Frequent failure in 74HC595 IC

[Prabhakarankft]

Hello Everyone.
Greetings. Recently I have developed a project with a lot of 74HC595D SMD IC.

A brief explanation about my project:
The actual purpose of my project is to control nearly 900-2000 solenoids( about 15V and 25mA each solenoid).

We have three separate boards to handle the user input and controls the solenoids.

MotherBoard--> Slave Board--> Driver Cards.

The real problem exists in the driver card. Distribution boards are sending the data to driver cards. Since my application is to control a large number of solenoids each driver's card is daisy-chained in connection.

The circuit inside the driver cards is so easy. We have a 74HC595 IC(P.No: 74HC595D,118 ) and ULN2803 (P.No: TBD62083AFWG). I am using 7805(P.no: L78M05CDT-TR, STM ) to give voltage to 74HC595 IC.

Operation wise everything is working fine. But many times 74HC595 IC is getting failure after some days or months sometimes. We will insert the driver cards directly to the solenoid box Like RAM ports in computers(Such a way solenoid boxes are designed- Each box contains 8 solenoids)

Distribution boards are having dedicated FRC headers with a latch to connect Driver cards. We using Flat Ribbon cable(14 way - Neltron brand 28AWG) to connect the distribution board and driver cards. 1 Connector will handle 16/26 number of driver cards. All driver cards are daisy-chained.

Clock pin, Latch Pin, and OE pins of 74HC595 ICs are connected parallelly and it is connected to 74LS244N( HEX Buffer) for isolation purposes. The data input for the driver card is directly coming from the MCU which is in the Distribution board. We found many failures in the 74HC595 IC and some times one/two channels in the hex buffer.

If We change 595 IC in the defected driver card, it is working properly. We need to know the cause of this problem. Kindly help me out.


Sorry for bbbbbbiiiigggggg storyyyyyy..

 

[stenzer]

Hi,

The circuit inside the driver cards is so easy. We have a 74HC595 IC(P.No: 74HC595D,118 ) and ULN2803 (P.No: TBD62083AFWG).

how does the connection of those ICs look like? I'm interested in the usage of the freewheeling diodes.

BR

 

[KlausST]

Hi,

Instead a long story you could post
* schematic
* wiring informations (length, star...)
* photos

Klaus

Btw: instead of HC595 plus ULNxxx you could use a power_595...there are several similar types from several vendors

 

[Prabhakarankft]

Thanks for showing your interest stenzer!
Circuit attached for your reference. We are using ULN2803 Thosiba IC. It is having inbuilt flywheeling diode.

--- Updated Oct 4, 2020 ---

Hello KlausST.

Schematics have 3 - 4 pages(Distribution and MotherBoards). So I thought explaining the problem will be easy to understand. I am attaching the driver card schematic for your reference

--- Updated Oct 4, 2020 ---

What would be the impact of ESD here? Sometimes I am suspecting "failures in 595" will be due to ESD issues.

During the installation time, there may be a chance of touching the boards with bare hands without any ESD protection!

For your consideration, We are using 14 way FRC Flat cables to connect all driver cards.

Distribution Board--------> Driver Card1--Driver Card2--Driver Card3--Driver Card4--Driver Card5--,--,--,--Driver Card 16

 

[KlausST]

Hi,

I don't think a textual description can replace a schematic.
...and no textual description can replace pictures of a PCB layout or a wiring.

I can try to explain how my car is build. All cars have a motor, four wheels, a gearbox, a chassis.... but how they are "joined" make how they work, how reliable the car is, what one can expect.

A HC595, a 7805, capacitors are just components
Each component will work according the datasheet (don't expect the problem is caused by the HC595),
but how they are "joined" = PCB layout, signal wiring, GND wiring, power wiring.....

So far .... I don't have enough useful informations to give good assistance.

Klaus

 

[stenzer]

Hi,

according to your initial post the solenoids are attached to ~15 V, but the freewheeling diodes are attached to 12 V. How does the actual connection looks like?

