• State Resolved
  • Locked Locked
  • Replies 9 replies
  • Subscribers 18 subscribers
  • Views 1698 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

SN74HC373 safe for 3.3V to 5V level conversion?

Peter Pan
Intellectual 290 points
Other Parts Discussed in Thread: SN74HC373, SN74HCT373

I tend to ramble, so let me ask my question simply, and then provide background after. If the SN74HC373 is powered by 5.0V, will a 3.3 V source from an be MCU reliably considered HIGH level.

The background is this.

I'm working with a  module that uses a TI mcu (CC2511F32) powered from a 3.3 volt regulated source, in a larger circuit that also has regulated 5.0VDC for the main power.  I'm using several of the MCU I/O pins configured as inputs,  to sense button presses (8 buttons present). All these inputs are internally pulled up to 3.3V, and one side of all the buttons is tied to 0V. So now I'd like to add a little LED next to each button that  lights when a button is pressed, and then stays lit. To do this with minimum parts, I'd like to employ a TI  SN74HC373 as follows.  All the "D" inputs of the chip will be tied to the button inputs. The, all the Q outputs will connect to low current LEDs through a resistor, tied to 5V VCC. Finally, I'll use another spare I/O line on the MCU to send a pulse to the LE pin on the IC, whenever I sense a new button hit. The desired result should be that regardless of what button (or combination of buttons) are pressed, the corresponding LEDs light and remain on.

But here's the issue . I'd like to power the 74HC373 with the 5V supply if I can, because the 3.3V regulator would be unacceptably taxed if I used it for this add on. So to do this without additional parts, the 3.3V present on all the pulled up MCU inputs, as well as 3.3V pulse I provide for the LE, will have to reliably be considered "HIGH" by the IC. Unfortunately its unclear from the data sheet whether or not this will work. The data sheet provides "Minimum High Level Input Voltage" of 1.5, 3.15, and 4.2 volts, for VCCs of 2,  4.5, and 6 volts, respectively.  Since 5VDC is such a common voltage, I'm a bit surprised it wasn't included.  At 4.5 V. it looks like 3.15 will work, but what about at 5V? Will my 3.3V be sufficient? Maybe I should add a diode in the VCC to the IC to drop it down to 4.5V?

I realize I'm cutting things close here, but again I want to accomplish the goal with as few parts as possible.

  • 0
    TI__Mastermind 19780 points

    Hi Randy,

    Your solution sounds simple enough employing a D flip flop to latch the LEDs on when the strobe to LE is registered by the IC.

     

    I am not the expert here but from what I can see the Vih for the device appears to be ~70% of VCC which is consistent with other CMOS devices (including the MCU you have).  Applying that to 5-V VCC means you'd need 3.5-V to guarantee reliable high levels.

     

    I'd do as you mentioned and throw a diode in series with the 5-V rail feeding the 74HC373 to bring VCC to the 74HC373 down to ~4.3-V.

     

    Regards,

  • 0
    TI__Mastermind 46480 points

    you will be operting out of spec with the 74HC373, However there is a simple solution. Use the SN74HCT373 instead. Then all will be good.

  • 0 in reply to Chris Cockrill
    Intellectual 290 points

    Thanks! I hadn't noticed that one. It draws a little more standby current, which won't be a problem with the powered 5V supply. However, there is a version of this project/product which operates from a LiPO cell, whose nominal voltage is 3.7V, and also operates my MCU down to about 3.0V.  I'd rather use the same part for both variations if possible. The HC version, from the spec, seemed to be usable down to a VCC of 2V, so even with my diode solution I should still be OK, even with the battery at its lowest usable output.  But why do you say I'd be operating out of spec with the HC version?

  • 0 in reply to Peter Pan
    TI__Mastermind 46480 points

    Ok , with the diode it would be in spec.

    CMOS parts will draw a lot more idle current if the input is not at Vcc or gnd. This might be a concern with a battery operated device. They way around this would be a dual VCC translator.

    You will need to prototype this to see what the Icc current might be under your conditions

    They way around this would be a dual VCC translator. however they dont come in a latch.

  • 0 in reply to Chris Cockrill
    Intellectual 290 points

    Thanks. I'll pay attention to that and be sure to check the current in both situations!  Of course in the battery case, it will be even more "in spec", though I may just make provision to eliminate the diode for that case (maybe replace my 805 footprint diode with a 0 ohm resistor).

