ADI's LT8645SIV#PBF chip uses the EN pin to control power on and off, causing it to burn out. Help~

木木夕了

ADI's LT8645SIV#PBF chip uses the EN pin to control power on and off, causing it to burn out. Help~ [Copy link]

 

LT8645SIV#PBF is a DCDC step-down chip from ADI, with a maximum input of 65V. Two of them burned out recently when I was using them, and I couldn't find out the problem. I came to the forum to ask for advice. P.S. The price of this chip has gone up a lot recently, and I can't afford to burn it...
As follows:
1. Attach the schematic diagram first: The peripheral circuit is relatively simple, with an input voltage of 30~58V and an output of 5V. A voltage divider network is formed by RA10 and RA13, and a switch is connected at H1 to control the enable pin of the chip. The idea is: when the switch is open, the chip EN pin is powered by the voltage divider of RA10 and RA13, and the chip is enabled to work; when the switch is closed, RA13 is short-circuited to ground by the switch, the chip EN pin is grounded, and the chip is turned off.

2. Problem: After adjusting the resistance value of RA13 in the schematic diagram, the chip was burned by using a switch to control the chip. However, the EN pin did not exceed the electrical properties in the manual.
3. The test conditions and phenomena are as follows: The simple description of the phenomenon is that when the power supply to the EN pin is greater than a certain voltage value, the switch is grounded, and the chip is burned.
Group 1: When RA13=8.2K,
the chip output is normal and the switch function is normal. And this combination has been used for a long time without any problems.
Group 2: When RA13=15K
① When the input voltage vcc=24v, the chip output is normal and the switch function is normal.
② When the switch is disconnected (that is, the EN pin is not grounded), add 50v voltage to vcc, and the chip output is normal. Then close the switch (ground), and the chip burns.
Group 3: When RA13=100K,
when the switch is disconnected (that is, the EN pin is not grounded), add 24v voltage to vcc, and the chip output is normal. Then close the switch (ground), and the chip burns.

4. Attached is the chip manual link, chip EN pin parameters, and official demo board schematic
data sheet: https://www.analog.com/media/en/technical-documentation/data-sheets/lt8645s-8646s.pdf
If you don’t want to look through the manual, I have cut out the key parameters.

RichSJ

Pin 30 should not be grounded. I wonder if this causes the internal error comparator to work in a confused manner during shutdown, resulting in burning.

Maybe no one has dared to try this before.

littleshrimp

Try removing CA13

木木夕了

RichSJ posted on 2022-6-25 10:31 Pin 30 should not be grounded. I wonder if this caused the internal error comparator to work in confusion during shutdown, causing it to burn out. Maybe no one dared to do this before...

PIN30 is GND. The 30 in the internal block diagram is the 8646 chip. Mine is 8645.

木木夕了

littleshrimp posted on 2022-6-25 11:37 Try removing CA13

Why remove CA13? I only have two of them... But the circuit does not exceed the maximum properties of the EN pin. I don't understand why it burns.

RichSJ

Mumu Xile posted on 2022-6-25 20:26 Why remove CA13? I only have two pieces... But the circuit does not exceed the maximum properties of the maximum EN pin. I don't understand why...

Sorry, my eyes are blurry.

Basically the same as changing the capacitor, when you increase the resistance by 1 to 13 times, the chip will burn out, and the RC constant will increase, so reducing C may be useful, but I think CA14 should be considered. But this is not a good solution, because it may increase the risk of switch jitter.

The problem may be that the shutdown signal is not steep enough during the Burst mode shutdown.

RichSJ

RichSJ posted on 2022-6-25 21:08 Sorry, I am dazzled. Basically the same as changing the capacitor, when you increase the resistance by 1 to 13 times, the chip will burn out and the RC constant will increase, so...

The safest approach is to use a circuit instead of a simple RC

不亦心

Mumu Xile published on 2022-6-25 20:26 Why remove CA13? I only have two of them... But the circuit does not exceed the maximum properties of the maximum EN pin. I don't understand why...

If it is a mechanical switch, it is normal to have voltage spikes, and you can use a transistor to control it. To verify whether it is caused by the switch, you can use an oscilloscope to check it, without soldering the chip.

木木夕了

RichSJ posted on 2022-6-25 21:08 Sorry, I am dazzled. Basically the same as changing the capacitor, when you increase the resistance by 1 to 13 times, the chip will burn out and the RC constant will increase, so...

I wonder if removing both CA13 and CA14 will solve this problem.

RichSJ

Posted by Mumuxi on 2022-6-26 02:20 If I remove both CA13 and CA14, I wonder if this problem can be solved

I think your project is related to electric vehicles. If so, safety is the primary requirement and unknown hidden dangers are not allowed.

If it is caused by Burst mode and shutdown (the manufacturer says that the power and EN are turned off at the same time), the power supply of the internal control circuit may be out of sync with the main power supply, which may cause damage.

Even if you remove the capacitor, there will still be jitter when manually switching on and off (the switch's shrapnel...) and there is still a chance of burning.

The previous person suggested using an oscilloscope to view it, but I don't think you can burn another chip to view it.

