Hi,
I am looking for IC for GFCI design for EV charger. I am looking for similar functionality as LM1851.
Hey Himit,
1.I admit it is not exactly straight forward because it will be dependent upon the resistor's survivable peak current over one line cycle (16.7ms) and this may not be something easily available. Technically the average power over one line cycle for a 10kΩ is 120Vrms^2/10kΩ = 1.44W, but you won't need anything close to this because the SCR will not be on continuously. Personally I have used 1/8W, 10kΩ resistors without damaging the resistor. You could consider the average power for a PWM (16.67ms on and 300ms off), average power = (170 pk * SQRT(2*ton/toff))^2/10kΩ = 157mW.
2. You can check that system is working by simply turning the system on (connecting power) and observing that the ALARM LED blinks once 250ms after. To observe all functionality, you simply need to probe the following pins with respect to GND: OUT and SCR. Refer to the application curves in datasheet on what the signal should look like for working systems.
3. You would need logic (or processor/MCU) to make the determination of self-test/real fault and perform the action. The logic would be: real fault if ALARM is not blinking; self-test fail if ALARM is blinking. The device was made with an automatic firing of SCR for any self-test fail to adhere to the safe standards of the UL 943. If the system is overall failing and cannot sense real faults anymore then it is dangerous.
You can disable the SCR self-test by floating the SEL pin, but the periodic self-test (FT pin) and continuous self-test (watchdog timers for SCRTST and PH pins) cannot be disabled.
4. Yes, everything that is connected to ground in the typical application circuit still must remain connected to AFE3010 ground.
5. If your external source is 20V, then you don't need the current-limiting resistors into VDD pin.
6. This relay probably cannot be safely activate directly by the SCR pin alone. The datasheet shows a 50ms minimum pulse duration. The SCR may only drive the relay for 1 line cycle (16.67 ms). Also, the nominal coil voltage is 12V. The SCR pin cannot go higher than 5V. You would need to create another stage to that drive relay coil to 12V for minimum 50ms upon being activated by the SCR signal.
Sincerely,
Peter
Hello Himit,
1.
Real Fault: SCR goes high for ~1 line cycle.
Self-test Failt: SCR goes high for ~1 line cycle and ALARM starts blinking one every second.
2. You need to essentially recreate the AFE3010EVM. If you don't need neutral-to-ground detection, then you don't need the 200 turn coil and you can float NGOUT pin. If you don't need the SCR self-test, then you can float the SEL pin and short SCRTST and PH pins together. Thus it would look like this:
3. For 240VAC, you will need to adjust some components. I have attached a calculator on how to adjust component power values for 240VAC.
4. I am not proficient with optocouplers, but it should be possible to translate the signal over and isolate the grounds of AFE3010 and MCU.
5. You can use PTT or SWOPEN pins to test circuit. You can use a simple normally-open switch to pull either pin low. Refer to Figures 11-14 on how device will respond when toggling these pins.
As for the thermal pad, it is not necessary to have connected to ground for general prototyping, but it should be connected to ground for any system validation as this is how the device was tested and characterized and is best practice for the package.
Sincerely,
Peter
Hey Himit,
No. The presence of a load does not matter to AFE3010 operation.
It could be either that the Vout signal is in phase with HOT-Neutral, or the feedback gain (set by R7) is not high enough.
I recommend using high-voltage differential probes to probe OUT-GND and LINE (HOT-Neutral) and see if you can get it to match Figure 17 except where OUT does not have any sinusoid component, but rather is a flat line at ~2.5V with a downward pulse that occurs every 3 seconds. Use a falling edge trigger at <2.3V on Vout to catch this. It should be pretty apparent.
What is happening is that the device cannot see this downward pulse at 1 second after power and thus it repeats the test every second until it fails 5 consecutive time. Once this happens it fires SCR and blinks ALARM.
If signals are in phase, then make sure PH is connected to HOT. If it is, then reverse the CT coil connections to the AFE3010. If signal are out of phase by ~180 (which is what you want), then increase R7 until you see Vout glitch get to ~1.9V and system stops failing self-fault test.
Note that there could be a chance the Q1 you have selected does not have a high enough beta and thus is not really pulling enough current, which could be why device is not seeing this fault. You will realize this if once you correct your system, you see device does not trip at 5mArms, but rather something higher. Which means you would have to decrease gain (decrease R7). If the fault re-appears, consider changing Q1 or reducing R8 until device stops failing self fault test.
Sincerely,
Peter
Hi Peter,
I have few updated after my further testing.
1. I am able to check signal on OUT for EVM and my prototype, the only difference is that in EVM there is downward pulse but in my prototype there is none. I have checked this again by swapping the pins for current transformer but still the result is same. As I don't know my system is in-phase or out of phase but still as per you suggestion I just tried to increase the value of R7 to 39K and check again. From this point LED on prototype board is not functioning at all. I have also attached one image to show how out signal looks like in this stage. I tried to reduce R7 to 36K but still the result of LED and pulse on OUT pin is same.
Why this is happening? What should I do next?
Thank you,
Himit
Hi Peter,
1. As of now I am connecting everything with 120VAC(Line and Neutral) and all the values in my design are according to that. 2 phase, 240VAC, I am discussing is next plan of action, once we are done with successful testing on 120VAC.
I am focusing on solving the issue for 120VAC now so we will discuss about 2 phase system after the successful testing.
1. What kind of resistive fault I can use? I have only one differential probe so how can I check 2 signals to verify, is it in phase or out of phase?
3. As the circuit I am using is already according to what you have suggested(Only thing I dont know is that if it is in phase or out of phase so I have swapped current transformer pins when I am increasing R7 but still it is same) and I have tried highest 43K and lowest 34.5K but I am only getting 2.5V signal without downward signal. As you have suggested, once I increase/decrease R7, voltage on OUT pin should reach around 1.9V, right?
Q1 is working turning on but is on for more period compare to EVM. Same to this the FT pin signal is also stay high for more period compare to EVM. Is this period more for first 5 seconds? Please find attached images for understanding.
Hi Peter,
Thank you for your update.
Please find below details as per my current testing. We have ordered 2nd differential probe and we will receive it shortly.
1. FT pin is on for around 2ms in EVM and around 7ms for our prototype. Due to this Q1 is on for same amount of time. As per my understanding FT pin generates fault test signal internally then why this time is different for EVM and prototype. I have attached images for all these signals with proper name for your understanding.
2. I am also attaching VDD signal for both EVM and prototype.
3. I have ordered same npn transistor as EVM and check result but it is same as before.
4. Just to make everything same as EVM I connected SEL pin with ground and connected SCR_TST pin with anode of SCR with 51K resistor. In this testing, LED starts blinking as soon as I power up the device.
