Author Topic: 30A precision current source  (Read 5484 times)

0 Members and 1 Guest are viewing this topic.

Offline ali_asadzadehTopic starter

  • Super Contributor
  • ***
  • Posts: 1967
  • Country: ca
30A precision current source
« on: April 06, 2019, 06:29:40 am »
Hi,
I want to generate from 1mA to 30A 50Hz sin wave currents for a system, it’s main use is to test and calibrate the DUT internal CT’s and current measurements circuits.

Here is a suggested circuit, That I have in my mind, I would generate a sin wave with a 16bit dac 0 – 3.3 V connected to my MCU, the output of the DAC would go to the positive input of the op-amp.

For simplicity I have only drawn 2 MOSFETs , But I plan to use 5 to 10 parallel MOSFETs, I want to use a PC power supply as the DC source of this circuit, I will use the 5V output for the current source, is this circuit Idea practical? And does this circuit generate a sin wave without DC bias?

Also, I thought Maybe I should change the 10mOhm sense resistor with a current sense IC like ACS712 from allegro, But the problem with them is that their zero current value is mid supply which is 2.5V, any Idea on how to use them, they would make it much simpler.

For the DAC Part I plan to use DAC8411IDCKR from TI,
Any Ideas?


ASiDesigner, Stands for Application specific intelligent devices
I'm a Digital Expert from 8-bits to 64-bits
 

Offline gf

  • Super Contributor
  • ***
  • Posts: 1425
  • Country: de
Re: 30A precision current source
« Reply #1 on: April 06, 2019, 11:27:12 am »
And does this circuit generate a sin wave without DC bias?

No, it is a (unidirectional) current sink, which draws current from DUT's pin 2.
But would you really want to feed a positive current into DUT's pin 2?
The latter would imply that the voltage at the DUT's pin 2 would need to be raised to > 5V, and the 5V voltage source would need to be capable to sink the current (I doubt that a PC power supply can sink significant amounts of current).
 

Offline ali_asadzadehTopic starter

  • Super Contributor
  • ***
  • Posts: 1967
  • Country: ca
Re: 30A precision current source
« Reply #2 on: April 06, 2019, 12:03:17 pm »
Quote
No, it is a (unidirectional) current sink, which draws current from DUT's pin 2.
But would you really want to feed a positive current into DUT's pin 2?
The latter would imply that the voltage at the DUT's pin 2 would need to be raised to > 5V, and the 5V voltage source would need to be capable to sink the current (I doubt that a PC power supply can sink significant amounts of current).

Thanks for the feedback, the DUT is just a simple short wire, it would goes across the center of CT,So does this solve the sink current to the PC power supply?

what's your Idea about the circuit? how should I change it?
ASiDesigner, Stands for Application specific intelligent devices
I'm a Digital Expert from 8-bits to 64-bits
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 20357
  • Country: gb
  • 0999
Re: 30A precision current source
« Reply #3 on: April 06, 2019, 01:44:43 pm »
No the circuit will not generate an AC current, without a DC bias.

What's the frequency range?

What's the maximum output voltage, at 30A?

If space isn't limited and the output voltage only needs to be 1V or so, then how about using a transformer to step-up the current, whist reducing the voltage?

1mA to 30A is a huge dynamic range. Your circuit will need multiple different range settings to give a decent level of accuracy.
« Last Edit: April 06, 2019, 06:42:54 pm by Zero999 »
 

Offline soldar

  • Super Contributor
  • ***
  • Posts: 3595
  • Country: es
Re: 30A precision current source
« Reply #4 on: April 06, 2019, 02:48:26 pm »
You could use an old arc welding machine which is just a stray field transformer (a transformer with a magnetic bypass or shunt in its core) which limits max current. Set it to 30 Amps or less although for very low currents you will probably need to use a resistor.
All my posts are made with 100% recycled electrons and bare traces of grey matter.
 

Offline OM222O

  • Frequent Contributor
  • **
  • Posts: 768
  • Country: gb
Re: 30A precision current source
« Reply #5 on: April 06, 2019, 04:03:20 pm »
something that most people forgot to point out: you can't just parallel 2 or even 10 mosfets from the same output of the op amp and hope it works.
They have different characteristics and one can turn on and do most of the work, while others are barely turned on. especially in their linear region which this application uses.

If you want to use parallel mosfets, they should each have their own op amp and ballast resistor.

Another point is: why would you want a bipolar output? that would complicate things a lot ... probably a simple push pull config won't be enough for your output and you need something like a diamond buffer.

