Author Topic: Voltage reference not working properly #SOLVED#  (Read 7326 times)

0 Members and 1 Guest are viewing this topic.

Offline FotatoPotatoTopic starter

  • Regular Contributor
  • *
  • Posts: 126
  • Country: us
  • It's probably in reverse...
Voltage reference not working properly #SOLVED#
« on: February 26, 2019, 01:25:02 am »
Hello everyone,

I'm currently building a programmable DC electronic load and for the most part, it works perfectly. But there is one problem and it is quite a big one. In my schematic (Linked below) I use an LTC1046 to generate -5v for the calibration pots and for the AD8630 opamp. Feeding on this is a TL431 voltage reference IC that is set to output -2.5v for the current calibration. But for some reason, it is outputting 0.36v which is causing the reported current draw from the Load to be higher than the actual current draw. In addition to that it is impossible for me to match the reported and actual current draw and no matter what, it is always about 100mA high on the reported current draw than it is actually drawing. 

I have double checked all of my connections and I cannot seem to figure out what is wrong with the TL431, any suggestions would be greatly appreciated!

Thanks  :D

Here is a PDF of the schematic. The -5v generation and -2.5v schematics are in the top right corner of the schematic sheet.
« Last Edit: March 02, 2019, 09:55:35 pm by FotatoPotato »
 

Offline Shock

  • Super Contributor
  • ***
  • Posts: 4323
  • Country: au
Re: Voltage reference not working properly
« Reply #1 on: February 26, 2019, 05:06:42 am »
When in doubt test it out (of circuit).
Soldering/Rework: Pace ADS200, Pace MBT350
Multimeters: Fluke 189, 87V, 117, 112   >>> WANTED STUFF <<<
Oszilloskopen: Lecroy 9314, Phillips PM3065, Tektronix 2215a, 314
 

Offline BNElecEng

  • Regular Contributor
  • *
  • Posts: 99
  • Country: gb
Re: Voltage reference not working properly
« Reply #2 on: February 26, 2019, 07:33:26 am »
Agreed. Remove the TL431. Measure the -5V and check it is ok. Also, check the -2.5V line on the board even though you've removed the voltage reference IC. There may be a short or something which is pulling the voltage higher than expected.
 

Offline Arznei

  • Contributor
  • Posts: 32
  • Country: de
Re: Voltage reference not working properly
« Reply #3 on: February 26, 2019, 08:07:54 am »
Another idea: the datasheet states there must be a minimum of 1mA going through the LM431 for proper operation of the internal reference voltage. With R1=2k2 you have only 1.136mA going into the node of 2.5V though you are loading it with 0.65mA via R29. That leaves less than 0.5mA for the LM431. Maybe try decreasing R1 to guarantee the 1mA?
« Last Edit: February 26, 2019, 09:35:46 am by Arznei »
 

Offline iMo

  • Super Contributor
  • ***
  • Posts: 5570
  • Country: va
Re: Voltage reference not working properly
« Reply #4 on: February 26, 2019, 11:01:23 am »
Try this:
Readers discretion is advised..
 

Offline StillTrying

  • Super Contributor
  • ***
  • Posts: 2850
  • Country: se
  • Country: Broken Britain
Re: Voltage reference not working properly
« Reply #5 on: February 26, 2019, 11:53:52 am »
"But for some reason, it is outputting 0.36v"

I don't think the too low 0.5mA through the 431 explains it, sounds like the 431 is wired wrong, perhaps backwards.

imo's looks like a good idea, but it's no longer a steady -2.5V reference from 0V.

Time travel: :)
Yes, it is "stable" -2.5V from -5V.

The -5V itself isn't regulated.
« Last Edit: February 26, 2019, 02:39:05 pm by StillTrying »
.  That took much longer than I thought it would.
 

Offline iMo

  • Super Contributor
  • ***
  • Posts: 5570
  • Country: va
Re: Voltage reference not working properly
« Reply #6 on: February 26, 2019, 12:54:59 pm »
..imo's looks like a good idea, but it's no longer a steady 2.5V reference from 0V.
Yes, it is "stable" -2.5V from -5V.
Readers discretion is advised..
 

Offline iMo

  • Super Contributor
  • ***
  • Posts: 5570
  • Country: va
Re: Voltage reference not working properly
« Reply #7 on: February 26, 2019, 01:23:46 pm »
..But for some reason, it is outputting 0.36v ..
On the first glance it looks like you have got it reversed. Double check the wiring again..
« Last Edit: February 26, 2019, 01:25:31 pm by imo »
Readers discretion is advised..
 

