Hi everyone,

First of all, I'm new on AAC. I'm a student in electronics.
I need to make as final project a high speed flash for a camera. So I thought to take as LED driver the "74121". First I wanted to pick the "555" but this one is too slow... The pulse width needs to be in nanoseconds so I thought this one is perfect. I read if I don't use a capacitor there will be the minimum pulse width (30-35ns), so if I'm using just the minimum R (=1.4k) this can work? (see schematic)
Thanks in advance!

Kind regards

 

Hi everyone,

First of all, I'm new on AAC. I'm a student in electronics.
I need to make as final project a high speed flash for a camera. So I thought to take as LED driver the "74121". First I wanted to pick the "555" but this one is too slow... The pulse width needs to be in nanoseconds so I thought this one is perfect. I read if I don't use a capacitor there will be the minimum pulse width (30-35ns), so if I'm using just the minimum R (=1.4k) this can work? (see schematic)
Thanks in advance!

Kind regards

Looks good. Enables are correct.

 

Most white LEDs take longer than that to wake up. Why so short?

ak

 

Most white LEDs take longer than that to wake up. Why so short?

ak

It will be used in an industrial process to take pictures of paper and to calculate the amount of dust in a paper factory. But I just need to make the flash for the camera I need 800 images/sec so my flash needs to be lower then 1µs.
I will not use usual white LED's because indeed, they are too slow and they are also not powerfull enough to light up the paper. I thought to use on the output a powerMOSFET or IGBT to switch on/off the LED's because the LED's I will use demands a lot of current.

 

The 74121 without the capacitor still has a capacitor, it's the parasitic capacitance of the circuit and wiring.
It's not going to be very stable, in terms of pulse width, as the parasitic capacitance will vary.

It's going to be a bit hard to get clean 50-100 nanosecond switching speeds, especially driving a high current MOSFET.
This is not an easy project.

 

The 74121 without the capacitor still has a capacitor, it's the parasitic capacitance of the circuit and wiring.
It's not going to be very stable, in terms of pulse width, as the parasitic capacitance will vary.

It's going to be a bit hard to get clean 50-100 nanosecond switching speeds, especially driving a high current MOSFET.
This is not an easy project.

The time of my flash needs to be lower then 1µs so 500ns is also good. So if I use a capacitor to get 500ns I need a capacitor of 1µF with R=1.4k. (formula: Tw = 0.33 * R * C = 500ns). Wouldnt be this more stable then? is this a good solution or how do I solve this best?

 

The time of my flash needs to be lower then 1µs so 500ns is also good. So if I use a capacitor to get 500ns I need a capacitor of 1µF with R=1.4k. (formula: Tw = 0.33 * R * C = 500ns). Wouldnt be this more stable then? is this a good solution or how do I solve this best?

Double check your math.

At 800 images/sec why do you need 500ns pulse?

C=1µF with R=1.4k does not give 500ns.

 

Double check your math.

At 800 images/sec why do you need 500ns pulse?

C=1µF with R=1.4k does not give 500ns.

Our teacher just said it needed to be lower then 1µs. I think the 800 images/sec was an example value he gave...
and sorry, I meant C=1nF and R=1.4K...

 

Our teacher just said it needed to be lower then 1µs. I think the 800 images/sec was an example value he gave...
and sorry, I meant C=1nF and R=1.4K...

Go back and discuss this with your teacher.
For 800 images/sec, 500µs pulses should be ok.

 

Your teacher is off by 1000:1. The inverse of 800 is 1.25 ms, so 1.24999 ms is the maximum flash duration. It might be that the intent is to use the flash duration to create the effect of a fast shutter speed, but that should be an explicit requirement. Also, cameras (film, vidicon, plumbicon, CCD, bucket brigade, CMOS, whatever) are integrators. They take time to accumulate photons. With a 1 us illumination time, the sensor noise could be greater than the dust you are trying to measure.

ak

 

If this is a printing press for example the paper may be moving at high speed so it could need the flashe to be much shorter than the time between pictures so the picture is not blurred by the distance the paper moves during the time of the flash. I think this is a case of not being given the FULL story by the OP.

Les.

 

Hey everyone,

Sorry for my late response but I did some research and asked my supervisor about the Images/sec. My flash needs to be that fast because it is fotographing paper on a speed (like @LesJones told). So meanwhile I ordered the IC SN74121 and did some tests. I tested like the image I already posted (just R= 1.4k). I tested with a capacitance of 100pF and resistor of 1.4K (100ns) and with a capacitance of 510pF and resistor of 1.4K (500ns).
So you can see, it's not like what I expected... The smallest output pulse width I get is +-1.5µs (C = 100pF and R= 1.4k). (tested the other schedules also but even more worse results of output pulse width)
The schedule and result can you see in the attachement.
Someone some suggestions to get a more shot output pulse width?
Thanks in advance!

Kind Regards

 

I noticed you are not using a 10:1 probe on the oscilloscope which you should.
At those high speeds the wire and input capacitance without a 10:1 probe will likely severely distort the signal.

You might try this CMOS high speed version of the 121.

Here's another possible solution.

 

I noticed you are not using a 10:1 probe on the oscilloscope which you should.
At those high speeds the wire and input capacitance without a 10:1 probe will likely severely distort the signal.

You might try this CMOS high speed version of the 121.

Here's another possible solution.

thanks for the tip! I made the circuit on a breadbord so I was thinking that maybe that also can influence the solution of my ouput width pulse. I will first try with the 10:1 probe and then look for an another monostable multivibrator. I will notify in this thread if I make some progress
Thanks in advance!

 

Your teacher is off by 1000:1. The inverse of 800 is 1.25 ms, so 1.24999 ms is the maximum flash duration. It might be that the intent is to use the flash duration to create the effect of a fast shutter speed, but that should be an explicit requirement. Also, cameras (film, vidicon, plumbicon, CCD, bucket brigade, CMOS, whatever) are integrators. They take time to accumulate photons. With a 1 us illumination time, the sensor noise could be greater than the dust you are trying to measure.

ak

High speed CCDs exist. We used them 25-years-ago for nano-second laser flash photolysis experiments. We were able to see get several dozen full absorption spectra in a 100 nano-second decay. So about 5 nSec per spectrum captured by the CCD.

Single digit pico-second is available today...
http://www.stanfordcomputeroptics.com/applications/iccd-methods/ultra-high-speed-imaging.html

.

 

I made the circuit on a breadbord so I was thinking that maybe that also can influence the solution of my ouput width pulse.

The certainly can be a big factor.
The breadboard should be closely laid out on a board with a ground plane and good local decoupling of all ICs.

A plug-in breadboard is a no-no.

 

Hi everyone!
I did the test again with a 1:10 probe and it works! I get a pulse of 100ns.
I added a resistor of 222 Ohm because I got first a little oscillation and with that resistor there is no oscillation.
The picture I added at this post is the result of 74121 with a resistor of 222 Ohm on Q.
The next step of my High speed flash is trying to switch all my powerfull LED's. My plan is too override my LED's so they can shine more bright.
I will post the whole scheme which I'm going to make. Any comments on the scheme?
Thanks for all the help!
Kind regards!

 

Hi everyone!
I did the test again with a 1:10 probe and it works! I get a pulse of 100ns.
I added a resistor of 222 Ohm because I got first a little oscillation and with that resistor there is no oscillation.
The picture I added at this post is the result of 74121 with a resistor of 222 Ohm on Q.
The next step of my High speed flash is trying to switch all my powerfull LED's. My plan is too override my LED's so they can shine more bright.
I will post the whole scheme which I'm going to make. Any comments on the scheme?
Thanks for all the help!
Kind regards!
View attachment 122112 View attachment 122111

Operation of irf540 is not specified at 5v gate-to-source. Normally, the gate is turns on with 10 volts or so. You may get fairly high resistance through that MOSFET if you drive the gate with only 5 v.

Other MOSFETS work fine with 5v on the gate. Check Datasheets.

 

Your output stage will go into constant-current limiting at around 8 A. Is this intentional?

Since Q1 and Q2 are emitter followers, they do not speed up the rise/fall times of the 121 output stage. Also, they decrease (by over 10%) the max Vgs available to enhance the FET in a circuit where that parameter already is questionably low. Why are they in there?

If Q1 and Q2 are there to increase the peak current drive to the FET gate capacitance, consider a different FET. The 540 is an old part; reliable as all get out, but newer devices have significantly lower gate capacitance for the same power capability, by as much as 80-90% lower. This might eliminate the driver stage and get you faster edge times.

ak

 

Your output stage will go into constant-current limiting at around 8 A. Is this intentional?

Since Q1 and Q2 are emitter followers, they do not speed up the rise/fall times of the 121 output stage. Also, they decrease (by over 10%) the max Vgs available to enhance the FET in a circuit where that parameter already is questionably low. Why are they in there?

If Q1 and Q2 are there to increase the peak current drive to the FET gate capacitance, consider a different FET. The 540 is an old part; reliable as all get out, but newer devices have significantly lower gate capacitance for the same power capability, by as much as 80-90% lower. This might eliminate the driver stage and get you faster edge times.

ak

I don't really get your answer. Do you mean if I search for an another power MOSFET, I can remove Q1 & Q2?
I used the totem pole with power mosfet circuit from a circuit which worked 2 years ago...