    My only issue now is that this is a fairly large chip, and the original PC board won't accommodate it, so I'll have to put it on a smaller "plug in" board with a header. You don't happen to know of a smaller package available for this chip do you? I'm only driving very low power (2mA) LEDs, and these chips are good to 20mA . So even if there were a lower power version in a smaller package, it might help me cram it into the existing PCB! :-)

  • 0 in reply to Chris Cockrill
    Intellectual 290 points

    Uh Oh! I think I'm in trouble! I was getting excessive current drawn, and so I powered everything from 3.3V to see what was going on. This circuit works, in that when the pulse is applied to LE, a LED will light and stay lit for each button pressed. The trouble is, with NO LEDs lit, a mA meter between the 3.3V source and the LED circuit reads a whopping 33 mA !!!   remember the LEDs are all in the off state. Is there any possible reason, aside from a defecive IC, that this could happen?  Here's my schematic...

    http://pixyland.org/elfintechnologies/publicPics/LedInd1.jpg

    Maybe I should post this question as a separate thread?

  • 0 in reply to Peter Pan
    Intellectual 290 points

    OK, disregard my last post. It was a bad chip, almost surely made bad by my poor handling (static or surge) cobbling a circuit together quickly. Hot-air lifting of the chip from my breakout board and substituting another now yields perfect operation with only uA quiescent current draw, even with my originally proposed combination using multiple voltages. Just odd that i was able to fry the first chip in such a way that it worked perfectly, aside from an obvious issue causing excessive current.


    As it turns out now, I may not need this IC though, because a midnight idea came to me that may result in an even lower parts count, ZERO! :-). The I/O lines on the TI mcu I'm using to sense my input buttons can instantly be switched from inputs to outputs. Therefore i may be able to alter the code so that these lines are only inputs during button sensing time, after which they could be turned into output and set to 0V, to drive a low power led. there might be some flickering, but then again that might be unnoticeable. Wish me luck. :-)

  • 0 in reply to Peter Pan
    TI__Genius 9860 points

    Mr. Pan,

    Glad to see you found a solution! Not sure the VOL of your MCU's outputs, but if it causes any issues with driving LEDs, then let us know and we can suggest some peripheral drivers.

    Out of curiosity, what is your end application? It's always interesting to see how our devices are being used.

    -Ryan

  • 0 in reply to C. Ryan Land
    Intellectual 290 points

    Oh its a long sordid tale Ryan. But first, I'm only driving small 2mA LEds, and most of the IO on that mcu are good for 4mA, so I'm good.


    Its a long complidcated story I hope has a good outcome. I'm a musician too, and I've developed this product using the Pololu Wixel, which used that MCU, which is now up to 4 major building blocks. Its purpose in a nutshell is short range remote control over effects racks, effects pedals, and up to 4 audio levels. There is a hand held and a floor pedal version of the transmitter, a main master receiver, and a pedal adapter. In truth its evolved so much over the course of the development,. I couldn't begin to tell you what a long list of features it has. After all, I'm not just a designer, I'm a user! This last addition, to put a LED next to each button as a visual confirmation of a button hit, is just getting down to the eye candy stage. But I sure was glad to come up with a solution that ivolved NO extara parts, NO PCB mods, and a lot less wiring than the 373 would have been.



    The reason it is a long sordid tale is because of an ongoing dispute I have with Pololu. You see, they SAID in their doc that they copied YOUR design, so getting FCC approval shouldn't be a problem. Indeed, if you start with something that has passed, you can leverage that certification and at least cut SOME of the testing cost. Turns out Pololu lied. In fac, the nearest design of their WIXEL to your test product didn't even go through FCC cert at the time they published their manual. It did indeed pass the FCC last fall, and Pololu said my conversations with them have made them decide to revisit the idea of having a test app with thier wixel module approved. But here we are a year later, and nothing has changed. that means I'm going to have to foot the whole bill for testing and certification, a real bummer. My product looks and functions great, but I may not be able to afford to complete this step without getting an investor involved.

    This isn't the start of the thread, but you can read most of my tale of wore on your forum, but starting here and scrolling to the beginning of the thread.

    http://e2e.ti.com/support/wireless_connectivity/f/964/p/359805/1276610.aspx#1276610