Offline Arznei

  • Contributor
  • Posts: 32
  • Country: de
Re: Voltage reference not working properly
« Reply #8 on: February 26, 2019, 03:40:18 pm »
That was also my first thought, though it shouldn't really matter. From my understanding of the block diagram of the device it should still keep its Anode-Cathode voltage at 2.5V the way it is wired up in the schematic. It actually makes more sense this way, since the desired outcome is a -2.5V reference from ground, not 2.5V from the -5V rail.

And since the device itself doesn't have a GND reference it shouldn't matter wether you decrease the Cathode voltage or increase the anode voltage with a resistor.

The current range for stable operation is 1mA to 100mA through the device. If a current below 1mA leads to the internal voltage reference to decrease that would lead to the internal transistor beeing overdriven. A fully driven BJT would explain the ~0.4V across the device which are beeing measured.
 

Offline FotatoPotatoTopic starter

  • Regular Contributor
  • *
  • Posts: 126
  • Country: us
  • It's probably in reverse...
Re: Voltage reference not working properly
« Reply #9 on: February 26, 2019, 11:55:59 pm »

Quote
Agreed. Remove the TL431. Measure the -5V and check it is ok.

I have checked the -5v rail and all is good on that end so no problems there. I also checked to see if there are any shorts that would pull the voltage up and I can't find any, so I don't think this is the problem.

Quote
sounds like the 431 is wired wrong, perhaps backward.

Again, I double checked the schematic and board and If you look at the schematic that I posted, the 431 is wired the same as IMO's LTSpice simulation which shows that it should be outputting -2.5. One thing about his Simulation that does confuse me tho is that he has the positive 5v going to the cathode and gnd which seems like it shouldn't work...  ???

Quote
The current range for stable operation is 1mA to 100mA through the device. If a current below 1mA leads to the internal voltage reference to decrease that would lead to the internal transistor beeing overdriven. A fully driven BJT would explain the ~0.4V across the device which are beeing measured.

I think you may be onto something here, the -2.5v is only used on a potentiometer to set the current calibration potentiometer, and that doesn't pull much if any current. So that might be one of the problems.

What I actually think is the problem (and this is super stupid on my part) is that the LM431 and the TL431 may be the "same" IC, but after googling it, it turns out they are very different. For one, the TL431 has a 1.25v internal reference while the LM431 has an internal 2.5v reference. Also, the voltage ranges and minimum current ranges are very different for the two parts. I'll post a pic of the chart I found. Anyway, I think I just pulled a big goof and used the wrong chop  :palm: I already ordered the LM431 and It should be here in the coming day or 2. Once it arrives I'll test it and post back if the issue has been resolved or not. Thanks so much for the feedback guys, I really appreciate it!  :)
 

Offline not1xor1

  • Frequent Contributor
  • **
  • Posts: 716
  • Country: it
Re: Voltage reference not working properly
« Reply #10 on: February 27, 2019, 06:31:38 am »
What I actually think is the problem (and this is super stupid on my part) is that the LM431 and the TL431 may be the "same" IC, but after googling it, it turns out they are very different. For one, the TL431 has a 1.25v internal reference while the LM431 has an internal 2.5v reference. Also, the voltage ranges and minimum current ranges are very different for the two parts. I'll post a pic of the chart I found. Anyway, I think I just pulled a big goof and used the wrong chop  :palm: I already ordered the LM431 and It should be here in the coming day or 2. Once it arrives I'll test it and post back if the issue has been resolved or not. Thanks so much for the feedback guys, I really appreciate it!  :)

TL431 and LM431 are both 2.5V and are both specified for a minimum current of 1mA (the 1.25 100µA version is LMV431)
I think the problem is wrong connections, wrong potentiometer value (lower than 10k) or resistor value (i.e. 22k instead of 2.2k) or dead/fake TL431. BTW I would also decrease the resistor value to 1-1.5k just in case of current greedy out-of-spec TL431.
 

Offline iMo

  • Super Contributor
  • ***
  • Posts: 5570
  • Country: va
Re: Voltage reference not working properly
« Reply #11 on: February 27, 2019, 07:27:23 am »
Again, I double checked the schematic and board and If you look at the schematic that I posted, the 431 is wired the same as IMO's LTSpice simulation which shows that it should be outputting -2.5. One thing about his Simulation that does confuse me tho is that he has the positive 5v going to the cathode and gnd which seems like it shouldn't work...  ???
The TL431's voltage at cathode (the bar) should always be positive against its anode (the triangle). In order to get 2.5V you have to connect the "REF" pin to the cathode. The regulated voltage is at the cathode (measured against the anode).
« Last Edit: February 27, 2019, 07:29:23 am by imo »
Readers discretion is advised..
 

Offline StillTrying

  • Super Contributor
  • ***
  • Posts: 2850
  • Country: se
  • Country: Broken Britain
Re: Voltage reference not working properly
« Reply #12 on: February 27, 2019, 08:50:47 am »
Again, I double checked the schematic and board and If you look at the schematic that I posted, the 431 is wired the same as IMO's LTSpice simulation

Yes, all the schems and sims are correct, it's more a case of checking the actual package's pins.

Could it be oscillating, at 2.5V and very low current it needs a cap on its output, 4u7 to 10u depending on which data sheet you look at. "Stability Boundary Conditions"

Edit: While I remember from yesterday visit, :) the DAC's data sheet says it should have a 100n very close to its power pins.
« Last Edit: February 27, 2019, 08:57:11 am by StillTrying »
.  That took much longer than I thought it would.
 

Offline FotatoPotatoTopic starter

  • Regular Contributor
  • *
  • Posts: 126
  • Country: us
  • It's probably in reverse...
Re: Voltage reference not working properly
« Reply #13 on: March 01, 2019, 03:48:27 am »
Ok, So I have spent the last 4 hours trying to debug this thing and I just can't seem to figure out what's wrong. When I set up the LM431 on it own on a breadboard it works fine and outputs -2.5v but when I connect it to the PCB the output changes again.

The original design was made by Louis over at Scullcom Hobby Electronics (He makes fantastic videos, check his channel out). I will add the PDF of his schematic to this post so that you can see it. The only differences I have in mine from his is that I opted not to have a trigger input, otherwise my schematic is (or should be) identical to his. I watched through his series of videos covering this design, particularly the one where he calibrates the load. In that video, he starts by calibrating the 4.096v reference (R31 on my schematic, RV5 on his), then the voltage readout (R5 on my schematic and RV2 on his) then the current null calibration (R29 on mine, RV3 on his), and lastly the current draw calibration (R26 on mine, RV4 on his). I tried to calibrate my load in the exact same order but, every time I change R29 then the 4.096v reference goes down to around 3.1v, and I can't get it back to 4.096v unless I change R29. Also, the -2.5v goes down to around -0.6v when I do this. In addition to that my voltage readout slowly goes out as the voltage increases, so at 5v it is perfectly accurate but at 30v it is 0.7v higher which is a big problem and no matter how much I adjust R5, nothing changes. Lastly, the current calibration doesn't seem to work on my board. Once I try to change the current null calibration, it throws off the current calibration and even if I don't touch the current null calibration, at low currents (500mA - 100mA) the readout on my load is very off. If I  set it to draw 100mA it will only draw 70mA but the load will display that it is drawing 120mA, and when I adjust R26 the actual current draw only goes up by about 7mA and the reported current draw doesn't change.

At this point I really don't know what to do |O, I have gone through both my schematic and His schematic to look for differences and I just can't seem to find any. I was hoping that you guys would be able to take a look and see if there is something that I'm missing or doing wrong. I would really appreciate it!

Thanks so much!

FYI: The file titled "DC Load" is my version and the other one is His.

Louis's Video:
 

Offline iMo

  • Super Contributor
  • ***
  • Posts: 5570
  • Country: va
Re: Voltage reference not working properly
« Reply #14 on: March 01, 2019, 08:17:25 am »
Try to feed the -5V rail from an external source (remove the LT1046). Double check the 4.096V rail is not overloaded (REF5040 is not a voltage regulator) by something in your schematics.
Readers discretion is advised..
 

Offline StillTrying

  • Super Contributor
  • ***
  • Posts: 2850
  • Country: se
  • Country: Broken Britain
Re: Voltage reference not working properly
« Reply #15 on: March 01, 2019, 11:34:10 am »
It looks like there are 2 unrelated voltage ref problems then. :o
The REF5040's +/- 10mA output should very easily drive the 10k pot and DAC without dropping to 3.1V.
Are you sure the +5V is always good.

.  That took much longer than I thought it would.
 

Offline Shock

  • Super Contributor
  • ***
  • Posts: 4323
  • Country: au
Re: Voltage reference not working properly
« Reply #16 on: March 01, 2019, 12:12:07 pm »
So after you have tested out of circuit the next thing to try is in circuit with the supply section disconnected from the rest of the circuit, obviously if there needs to be a minimum load take that into consideration as well.

If that checks out, isolate it from the circuit it's powering and then feed that circuit from a bench supply (or alternate source) instead of using the supply circuit.

Anyway that is how to isolate a supply problem and also discover if it's a load based problem.
Soldering/Rework: Pace ADS200, Pace MBT350
Multimeters: Fluke 189, 87V, 117, 112   >>> WANTED STUFF <<<
Oszilloskopen: Lecroy 9314, Phillips PM3065, Tektronix 2215a, 314
 

Offline FotatoPotatoTopic starter

  • Regular Contributor
  • *
  • Posts: 126
  • Country: us
  • It's probably in reverse...
Re: Voltage reference not working properly
« Reply #17 on: March 02, 2019, 12:03:51 am »
So I did some probing and the +5v is perfectly stable, 0 ripples and 0 instability. The next thing I checked was the -5v and it seems that when I adjust R29 on my schematic, not only does the 4.096v and -2.5v change but the -5v goes down to around -4.57v which isn't as severe as the other regulators but it is still enough to throw the calibration off. It seems to me that R29 is the source of the problems which is a  big issue because I need it to calibrate the load. I also tested the -5v and -2.5v on a breadboard and everything seems to be working fine. Also once I remove the LM431 and the -2.5v is removed from R29, the 4.096v ref goes back to normal  :wtf:. Still not sure what to do because it seems that I would have to remove essential parts of the circuit which would render my load useless and wildly uncalibrated...  |O
 

Offline iMo

  • Super Contributor
  • ***
  • Posts: 5570
  • Country: va
Re: Voltage reference not working properly
« Reply #18 on: March 02, 2019, 12:31:33 pm »
1. is the value of R29=10k?
2. is the value of R28=4M7?
3. it could be at a specific R29 position the whole stuff starts to oscillate wildly (look at it with o'scope). You may also verify that easily - remove the R28 from your circuit, and wire the 4M7 from R29's wiper to GND, for example. Turn your R29 up/down - my bet the 4.096V and -2.5V will stay stable (assuming the R29 is not defunct per se).

« Last Edit: March 02, 2019, 12:59:57 pm by imo »
Readers discretion is advised..
 

Offline StillTrying

  • Super Contributor
  • ***
  • Posts: 2850
  • Country: se
  • Country: Broken Britain
Re: Voltage reference not working properly
« Reply #19 on: March 02, 2019, 01:52:17 pm »
"Still not sure what to do because it seems that I would have to remove essential parts of the circuit which would render my load useless and wildly uncalibrated..."

When you're fault-finding on a new circuit with lots of interacting parts it's quite normal to spend nearly all of the time "not sure what to do"! You just have to eliminate details one by one by very careful checking, while keeping an overall view of how it works so that you're hopefully concentrating on the most likely.

"It seems to me that VR29 10k is the source of the problems"

I agree with imo's checks immediately above, but unless you've got it on very long leads I don't think it's the problem, it should just be a very contant 0.66 mA load between the 4.096V and -2.5V.

Are you sure the output of U2 U7 OPA277 always stays ~0V or above, especially during calibration. If its output tries to go lower than about -0.3V it will start to draw a lot more than its 0.8 mA from the -5V. It's about the only place I can see where the varying load on the -5V would be coming from.

« Last Edit: March 02, 2019, 04:05:46 pm by StillTrying »
.  That took much longer than I thought it would.
 

Offline iMo

  • Super Contributor
  • ***
  • Posts: 5570
  • Country: va
Re: Voltage reference not working properly
« Reply #20 on: March 02, 2019, 03:20:41 pm »
..Are you sure the output of U2 OPA277 always stays ~0V or above, especially during calibration. If its output tries to go lower than about -0.3V it will start to draw a lot more than its 0.8 mA from the -5V. It's about the only place I can see where the varying load on the -5V would be coming from.
Good point - when the OPA277' output goes below -0.3V the AD8630's input lower side clamping diode starts to conduct, and the current from the -5V source flows via OPA277's output to AD8630's input, then via the clamping diode to GND (assuming the V- of the AD8630 is GND). It could be many mAmps.
Double check - put a 100k (for example) resistor between the OP277 output and the AD8630 input as a test.
PS: if it helped you may leave the 100k resistor there as the input current of that AD is less than 100pA - that creates <10uV offset.
« Last Edit: March 02, 2019, 04:08:57 pm by imo »
Readers discretion is advised..
 

Offline FotatoPotatoTopic starter

  • Regular Contributor
  • *
  • Posts: 126
  • Country: us
  • It's probably in reverse...
Re: Voltage reference not working properly
« Reply #21 on: March 02, 2019, 04:22:28 pm »
So I just did some probing and here are my results:

OPA277 output: 0.645v ± 0.01v
LM7805 out: 5.027v ± 0.02v
LTC1046 out: -4.882v ±0.24v
LM431 out: -2.654v (not connected to anything) and -0.186v (Connected to the rest of the circuit, voltage depends on the value of R29)
REF5040 out: 3.479v ± 0.8v depending on what the value of R29 is
Value of R29: 9.952K \$\Omega\$
Value of R28: 22.85K \$\Omega\$

R28 is very weird considering it IS a 4.7M resistor, I triple checked the color bands on it and it is a 4.7M resistor but for some unexplainable reason, it is 22.8K...  :wtf:

I think we are getting close to a solution and I definitely seem to be the circuitry surrounding R29 and R28. I am going to replace R28 and see what happens but I doubt it will make a difference.

 

Offline iMo

  • Super Contributor
  • ***
  • Posts: 5570
  • Country: va
Re: Voltage reference not working properly
« Reply #22 on: March 02, 2019, 04:44:11 pm »
Quote
So I just did some probing and here are my results:

OPA277 output: 0.645v ± 0.01v   <<<<< Still you have to limit its output current with a resistor - see above >>>>>
LM7805 out: 5.027v ± 0.02v
LTC1046 out: -4.882v ±0.24v
LM431 out: -2.654v (not connected to anything) <<<< 2.65 not good afaik >>>> and -0.186v (Connected to the rest of the circuit, voltage depends on the value of R29)
REF5040 out: 3.479v ± 0.8v depending on what the value of R29 is <<<< it must be 4.096 always >>>>>
Value of R29: 9.952K \$\Omega\$
Value of R28: 22.85K \$\Omega\$      :palm:

R28 is very weird considering it IS a 4.7M resistor, I triple checked the color bands on it and it is a 4.7M resistor but for some unexplainable reason  <<<<< DOUBLE CHECK ALL 40 RESISTORS >>>>, it is 22.8K... 
« Last Edit: March 02, 2019, 04:48:19 pm by imo »
Readers discretion is advised..
 

Offline FotatoPotatoTopic starter

  • Regular Contributor
  • *
  • Posts: 126
  • Country: us
  • It's probably in reverse...
Re: Voltage reference not working properly
« Reply #23 on: March 02, 2019, 04:53:16 pm »
Just replaced R28, no difference
 

Offline StillTrying

  • Super Contributor
  • ***
  • Posts: 2850
  • Country: se
  • Country: Broken Britain
Re: Voltage reference not working properly
« Reply #24 on: March 02, 2019, 05:07:04 pm »
Something else to try.

What value have you reduced R1 2k2 supplying the LM431 to, I don't think ~1k is low enough.

The -2.5V supply needs 1mA for the 431(they usually work fine @ 0.7mA), 0.8mA for the OPSA277 and 0.66 mA for the VR29 10k pot = 2.46 mA.
Assuming the -5V is only -4.8V, R1 has to be (4.8-2.55)/2.46 = 915R, 820R should do.
 :palm:
(4.8-2.55)/1.66 = 1.4k or slightly lower for R1.


Edit: While you're there add a 1u to 10u to the -2.5V supply. :)

"REF5040 out: 3.479v ± 0.8v depending on what the value of R29 is"

It's still a mystery why the REF5040's 10mA 4.096V output isn't more than enough to power the DAC, unless the DAC is oscillating, or it's output is shorted.
« Last Edit: March 02, 2019, 05:57:01 pm by StillTrying »
.  That took much longer than I thought it would.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf