Forums

Precision current source

Started by George Herold September 23, 2015
Hi all, I've using a voltage reference->Resistor-> 
opamp feedback to make a current source.  
Like figure 10 here.
http://www.ti.com/lit/ds/symlink/ref102.pdf
The reference is the REF102, but I'm using an opa134
for the opamp.  (Currents down to 10 nA so I wanted a fet
opamp.)  

Now someone wants to use it where noise might be more
of an issue, so I figured the first thing was to measure the 
noise.  

Well, the dang thing is singing at ~1.8 MHz.  
Not that loud... about 100 uV p-p, but still 
this seems like bad form.  
Any thoughts on how I might squash this?  

TIA 
George H. 
(Then maybe I'll ask about how to filter the noise further..
I was thinking about a cap. multiplier hanging on 
the output)     
On Wednesday, September 23, 2015 at 10:49:18 AM UTC-4, George Herold wrote:
> Hi all, I've using a voltage reference->Resistor-> > opamp feedback to make a current source. > Like figure 10 here. > http://www.ti.com/lit/ds/symlink/ref102.pdf > The reference is the REF102, but I'm using an opa134 > for the opamp. (Currents down to 10 nA so I wanted a fet > opamp.) > > Now someone wants to use it where noise might be more > of an issue, so I figured the first thing was to measure the > noise. > > Well, the dang thing is singing at ~1.8 MHz. > Not that loud... about 100 uV p-p, but still > this seems like bad form. > Any thoughts on how I might squash this? > > TIA > George H. > (Then maybe I'll ask about how to filter the noise further.. > I was thinking about a cap. multiplier hanging on > the output)
Current ranges from 10 nA to 1 mA. There is no noticeable change in the oscillations as the current is changed. (Which kinda bothers me.) George H.
On Wed, 23 Sep 2015 07:49:09 -0700 (PDT), George Herold
<gherold@teachspin.com> wrote:

>Hi all, I've using a voltage reference->Resistor-> >opamp feedback to make a current source. >Like figure 10 here. >http://www.ti.com/lit/ds/symlink/ref102.pdf >The reference is the REF102, but I'm using an opa134 >for the opamp. (Currents down to 10 nA so I wanted a fet >opamp.) > >Now someone wants to use it where noise might be more >of an issue, so I figured the first thing was to measure the >noise. > >Well, the dang thing is singing at ~1.8 MHz. >Not that loud... about 100 uV p-p, but still >this seems like bad form. >Any thoughts on how I might squash this? > >TIA >George H. >(Then maybe I'll ask about how to filter the noise further.. >I was thinking about a cap. multiplier hanging on >the output)
Nasty twist on Howland. Might try capacitor from Vout to Trim. ...Jim Thompson -- | James E.Thompson | mens | | Analog Innovations | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | San Tan Valley, AZ 85142 Skype: skypeanalog | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at http://www.analog-innovations.com | 1962 | I love to cook with wine. Sometimes I even put it in the food.
On Thursday, 24 September 2015 00:49:18 UTC+10, George Herold  wrote:
> Hi all, I've using a voltage reference->Resistor-> > opamp feedback to make a current source. > Like figure 10 here. > http://www.ti.com/lit/ds/symlink/ref102.pdf > The reference is the REF102, but I'm using an opa134 > for the opamp. (Currents down to 10 nA so I wanted a fet > opamp.) > > Now someone wants to use it where noise might be more > of an issue, so I figured the first thing was to measure the > noise. > > Well, the dang thing is singing at ~1.8 MHz. > Not that loud... about 100 uV p-p, but still > this seems like bad form. > Any thoughts on how I might squash this? > > TIA > George H. > (Then maybe I'll ask about how to filter the noise further.. > I was thinking about a cap. multiplier hanging on > the output)
The obvious issue with the REF102 circuit is that there's no capacitor in parallel with R3 - typically one uses a couple of pF - to compensate for the input capacitance to ground at the inverting input (typically another couple of pF). The need for it varies with the bandwidth of the op amp (and the input capacitance) but leave it out can give you the kind of oscillation you mention. The major source of noise in a voltage reference is often the reference voltage generator, which is why the REF102 circuit has 7k resistor which you can use with a capacitor to low-pass filter the noise coming out of the voltage reference. 6.2V zener-based references are essentially avalanche diodes and can be quite noisy at low currents. AOE3 is a full bottle on the subject, and has a couple of very useful pages of advice. -- Bill Sloman, Sydney
On Wed, 23 Sep 2015 07:49:09 -0700 (PDT), George Herold
<gherold@teachspin.com> wrote:

