Author Topic: Why is this circuit's transient response settling twice  (Read 1836 times)

0 Members and 1 Guest are viewing this topic.

Offline nikola22Topic starter

  • Contributor
  • Posts: 10
  • Country: us
Why is this circuit's transient response settling twice
« on: April 22, 2020, 01:36:50 am »
currently working on an improved Howland Current Source Architecture that I have built using this schematic. (R1=R2=R3=R4=10k & R5 = 10). configured with a gain of 1 and a transfer function of Io = Vin/R5 However, in the lab I am noticing an odd behavior with my transient response. It is actually settling twice! Instead of using a BJT as a current buffer I am using the TCA0372 as a voltage buffer. (https://www.onsemi.com/pub/Collateral/TCA0372-D.PDF).

Would someone please be able to explain why I am seeing this transient response? Where this circuit is apparently settling twice? The yellow trace is VL and the Orange Trace is Iout. The input is a 0 to -300mV step input voltage. Appreciate any help!

The supply voltages are +/-12V. The load is either a resistive box of 85 ohms, or 85 ohms and 150mH.


« Last Edit: April 22, 2020, 09:25:23 pm by nikola22 »
 

Online Jay_Diddy_B

  • Super Contributor
  • ***
  • Posts: 2766
  • Country: ca
Re: Why is this circuit's transient response settling twice
« Reply #1 on: April 22, 2020, 02:16:02 am »
Hi,

What is the load?

Those waveforms are consistent with an inductive load.

Try a resistive load.

Jay_Diddy_B
 

Offline nikola22Topic starter

  • Contributor
  • Posts: 10
  • Country: us
Re: Why is this circuit's transient response settling twice
« Reply #2 on: April 22, 2020, 02:19:48 am »
This photo was taken with an inductive load, but I believe I noticed the same behavior with my resistive load testing.
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22436
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Why is this circuit's transient response settling twice
« Reply #3 on: April 22, 2020, 07:03:07 am »
What supply voltages?

What does the intermediate node (op-amp output / buffer input) look like?

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline nikola22Topic starter

  • Contributor
  • Posts: 10
  • Country: us
Re: Why is this circuit's transient response settling twice
« Reply #4 on: April 22, 2020, 03:51:48 pm »
The supply voltages are +/-12V. The load is either a resistive box of 85 ohms, or 85 ohms and 150mH.

Here are waveforms from the non-inverting input terminal and the output of the buffer and the intermediate node.


« Last Edit: April 22, 2020, 04:45:21 pm by nikola22 »
 

Offline duak

  • Super Contributor
  • ***
  • Posts: 1048
  • Country: ca
Re: Why is this circuit's transient response settling twice
« Reply #5 on: April 22, 2020, 04:43:53 pm »
Nikola22,

[Edited]:  My first impression is that If the first op-amp has rail to rail outputs and the supply voltages are fairly low, I think you'll see those waveforms because the output amp does not have the output swing needed to follow the intermediate voltage.  When I think about the supply voltages and the millisecond time scale, this doesn't make sense.

What part is the first op-amp?

Are you sure the inputs of the op-amp are connected as shown?  ie., could they be swapped?

What happens when you apply the input signal to R1 instead of R3?
« Last Edit: April 22, 2020, 06:22:32 pm by duak »
 

Offline nikola22Topic starter

  • Contributor
  • Posts: 10
  • Country: us
Re: Why is this circuit's transient response settling twice
« Reply #6 on: April 22, 2020, 04:46:57 pm »
the first op-amp is an OPA4172. +-/12V supply rails. The inputs are connected properly. the circuit is on an existing PCB, so cant easily swap inputs  :(
 

Offline duak

  • Super Contributor
  • ***
  • Posts: 1048
  • Country: ca
Re: Why is this circuit's transient response settling twice
« Reply #7 on: April 22, 2020, 06:29:08 pm »
The tolerance on the resistor values is quite tight.  Have you checked their values?