Another point: 30A at 1mA increments is 30000 counts. a 16 bit dac is 65000 counts. that would imply an INL + DNL of 1LSB max, or your results will not be accurate.you either need to buy a higher resolution dac (18 or 20 bits) or make sure the 16 bit one is damn near perfect.
 

Offline soldar

  • Super Contributor
  • ***
  • Posts: 3595
  • Country: es
Re: 30A precision current source
« Reply #6 on: April 06, 2019, 05:02:35 pm »
30A at 1mA increments is 30000 counts. a 16 bit dac is 65000 counts. that would imply an INL + DNL of 1LSB max, or your results will not be accurate.you either need to buy a higher resolution dac (18 or 20 bits) or make sure the 16 bit one is damn near perfect.

30A at 1 mA increments is meaningless and impossible. Regulating 30A with a precision of 1mA is not really practical in any way.

If you use a regular pot-meter and roughly assume ranges with a ratio of 3 you would need like 9 ranges (in mA): 1- 3 -10 - 30 - 100 - 300 - 1000 - 3000 - 10000 -30000

An that would give you a reasonable resolution of maybe 1 or 2%.
All my posts are made with 100% recycled electrons and bare traces of grey matter.
 

Online SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15795
  • Country: fr
Re: 30A precision current source
« Reply #7 on: April 06, 2019, 05:10:08 pm »
I doubt you will get down to 1mA with such a circuit and such a small shunt resistor.

It's clearly not a bipolar current source (it's a simple current sink). I'm guessing you implied you want a bipolar source (and several here have understood so, probably because of your 'DC bias' remark) whereas I don't see it clearly stated in your post unless I missed something. With a 0-3.3V input sine signal, you'll just get a current sink between Imin (won't be 0, there will be a significant offset) and 3.3/Rshunt basically. That could be what you wanted after all. Unclear. Also, with 5V supply for your DUT, make sure you'll have enough headroom - we don't know anything about your DUT.

Also, paralleling transistors in this way won't work properly as explained above.
Obviously if you parallel several current sinks as OM2220 suggested, you'll have to set your shunt resistors appropriately since all the sinks will add to the overall sunk current. Also obviously, the errors will add up.

With such a wide current range, I suggest paralleling several current sinks as above, each with a different current range, and ideally controlled independently.

Why would a current sensor be simpler than a basic shunt resistor? And besides the DC offset, it probably has enough phase shift (I haven't read the DS enough to figure that out) to render it unsuitable here.
 

Offline gf

  • Super Contributor
  • ***
  • Posts: 1425
  • Country: de
Re: 30A precision current source
« Reply #8 on: April 06, 2019, 06:00:05 pm »
With a 0-3.3V input sine signal, you'll just get a current sink between Imin (won't be 0, there will be a significant offset) and 3.3/Rshunt basically.

Exactly, so 1 mOhm does not fit either with the desired 30A at 3.3V input voltage. 30A would rather require 110 mOhm for a voltage drop of 3.3V (but then the sense resistor would dissipate 100W :phew:, and there were only 1.7V headroom for the FET and the DUT).
 

Offline nctnico

  • Super Contributor
  • ***
  • Posts: 28429
  • Country: nl
    • NCT Developments
Re: 30A precision current source
« Reply #9 on: April 06, 2019, 06:39:43 pm »
something that most people forgot to point out: you can't just parallel 2 or even 10 mosfets from the same output of the op amp and hope it works.
They have different characteristics and one can turn on and do most of the work, while others are barely turned on. especially in their linear region which this application uses.

If you want to use parallel mosfets, they should each have their own op amp and ballast resistor.
Agreed. Also in the linear operating mode MOSFETs have a negative temperature coefficient so they are prone to thermal runaway just like transistors. To make things worse the variation between MOSFETs is much bigger compared to transistors. All in all I'd rather use transistors.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 20357
  • Country: gb
  • 0999
Re: 30A precision current source
« Reply #10 on: April 06, 2019, 07:11:39 pm »
How about a Howland current pump driving a transformer?

A bipolar power supply is idea, but it can be bridged to work from a single supply.

This design is only 3A, but add a 10:1 transformer and you have 30A on the secondary, albeit at a lower maximum voltage, which I presume won't matter has it's only driving a current transformer.

The voltage between the feedback and output nodes follows the voltage on the in node. Because R7 and R8 form a potential divider, the voltage on R5 and R6 is double the input voltage.

