max232epe gets hot , real hot

[richard]

has anybody got any clues why my max232epe's get hot and draw excess current nearly 400ma but only sometimes at other times they are fine . several power ups are sometimes required before the chips will run properly.
there are no floating inputs and there is no external load on the +/-`10v supplies . I have tried tant / electro and npo caps they all have the same issue. if I replace the epe version with a 232n all is good problem gone (except you can't buy them any more). the max232epe's appear to be genuine Maxim devices and were supplied by a reputable supplier . I know the epe's have no internal pullups on the cmos / ttl inputs but none of these are left floating (once the pic sets the tris regs anyway). I'm mystified

[Archangel]

If it's getting hot, obviously that power's going somewhere, I'm guessing you have an input / output state conflict between it and whatever is connected to it. I suggest some current limiting resistors and trace out the logic to resolve this issue, or it could just be fried

[richard]

If it's getting hot, obviously that power's going somewhere, I'm guessing you have an input / output state conflict between it and whatever is connected to it. I suggest some current limiting resistors and trace out the logic to resolve this issue, or it could just be fried

that's not the case a max232n in the same socket works fine , a max232epe works sometimes

[Archangel]

that's not the case a max232n in the same socket works fine , a max232epe works sometimes

. . . and you bought these from a reputable source?

[HenrikOlsson]

Hi Richard,
I haven't looked into the difference between N and EPE but are they both supposed to have the same value capacitors as you're using? I know some needs 1uF while others need 0.1uF for example.
Another possibility is bad chips, where did you get them? EBAY?

/Henrik.

[richard]

Another possibility is bad chips, where did you get them? EBAY?

chips came from three separate supliers (farnell aka element14 ,rs components and a local supplier) they all look identical so I can't tell which is which any longer ,
the rs232c inputs are floating and the rs232 outputs are not connected to anything for the test process .according to the data sheet the cmos/ttl level inputs have no internal pullup and need to be not left floating , the rs232 inputs have internal pull down resistors so I'm assuming these can be left o/c (surely). I got a bunch of soic (smd) pack ones on ebay too but these have not been tested yet.
I have only tried 1uF caps but three different types. I'm at a loss to understand why they sometimes fail to work but after a few power on/off attempts they work (for days ) until the next power up

[HenrikOlsson]

Hi,
Well, with chips from three different, well known, suppliers fake/bad chips probably isn't the cause.
Thinking out loud:
What does the supply voltage look like, ristime etc? Perhaps the internal chargepumps doesn't start up properly or something.

/Henrik.

[richard]

that seems to be the issue, the slightest waver in vcc as the charge pumps fire up stalls the osc and then the excess current causes the ps to hiccup then the chip just sits there getting hot (it only takes a millisec ) . I can get it to fail at every switch on now . I will try moving the bypass cap closer to the vcc pin . these epe's are touchy

[richard]

bypass did not help, could it be contact bounce on the power switch ? . I get the same result using the 3amp regulated bench supply

[Ioannis]

I would mark the chips, say by a number, and see if the same chip is always the one to heat up. If this is true, then you know what to do...

Ioannis

[richard]

I have tried 5 epe with the same result all are subject to failure to operate ( none have been destroyed yet) and 5 of the 232n variety they never fail . if I leave the sw on and turn the bench supply on/off the chips always work ,if the sw is used to turn the board on then it may or may not work. increasing the ps smoothing cap made things worse . I have tried driving inputs high and low at power up nothing seems to satisfy these beastly chips

[Dave]

I personally would also check the value of the caps I was using as well as the polarity as connected to the MAX-232. I have used these level converters for over 20 years and have never had a problem with them. Although now that I remember, I did have one that did not have an very robust negative supply.

[richard]

I was just looking at a schematic for a microchip usb/serial debugger header and noticed they only use 0.1uF caps on pins 2 and 6 (the +- 10v rails) in lieu of the recommended 1.0uF or bigger by the maxim datasheet. tried that, it has improved the charge pump startup to almost 100% .
fingers crossed

ps the max232n seem to switch the caps @19khz the max232epe @640khz they are quite different animals

[Ioannis]

Seems logical, higher frequency, lower capacity.

I believe this was the problem.

Ioannis