Go Down

Topic: Using the MAX485 (Read 2323 times) previous topic - next topic

pgmartin

Jun 06, 2011, 03:09 am Last Edit: Jun 06, 2011, 02:15 pm by pgmartin Reason: 1
I´m using the MAX485CPA in a project, connecting it to 5V Vcc directly. This is way under the maximum rating of 12V from the datasheet. They work fine, but the ones working as drivers/emitters are warm to the touch. Does anyone have experience with them? Is this normal?
Should I use resistors for the power suply and the DE and RE pins?
Regards

pwillard

I think at this point it would be good to see your design first.

pgmartin

This is a basic blueprint of my design (sorry, but I'm not the best using fritzing).
As you can see, the DI pins ar connected directly to digital outputs, the DE and RE ar both high, connected to VCC, and the resistors are 120 ohm terminators.

tjbaudio

At first glance your circuit looks the same as mine.  I am having the same problem and have contacted Maxim for help.  Have not gotten down to a solution yet.   I was concerned I got some counterfeit chips.  Where did yours come from?  The ones I got from a reseller are the ones I have the most problem with.  They look a little different from the ones I got direct from Maxim.  The ones from Maxim draw about 1/2 what the suspect chips do.

I am finding that with out the termination resistor (120 ohm) the chips draw next to nothing and stays cool.  However adding a 120 ohm resistor causes the current to jump to 50mA on the suspect chip and over 100mA when driving 60 ohm (both ends of the cable terminated.)  At those current draws the chip gets warm and hot.

I also need to find a solution to this.  Even the factory fresh chips are drawing more than they should but not as bad.  That was today's project to figure out.

avenue33

Hi!

At what bauds rate are you working?

tjbaudio

In my case zero.  In my project the displays are latching so no signal is needed most of the time.  I am beginning to wonder if that is part of the problem.  In other words does the chip NEED signal if it is in transmit mode? If that is the case I will have to add a latch pin and turn the transmit mode on and off.

pgmartin

#6
Jun 06, 2011, 06:15 pm Last Edit: Jun 06, 2011, 06:21 pm by pgmartin Reason: 1
I bought my chips in ebay, so they might be fakes.
Maybe I should try disconecting the termination resistors, but this doesn't look like a neat solution.
I'm also using CAT5e cable that it's almost 200 feet long.
So long they work, and the thing is that only the ones acting as drivers get hot, the ones receiving are cool.

I'm sending shiftout signals with a 3 microsecond delay, so I asume something like 333KBps. But I also use one as a latch, so the tranfer rate is way slower, and all of them get equally hot.

pgmartin

I've found this article http://lms.uni-mb.si/~meolic/students/mpk/designingRS485.pdf

In figure 1 I see that the multiple notes share a ground connection via 100 ohm resistors. I didn't do anithing like that, but I still have a pare twisted pair available. Should I try it?

Ther are also Biasing resistors. What is their role?

tjbaudio

That is a good writeup.  Read it and the biasing resistors are fully explained.  In my case I think I am going to go with termination only on the receive end and no biasing.

pgmartin


That is a good writeup.  Read it and the biasing resistors are fully explained.  In my case I think I am going to go with termination only on the receive end and no biasing.


Yeap. I read it after posting the last message an now I understand a lot more.

Tell me if omitting the terminators on the emitter side helped.

tjbaudio

I know the system works with out terminators.  The max distance drops to some thing like 100 ft.  Tho depending on how sensitive the receiver is to noise I may not have a problem even further out.  Or it could be a diagnostic nightmare as things drift in and out of tolerance.

Is your project a large one or just a one off?  I ended up getting some PCBs printed for mine.

pgmartin


Is your project a large one or just a one off?  I ended up getting some PCBs printed for mine.


Mine is just a prototype. Still no custom made PCB.

Yesterday I came to a different approach: I added a much larger heatsink to the 7805 powering the sistem. And suddenly the temperature on the 485s dropped. Looks like heat was traveling throug the cables and traces  :smiley-sweat:
Now I can feel that the hotter one is the one transmitting more data. Next experiment will be trying to lower the data rate.

pgmartin

OK, now they look that they are takeing heat again. But this is a very subjective measurement, since I´m just using my fingertip-thermometer  XD
Anyway, I´ve noticed that me and tjbaudio are wiring the DE and RE pins directly to VCC. No resistor, no disengaging of it ever. Well, I just disconected this, and the communication keeps working! Being this no solution, I kept this connected, just in case.
But, I measured the current drain of the 3 MAX485 together and it totals...160mA!!!
So I´m in the same current range as tjbaudio. Is that normal? I wasn´t able to find, or understand, any tipical curent draw while transmitting in the datasheet.
I did try to lower de transfer rate by increasing the pause between bits sent and nothing changed.

Now i found that part in the datasheet:
Continuous Power Dissipation (TA = +70°C)
8-Pin Plastic DIP (derate 9.09mW/°C above +70°C) ....727mW

Does this mean that this IC can dissipate up to 727mW. In that case, I dont have to worry and just let them be hot  ]:D

Go Up