Ensure:
R1 = R3
R2 = R4
R5 = R6

R1 & R2 configure the gain.
R5 & R6 double as emitter ballast resistors and current sensing resistors.
X1 can be replaced with a resistor to reduce the output current.
 
The following users thanked this post: Mechatrommer

Offline coromonadalix

  • Super Contributor
  • ***
  • Posts: 7001
  • Country: ca
Re: 30A precision current source
« Reply #11 on: April 06, 2019, 07:27:18 pm »
gonna sound stupid   but this schem is an audio amp ?? 

Adding a transformer wont help at all,  you have core losses, hysteresis  etc ...  you're creating problems and affect any precision you want ??

You want 30 amps at 50 hz, but at what voltages ??  1v rms or pk to pk ?


there is an really powerfull op amp, an lm12

http://www.ti.com/lit/an/snoa718/snoa718.pdf


I saw your howlad pump principles
https://www.allaboutcircuits.com/technical-articles/the-howland-current-pump/
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 20357
  • Country: gb
  • 0999
Re: 30A precision current source
« Reply #12 on: April 06, 2019, 08:15:59 pm »
gonna sound stupid   but this schem is an audio amp ?? 
Yes, it is the same topology as an audio amplifier, except the output is constant current, rather than voltage.

Quote
Adding a transformer wont help at all,  you have core losses, hysteresis  etc ...  you're creating problems and affect any precision you want ??
Please explain how?

Transformers are commonly used for measuring current and as long as they're not overloaded, don't affect the accuracy at all. I've recently used one in a current measuring application and found the secondary current was within 1% of the primary, which was within the bounds of accuracy of my oscilloscope. The usual configuration is a one turn primary with many on the secondary, to reduce the current so it can be sensed using a much higher value resistor (known as the burden) than would otherwise be practical, but it will work in reverse: more turns on the primary, than secondar. To increase the knee voltage, the primary can be more than one turn, although I think that will increase the phase shift, but it's probably a non-issue in this application.
https://en.wikipedia.org/wiki/Current_transformer
 

Offline ali_asadzadehTopic starter

  • Super Contributor
  • ***
  • Posts: 1967
  • Country: ca
Re: 30A precision current source
« Reply #13 on: April 07, 2019, 06:27:16 am »
Thanks all for your suggestions and hints.


Quote
It's clearly not a bipolar current source (it's a simple current sink). I'm guessing you implied you want a bipolar source (and several here have understood so, probably because of your 'DC bias' remark) whereas I don't see it clearly stated in your post unless I missed something. With a 0-3.3V input sine signal, you'll just get a current sink between Imin (won't be 0, there will be a significant offset) and 3.3/Rshunt basically. That could be what you wanted after all. Unclear. Also, with 5V supply for your DUT, make sure you'll have enough headroom - we don't know anything about your DUT.

As I mentioned the DUT is a simple wire wrapping 1 turn inside a CT.


Quote
No the circuit will not generate an AC current, without a DC bias.

What's the frequency range?

What's the maximum output voltage, at 30A?

If space isn't limited and the output voltage only needs to be 1V or so, then how about using a transformer to step-up the current, whist reducing the voltage?

1mA to 30A is a huge dynamic range. Your circuit will need multiple different range settings to give a decent level of accuracy.

The frequency range is from 40Hz to 400Hz, the maximum open output voltage is not important, the lower the better, because we have less dissipation.


Here Came something completely different to my mind! I can buy dirt cheap 5V 40A power supplies, something like this,

Also for the DAC, I decided to switch to an audio codec part, with 24Bit resolution (CS4344), This time the output is bidirectional, also I have managed to use a current sense IC instead of shunt resistor,
Any Idea's about this new design?




ASiDesigner, Stands for Application specific intelligent devices
I'm a Digital Expert from 8-bits to 64-bits
 

Offline beanflying

  • Super Contributor
  • ***
  • Posts: 7375
  • Country: au
  • Toys so very many Toys.
Re: 30A precision current source
« Reply #14 on: April 07, 2019, 06:59:59 am »
There is a series of commercial products available for what your spec is. Search for Transconductance Amplifiers or Current Calibrators should get you some hits.

eBay auction: #113702013768 Manual and Schematics Description here https://valhallascientific.com/wp-content/uploads/2018/09/2555A_Manual.pdf
Circuit Diagram and description of the 10A Cousin here https://valhallascientific.com/wp-content/uploads/2018/09/2500EP_Manual.pdf


Or there is several Fluke offerings designed to be run in conjunction with their AC Calibrators.
« Last Edit: April 07, 2019, 07:06:16 am by beanflying »
Coffee, Food, R/C and electronics nerd in no particular order. Also CNC wannabe, 3D printer and Laser Cutter Junkie and just don't mention my TEA addiction....
 

Offline Doctorandus_P

  • Super Contributor
  • ***
  • Posts: 4004
  • Country: nl
Re: 30A precision current source
« Reply #15 on: April 07, 2019, 09:22:59 am »
I would probably build this from an audio amplifier and a resistor to measure the current, or measure the current by putting a DMM in series.

You can also easily relax the current you have to drive by putting a few windings on your current transformer.
I.E, if you put 5 windings on the transformer, you only have to generate 30/5 = 6A.

Also think about what sort of accuracy you want.
1mA resolution in a 30A range is 1:30000, which would require a 15 bit adc.
An easy solution is to make 2 or even 3 measurement ranges.
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 20357
  • Country: gb
  • 0999
Re: 30A precision current source
« Reply #16 on: April 07, 2019, 10:38:21 am »
Thanks all for your suggestions and hints.


Quote
It's clearly not a bipolar current source (it's a simple current sink). I'm guessing you implied you want a bipolar source (and several here have understood so, probably because of your 'DC bias' remark) whereas I don't see it clearly stated in your post unless I missed something. With a 0-3.3V input sine signal, you'll just get a current sink between Imin (won't be 0, there will be a significant offset) and 3.3/Rshunt basically. That could be what you wanted after all. Unclear. Also, with 5V supply for your DUT, make sure you'll have enough headroom - we don't know anything about your DUT.

As I mentioned the DUT is a simple wire wrapping 1 turn inside a CT.
Is it possible to access the CT or is it inside a case?

If so, then how about just wrapping more turns through the middle?

Alternatively, get another current transformer and run it in reverse.


Quote
Quote
No the circuit will not generate an AC current, without a DC bias.

What's the frequency range?

What's the maximum output voltage, at 30A?

If space isn't limited and the output voltage only needs to be 1V or so, then how about using a transformer to step-up the current, whist reducing the voltage?

1mA to 30A is a huge dynamic range. Your circuit will need multiple different range settings to give a decent level of accuracy.

The frequency range is from 40Hz to 400Hz, the maximum open output voltage is not important, the lower the better, because we have less dissipation.


Here Came something completely different to my mind! I can buy dirt cheap 5V 40A power supplies, something like this,

Also for the DAC, I decided to switch to an audio codec part, with 24Bit resolution (CS4344), This time the output is bidirectional, also I have managed to use a current sense IC instead of shunt resistor,
Any Idea's about this new design?
No that circuit will only produce magic smoke. Q3 and Q4 are connected backwards and will just short Q1 and Q2 to 0V when they conduct.

IF a dual power supply is an issue, the circuit I posted previously can be bridged.
« Last Edit: April 07, 2019, 02:13:49 pm by Zero999 »
 

Offline ali_asadzadehTopic starter

  • Super Contributor
  • ***
  • Posts: 1967
  • Country: ca
Re: 30A precision current source
« Reply #17 on: April 08, 2019, 05:37:58 am »
Quote
No that circuit will only produce magic smoke. Q3 and Q4 are connected backwards and will just short Q1 and Q2 to 0V when they conduct.
OH, I see that now, thanks.

Quote
IF a dual power supply is an issue, the circuit I posted previously can be bridged.
Did you draw your circuit in LT spice? would you share the design file? also can it be done with MOSFET's? and finally if I use some DC/DC converter to lower the voltage even lower like 2.5V or 1.8V is it doable? I mean using something like LMZ31530 from TI so in that way the heat dissipation would be way lower.


Quote
Is it possible to access the CT or is it inside a case?

If so, then how about just wrapping more turns through the middle?

Alternatively, get another current transformer and run it in reverse.
No it's not possible, because this device main use would be to calibrate a system that we are producing.