What happens when R1 is paralled with 1M0? 100K?  a Short?
 

Offline SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15800
  • Country: fr
Re: Why is this circuit's transient response settling twice
« Reply #8 on: April 22, 2020, 06:35:31 pm »
This photo was taken with an inductive load, but I believe I noticed the same behavior with my resistive load testing.

You may have noticed something you didn't expect with a resistive load, but it's impossible that it was "the same behavior", as obviously the voltage output entirely depends on the load's impedance. Could you please take the same scope captures with a pure resistive load?

 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22436
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Why is this circuit's transient response settling twice
« Reply #9 on: April 22, 2020, 06:38:28 pm »
Also looks suspiciously like your probe compensation and gain may be set wrong?

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline nikola22Topic starter

  • Contributor
  • Posts: 10
  • Country: us
Re: Why is this circuit's transient response settling twice
« Reply #10 on: April 22, 2020, 09:19:35 pm »
Sorry, I was not trying to imply they were exactly the same, just that they had similar output performance. I have taken photos of both an inductive (150mH/85ohm load), and a purely 85ohm load. Inductive is the first image, and purely resistive is the 2nd image.




« Last Edit: April 22, 2020, 09:25:44 pm by nikola22 »
 

Offline SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15800
  • Country: fr
Re: Why is this circuit's transient response settling twice
« Reply #11 on: April 22, 2020, 09:46:32 pm »
What is "Vout" exactly? The voltage across the load, or the potential at the voltage buffer? With your initial "yellow is VL" statement, it was a bit unclear. But it does look like it's the second option, as otherwise obviously the current and voltage would be exactly proportional for a pure resistive load.
 

Offline nikola22Topic starter

  • Contributor
  • Posts: 10
  • Country: us
Re: Why is this circuit's transient response settling twice
« Reply #12 on: April 22, 2020, 09:53:40 pm »
Its the voltage at the load. The current is set by the voltage across R5, so Vout - Vbuffer its a howland current source topology. Iout =( Vout-Vbuffer) / R5.
« Last Edit: April 22, 2020, 09:55:45 pm by nikola22 »
 

Offline duak

  • Super Contributor
  • ***
  • Posts: 1048
  • Country: ca
Re: Why is this circuit's transient response settling twice
« Reply #13 on: April 22, 2020, 10:28:43 pm »
When I see a 10 to 20 ms settling time i think either C1 is the wrong value or maybe there's some thermal issue.  Could there be a problem with the VCC and VEE connections to the TCA0372?  Maybe it's not powered and the OPA4172 is actually sourcing/sinking all the load current through the ESD diodes of the TCA0372.  The OPA4172 then heats up and limits current.  It's sort of in line with the short circuit current graph on p.18. Is the OPA4172 getting hot?

BTW, does an input pulse of the opposite polarity act differently?
 

Offline SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15800
  • Country: fr
Re: Why is this circuit's transient response settling twice
« Reply #14 on: April 22, 2020, 10:36:44 pm »
Yep.
Also, the scope capture with a resistive load looks suspect. If the actual circuit is exactly like on the schematic, any current flowing through R5 should also be flowing through the resistive load, with only negligible current flowing through R2 and R4 (where else could it also be flowing?). But as we can see on the capture, the voltage across the resistive load looks nothing like the voltage across R5 (proportionally) after ~20ms, which is highly suspect. Some incorrect wiring and/or part overheating is indeed possible. Did any of the opamps feel hot to touch? I'm also trying to figure out if an excessive unbalance of the 10k resistors could explain this.


« Last Edit: April 22, 2020, 10:39:00 pm by SiliconWizard »
 

Offline nikola22Topic starter

  • Contributor
  • Posts: 10
  • Country: us
Re: Why is this circuit's transient response settling twice
« Reply #15 on: April 22, 2020, 11:15:39 pm »
This is a more complete view of the schematic, ive tried removing R9, and the snubber component, and the buffer op-amp, none of these seem to change the improper transient response. BTW the resistors are 0.1% tolerance, I don't think the issue lies there, unfortunately :(

« Last Edit: April 22, 2020, 11:54:55 pm by nikola22 »
 

Offline nikola22Topic starter

  • Contributor
  • Posts: 10
  • Country: us
Re: Why is this circuit's transient response settling twice
« Reply #16 on: April 22, 2020, 11:54:00 pm »
I don't believe there is a thermal issue, as I can run this circuit all day, and it will give me the correct output current of ~50mA, but only on transient do I see unexpected behavior. There is a proper connection to the rails of the op-amp, and I have even removed the buffer at one point to measure the transient response. The problem is still the same... A pulse of opposite polarity, acts exactly the same EXCEPT for a very small pulse of approximately +/- 10mV, then it seems to have a proper transient response (critically damped ...)
 

Offline SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15800
  • Country: fr
Re: Why is this circuit's transient response settling twice
« Reply #17 on: April 23, 2020, 12:35:13 am »
Are you sure the 47pF capacitor is not actually 47nF?

I still don't get the pure resistive load case. How can the current and the voltage across the load not be proportional (which they don't seem to be during the first 20ms)? Where does the current flow?

This 20ms time constant also seems suspect. It's very long (that's why I asked about the cap.) With an inductive load, you can expect some ringing, but not long steps like this.

Are you also 100% sure about your input voltage? That may sound obvious, but I don't think we have actually seen it on your scope captures. It could have a nasty transient maybe?

Anyway, I'll attach an LTSpice sim file if you or anyone's interested in simulating this. Not with the same opamps, but I've failed to get anything close to what you are observing here. There must be something we don't know or are overlooking here...
« Last Edit: April 23, 2020, 12:56:01 am by SiliconWizard »
 

Offline nikola22Topic starter

  • Contributor
  • Posts: 10
  • Country: us
Re: Why is this circuit's transient response settling twice
« Reply #18 on: April 23, 2020, 12:55:46 am »
My simulation showed exactly the same! I am utterly confused... Now I am wondering if its because of the input diodes on the OPA4172, wondering if that its in the simulation model.

The 12V rails look clean, unfortunately problem does not seem to be there... :(
« Last Edit: April 23, 2020, 02:36:00 am by nikola22 »
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22436
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Why is this circuit's transient response settling twice
« Reply #19 on: April 23, 2020, 06:25:10 am »
Pic of the breadboard?

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15800
  • Country: fr
Re: Why is this circuit's transient response settling twice
« Reply #20 on: April 23, 2020, 03:50:41 pm »
At this point, I'm definitely suspecting either the input signal generation, or the power supplies, or both.

Could it be that the OP is actually testing the transient behavior of this circuit by switching on the power supplies instead of testing that when those are fully stabilized? Sorry if that sounds silly and like I'm not trusting the OP, but it's just that it could well explain the behavior.
 

Offline nikola22Topic starter

  • Contributor
  • Posts: 10
  • Country: us
Re: Why is this circuit's transient response settling twice
« Reply #21 on: April 23, 2020, 04:17:21 pm »
It was similar to that SiliconWizard! Im a dope, I was using an autoranging multimeter, and the impedance was changing mid-transient!!! Got it worked out now! Thanks for all the help!
 
The following users thanked this post: thm_w, Wimberleytech

Offline SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15800
  • Country: fr
Re: Why is this circuit's transient response settling twice
« Reply #22 on: April 23, 2020, 04:53:09 pm »
Glad you sorted it out. And yes, auto-ranging multimeters can be a trap. Likewise, using auto-range for measuring the current of a circuit with intermittent power draw profiles (such as MCU with sleep/active cycles) can lead to spurious MCU resets (due to ther burden voltage - the shunt resistor automatically changing) which can be hard to figure out until you realize it's the fricking meter.

Those are situations in which disabling auto-ranging is called for.

And now the conclusion is that your circuit has no inherent fault and could be useful for other people.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf