Author Topic: Bias current sources  (Read 2088 times)

0 Members and 1 Guest are viewing this topic.

Offline aheidTopic starter

  • Regular Contributor
  • *
  • Posts: 245
  • Country: no
Bias current sources
« on: January 02, 2019, 10:13:23 pm »
The attached schematic was pulled from an opamp datasheet. It shows a few bias current sources and sinks.

If I were to implement the circuit as shown, how could I implement those current sources and sinks? Not that I am going to implement the circuit, I'm just curious how this is done in practice.

I've read about the Widlar current source and Wilson current source, would either be a viable option? If so, for a sink, would I switch the NPNs for PNPs and swap the rails or?

 

Offline aheidTopic starter

  • Regular Contributor
  • *
  • Posts: 245
  • Country: no
Re: Bias current sources
« Reply #1 on: January 03, 2019, 01:18:33 am »
Aaand I forgot to add the image of course  |O
 

Offline Wimberleytech

  • Super Contributor
  • ***
  • Posts: 1134
  • Country: us
Re: Bias current sources
« Reply #2 on: January 03, 2019, 02:31:53 am »
The attached schematic was pulled from an opamp datasheet. It shows a few bias current sources and sinks.

If I were to implement the circuit as shown, how could I implement those current sources and sinks? Not that I am going to implement the circuit, I'm just curious how this is done in practice.

I've read about the Widlar current source and Wilson current source, would either be a viable option? If so, for a sink, would I switch the NPNs for PNPs and swap the rails or?

Short answer: Yes, either could be used.  Which,...depends on a number of things...many outlined in the wiki article.

For a current sourced from Vcc, use PNP.  For current sunk to ground, use NPN.
 

Offline aheidTopic starter

  • Regular Contributor
  • *
  • Posts: 245
  • Country: no
Re: Bias current sources
« Reply #3 on: January 06, 2019, 02:59:43 am »
Excellent, thanks!
 

Offline exe

  • Supporter
  • ****
  • Posts: 2647
  • Country: nl
  • self-educated hobbyist
Re: Bias current sources
« Reply #4 on: January 06, 2019, 11:40:18 am »
I like jfet current sources due to their simplicity. Just a jfet and a resistor.

The catch is for each jfet the resistor has to be selected individually.
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 17427
  • Country: us
  • DavidH
Re: Bias current sources
« Reply #5 on: January 06, 2019, 01:23:39 pm »
It is not shown in the simplified schematic but those are current mirrors with multiple outputs and there is a separate series circuit between the top mirror and the bottom mirror which generates a temperature compensated constant current so only one current reference is required.
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 20357
  • Country: gb
  • 0999
Re: Bias current sources
« Reply #6 on: January 06, 2019, 07:25:49 pm »
It is not shown in the simplified schematic but those are current mirrors with multiple outputs and there is a separate series circuit between the top mirror and the bottom mirror which generates a temperature compensated constant current so only one current reference is required.
Yes, a J-FET has been used in the past but no doubt modern ICs use something a bit more sophisticated.
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 17427
  • Country: us
  • DavidH
Re: Bias current sources
« Reply #7 on: January 06, 2019, 11:51:30 pm »
It is not shown in the simplified schematic but those are current mirrors with multiple outputs and there is a separate series circuit between the top mirror and the bottom mirror which generates a temperature compensated constant current so only one current reference is required.

Yes, a J-FET has been used in the past but no doubt modern ICs use something a bit more sophisticated.

The 741 just used a resistor but the 301 used a bipolar constant current generator and the 301A used a temperature compensated constant current generator with a JFET which is actually a pinch resistor.  Modern parts including the 324/358 are very much like the 301A so not much has changed.  Linear Technology was good about publishing full schematics for their early parts which are modern in every way so check them out for examples.

If your application has fixed supply voltages, then just use a resistor.

 

Offline aheidTopic starter

  • Regular Contributor
  • *
  • Posts: 245
  • Country: no
Re: Bias current sources
« Reply #8 on: January 07, 2019, 11:30:00 pm »
Thanks a lot for the additional input. I'll go study some LT datasheets then! :)
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 17427
  • Country: us
  • DavidH
Re: Bias current sources
« Reply #9 on: January 08, 2019, 02:59:52 am »
I checked a few more old schematics and every JFET in the bias control circuit is actually a three lead space saving pinch resistor.  Pinch resistors have a structure sort of like a JFET which contributes to very high resistance per square.  I guess they might be considered a very leaky JFET with an excessively long channel.
 
The following users thanked this post: aheid


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf