Author Topic: Buffering - optimizing for simplicity vs. cost  (Read 3497 times)

0 Members and 1 Guest are viewing this topic.

Offline TimeTopic starter

  • Frequent Contributor
  • **
  • Posts: 725
  • Country: us
Buffering - optimizing for simplicity vs. cost
« on: August 02, 2011, 03:29:38 pm »
I was hoping to tap into the vast knowledge of the EEVBlog forum community for help regarding simple buffering circuitry.

I am wanting to properly buffer some signals that control a driver that controls a very large set of IGBTs.

http://www.igbt-driver.com/fileadmin/Public/PDF/Products/ENG/SCALE-2/Cores/2SC0435T/2BB0435T/2BB0435T_Manual.pdf

Page 11 shows what I am trying to implement.

I understand and know how to do this easily with operational amplifiers in voltage follower configuration and inverting amplifier configurations but I don't know how to choose the best modern day amplifier that best fits this application. 

Maybe I don't want to use a typical op-amp - maybe there are simple devices out there that do these things in a nice neat little package that consists of input, output, and supply rails??

Considerations:
  • Money is not a major issue as long as its sensible.  I don't mind paying 3 dollars for all in one parts as opposed to 20 cents for a solution that uses more components.
  • The non-inverting buffers take a signal from a twisted pair t-line and put their signals into the driver board which has an input impedance of about 3.5-4.6 kOhm.
  • The inverting buffers will drive a signal onto a twisted pair t-line maybe 30 feet long into a high impedance input of a DAQ.  These status pins on the driver board just produce a low impedance upon being triggered by a status fault in the IGBT driver board.  They are normally high impedance, so hence the inverting configuration.  Upon fault I expect about 90 ms of recovery time so it will essentially make the inverting buffer (is this even a buffer anymore?) produce a 90 ms pulse.
  • The control PWMs for the non-inverting buffers are just going to be single 1-2s long pulses every minute or so.  So no real frequency.  Rise times are not a major issue since the inputs are schmitt-triggered (I am assuming rise time characteristics are negligible without really knowing for sure, feel free to correct me).
  • My power supply voltage is 15V.  This allows me to eliminate D1 and D2, I believe.  If I am mistaken on this maybe someone will correct me.

If there is a detail I left out please ask.

I just need a solution that is simple and quick.  Something to have the probability of producing the least amount of trouble shooting blips in the implementation process so things can get up and running without too much time spent here.  If anyone can help or guide me in finding the simplest solution, as opposed to cheapest, it would be much appreciated.
-Time
 

Online Conrad Hoffman

  • Super Contributor
  • ***
  • Posts: 1981
  • Country: us
    • The Messy Basement
Re: Buffering - optimizing for simplicity vs. cost
« Reply #1 on: August 02, 2011, 06:37:11 pm »
Have a look at the Analog Devices SSM-2142 and companion part numbers. Might do what you need.
 

Offline TimeTopic starter

  • Frequent Contributor
  • **
  • Posts: 725
  • Country: us
Re: Buffering - optimizing for simplicity vs. cost
« Reply #2 on: August 02, 2011, 08:50:09 pm »
Ok, thanks.  The solution doesn't necessarily have to be a single chip type thing but it would be nice.

I just want everyones opinion.
-Time
 

Offline TimeTopic starter

  • Frequent Contributor
  • **
  • Posts: 725
  • Country: us
Re: Buffering - optimizing for simplicity vs. cost
« Reply #3 on: August 04, 2011, 02:10:08 pm »
So that response shed some light on what I needed.  I am going to use line drivers.

For the non inverting buffer I was planning to use: http://www.onsemi.com/pub_link/Collateral/MC14049B-D.PDF

My next question is: Should I put resistors to ground on all the inputs parallel to the signal line?
-Time
 

Offline Bloch

  • Supporter
  • ****
  • Posts: 456
  • Country: dk
Re: Buffering - optimizing for simplicity vs. cost
« Reply #4 on: August 04, 2011, 03:17:58 pm »
So that response shed some light on what I needed.  I am going to use line drivers.

I must miss some thing but it will help if you gave some info about input / output

Quote
For the non inverting buffer I was planning to use: MC14049B

On page 11 is "just" a level shifter / buffer. It don't have to be a inverter if you make software / logic.

Quote
My next question is: Should I put resistors to ground on all the inputs parallel to the signal line?

Yes. I cant see there the "wire" is going and they recommend it so why not.
 

Offline Bloch

  • Supporter
  • ****
  • Posts: 456
  • Country: dk
Re: Buffering - optimizing for simplicity vs. cost
« Reply #5 on: August 04, 2011, 03:34:01 pm »
Hi Time

Now i have read it all again.

I think the MC14049B is ok for the job.

You will only need one.

If you go the rute off have "true" hardware signals then just use one off the spar buffers. And simple invert it one more time. They are so fast i don't think it will be a problem with the extra "delay"
 

Offline TimeTopic starter

  • Frequent Contributor
  • **
  • Posts: 725
  • Country: us
Re: Buffering - optimizing for simplicity vs. cost
« Reply #6 on: August 04, 2011, 06:09:55 pm »
That is what I was going to use for the non-inverting portion.  I was going to use http://www.nxp.com/documents/data_sheet/HEF40240B.pdf for the inverting need.  I have 8 lines that need inverting and 4 lines that are need non inverting.
-Time
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf