MAX232/MAX3232 heating up.

I guess that your problem could perhaps be solved by shielding/grounding approaches.
It is more than notorious that the actual assembly with wire-up connections, RF transmitter too close, unconnected pins, and external cable is susceptible to EMI.

Concerning to grounding, you has not mentioned anything about grounding the unused rails of the breadboard, as mentioned before, whether this improved something or not. Concerning to shielding, it could be advisable to use shielded DB connectors, as well as to electrically connect its metallic case of the DB connector to the earth of the PCB ( in your current assembly, the chassis of the breadboard ), once this way all the common mode potentials caught by the antennas effects of your apparatus would be drained to the Earth.

Hi,

Is it possible to provide a schematic

Click to expand...

No, You really don´t need a schematic for two resistors. ;-)
* One from pin 10 to VCC
* The other from pin 11 to VCC

Klaus

KlausST said:

Hi,


No, You really don´t need a schematic for two resistors. ;-)
* One from pin 10 to VCC
* The other from pin 11 to VCC

Klaus

Click to expand...

ok...

To make sure ....

MAX232 Pin 11 is the pin that is currently connected to Arduino MEGA2560 pin number 18 which is TX1. So I should connect a 100k resistor from VCC to the trace that is connecting Arduino TX1 to Pin 11.

MAX232 Pin 10 is the pin that is currently unused... So instead of connecting it to ground??? I should connect it via a 100k resistor to VCC?


Should I do anything to the pin 9 and 8?! Connect them to anything?


====

I am not just sure how do you want me to use the zenner... and where.... and what value Zener diode?

I think Zener between Vcc and Gnd rail is going to protect the IC, but isn't it going to fry in a matter of milli second because of high current if it gets activated during overvoltage?

I can understand that a Zener diode in series with Vcc can help during undervoltage.... since it only conducts when the minimum required voltage is met.... and it acts like a voltage regulator! Please forgive me if I am asking a n00bish question!

Also thank you all for your patience!!! :thumbsup:

- - - Updated - - -

andre_teprom said:

I guess that your problem could perhaps be solved by shielding/grounding approaches.
It is more than notorious that the actual assembly with wire-up connections, RF transmitter too close, unconnected pins, and external cable is susceptible to EMI.

Concerning to grounding, you has not mentioned anything about grounding the unused rails of the breadboard, as mentioned before, whether this improved something or not. Concerning to shielding, it could be advisable to use shielded DB connectors, as well as to electrically connect its metallic case of the DB connector to the earth of the PCB ( in your current assembly, the chassis of the breadboard ), once this way all the common mode potentials caught by the antennas effects of your apparatus would be drained to the Earth.

Click to expand...

Thank you.

Just tested and connected all of the unconnected pins to the ground as you suggested.. still sometimes it doesn't work...

What are shielded DB connectors exactly look like?

Do you mean I should connect the chassis (metalic frame) of the DB connector with jagged pins to the ground? If that... I have already done that... I am also using a shielded cable for RS232.


The good news... I just tested a 39 ohm resistor... and so far, after turning it on and off for 50 times I wasn't able to reproduce the heating problem!! I don't have 20 ohm resistor around!

Hi,

Pin10 and pin11 must never be left floating.
If it is otherwise connected to anything always valid, then it is O.K.
But if there ever is the chance of being unconnected, then a 100k safes from floating.
A 100k won't harm in either way.

Pin8 and pin9 don't need to be connected to anything.

Zener:
Knee voltgage must be a bit above VCC. Cathode to VCC, anode to GND. On overvoltage current should flow to protect the RS232 chip.
A short pulse will not kill the diode, but without tbe diode the overvoltage may kill an IC. Onky a long duration or high power pulse may kill the zener.

Klaus

andre_teprom said:

I guess that your problem could perhaps be solved by shielding/grounding approaches.
It is more than notorious that the actual assembly with wire-up connections, RF transmitter too close, unconnected pins, and external cable is susceptible to EMI.

