1. Home
  2. Electronics Design
  3. switching current source

switching current source

Hi everybody I need to make a switching current source similar to this:

formatting link
The trick is that I need a PWM controller that will "fire first shot" within 1us from control signal transition. Can anybody recommend a chip that does it? Manufacturers do not usually specify it, the controller I am using in the design takes up to 40us to "wake up". Another option: suppose the PWM controller needs up to N seconds to "wake" up. I will be happy if it can wait MORE than N seconds for the current to build-up.

Thanks Michael

Explanation. I need to design a current source to drive an arc triggered by the spark. Trigger signal is ~1.5us long (spark ignites somewhere in the middle). The current control switch must be ON at the moment as this is the time when the current can flow. Thanks Michael

Reply to
Michael
Loading thread data ...

Your "fire first shot within 1us" solution won't work, because it takes longer than 1.5us to establish the inductor current, and even longer than that to establish the PWM's control-loop current setting. My approach in scenes like this, where a fast current availability is required, is to create a standing current into a dummy load, and then use a steering diode or MOSFET switch to direct the steady current into the desired load at the right point in time.

--
 Thanks,
    - Win
Reply to
Winfield Hill

Try one of these....

formatting link

You'll need LTspice

It's a nasty self oscillating thing that sets the inductor current with a mosfet to dump the current into the load...... like what Win is suggesting.

DNA

Reply to
Genome

Thanks!! This approach will probably not work (for me).

  1. The circuit is handheld, battery powered, and very small (1 or 2 extra watts dissipation might be a problem)
  2. Time needed for "conventional" PWM controller to wake up is unknown (e.g.: LM3478 I am using, needs anywhere between 10 and 40us - I took a quick look at one board, it is not statistics). The extra power dumped into the dummy load (or short) might require more of a DC-DC converter, battery.....
  3. The voltage across the load (spark) is not well_known/stabe (~
50-100V - ???)...

I have a circuit that works but is too sensitive to noise (and I have plenty of it on ~2.5" square board). I came up with the solution yesterday (I hope it will work) - buck converter with hysteretic control. Switch and and diode current signals are added up. The feedback comparator has enough hysteresis to keep switching frequency at ~100kHz (the way it is done now). Oh well... :o)

Genome wrote:

Reply to
Michael

Thanks!! This approach will probably not work (for me).

  1. The circuit is handheld, battery powered, and very small (1 or 2 extra watts dissipation might be a problem)
  2. Time needed for "conventional" PWM controller to wake up is unknown (e.g.: LM3478 I am using, needs anywhere between 10 and 40us - I took a quick look at one board, it is not statistics). The extra power dumped into the dummy load (or short) might require more of a DC-DC converter, battery.....
  3. The voltage across the load (spark) is not well_known/stabe (~
50-100V - ???)...

I have a circuit that works but is too sensitive to noise (and I have plenty of it on ~2.5" square board). I came up with the solution yesterday (I hope it will work) - buck converter with hysteretic control. Switch and and diode current signals are added up. The feedback comparator has enough hysteresis to keep switching frequency at ~100kHz (the way it is done now). Oh well... :o)

Genome wrote:

Reply to
Michael

Hallo Michael,

What do you mean by "wake up"? I have used the LM3478 in designs and it's very fast. The thing that is (on purpose) slow is the soft start which can take a few msec. To shut it down takes 30usec AFAIR.

--
Regards, Joerg

http://www.analogconsultants.com
Reply to
Joerg

Hi Joerg

The time interval between pulling frequency setting resistor to ground (pin7) and "Drive" pin (#6) going up first time varies between 50 and

290us (measured for one sample only). Thanks Michael

Joerg wrote:

Reply to
Michael

Hello Michael,

It's been too long ago that I did that design but AFAIR I did not see such a delay (I also used that dormant mode function). It might be worth contacting National about it, they are very responsive. There is also a forum on their site and typically you'll find an answer from them same day.

Do you really need the

Reply to
Joerg

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.