if you have 30 rails sequencing, it can look fairly complex on the screen, so sometimes it's easier just to hide certain rails. The next one over is just the lock function. What this does is, click on it, it locks out. So now when I hover over it, I no longer get the drag and drop toolbox. This is useful if you have multiple rails that sequence exactly at the same point in time or the same soft start ramp. It stops you from clicking on the wrong drag box by mistake.
The next box is sequencing box. Now, automatically you are sequencing everything from a time based standpoint. But one of the benefits of the digital power devices is a DDC Bus that links every single one of the products. What this allows you to do is now have an event-based sequencing. The benefit of this is, if you have one rail that powers up first, maybe the core, and you want the I/O to come up later, if the core never makes it, maybe it faults from an overcurrent event, you don't want to have any later rails sequence up. The devices will automatically communicate across DDC to ensure that one faults, the next device will not start up.
So, let's show you how this works. If I set back all the delays to exactly the same, no delay time, 2ms rise. Clicking on the sequencing view, you can now see that the first rail comes up which is the I/O rail. As soon as the power grid threshold is reached, the next device will start sequencing. So now, it's an event-based. It waits for the power grid threshold to occur before the next rail starts. If we want to change the order of these two devices, you can use these up/down arrows. So maybe we want the core to sequence first. And now you can see, once you reach the power grid threshold for the core, the I/O rail will start to sequence up.
Now you can add additional delays into the system. For instance, this Ton delay for the I/O is no longer based off the initial enable signal. It's delayed from when it receives the start based on the power grid threshold of the core. So if I add in 2ms, you can see that the power grid thresholds reach about 1.8ms. Two milliseconds after that point, the core will start ramping up. If you have more rails in the system that you want to be event-based sequence, you just click the check box and keep adding them into the group. It's actually possible to setup different group patterns. So maybe you have five rail sequenced one way event-based and another five sequenced in a different manner.
This works for not only start up, but also for shut down. If you want to control the soft stop, if you look on the top right, you see this little radio button. Clicking on it, you can see the exact same behavior works but in reverse. One rail will shut down, and then the second rail. And, again, you can set up individual timing parameters for the delays and the fall times. So you can see, as I set back both delays to zero, the first rail starts dropping. Once it reaches its power grid threshold, the falling threshold, then it sends a signal across the DDC Bus to the core, and then that rail will begin its shutdown process.
The next box on the screen that you see is a track pin. By clicking on this, we now set up the rail to track an external voltage. There's a drop-down box right next to it that selects exactly how this operation will work. You can either track 100%, which means the output voltage for the rail will track whatever is being applied on that track pin, up to the desired output voltage that you have set. The other option is a 50% tracking mechanism. What this means is, whatever is applied to the track pin, the device will follow by 50%. This is useful for DDR termination applications where you want the output voltage to track exactly one-half of that input voltage.
For more information on the PowerNavigator GUI and see for yourself the benefits it provides, please download your own version from the PowerNavigator product page.