RichSJ

RichSJ posted on 2022-6-26 04:00 I think your project is related to electric vehicles. If so, safety is the primary requirement, and unknown hidden dangers are not allowed. If it is because of Bu ...

The 8th floor mentioned that a three-stage tube circuit can be used.

RichSJ

RichSJ posted on 2022-6-26 04:02 The 8th floor mentioned that a three-stage tube circuit can be used.

By the way, when using your project, can you not shut down in Burst Mode? If you change to other modes before shutting down, the risk will be smaller.

littleshrimp

Mumu Xile posted on 2022-6-25 20:26 Why remove CA13? I only have two pieces... But the circuit does not exceed the maximum properties of the maximum EN pin. I don't understand why...

I have carefully read your circuit and the manual of LT8645, but cannot find the direct cause of the chip burning out. According to your description that increasing RA13 will cause the chip to "burn out", I guess the problem may be related to the two capacitors CA13 and CA14.

According to your description, your previous circuit has been in normal use, so it can be basically ruled out that static electricity caused damage to the chip through the EN/UV pin. In addition, because you increased the resistance of RA13, I am not sure whether the voltage of the EN/UV pin fluctuates outside the hysteresis range of the comparator due to interference or other reasons at a certain moment, causing the chip to be frequently turned on and off, which will cause damage to the chip.

Your actual testing process and methods require you to provide more information, such as whether the chip burnout was caused in the actual usage environment, or was tested in an experimental environment using a regulated voltage, what kind of load was used, whether the switch was directly controlled by hand, etc. (Because the test data is small, the cause of static electricity cannot be completely ruled out).

Another question is what does the "burn" you mentioned look like? Does the chip emit smoke directly, or does the chip just have no output? By measuring and comparing the bad chip with the good chip, which pins have the largest differences in parameters?

木木夕了

RichSJ posted on 2022-6-26 04:00 I think your project is related to electric vehicles. If so, safety is the primary requirement, and unknown hidden dangers are not allowed. If it is because of Bu ...

Thank you very much for your reply

This is the waveform of the EN pin when the switch is in action (no load is connected).

木木夕了

littleshrimp posted on 2022-6-26 13:10 I have carefully read your circuit and the LT8645 manual, and cannot find the direct cause that may cause the chip to burn out. According to your suggestion, increasing RA13 will cause the chip to "...

Thank you very much for your reply

① When RA13=8.2K, the project was still in the testing stage. It had been tested dozens of times before, and high and low temperature environment tests were also conducted during the period. No abnormal situation occurred. The chip burned out when the RA13 resistor was adjusted a few days ago.

②The environment in which the chip burned was as follows: The output was not loaded (RA1 was not soldered in the schematic diagram); the switch used a Honeywell toggle switch (manually turned on); the input used a regulated power supply, and the output was limited (output 50V/0.1A); the testers wore anti-static wristbands during the experiment.

③ Experimental phenomenon: The voltage divider resistor of RA13 was readjusted. When the switch was not closed, the voltage and ripple measured were qualified. So the switch function was tested. As soon as the switch was grounded, the chip was abnormal. The manifestation was: after the switch was grounded, the input and output were short-circuited, and the voltage-stabilized power supply was directly shut down when the set current exceeded (so the chip did not smoke); after the chip was removed, the output and input pins of the chip were measured to be short-circuited with the GND pin of the chip.

④ Waveform of EN pin when RA13=15k, VCC=50V (without chip)

木木夕了

Bu Yixin posted on 2022-6-25 22:00 If it is a mechanical switch, it is normal to have voltage spikes, and you can use a transistor to control it. To verify whether it is caused by the switch, you can use an oscilloscope to check it out...

Hi, tested.

木木夕了

littleshrimp posted on 2022-6-26 13:10 I have carefully read your circuit and the LT8645 manual, and cannot find the direct cause that may cause the chip to burn out. According to your suggestion, increasing RA13 will cause the chip to "...

Hello, I retested the EN pin and adjusted the oscilloscope parameters. When the switch was closed, the EN pin ringed . I don't know if it was because of this reason...

木木夕了

RichSJ posted on 2022-6-26 04:00 I think your project is related to electric vehicles. If so, safety is the primary requirement, and unknown hidden dangers are not allowed. If it is because of Bu ...

Hello, the latest test waveform has been updated. I don't know if it is because of this reason..

littleshrimp

Posted by Mumu Xile on 2022-6-26 17:19 Hello, I retested the EN pin and adjusted the oscilloscope parameters. When the switch was closed, the EN pin ringed. I don't know if it's because of this...

From the waveform you measured, it can be
seen that the voltage fluctuation of the EN pin will also increase when the resistance increases. It is possible that the chip is burned out due to excessive negative pressure. Can you find out whether the cause of the oscillation is that the switch uses a relatively long wire?

木木夕了

littleshrimp posted on 2022-6-26 20:19 From the waveform you measured, when the resistance increases, the voltage fluctuation of the EN pin will also increase. It is possible that the chip is burned out due to excessive negative pressure. You can find the cause of the vibration...

① Switch wire 50cm

② After removing CA13 and CA14, the instantaneous oscillation of the switch grounding still exists.

③ After removing CA13 and CA14 and replacing RA11 with a 10k resistor, the oscillation disappears. The specific waveform is as follows