BR

 

[danadakk]

What you might consider doing is use your DSO and set it up for single
shot trigger, leading edge trigger, set trigger value for a pin allowed V +
~ .5V. So 595 Vcc is 5V, set trigger to 5.5, and see if you get any triggers.
Do same with its outputs. Also trigger outputs at Vss - .5, see if you get any
triggers. Basically you are looking for a V that exceeds 595 allowed pin
voltages.

Do both inputs and outputs.

Note you probe has to be comped and its ground lead short and as close
to 595 ground pin as possible.

Inputs that are daisy chained across several cards, if power transients are
causing transients due to L in traces/leads suspicious.

Regards, Dana.

 

[Prabhakarankft]

Hi,

according to your initial post the solenoids are attached to ~15 V, but the freewheeling diodes are attached to 12 V. How does the actual connection looks like?

BR

Dear Stenzer
The actual working voltage is 15V. Solenoids are working at the same voltage. I could not find the 15V symbol(I thought it is not in the component library). So Instead of 15V, I have used the 12V symbol. I apologize for the wrong information I have conveyed

 

[stenzer]

Hi,

... and some times one/two channels in the hex buffer.

the HEX buffer is not shown in your partial sketch, and it seems it is also causing troubles/ is affected. Please provide the missing parts of your schematic and if possible a picture of your complete system. It is hard to get an overview only by the textual explanation.

BTW, the 74HC595D has a higher HBM ESD rating compared to the 74LS244N (MM not stated for the HEX buffer), thus the 74HC595D should be less prone to ESD (HBM). Therefore, the interconnection as well as the layout itself is of special interset to determine potential failure (ESD) causes.

BR

 

[Prabhakarankft]

Instead a long story you could post
* schematic
* wiring informations (length, star...)
* photos

Klaus

Btw: instead of HC595 plus ULNxxx you could use a power_595...there are several similar types from several vendors
[/QUOTE]
Hello,
Does power_595 comes with inbuilt free wheeling diodes?

--- Updated Oct 23, 2020 ---

Hi,



the HEX buffer is not shown in your partial sketch, and it seems it is also causing troubles/ is affected. Please provide the missing parts of your schematic and if possible a picture of your complete system. It is hard to get an overview only by the textual explanation.

BTW, the 74HC595D has a higher HBM ESD rating compared to the 74LS244N (MM not stated for the HEX buffer), thus the 74HC595D should be less prone to ESD (HBM). Therefore, the interconnection as well as the layout itself is of special interset to determine potential failure (ESD) causes.

BR

Hello,
To get a better idea I have created a new schematic with only one port(Channel). I explained all the stories about my project. The attached schematic is not exactly the same which we are using now. Just to explain how We have connected Hex buffer with 74HC595 and ULN2803.

I am suspecting many things for the failure of 595 and Hex buffer.
1. Output from the Hex buffer(74LS244N) is 3.6V and it is powered by 5V. I have checked the datasheet also. They have mentioned VOH is 3.6V. So SCK, RCLK, and OE pins of 74HC595 are working at 3.6V(Output from Hex buffer). All SCK, RCLK, OE pins are paralleled and data pins are connected with a daisy chain

2. In the data sheet of 74LS244N Vin is just 2V nominal. It is a non-standard value. Can anyone clarify this?? IO's connected from MCU to Hex buffer is 5V. In between Hex buffer and MCU, we have 22E Resistor

3. Any noises will be added from the Flat ribbon cable. Flat cable is 14 ways and it is carrying 15V DC for solenoids as well as data, clock for 595.

4. 595 controlling pins are working at 3.6V. data in pin working at 5V. 595 powered with 5V. Working with 2 different input voltages causes any problem??

 

[stenzer]

Hi,

please upload a larger version of your schematic, it is impossible to read anything.

4. 595 controlling pins are working at 3.6V. data in pin working at 5V. 595 powered with 5V. Working with 2 different input voltages causes any problem??

