Hello, I am Chance Dunlap. I'm the business unit manager of the Zilker Labs Digital Power Products at Intersil. And in this video, I'm gonna show you how to setup the output current measurements with the ZL8800. The ability of this device is to have accurate current measurements of the output current at all times. Now typically, with most devices where you're measuring output current, you may have to put a lossy element in place.
The benefit of the ZL8800 is you get accurate measurements without inserting any components. So the device is able to measure the output current all the way from, as low as .6 volts output up to 5 volts using DCR current sensing. This just puts an RC around the inductor that's already in the circuit. So it's like you're using the parasitic resistance of this inductor as the sensing element. What we get out is a differential current sense line that's then fed back to our ADC, and we'll walk through exactly how to set up the ADC parameters in a minute.
The benefit of this circuit, as mentioned, is high efficiency. The one downside of using an inductor DCR is own inductors get specified with a tolerance, maybe a couple percent up to 5%, 10%, depending on the inductor you use. If you're looking for utmost accuracy, one solution is to use a resistor instead. So looking at this diagram here, you can see the differential sensing allows you to place a series resistance in the output just after the inductor. And that way, you can use a resistor that's very small, order of a couple of milliohms, that has a tolerance of less than 1%, just for a little bit more improved initial accuracy.
The one other consideration that you have to take into account when using other than the inductor DCR method or the resistor sensing element is both of those parameters will vary with temperature. And we look at the inductor, it's got a varied base of copper because it's just a copper mild core. Now, copper will vary about 3600 ppm per degree C. That gives you a considerable error across temperature. Now, a resistor is way much better, but it's not perfect. It's not flat. It's going to vary maybe about 200 ppm per degree C.
In order to compensate for this temperature variation of the inductor or the resistor, the ZL8800 allows you to use two external temp sensors, one for each channel. What this is is it just uses a Vbe junction of an NPN. So we can put a small little 3904 directly against the inductor and just connect the collector and the base together, and they even have a differential line back to the device. The 8800 will inject multiple currents through this device and measure the temperature.
So as the inductor varies, this Vbe junction will change and the device will compensate appropriately. This can be used either for the inductor or the resistive method. And the 8800 allows you to put in the exact temp code desired. So even if you see it offset, something different from here, you can always put that into the device and will compensate appropriately. So this gives you a gain setting across temperature.
Now that we have a DCR-based current sensing scheme set up, let's take a look at how the Isense signal goes back into the 8800 and how we set up the ADC and all the blanking parameters for the current measurement. So let's take a look through the signal chain of how that output current translates through the ADC and into a measurement from the device. Now, we know we're measuring effectively the current of the output through this DCR of the inductor. So what you're gonna see from a sense voltage across this capacitor is purely that DCR resistance times by the output current. So you can see this voltage, the DCR times the output current right at the input of the 8800.
Now, we don't go directly into an ADC. On the 8800, there's a PGA gain amplifier in the front end. This allows us to use a very good low DCR resistance. So if you're using an inductor with a DCR of maybe .2 milliohms, you'll wanna use the lowest gain settings. Now, the three different settings is low, medium, high, ±25 millivolts, ±35 millivolts, or at the highest gains settings, ±50 millivolts. So there's a wide variety. Now what this means is you've got to make sure that the voltage you see at the device is always less than this voltage. So you have to take into account not just the output current reflected across the resistance of the DCR but also the ripple current and allowing room for temperature variation and overcurrent spikes.
Now, you will notice we did say ±50 millivolts and ±25. The reason why it's a negative voltage is the device does support negative current. So if you have a situation where you're running reverse current through the inductor, you do get kilometric feedback. And in the section where we talk about full protection, you'll see that we have overcurrent and undercurrent volts. But now the next stage, as we look through the 8800 after the PGA gain, we do have a 10-bit ADC. So we get very high resolution of the output current measurement, and this is digitized continuously.
So if we look at the plot of the current waveform coming in the device, what you would expect is the inductor current would ramp up during the on-time and then ramp down during the off-time. Now, continuously through this period, the ADC is constantly sampling. Now because of the switching noise that occurs on both transitions and the short time period, especially for applications if you're running 12 volts in to 1.2 volts, that's only 10% duty cycle. I mean, this is a very small period of time. As a result, with the 8800, you can select whether you want up-slope sampling or down-slope sampling. The default is down-slope sampling, looking at this period of the waveform.
Now, the device will go through and sample continuously and will average out those periods to get a very accurate output current measurement. And we did mention noise before, you also have a selectable blanking time. So you can blank the first few measurements in the device and this is a selectable filter. That way, by avoiding these points and beginning in the end, you can still get the accurate measurement without any switch transitions affecting it.
We can take a look, with the boards on the bench using the GUI, to see how this works and reacts versus little changes. To demonstrate how the output current measurement works, I'm using the ZL8800 dual channel evaluation board. Now, with this evaluation board, I have a single ZL8800 IC and then two independent output stages. So this is just the DrMOS inductor, and the output capacitors. And there's two different terminal blocks for each output because the 8800 is capable of regulating two output voltages simultaneously. The hardware is completely independent from each other.
Based on the enable switches right now, I only have one of the channels enabled. So as we look in the GUI, we can see that channel zero is enabled with a 1.2-volt output voltage, channel one is not enabled, hence the power good flag is shown as red in the GUI. As you can see in the GUI, the output current measurement's reading zero amps, which is what you would expect with the load box being off right now. But as I turn on the electronic load, as I'm applying 5 amps, 10 amps, we'll see the GUI reflect the change in output current. So as you can see, we're applying a 5-amp load on load box and the GUI correspondingly is showing about a 5-amp measurement on the ZL8800. So you get a fairly good linearity, and let's keep on increasing the output current.
So right now, we just increased the load box. So we're pulling around 9 amps. However, you know it's showing in the load box that we're down to .8 volts, and that's really because these thick cables are causing an IR drop because I've got a lot of cable length to the board. However, when you look in the GUI, it's still showing it's accurately regulating the output voltage exactly at 1.2 volts. That's really the benefit of remote sense. You have long cable lanes and a lot of IR drop on the board, the remote sense can easily compensate for that.
Now, the remaining problem though is with output current measurement, you may get a very stable measurement at the load current, but you have to worry about the temperature effects. We showed earlier then you can enter in temperature to compensate for the inductor's variation. Now, looking at the GUI, you can sit there and see that there are two measurements, the internal measurement which is the temperature of the silicon and also the external temperature which is the temperature measurement taken right by the inductor. So this board uses a very small 3904 transistor right by the inductor so as we change the inductor temperature, it will compensate for the DCR variance and therefore, we're always gonna get a consistent output current response.
To demonstrate this, let's have a can of freeze spray and I'm just gonna spray the inductor to cool it down rapidly. Now, that transistor is close enough that will also be affected by the temperature change and correspondingly, it would change the DCR value that the GUI is measuring. So quick blast with the freeze spray and it's probably dropped by the GUI a good 20, 30 degrees, and you can see that the output current measurement is still exactly the same as it was.
So in conclusion, the 8800 has extremely accurate output current measurements. It is able to compensate across temperature, and with a temp at 80 C, is able to resolve even small output current changes. In the end, you get a product that's always able to give you accurate measurements of what your load is doing at any point in time. For more information on the telemetry and additional features of the ZL8800, please continue watching the video series.