The 74HC595 output high level is only half of the supply voltage

懒猫爱飞

The 74HC595 output high level is only half of the supply voltage [Copy link]

 

Background of the problem: I was bored and made a little gadget, but the MCU did not have enough IO ports, so I added a serial expansion IC: 74HC595A. Description of the problem: 1) The system power supply voltage is 3.3V, and the power supply voltage of 74HC595A is also 3.3V 2) 74HC595 has been checked and there is no cold soldering 3) Theoretically there is no problem with the driver 4) When the output is set to high, 74HC595A only outputs 1.65V, which is half of the power supply voltage? 5) Whether setting one port to output high or all 8 ports to output high, the output is only 1.65V 6) The drive voltage following 74HC595A has been disconnected, and the measured voltage is the voltage on the output pin of 74HC595A 7) As I understand it, the output high level should be the same as the system power supply voltage, which is 3.3V. But how does the output only reach 1.65V? Dear experts, have you ever encountered a similar situation? What is the main reason? I would like to ask for your advice. I have been confused for a whole day yesterday. I have tried several methods but none of them worked. I mainly tried the following methods: 1) I thought there was a problem with the driver code at first, but I checked it again and again and there was no problem. I carefully looked at the timing diagram and there was no problem either. Moreover, 74HC595A has high and low level theories, but the high level voltage is wrong 2) Checked the hardware voltage and carefully checked the three driver ports, data, clock, and lock signals. There was no problem connecting to the MCU, and the IO port settings were correct 3) Checked if there was any cold soldering 4) I changed an IC, but the problem persisted 5) I changed an MCU, but the problem persisted 6) I am going crazy and don’t know where to start…

dcexpert

I think it's either a problem with the 595 or a problem with the power supply.

懒猫爱飞

dcexpert posted on 2018-8-20 09:14 I think it is either a problem with 595 or a problem with the power supply.

1) The power supply is a desktop power supply, using the 3.3V output of LM1117 step-down. The system power consumption is relatively low, so there should be no problem with the power supply. 2) I don't know if there is a problem with 74HC595, but I changed one and it's still the same. Is it a batch problem?

dcexpert

After the basic problems are eliminated, it is difficult to analyze. You may need to find an oscilloscope to look at the waveforms of the power supply and input and output. The output voltage is exactly half, so will there be self-excitation?

IC90

It is recommended to use a logic analyzer to capture the timing of the input signal, Enable, Reset, shift clock and latch clock.

bobde163

It is best to use an oscilloscope to look at the waveform. If you are continuously inputting square waves with a duty cycle of 50%, the voltage measured with a multimeter will be half the voltage. I guess there is something wrong with your program. If you don’t have an oscilloscope, you can also use a multimeter to measure whether the signal output by the microcontroller is also less than 3.3V. If so, then it can be determined that the problem is with your microcontroller.

懒猫爱飞

IC90 posted on 2018-8-20 09:39 It is recommended to use a logic analyzer to capture the timing of the input signal, Enable, Reset, shift clock and latch clock.

OE - directly connected to GND RESET - directly connected to VCC DATA - output high-bit shift clock first - transfer 1BIT to a clock latch clock - after transferring 8BIT data

懒猫爱飞

bobde163 posted on 2018-8-20 09:43 It is best to use an oscilloscope to look at the waveform. If you are continuously inputting square waves with a duty cycle of 50%, the voltage measured with a multimeter will be half the voltage, ...

1) I don't have an oscilloscope. 2) During breakpoint simulation, the voltage levels of the clock and data input pins are 3.3V without any problems. 3) The output is incorrect

懒猫爱飞

Well, if this document is correct, I think I know the reason. If it is not correct, I have to continue to look for other reasons. Because I have never used this IC before, I looked at the data sheet of On Semiconductor, and did not find similar keywords in the manual. And my circuit is connected to an NMOS. Because the NMOS has not been turned on, I removed it and measured the pins directly, which caused the above problem. Well, if you often walk by the river, you will finally get your shoes wet. Finally, thank you all for your enthusiastic help! In addition, existence is reasonable, and what you see is not necessarily the most real. You must see the essence through the phenomenon^_^ Regarding this document on Baidu, I will continue to test and verify it when I go back...

dcexpert

Lazy Cat Loves Flying posted on 2018-8-20 10:32 Ok, I think I know the reason because I have never used this IC before, and then I looked at the data sheet of On Semiconductor, and there is no...

I have never heard that 74HC595 has an open-drain output model.

chunyang

Lazy Cat Loves Flying published on 2018-8-20 10:32 Well, I think I know the reason. I haven't used this IC before, so I looked at the data sheet of On Semiconductor. In the manual, there is no...

595 is not an open-drain output, but a conventional push-pull output. If it is an open-drain output, there is no external pull-up, and it can only measure a low level. Everything is subject to the device manual. Anything that contradicts the device manual and is not an errata sheet can only be said to be nonsense, especially in Chinese. The phenomenon you encountered must be due to improper use. It may be a hardware problem or a software problem, or even a problem with both.

懒猫爱飞

chunyang posted on 2018-8-20 10:47 595 is not an open-drain output, but a conventional push-pull output. If it is an open-drain output, without an external pull-up, it can only measure a low level. Everything depends on the device manual...

Well, I understand this. I will continue to measure it when I go back tonight.

懒猫爱飞

dcexpert posted on 2018-8-20 10:40 I have never heard of 74HC595 having an open-drain output model.

I did not find similar keywords in the manual, and I have reservations about the correctness of the documents uploaded online.

bobde163

Lazy Cat Loves Flying posted on 2018-8-20 09:55 1) I don’t have an oscilloscope, 2) When simulating at breakpoints, the voltage levels of the clock and data input pins are 3.3V and there is no problem. 3) The output is incorrect

You can't use breakpoint simulation. You should measure the data input pin when the program is running normally. If the data input pin is really outputting square waves, the voltage measured by the multimeter will definitely be less than 3.3V. Or when simulating with breakpoints, after sending data to 595, pause the breakpoint and measure whether the output pin of 595 is 3.3V.