Concerning to grounding, you has not mentioned anything about grounding the unused rails of the breadboard, as mentioned before, whether this improved something or not. Concerning to shielding, it could be advisable to use shielded DB connectors, as well as to electrically connect its metallic case of the DB connector to the earth of the PCB ( in your current assembly, the chassis of the breadboard ), once this way all the common mode potentials caught by the antennas effects of your apparatus would be drained to the Earth.

Click to expand...

Dear Andre and other users.

In quest to save this problem completely, I see that a member on another forum, said this as a warning to the person who ran into the same problem:

When you design the circuit you didn't consider ESD considerations. I am sure that the shield of DB9 its connected direct to GND or not connected. I always use 100 ohm serial,100nF to GND for each pin from DB9 to IC. 1 uF capacitors use as multilayer ceramic low ESD type instead Tantalum. Less space, lower price better performance.

Note: Between the DB9 shield to board's GND connect via Ferrite bead parallel 1000pF/2kV 0805 capacitor.

So I think that this guy believes that connecting the shield of DB9 to directly to the GND is not a good solution and should be done via a 1000pF capacitor....! What do you guys think!? Do I need to connect the capacitor in series, between the shield and gnd?

Hi,

So I think that this guy believes that connecting the shield of DB9 to directly to the GND is not a good solution and should be done via a 1000pF capacitor....! What do you guys think!? Do I need to connect the capacitor in series, between the shield and gnd?

Click to expand...

I think it is a good idea.
The RC is a path for an ESD pulse..but it prevents from ground loops.
It allowes DC offset and it has damping character, that means it prevents oscillations.

Klaus

uart_rx_tx said:

So I think that this guy believes that connecting the shield of DB9 to directly to the GND is not a good solution

Click to expand...

Me too.

By the way, I haven't said to you connect the shielding contact of the DB9 cable to the GND.

When I said grounding at post #21, I was rather referring to "earthing", meaning that - considering that the shielding mesh of the cable is connected to the connector's frame - the metalic case of the connector is what should be connected to the Earth, which is also supposed to be connected to the chassis of the equipment ( in your assembly, the plate of the breadboard )

In other hand, at the post #10, in fact, the term grounding was used in the exact sense of connecting to GND the unused nearby rails of the breadboard.

Concerning to the decoupling capacitor, it is always recomended to add ones between the GND and the Earth, as you quoted above.

Thank you very much everybody...


I searched the term "ferrite bead" in the google and I found out what it is!! I didn't know the name, and honestly didn't pay attention to it before. But I now for sure know that the original cable from the sensor doesn't have a ferrite bead!

Do you think that it is an important thing to implement in my design? I can easily replace the cable!

I also wonder whether this cold have been the culprit!

I think that ferrite beads are more suited for high levels of EMI. In your case, even the smallest noise is able to sensitize the high impedance inputs of the IC. A pull-up resisor as Klaus mentioned, in addition to extra minor precautions on the grounding and shielding scope could pretty suffice.

I had questiopns about the max232 about 5 years ago - https://www.edaboard.com/threads/180893/
There is a lot of noise made by the charge pumps on this chip - but I can't see a reason why one should heat up unduly.

Reading this thread one thing does come to mind - the fact that cycling the power causes a change.
I can't see any mention in the thread that you have tried changing the caps. It just occurs to
me you may have a bad one causing the charge pump to go wonky on power cycling.

just a thought.

P.S. It may be worth making sure you havn't got an RS232 cable grounded at both ends.

123jack said:

I can't see any mention in the thread that you have tried changing the caps. It just occurs to
me you may have a bad one causing the charge pump to go wonky on power cycling.

Click to expand...

Your remark brought me something to mind: Given that he mentioned have assembled the circuit in a breadboard, perhaps the the external capacitors of the IC are not placed at an adequate distance. A picture would help to confirm or not this possibility.