You are using the HC version which has CMOS input levels, thus for a 5 V supply voltage your minimum high voltage level is 5 V \[\cdot\] 0.7 = 3.5 V, which is pretty much on the edge. Have you ever tried the HCT version which has a minimum high level voltage of 2 V ?

BR

 

[KlausST]

Hi,

(Please use the QUOTE funtionality of the forum software)

not the classical free wheeling diode, but overvoltage clamp. (and include overcurrent protection)
See https://www.ti.com/lit/gpn/TPIC6C595
also NPIC6C595, STPIC6C595, ...
I think it´s compatible to HC595 from the serial interface side.

but there are many other open drain power drivers with serial inteface.
(maybe not HC595 inteface compatible)
some have additional features like open load detection, read input state, thermal shutdown...
like BD8LA700, NCV7240DPR2G, MC33996 ... just do a search at the IC manufacturers or distributors internet sites.

Klaus

 

[Prabhakarankft]

Hi,

please upload a larger version of your schematic, it is impossible to read anything.



You are using the HC version which has CMOS input levels, thus for a 5 V supply voltage, your minimum high voltage level is 5 V \[\cdot\] 0.7 = 3.5 V, which is pretty much on the edge. Have you ever tried the HCT version which has a minimum high-level voltage of 2 V?

BR

Hi Stenzer,
Thanks for your prompt response. If We use TTL level IC(74HCT595) need to change Hex buffers? CMOS Level ICs are having many more advantages over TTL level. Are there any other reasons for your suggestion?

Really I am confused what is the use of Hex buffer here. As of my knowledge, it will give high input impedance and isolate MCU pins for the load(So that We can connect N number of CLK/RLCK/OE input of 595 - to save IO pins- daisy chain

What are all the factors that need to be considered while choosing a Hex buffer? CMOS devices are having a high fan-in capacity as per the google results.

--- Updated Oct 23, 2020 ---

Hi,




how does the connection of those ICs look like? I'm interested in the usage of the freewheeling diodes.

BR

I am using ULN2803(P.No: TBD62083AFWG ) for driving solenoids 15V . It is having inbuilt free wheeling diodes/Clamp diode .

 

[KlausST]

Hi,

CMOS Level ICs are having many more advantages over TTL level.

It depends. What advantages do you think of?

Klaus

 

[stenzer]

Hi,

If We use TTL level IC(74HCT595) need to change Hex buffers?

no, the HCT accepts TTL input voltage levels and provides CMOS compatible output levels as long you are not sourcing a too high output current (V_OH,min = 3.7 V @ 6 mA). Simply compare the output values in [1].

Really I am confused what is the use of Hex buffer here.

Me too, as I do not see how it is used until now. Please update the schematic in post #10 to a larger version.

[1] https://assets.nexperia.com/documents/data-sheet/74HC_HCT595.pdf

BR

 

[FvM]

In post #4, the indiactor LEDs are supplied with Vcc (5V) which causes violation of the maximal reverse LED voltage. Please correct.

Damage to the HC595 can be caused by input voltage outside maximum ratings. HCMOS has substrate diodes clamping the inputs agaist GND and VCC, but with limited current capability. Series resistor at all logic inputs can protect the IC, in extreme cases additional external clamping diodes. The 5V node should be protected by a zener or TVS diode against being pulled high.

 

[Prabhakarankft]

In post #4, the indicator LEDs are supplied with Vcc (5V) which causes a violation of the maximal reverse LED voltage. Please correct.

Dear FVM,
In that schematic Supply voltages for LED is not 5V as u mentioned. It 15V coming from a Pin of Header. For easy connection, I have labeled those pins as VCC. But it is 12V. Reverse voltage will not mix up with 5V supply at any time.

VCC(15V DC)----> 12K Resistor---> LED--->Collector of ULN2803

 

[KlausST]

Hi,

But it is 12V

???

Klaus

 

[Prabhakarankft]

???

Oh My god. It is typo.. 15V

 

[FvM]

Holy confusion. You could have placed the 12V net label at least...
Anyway, my major point was the logic input protection.