• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Answers 1 answer
  • Subscribers 20 subscribers
  • Views 187 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.

TPUL2G123: Pulse generator IC < 10nS

JLF1000
Prodigy 15 points
Part Number: TPUL2G123

Tool/software:

Hi,

I need to design a Cct that can generate a one shot pulse from a rising edge or longer input pulse. The Cct must allow the Pulse width to be varied from 5ns to 500ms.

I've looked at various discrete solutions such as a BJT Monostable Multivbrator CCt that uses a switch to increase the capacitance and hence increase the pulse width. See LT spice image. The Cct simulates and works across the desired range in LT spice, but when breadboarding the CCt using RF BJT BFU550D the circuit fails to work at pulse widths below the uS range.  

When a step is applied at the input the voltage at the node 'base' should drop from a Vbe (0.8V) to a negative voltage which in turn switches off Q2 and causes the output to rise to the the Rail voltage see screenshot below. In practice the base node is 'stuck' at vbe and hence the Cct isn't working.

After some head scratching I've noticed the rising edge of the input signal is in the us range (I'm using a pretty old signal generator - Black Star Jupiter 2000), If I change the input signal rising edge to say 2us in my sim the Cct also stops working as the base node voltage doesn't fall below the Vbe threshold of Q2.

So the Cct requires a very fast rise time on the input signal to generate <10ns Pulse widths. This means that between matching circuits on the same PCB there may be differences in the pulse widths seen due to the difference in tracking to each pulse generator Cct.

So instead I started to look for an IC that can do this for me. So far I've not found anything that can generate pulse widths in the range I need. The TPUL2G123 can generate pulses from 815ns to 800ms which is great for the upper us and ms range but doesn't get me down to 5ns min pulse width required. This device from Analog DS1040 that can generate 5nS pulses but the range is too small for it to be useful in my application. 

Do TI offer any MMV that works in the ns range with a wide enough range to get to say 815nS. This way I could use one IC to generate the long pulses 1us - 500ms (TPULG123) and another to generate the short pulses 5ns - 800ns ?

I know the easiest way to do this would be to use a MCU such as STM32 and a HR Timer to generate the short pulses but we don't have any programmers to set this up so it must be done in hardware.

Thanks In advance 

  • 0
    TI__Mastermind 38535 points

    Hi,

    We do not have an MMVs that will work within this range. 

    We do have our TPLD devices, but I will have to see if we can actually get down to this range of pulse. I will update you within the week. 

  • 0 in reply to Albert Xu1
    Prodigy 15 points

    Thanks Albert I look forward to your response.

  • 0 in reply to JLF1000
    TI__Mastermind 38535 points

    Hi,

    I actually talked to the systems engineer that worked on this device. Table below is from section 5.7

     

    These are pulse width values with Cext = 0. I am not entirely sure what Vcc you would be using, but lets say Vcc = 3.3V. A typical output pulse would be around 95ns. 

    Realistically, if you lower Rext to 1k instead, you can get a width that is 10x smaller. Resulting in ~8-10ns Only issue is while this may be achievable, it might vary more than specced (normal variation of 5-10%). 

  • 0 in reply to Albert Xu1
    Prodigy 15 points

    HI Albert,

    Thanks for the quick reply on this. Looking at the datasheet for this part (TPUL2G123) I see that there is a min & max requirement for Rext & Cext.

    The recommended operating table states Rext min = 6.5k 

    .

    The Abs max table states Rext min = 1k

    This may be a suitable solution If I can get it to produce 8-10ns Pulses in practice. I was planning to use VCC = 3.3V, but looking at the table maybe a 5V supply would be better as min Pulse width decreases with higher VCC. 

    As it's a test PCB we can get away with using 1k Rext i.e outside of the recommended operating range although I wouldn't want to do this for a production board.. The 10% variation is ok for this application. I'll order some of these devices and test the theory on the bench

    Thanks again for your help with this.

    JLF

  • 0 in reply to JLF1000
    Prodigy 15 points

    I've just noticed that the IC TPUL2G123 IC is pre-production and there are no samples/stock available on the website. Do any of the other pulse generators have similar performance ? Is there any chance to get samples of this device ?

  • 0 in reply to JLF1000
    TI__Mastermind 38535 points

    Hi JLF,

    1. I checked and went through the process myself and you should be able to order samples. Can you try again? I see inventory of 3000+ currently.

    2. Regarding min/max Rext, compared to other specs such as voltage, there is potential damage to the device if you go outside of the range. However, the min/max value for Rext only pertains to the accuracy of the device in this case. Hence why I note that the variation may be a lot larger than 10% if you go under the recommend 6.5kOhms.