Please don’t take this mail as a beginner mail. I have really a serious problem that turns my twelve months product design to a waste.
I designed a an RS232 interface board and used MAX232CWE+ 16 wide SOIC
I used 1uF/50V SMD Case B Aluminum Electrolytic capacitors for MAX232. I use a generic 100nF SMD capacitor as a decoupler of MAX232.
You can see my sch. and layout parts below;
Here is my problem,
When I connected my system to a PC serial port via an RS232 cable (19200 baud and no parity, etc);
1. There is no problem first. System is communicating correctly.
2. I started to disconnect and connect the serial port periodically while my finger was on the MAX232 to measure the heating J
3. In some way MAX232 is starting to overheat and the serial line is starting to not working.
4. If I close my system while MAX232 is heating, waits MAX232 get cold for nearly 1 minute and plug the system again there is no problem. MAX232 is working well.
5. If I don’t close the system, MAX232 is damaged because of the heating.
6. I dumped the problem with an oscilloscope. When I send some bytes to the system via a PC I should see nearly 12,13V if there is no problem but when the IC is damaged, I see only 1V at the R1IN pin (13. Pin) of MAX232. So I’m sure that MAX232 R1IN pin was collapsed the PC transmit signal. Please note that none of five capacitors for MAX232 was not damaged or heating up.
7. I changed the MAX232 IC and all of the rouitine is starting again. I did this nearly 10 times J
8. Please also note that I double checked the capacitors polarity. There is no problem for all
9. In maxim site It tells about an ERRATA for MAX232CWE+( Product Change Notice 801-ERRATA) ; However I couldn’t find this document. I suspected that this IC has a BUG that I encountered
10. I’ve searched the internet for 1 week and there are same problems but all of explanations are about look at the directions of the caps, etc.
11. I found a forum thread on TI forums below and a person named David connected a 20 ohm series to MAX232 VCC and solved his problem. It is not a good but may be a reasonable solution
12. I know there is a TI employee could check my problem in detail. If you do I’ll be very glad
Waiting for your urgent reply
In reply to saltan:
I have tried to get the MAX232 to overheat but I was not successful. Can you return a couple devices to your distributor as defective for failure analysis. It would be good to return samples that have the overheat issue but still work after cooling off. Also return some that are permanently damaged (don't work when cool).
The C3 capacitor can be placed in either location for either device. It does not affect the charge pump action.
In reply to Ron Michallick:
I've connected the C8 negative pin to VCC and measured the VCC, V+, V- for damaged and not damaged IC at run time. Please see below
NOT DAMAGED MAX232 ON BOARD
V+ : 9.3V
V- : -8.71V
Some Pins of MAX232: Squarewave signals
DAMAGED MAX232 ON BOARD
V+ : 4.3V
V- : 1.2V
Some Pins of MAX232: NO Squarewave signals
Heating: LIKE HELL
I'm pretty sure that the VCC is stable
Do you think that the below switching freqencies are normal
MAXIM MAX232CWE+ : nearly 600KHz at pin 3
TI MAX232 : nearly 45KHz at pin 3
I read to many things on the net about charge pump capacitors. In short;
- if you use fast switching you need lower capacitance
- if you need low switching you need higher capacitance
From this point of view, capacitors suggestions for "600KHz and 45 Khz" are same and 1uF. for both TI and MAXIM ICs ???
Do you have any comments for switching frequencies and capacitor selection
45kHz is too low. I do not have data on hand for the nominal frequency of the charge pump. I will check some samples this week.It takes a lot of current to pull the V- pin positive. There is a Schottky diode to ground on the V- pin.
On devices that overheat but are not destroyed, is the V- pin also positive? It seems that the MAX232 is latching up. Other users who added a VCC resistor may have prevented latch up by limiting the (SCR) hold current so the latch is unsustainable.
I do not know where the trigger for the latch originates, but the trigger is certainly external to the device.
Have you considered upgrading to the MAX202? It is the same cost and is a drop in alternative to the MAX232. Both device have the same function, but the internal schematics are very different.
Thanks for your answer. Please see below for my answers;
45kHz is too low.
I see nearly 45kHz with an oscilloscope.
I do not have data on hand for the nominal frequency of the charge pump. I will check some samples this week.
It takes a lot of current to pull the V- pin positive. There is a Schottky diode to ground on the V- pin.
On devices that overheat but are not destroyed, is the V- pin also positive?
Yes V- pin also positive for both TI, or MAXIM ICs. I tried them on the same PCB.
It seems that the MAX232 is latching up. Other users who added a VCC resistor may have prevented latch up by limiting the (SCR) hold current so the latch is unsustainable.
I do not use MAX202 But I can. Pins are ok but footprints are a little different. I may use it.
You are correct, the MAX232 does run near 45kHz. It is the MAX202 and entire MAX/TRS32xx series that runs over 100kHz.
The V- should always be negative in operation. I tried forcing V- positive (up to 1Amp), but it ran normally once, I stopped the abuse.
I have not been able to get any MAX232 to latch-up (overheat) in the several times I have tried. Please try the MAX202 and let me know the result.
Dear Ertan Bey,
When you design the circuit you didn't consider ESD considerations. 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.
When you do this please share the results with us.
BTW - There in mounting hole on trace to Pin# 8 of DB9 I hope its not plated hole and you have short to GND
Haim Rodrik - CEO/CTO
jTrack M2M Tellocation Systems LLC
Note: Between the DB9 shield to board's GND connect via Ferrite bead parallel 1000pF/2kV 0805 capacitor.
if you need more assistance feel free to use rodrik5000
In reply to Haim Rodrik:
I have reached this forum in search for a solution for the same latch-up situation.
In my case I think I have found the scenario that causes the latch-up - every time the analog side RS232 pins of the chip are connected to a powered serial port (e.g. connecting a board with the MAX232 to a running PC's serial port) without power applied to it's power pins (i.e. the board not powered), once I apply +5V to the chip it immediately enters the latch-up condition. I have narrowed this to the TxD input - because it's driven with a negative voltage (negative idle voltage on RS232 line), the combination of rising VCC and negative input on TxD causes in my opinion the latch-up condition. This seems like a bug of these chips related to the silicon implementation.
Can TI tech support check this analysis and try t simulate it too?
All content and materials on this site are provided "as is". TI and its respective suppliers and providers of content make no representations about the suitability of these materials for any purpose and disclaim all warranties and conditions with regard to these materials, including but not limited to all implied warranties and conditions of merchantability, fitness for a particular purpose, title and non-infringement of any third party intellectual property right. TI and its respective suppliers and providers of content make no representations about the suitability of these materials for any purpose and disclaim all warranties and conditions with respect to these materials. No license, either express or implied, by estoppel or otherwise, is granted by TI. Use of the information on this site may require a license from a third party, or a license from TI.
TI is a global semiconductor design and manufacturing company. Innovate with 100,000+ analog ICs andembedded processors, along with software, tools and the industry’s largest sales/support staff.