• State TI Thinks Resolved
  • Locked Locked
  • Replies 6 replies
  • Answers 1 answer
  • Subscribers 33 subscribers
  • Views 1017 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

MAX3232: Serial communication error

user4652530
Prodigy 50 points
Part Number: MAX3232


We are using TI's RS232C transceiver MAX3232IPWR.

There are serial communication error with some of devices.

We find a problem with charge pump circuit inside the device.

Almost devices' charge pump circuits are running about 700Hz without comunication signal.

And when comunicate, the frequency inclrease.

Of course, serial comunication are fine.

But, some of devices' charge pump circuit are running about 38.7Hz without comunication signal.

And alco when comunicate, the charge pump circuit seems stop working until V+/V- level become some certain level about +/-3V.

Our circuit : 

Capacitors' value are same as on the datasheet : 

Measurement environment for waveform 1 and 4 : 

Waveforms : 

1 and 2 : 

RS232_Tx's voltage become low when error, because of V+ is too low when charge pump circut just start working.

So, PC recieved oposite sign signal then serial comunication failure happen.

3 :  V+/V- without comnunication signal : 

Other board without problem : 

4 and 5 : 

Question : 

1. Could you tell me the factors that cause the oscillation frequency to change like this?

2. Also, could you tell me the expected value of the oscillation frequency when the capacitor constant is set to the recommended value in the datasheet?

3. How can we solve this problem?

  • 0
    TI__Guru 53062 points

    Hi,

    So before I take a deeper look into this can you take a picture of the top of the IC that has a strange behavior and post it - I want to verify the part markings to see if this is a legitimate device as I think it may be counterfeit. Ultimately there is nothing concerning with your circuit if the values are as you've indicated - and in situations like this there could be a rather high chance of it being counterfeit - so I want to verify that first.

    Please let me know.

    Best,

    Parker Dodson

  • 0 in reply to Parker Dodson
    Prodigy 50 points

    Hi Parker,

    Thanks for your repy.

    Photo of failure one : 

    It looks dirty since we've manually soldared for figure out problem.

    But marking are the same as good one as next photo.

    MB3232I

    29K G4

    AECQ

    Photo of good one : 

    And I've asked our purchaser to check route where we buy those.

    They answered said those from "Chip One Stop" part of "Arrow".

    Regards,

    Jun

  • 0 in reply to user4652530
    Prodigy 50 points

    We've asked to Chip One Stop and find out.

    LOT：5000376MY2

  • 0 in reply to user4652530
    TI__Guru 53062 points

    Hi Jun,

    Thank you so much for the information - these part markings look fine; I think they are valid devices. 

    For your direct questions ( I have also included some questions/suggestions in bold under each answer) :

    1. Frequency changes in the charge pump have to do with loading and the capacitive feedback network - or if the load is too low impedance (< 3K) it could have issues maintaining its voltage - but since this issue is that the charge pump seems to be slow to react I don't think its a loading issue - however it would be good to know if there is any strange loading - if its just another RS-232 device I don't think its a loading issue. 

    q1) Has customer tried replacing the caps on the problematic board and retesting? 

    q2) Is it possible to test the 5V configuration (C1 = 47nF; C2=C3=C4 = 330nF) of capacitors to see if there is improvement - this could improve the response time of the charge pump. 

    q3) Can you have customer measure VCC on the problematic board as well - I want to ensure that the main power supply is looking ok. 

    2.  We don't characterize this value - the charge pump is pretty fast but the ripples you see is based off of the feedback network. Essentially we don't have a guaranteed value for this

    3. It depends on the cause - I have two suggestions right now (same as under point 1) besides ensuring that the load is > 3k (RS-232 transceivers have a minimum input impedance of 3k) 

    Suggestion 1: Replace caps on bad board and retest - a bad capacitor could result in this issue seen 

    Suggestion 2: If replacing caps didn't fix issue - try the 5V setting instead of the 3.3V/5V as its response time should be faster due to the lower capacitance and still will work at its 5V value. 

    Please let me know if these tests can be performed and if yes the results would be very helpful!

    Best,

    Parker Dodson 

  • 0 in reply to Parker Dodson
    Prodigy 50 points

    Hi Parker,

    Thanks for your reply.

    q1, q2 ), Suggestion 1 and 2:

    Yes, we've replaced caps and we've tried some other combination as below, too;

    1) C1 = 47nF; C2=C3=C4 = 220nF

    2) C1=C2=C3=C4=1uF

    Both results are no improvement.

    For 2), it will increase time before failure, but still same situation as original failure.

    We may try some other conbination as you suggested.

    But, basically we think this failure situation is nothing about capacitors issue.

    q3) 

    We've checked VCC for both of failure one and good one, both of them no problem.

    We've found similar situation on this "E2E Deisign support" as below.

    https://e2e.ti.com/support/interface-group/interface/f/interface-forum/669470/max3232-asking-for-the-max3232-abnormal-waveform
    https://e2e.ti.com/support/interface-group/interface/f/interface-forum/660979/max3232-max3232-output-waveform-issue

    Both problem seems never solved.

    So, I guess MAX3232 has design failure about on chip charge pump circuit and some very low percentage devices may make actual failure.

    Regards,

    Jun

  • 0 in reply to user4652530
    TI__Guru 53062 points

    Hi Jun,

    Alright - so since switching capacitors didn't work - I don't think its a loading issue or you'd never have a proper voltage - and it seems to be slow not unable to get to the right voltage. I would highly advise looking into a device return because it could be a faulty charge pump - while this could be the result of some electrical overstress - its not very likely. Please see page here on returns: https://www.ti.com/support-quality/additional-information/customer-returns.html - Arrow is an authorized distributor in most regions (except Japan) so I'd try to return the bad unit to them so we can get it back and see if there is an issue that we are missing.  I will say the other threads you linked are pretty old and one of them seems to be a firmware issue not hardware - so if there is a problem its rare and on the first post I am not sure if its unsolved as no response doesn't indicate no solution - it doesn't indicate a successful solution either though - so its more of a gray area. 

    The biggest issue that I see is that the charge pump fails to regulate an output voltage under a standard load and due to the variability of the charge pumps output voltage - this fluctuation can cause communication failure (bit loss/bit error). I'd also explain the loading situation to them. The more information you provide on the failure the more information for the return to hopefully pin down what the issue is.

    Please let me know if you have any other questions! 

    Best,

    Parker Dodson