>Hi all, I've using a voltage reference->Resistor-> >opamp feedback to make a current source. >Like figure 10 here. >http://www.ti.com/lit/ds/symlink/ref102.pdf >The reference is the REF102, but I'm using an opa134 >for the opamp. (Currents down to 10 nA so I wanted a fet >opamp.) > >Now someone wants to use it where noise might be more >of an issue, so I figured the first thing was to measure the >noise. > >Well, the dang thing is singing at ~1.8 MHz. >Not that loud... about 100 uV p-p, but still >this seems like bad form. >Any thoughts on how I might squash this? > >TIA >George H. >(Then maybe I'll ask about how to filter the noise further.. >I was thinking about a cap. multiplier hanging on >the output)
This is stable: https://dl.dropboxusercontent.com/u/53724080/Circuits/Current_Sources/Current_Sink.JPG or use a bootstrap source https://dl.dropboxusercontent.com/u/53724080/Circuits/Current_Sources/Isrc_Boot.JPG https://dl.dropboxusercontent.com/u/53724080/Circuits/Current_Sources/Ramp.JPG You could lowpass filter the bootstrap voltage, to reduce noise, but that decreases the current source impedance at higher frequencies. Ultimately the Johnson noise will getcha.
On Wednesday, September 23, 2015 at 11:03:41 AM UTC-4, Jim Thompson wrote:
> On Wed, 23 Sep 2015 07:49:09 -0700 (PDT), George Herold > <gherold@teachspin.com> wrote: > > >Hi all, I've using a voltage reference->Resistor-> > >opamp feedback to make a current source. > >Like figure 10 here. > >http://www.ti.com/lit/ds/symlink/ref102.pdf > >The reference is the REF102, but I'm using an opa134 > >for the opamp. (Currents down to 10 nA so I wanted a fet > >opamp.) > > > >Now someone wants to use it where noise might be more > >of an issue, so I figured the first thing was to measure the > >noise. > > > >Well, the dang thing is singing at ~1.8 MHz. > >Not that loud... about 100 uV p-p, but still > >this seems like bad form. > >Any thoughts on how I might squash this? > > > >TIA > >George H. > >(Then maybe I'll ask about how to filter the noise further.. > >I was thinking about a cap. multiplier hanging on > >the output) > > Nasty twist on Howland. Might try capacitor from Vout to Trim.
That was a good idea, but did nothing. George H.
> > ...Jim Thompson > -- > | James E.Thompson | mens | > | Analog Innovations | et | > | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | > | San Tan Valley, AZ 85142 Skype: skypeanalog | | > | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | > | E-mail Icon at http://www.analog-innovations.com | 1962 | > > I love to cook with wine. Sometimes I even put it in the food.
On Wednesday, September 23, 2015 at 11:17:46 AM UTC-4, Bill Sloman wrote:
> On Thursday, 24 September 2015 00:49:18 UTC+10, George Herold wrote: > > Hi all, I've using a voltage reference->Resistor-> > > opamp feedback to make a current source. > > Like figure 10 here. > > http://www.ti.com/lit/ds/symlink/ref102.pdf > > The reference is the REF102, but I'm using an opa134 > > for the opamp. (Currents down to 10 nA so I wanted a fet > > opamp.) > > > > Now someone wants to use it where noise might be more > > of an issue, so I figured the first thing was to measure the > > noise. > > > > Well, the dang thing is singing at ~1.8 MHz. > > Not that loud... about 100 uV p-p, but still > > this seems like bad form. > > Any thoughts on how I might squash this? > > > > TIA > > George H. > > (Then maybe I'll ask about how to filter the noise further.. > > I was thinking about a cap. multiplier hanging on > > the output) > > The obvious issue with the REF102 circuit is that there's no capacitor in parallel with R3 - typically one uses a couple of pF - to compensate for the input capacitance to ground at the inverting input (typically another couple of pF). > > The need for it varies with the bandwidth of the op amp (and the input capacitance) but leave it out can give you the kind of oscillation you mention. > > The major source of noise in a voltage reference is often the reference voltage generator, which is why the REF102 circuit has 7k resistor which you can use with a capacitor to low-pass filter the noise coming out of the voltage reference. > > 6.2V zener-based references are essentially avalanche diodes and can be quite noisy at low currents. > > AOE3 is a full bottle on the subject, and has a couple of very useful pages of advice. > > -- > Bill Sloman, Sydney
Thanks Bill, I think this was Jim T's suggestion for some C from out to the trim. George H.
On Wednesday, September 23, 2015 at 11:35:24 AM UTC-4, John Larkin wrote:
> On Wed, 23 Sep 2015 07:49:09 -0700 (PDT), George Herold > <gherold@teachspin.com> wrote: > > >Hi all, I've using a voltage reference->Resistor-> > >opamp feedback to make a current source. > >Like figure 10 here. > >http://www.ti.com/lit/ds/symlink/ref102.pdf > >The reference is the REF102, but I'm using an opa134 > >for the opamp. (Currents down to 10 nA so I wanted a fet > >opamp.) > > > >Now someone wants to use it where noise might be more > >of an issue, so I figured the first thing was to measure the > >noise. > > > >Well, the dang thing is singing at ~1.8 MHz. > >Not that loud... about 100 uV p-p, but still > >this seems like bad form. > >Any thoughts on how I might squash this? > > > >TIA > >George H. > >(Then maybe I'll ask about how to filter the noise further.. > >I was thinking about a cap. multiplier hanging on > >the output) > > This is stable: > > https://dl.dropboxusercontent.com/u/53724080/Circuits/Current_Sources/Current_Sink.JPG > > or use a bootstrap source > > https://dl.dropboxusercontent.com/u/53724080/Circuits/Current_Sources/Isrc_Boot.JPG > > https://dl.dropboxusercontent.com/u/53724080/Circuits/Current_Sources/Ramp.JPG > > You could lowpass filter the bootstrap voltage, to reduce noise, but > that decreases the current source impedance at higher frequencies. > Ultimately the Johnson noise will getcha.
Yeah.. I don't really want to do a different circuit. I'm wondering if my feedback opamp (opa134) is just too fast. I'll see if I can dig up a slower FET. George H.
On 09/23/2015 10:58 AM, George Herold wrote:
> On Wednesday, September 23, 2015 at 10:49:18 AM UTC-4, George Herold wrote: >> Hi all, I've using a voltage reference->Resistor-> >> opamp feedback to make a current source. >> Like figure 10 here. >> http://www.ti.com/lit/ds/symlink/ref102.pdf >> The reference is the REF102, but I'm using an opa134 >> for the opamp. (Currents down to 10 nA so I wanted a fet >> opamp.) >> >> Now someone wants to use it where noise might be more >> of an issue, so I figured the first thing was to measure the >> noise. >> >> Well, the dang thing is singing at ~1.8 MHz. >> Not that loud... about 100 uV p-p, but still >> this seems like bad form. >> Any thoughts on how I might squash this? >> >> TIA >> George H. >> (Then maybe I'll ask about how to filter the noise further.. >> I was thinking about a cap. multiplier hanging on >> the output) > > Current ranges from 10 nA to 1 mA. There is no noticeable > change in the oscillations as the current is changed. > (Which kinda bothers me.) > > George H. >
Resistor in series between op amp output and -in Series RC between -in and +in Adjust to give the op amp a unity gain cross of ~500 kHz. This works by increasing the noise gain. Cheers Phil Hobbs -- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC Optics, Electro-optics, Photonics, Analog Electronics 160 North State Road #203 Briarcliff Manor NY 10510 hobbs at electrooptical dot net http://electrooptical.net
On Wednesday, September 23, 2015 at 12:25:30 PM UTC-4, Phil Hobbs wrote:
> On 09/23/2015 10:58 AM, George Herold wrote: > > On Wednesday, September 23, 2015 at 10:49:18 AM UTC-4, George Herold wrote: > >> Hi all, I've using a voltage reference->Resistor-> > >> opamp feedback to make a current source. > >> Like figure 10 here. > >> http://www.ti.com/lit/ds/symlink/ref102.pdf > >> The reference is the REF102, but I'm using an opa134 > >> for the opamp. (Currents down to 10 nA so I wanted a fet > >> opamp.) > >> > >> Now someone wants to use it where noise might be more > >> of an issue, so I figured the first thing was to measure the > >> noise. > >> > >> Well, the dang thing is singing at ~1.8 MHz. > >> Not that loud... about 100 uV p-p, but still > >> this seems like bad form. > >> Any thoughts on how I might squash this? > >> > >> TIA > >> George H. > >> (Then maybe I'll ask about how to filter the noise further.. > >> I was thinking about a cap. multiplier hanging on > >> the output) > > > > Current ranges from 10 nA to 1 mA. There is no noticeable > > change in the oscillations as the current is changed. > > (Which kinda bothers me.) > > > > George H. > > > > Resistor in series between op amp output and -in > Series RC between -in and +in > > Adjust to give the op amp a unity gain cross of ~500 kHz. > > This works by increasing the noise gain.
Hmm OK thanks.. how to reduce the BW of a unity gain buffer... I didn't know that. Say R = 10 k C ~100pf.. and play around there. I found a 1 MHz opamp (vs 8MHz for opa134 before) and that reduced the signal by a factor of two or so. To be honest, this might be my measurement error. I was trying to filter the oscillations with an RC low pass before the load and couldn't... Scratch scratch... George H.
> > Cheers > > Phil Hobbs > > -- > Dr Philip C D Hobbs > Principal Consultant > ElectroOptical Innovations LLC > Optics, Electro-optics, Photonics, Analog Electronics > > 160 North State Road #203 > Briarcliff Manor NY 10510 > > hobbs at electrooptical dot net > http://electrooptical.net