help ...to control pulse width of a -15vdc pulse train

[fringe]

Help needed to control the pulse width of a (negative) -15vdc pulse train at 20khz..

Basically, the pulse is as described in the picture attached which is of the pulse itself all details are within picture.



Pulse width is 6 micro-sec, and i want to be able to reduce increase or reduce it
by around 30% of current value

I will not need to vary the frequency only the pw..
I would like to use something very simple... I have been searching everywhere.

However, most designs are base on negative going pulses on a +dc supply and not on a wave form that actually goes positive from a negative -15vdc supply.....

please idea's would be great appreciated.
Ps: The pulse train controls a fet driven NEG Dc chopper circuit (high current)

Thanks to all who take the time to at least read this post.

[Hero999]

I assume you're doing this the old way i.e. using logic rather than a microcontroller which might be easier.

Stretching the pulse is pretty easy, you need a monostable which is falling edge triggered, a Schmitt trigger or similar will work. A CD40106 or 74C14 can be used, see the last page of the following application note.
http://www.fairchildsemi.com/an/AN/AN-140.pdf


Technically shrinking the pulse is impossible unless it's a repeating waveform with a constant pulse width and memory is used, i.e. the length of the pulse needs to be known so the circuit would first have to measure it, then calculate the length of the next pulse, assuming the incoming pulse is the same as the previous one. Although it's probably not necessary if all you want is a varying pulse width from a constant stream of pulses, all you need is a positive edge triggered monostable.

Don't let positive and negative pulses and power supplies confuse you. Assuming all the voltages in your circuit are negative, you can just think of it as using a positive power supply and invert the input and output waveforms i.e. your 20kHz stream of negative 6

[fringe]

Thanks for the reply,

The pulse train (20 kHz negative pulses with respect to gnd) remain constant in frequency period is 50 micro-secs'

The width is automatically adjusted between 6 and 20 usec by the current PWM..Which is (completely epoxy bonded) so no change can be made at that level.

1. Firstly, I need to reduce the minimum pulse width from 6 to around 4.5 usecs to reduce to storage voltage on the chopper capacitors when the unit is in standby mode, this must be done without affecting the freq...

2. Secondly, The unit will return to normal operation and bypass the circuit I am designing when a load is applied to the chopper output..............(This part I have already designed)

I have all the tools at my finger tips.... I did first try to simulate 555 pwm making VCC = VEE (0 volts) and GND = VSS (-15volts)
However, multisim v11 didn

[Kevin Weddle]

If the load is a high impedance FET, then it can be done more easily. If the load is low impedance, you should change the circuit which generates the pulse. Is the load a high impedance?

[Hero999]

I believe from your post you are saying I can connect a 555 VCC = 0vdc and Gnd = -15vdc and all the rest should work as normal.. Is that right?

It will work but you'll need to make some modifications to the 555 circuit:

The input pulse is longer than the output pulse so you'll need to AC couple the input to the source otherwise the output will remain high after the delay has finished.

The output needs to be inverted. Remember the 555 goes high for a period when the input goes low so in this case when the input is 0V the output will be nearly 0V for the period before returning to -15V. Adding a single NPN transistor to the output will get round this problem.

I'd be tempted to go for a logic IC such as the CD4011 rather than the 555 as it will give you both inverting and non-inverting outputs. Note that the CMOS gates do have a high output impedance and any loading will interfere with the operation of the circuit. You could parallel the remaing two gates and use them to buffer the output to supply slightly more current.

[fringe]

Firstly, thank you Kevin and hero for taking the time....

Kevin",If the load is a high impedance FET, then it can be done more easily. If the load is low impedance, you should change the circuit which generates the pulse".

Yes, it does have a high impedance fet- the fet is the main driver for the chopper . Basically it chops the negative side of a generated then rectified dc voltage  , thus producing a high current DC chopped output (40vpp 20khz )

Changing the Pulse drive circuit is not and option on the OEM board ...
I will put up a diagram of the output circuit later today... but work must come first

Hero, as for the CD4011 i will investigate the option possibly build both,

[Hero999]

Whoops I got the VDD and VSS the wrong way round.

VDD = 0V and VSS = -15V, I've edited the schematic.