Author Topic: Current Sink with CFBOA  (Read 1366 times)

0 Members and 1 Guest are viewing this topic.

Offline stcosoTopic starter

  • Regular Contributor
  • *
  • Posts: 107
  • Country: it
Current Sink with CFBOA
« on: August 27, 2019, 08:37:48 pm »
Hi everyone!

I'm trying to build a litttle amp to drive (fast) some power LEDs from a function generator with square wave.


I've read a little about using VFB op amps to build current sink for this purpose and their compensation for capacitive loads. I haven't found much about using CFB op amps to do this. I've attached a schematic.



The circuit attached works good enough but i've noticed that varying the supply voltage for the LEDs (in the schematic i've only reresented one but there are 5 of them in series with a total Vf=17V and maximum current of 600mA) from 20 to 24 volts may induce some oscillation in the output waveform.  I imagine that this is caused by the variation of capacitance in the FET that i'm using.

In VFB would be "easy" to compensate for the FET capacitance... but in CFB? I'm using a ths3491 and a pd55015-e.

With this circuit i'm obtaining 25ns rise/fall time (measuring the current at the shunt resistance)... can this be improved in your opinion?


Thanks in advance for every comment.

 

Offline capt bullshot

  • Super Contributor
  • ***
  • Posts: 3033
  • Country: de
    • Mostly useless stuff, but nice to have: wunderkis.de
Re: Current Sink with CFBOA
« Reply #1 on: August 27, 2019, 09:22:50 pm »
According to this article:
https://www.analog.com/en/analog-dialogue/articles/compensating-current-feedback-amplifiers.html
compensation should work similar to VFB amps.
Edit: forget it, that's something else - and me posting too fast  ;)

Anyway, 25ns already is pretty fast, don't see an obvious point to improve that, but didn't go into details ...
« Last Edit: August 27, 2019, 09:34:20 pm by capt bullshot »
Safety devices hinder evolution
 
The following users thanked this post: stcoso

Offline stcosoTopic starter

  • Regular Contributor
  • *
  • Posts: 107
  • Country: it
Re: Current Sink with CFBOA
« Reply #2 on: August 27, 2019, 09:30:16 pm »
Actually i've read that page some days ago  ;D. Interesting insights on how this amplifier works...
 

Offline capt bullshot

  • Super Contributor
  • ***
  • Posts: 3033
  • Country: de
    • Mostly useless stuff, but nice to have: wunderkis.de
Re: Current Sink with CFBOA
« Reply #3 on: August 27, 2019, 09:34:59 pm »
Compensation should work by varying the series resistance in the "-" input (R2).
Safety devices hinder evolution
 

Online wraper

  • Supporter
  • ****
  • Posts: 17952
  • Country: lv
Re: Current Sink with CFBOA
« Reply #4 on: August 27, 2019, 09:44:50 pm »
With this circuit i'm obtaining 25ns rise/fall time (measuring the current at the shunt resistance)... can this be improved in your opinion?
Are you trying to make radio transmitter?
 

Online wraper

  • Supporter
  • ****
  • Posts: 17952
  • Country: lv
Re: Current Sink with CFBOA
« Reply #5 on: August 27, 2019, 09:53:02 pm »
Driving power LEDs with "THS3491 900-MHz, 500-mA High-Power Output Current Feedback Amplifier" and RF mosfet is just insane.
« Last Edit: August 27, 2019, 09:55:13 pm by wraper »
 

Offline stcosoTopic starter

  • Regular Contributor
  • *
  • Posts: 107
  • Country: it
Re: Current Sink with CFBOA
« Reply #6 on: August 28, 2019, 09:11:52 am »
With this circuit i'm obtaining 25ns rise/fall time (measuring the current at the shunt resistance)... can this be improved in your opinion?
Are you trying to make radio transmitter?



Actually no. It's for an experiment... I'm not planning to light my kitchen with sub nanoseconds LEDs pulse if that's is your concern.  :)
 

Online wraper

  • Supporter
  • ****
  • Posts: 17952
  • Country: lv
Re: Current Sink with CFBOA
« Reply #7 on: August 28, 2019, 03:24:48 pm »
It's unlikely that you could light up usual power LED this fast without some insane driver. The fact there is current, does not mean that voltage on LED has actually risen to lit it. Don't forget that LED has capacitance. I just took some random 1W LED and it has 1.2nF capacitance. As you also mentioned fall time, it becomes obvious that even if rise time is good, LED won't go off this fast with current circuit as it contains charge.
« Last Edit: August 28, 2019, 03:34:03 pm by wraper »
 

Offline stcosoTopic starter

  • Regular Contributor
  • *
  • Posts: 107
  • Country: it
Re: Current Sink with CFBOA
« Reply #8 on: August 29, 2019, 09:51:13 am »
The blu trace is the photodiode picking up the emitted light. Yellow is current (at the shunt).
It's 45ns on the photodiode...  I posted in "beginners" because I'm asking if I can do better bad eventually how to do better. :)
« Last Edit: August 29, 2019, 09:53:26 am by stcoso »
 

Offline MarkF

  • Super Contributor
  • ***
  • Posts: 2764
  • Country: us
Re: Current Sink with CFBOA
« Reply #9 on: August 29, 2019, 05:31:00 pm »
I would suggest adding a small capacitor in the feedback loop (C12 = 10nF)
And, the series gate resistor seems a bit small (change R4 = 100)
I would also add a small capacitor to the input (C13 = 1nF)

This will slow things down some and help with the oscillation.

« Last Edit: August 29, 2019, 06:01:43 pm by MarkF »
 

Offline stcosoTopic starter

  • Regular Contributor
  • *
  • Posts: 107
  • Country: it
Re: Current Sink with CFBOA
« Reply #10 on: August 29, 2019, 06:44:36 pm »
I tried adding a "C12" in the past, to stop oscillation... It made the problem worse  ;D. I'm not an EE but capacitors between output and input of a CFBOA introduces poles in the loop response making things unstable.

Wouldn't C13 form a low pass filter ( with the output resistor of my func gen) ?

Anyway... I've redone my pcb with a little bit more care and seems like the oscillation are gone.


Would still love to see if this LEDs can be driven faster than this ;D
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf