MAX232 problems

[E Kizer]

Occasionally I will search the forum for serial routine ideas. These searches always bring up old posts from people who have overheating issues with their RS232 interface IC’s. In every post I have seen here, that included a schematic, the schematic is wired incorrectly. Instead of posting how to correctly wire an interface chip, the threads sometime gravitate to discussions on fake chips from China. Though fake chips are a problem, the fix is to observe proper wiring of TTL devices.

TTL wiring rule number one;

All inputs MUST go somewhere.

Leaving an input floating can cause wild oscillations, overheating issues and massive current draw issues. The MAX232 is a two channel device with two TTL inputs. Every schematic posted in this forum shows only one input tied to something. The other is left floating. This is wrong. Tie that other, unused, input to either ground or 5 volts.

Another point of confusion I see is that “MAX232” is taken as literally, when it is really being used generically . When shopping for a new IC, it is better to search for a “RS232 interface IC”. A Maxim Integrated MAX232CPE from Mouser.com cost $3.20. I prefer to use a Texas Instruments TRS202ECN from Mouser.com which costs $1.53. I get the same performance at half the price. There’s no need to turn to China for affordable interface ICs.

So if you happen across a post that includes a schematic that shows an RS232 interface chip that is wired incorrectly, an input left floating, reply to that post to do something with that unused input. Doing that will save someone hours of trouble-shooting time.

[sayzer]

What is it that you did not understand on this circuit?

[Demon]

I always check the datasheet and check at a sample circuit there. Mainly to make sure what to use as pull-down/up resistor values (some people approximate in their circuits).

I know you can get away with approximations in resistors and capacitors in a LOT of cases. I just prefer to err on the side of caution, usually cause I end up on the side of user-error in many cases regardless.

Robert


EDIT: Take the MAX232, I'll even try to find the datasheet for that given manufacturer, just in case.

I know many chips "might" follow industry standards for a given chip, like a MAX232 (just an example), but again, I wouldn't trust a circuit I invested a gazillion hours putting together, only to have it "oscillate wildly" cause one manufacturer decided to be the lone ranger.

I'm not going to put anything in production, sadly, but I still want to be able to trust my crap when I do get it working, one day, eventually...

[richard]

TTL wiring rule number one;

All inputs should go somewhere.




CMOS wiring rule number one;

All inputs must go somewhere.


The problem with sayzer's circuit is that it will not work for max232epe chips , they will get hot real hot

the switching caps need to be ten times smaller 0.1uF

http://www.picbasic.co.uk/forum/showthread.php?t=19548

[sayzer]

TTL wiring rule number one;
....

The problem with sayzer's circuit is that it will not work for max232epe chips , they will get hot real hot

the switching caps need to be ten times smaller 0.1uF

http://www.picbasic.co.uk/forum/showthread.php?t=19548

Using ST232 and MAX232, I use 1uf elect type for many years now. Never had a heat problem.
For 232epe, some additional care may be required as Richard mentioned.

[Art]

Did you actually bother to look at a data sheet for any of the devices you are talking about?

The inputs of all MAX232 devices (at least authentic ones), are internally pulled low,
so that when left floating, their respective output is high, which is the serial idle state.
The RS232 inputs can also be left disconnected, and live inputs can be plugged in before the chip is powered.

Not bad what two minutes of research can uncover.

Similarly with Pics, you can turn on weak internal pull-ups and leave their inputs floating as well.

[Art]

Old thread I know, but some of the posts mentioned are probably going to be mine,
but when I say MAX232, I actually mean Maxim/Dallas MAX232, which for hobbyists,
are, or at least were, the same price as Pics... zero dollars as free samples from the respective companies.

[E Kizer]

Nice to hear from you Art.

I just wanted to make a small comment.

The RS232 inputs of the MAX232 (pins 8 & 13) are pulled low internally with 5k ohm resistors. In fact any and all chips that have a RS232 input will have a pull down resistor ranging from 3k to 7k in accordance with the RS232 specification on their RS232 inputs.

What I was referring to was the TTL (CMOS) inputs for the RS232 driver. In every drawing on this and other forums the RS232 transceiver chip is drawn with only one input going somewhere, the other is left floating ( note; so using a MCU's weak pull up feature wouldn't work since there is no connection made).

There was a post of people experiencing problems with overheating of their chips. In almost all cases of overheating of any type of TTL/CMOS be it transceiver, Gate, Counter, etc it's because an input was left floating.

The Maxim MAX232 has internal pull UP (400k ohms) resistors on the driver inputs pins 10 & 11. It should be noted that there are many many variations of the RS232 transceiver chip. And not everyone uses the MAX232. Some users might prefer a chip that uses a smaller capacitor for the charge pump. Some just like a cheaper chip.

Personally I use the (cheaper) Texas Instrument TRS202. It has all the same features as the MAX232 except it doesn't have internal pull up resistors on pins 10 & 11. From the TI TRS202 spec sheet; "8.3.2, RS232 Driver, The driver inputs do not have internal pull up resistors. Do not float the driver inputs."

So you are correct in saying the Maxim/TI MAX232 doesn't need external Pull up/down resistors on the driver input. But not every RS232 transceiver chip offers that feature. Yes, read the technical documents. When it comes to working with RS232 transceiver chips research if the TTL inputs have internal pull ups. But don't assume they do.

I leave it in you capable hands to help the users out there that are experiencing overheating with their IC's. There are way too many variations of RS232 transceivers for me to keep up with. Personally I still follow the TTL/CMOS rule that every input must go somewhere ie hi/lo.

Thanks for your comments.

ed

[Art]

To be honest, I’ve never read such a post complaining about heat from any of them, but at the same time, am not here a whole lot these days.

If a chip was overheating while still functioning, oscillation from a floating input is the first thing I’d suspect,
but in modern times, anything other than stock logic is becoming more & more likely to include diode clamping, ESD/transient protection, etc.
Stock logic has become a thing of the past for most people other than me, and your typical PGA that replaces them will usually have internal pull-ups as well.

In modern times, I’d certainly replace rule number one with “Study the data sheet”.

[Dave]

If the parts you are using are original MAX232's then first I would look for the capacitor polarity's being correct. The original MAX232 devices required a 1 uF. cap for each of the up and down converters.