123jack said:

I had questiopns about the max232 about 5 years ago - https://www.edaboard.com/threads/180893/
There is a lot of noise made by the charge pumps on this chip - but I can't see a reason why one should heat up unduly.

Reading this thread one thing does come to mind - the fact that cycling the power causes a change.
I can't see any mention in the thread that you have tried changing the caps. It just occurs to
me you may have a bad one causing the charge pump to go wonky on power cycling.

just a thought.

P.S. It may be worth making sure you havn't got an RS232 cable grounded at both ends.

Click to expand...

What happens if the chassis of the cable is grounded at both sides?

I will change the capacitors and will notify you.

I have also removed the Vcc resistor, and tested the circuit with the 100k resistor connected between Vcc and 11 and 10 pins... it still heats up.

- - - Updated - - -

andre_teprom said:

Your remark brought me something to mind: Given that he mentioned have assembled the circuit in a breadboard, perhaps the the external capacitors of the IC are not placed at an adequate distance. A picture would help to confirm or not this possibility.

Click to expand...

I don't have access to the circuit for several days... but I have posted some videos on this thread: OpAmp driven RX/TX activity LED for UART.
You might find out if I have done anything wrong of not... I am not sure.

uart_rx_tx said:

I have posted some videos on this thread: OpAmp driven RX/TX activity LED for UART.

Click to expand...

Impossible to identify something in details there.

uart_rx_tx said:

What happens if the chassis of the cable is grounded at both sides?

Click to expand...

You can set up a ground loop.
**broken link removed**)

The paragraph on signal cables gives a good intro I think.

123jack said:

You can set up a ground loop.
**broken link removed**)

The paragraph on signal cables gives a good intro I think.

Click to expand...

Thank you,

Does it also apply if both sides of the cable (modem{sensor} and receiver{uC}) are using the same ground? Because I am supplying the 5V voltage from the same lead-acid battery (through a regulator)!

By the way, the sensor's wire is Cooner Wire with a Lemo Connector.

- - - Updated - - -

This one also shows the power connector that directly goes to the cable (the bottom one)

Ground loop problems are usually just caused by the cable shielding being grounded at both ends.
Just disconnecting one end usually fixes any issues. (I once had an industrial robot go very barmy because of this)
Normally the signal connections aren't an issue in practice unless the design is really poor
or cable runs are very long or wrapped around mains cable.

Induced current from magnetic fields "floating about equipment or cable" can be a bit of a black
art but from what I read here I dont think it is a major issue for you? Unless you're working next
to an arc welder or something.

If everything is battery powered you shouldn't have a problem.
If you just check your cable at each end that should eliminate the issue for you.
One end grounded - the other not - then forget the issue.

**broken link removed** circuit has been used for over 10 years.



It may be used freely.

maruker said:

**broken link removed** circuit has been used for over 10 years.

View attachment 128024

It may be used freely.

Click to expand...

Thank you,

What is the purpose for putting the 100 ohm resistor in tx and rx lines? What are the pros and cons of using this resistor?

And also what is the purpose for putting the R2 resistor between 12 (TTL_RX) and 5V? What are the pros and cons of using this resistor?

uart_rx_tx said:

What is the purpose for putting the 100 ohm resistor in tx and rx lines?

Click to expand...

The 100R resistors at above schematic weren't placed at the rs232 side, but at the logic side, although I would prefer putting them at the part of the circuit which interfaces with the external "world". In fact, they were intended to protect the IC against spikes, somehow limiting the inrush current.

Hi,

My opinion:

What is the purpose for putting the 100 ohm resistor in tx and rx lines? What are the pros and cons of using this resistor?

Click to expand...

pro: limits accidental input currents. Prevents from latch up
con: costs. no technical con

And also what is the purpose for putting the R2 resistor between 12

Click to expand...

pro: As dicussed before. It prevents TTL_RX from accidental floating. (cable disconnect, unconfigured IO port microcontroller ...)
con: costs. no technical con. (As long as the microcontroller has enough drive_low_current. Should be fine)

Klaus