There comes another Idea to my Mind too, Can I use the other side of the Bridge mosfets to just conduct and make a short during each sin wave cycle, so in this way I can lower the drive voltage to something even lower, like 1V ! and the total dissipation would be 30W ;)
ASiDesigner, Stands for Application specific intelligent devices
I'm a Digital Expert from 8-bits to 64-bits
 

Offline gf

  • Super Contributor
  • ***
  • Posts: 1425
  • Country: de
Re: 30A precision current source
« Reply #18 on: April 08, 2019, 06:32:53 am »
There comes another Idea to my Mind too, Can I use the other side of the Bridge mosfets to just conduct and make a short during each sin wave cycle, so in this way I can lower the drive voltage to something even lower, like 1V ! and the total dissipation would be 30W ;)

Note, the MOSFET you had selected has a specified ON resistance of up to 77 mOhm. At 30A this gives already a voltage drop of up to 2.3V at a single, fully condicting MOSFET (and you need to drive the gate with > 10V in order to get this state -> btw, keep in mind the required output voltage swing of the driving opamp). So with two ones in series (bridge), there are only 0.4V left for the DUT, at 5V anyway.

Btw, what is on the other side of the CT? The impedance at the secondary side eventually gets transformed to the primary side, therefore the impedance which terminates your AC current source may be significantly higher than just the resistance of the wire (and may require a higher voltage then).
« Last Edit: April 08, 2019, 06:35:24 am by gf »
 

Offline pansku

  • Regular Contributor
  • *
  • Posts: 60
  • Country: fi
Re: 30A precision current source
« Reply #19 on: April 08, 2019, 03:21:41 pm »
I once did some planning for a similar project, but never got to the point of simulation. However I suggest that you could look at this paper for some inspiration: A Wide-Band Transconductance Amplifier for Current Calibrations

That circuitry is good for 8A Rms and has over 100kHz of -3dB bandwidth.
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 20357
  • Country: gb
  • 0999
Re: 30A precision current source
« Reply #20 on: April 08, 2019, 09:01:58 pm »
Quote
No that circuit will only produce magic smoke. Q3 and Q4 are connected backwards and will just short Q1 and Q2 to 0V when they conduct.
OH, I see that now, thanks.

Quote
IF a dual power supply is an issue, the circuit I posted previously can be bridged.
Did you draw your circuit in LT spice? would you share the design file? also can it be done with MOSFET's? and finally if I use some DC/DC converter to lower the voltage even lower like 2.5V or 1.8V is it doable? I mean using something like LMZ31530 from TI so in that way the heat dissipation would be way lower.


Quote
Is it possible to access the CT or is it inside a case?

If so, then how about just wrapping more turns through the middle?

Alternatively, get another current transformer and run it in reverse.
No it's not possible, because this device main use would be to calibrate a system that we are producing.


There comes another Idea to my Mind too, Can I use the other side of the Bridge mosfets to just conduct and make a short during each sin wave cycle, so in this way I can lower the drive voltage to something even lower, like 1V ! and the total dissipation would be 30W ;)
Yes, I drew it in LTSpice. I can't upload the design file because I don't have access to that PC at the moment.

How do you plan to calibrate this piece of test equipment? You need calibrated current meter.

Using a transformer to boost the output current is the only sane way to do this. I suggest getting another current transformer, run in reverse and measure the output current using a calibrated meter.
 

Offline ali_asadzadehTopic starter

  • Super Contributor
  • ***
  • Posts: 1967
  • Country: ca
Re: 30A precision current source
« Reply #21 on: April 09, 2019, 05:55:50 am »
Quote
Yes, I drew it in LTSpice. I can't upload the design file because I don't have access to that PC at the moment.

How do you plan to calibrate this piece of test equipment? You need calibrated current meter.

Using a transformer to boost the output current is the only sane way to do this. I suggest getting another current transformer, run in reverse and measure the output current using a calibrated meter.

Thanks for the tips, 0.1% accuracy is good enough for me, so maybe I use a 5.5 digit or a 6.5 digit bench mulitmeter to calibrate it, although it can be done at most to 10A,

Also this device would be a 4 channel one, because it would be used for a 3 phase system including earth, and I can not use another CT because I should be able to calibrate the phases of the channels too. including another CT would add phase shift into each channel output and would complicate things a lot more, I can use better Mosfets with much lower RDs on too for my last suggestion, so I can reduce the output voltage lower to reduce the dissipation.
ASiDesigner, Stands for Application specific intelligent devices
I'm a Digital Expert from 8-bits to 64-bits
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 20357
  • Country: gb
  • 0999
Re: 30A precision current source
« Reply #22 on: April 09, 2019, 08:10:05 am »
Phase shift won't a problem, if it's the same for all phases.

Using MOSFET doesn't gain you much, because anything operating in its linear mode needs to drop some voltage. You could use a class D amplifier to save power, but no doubt the high frequency noise will be a concern.

I suggest you invest in a clamp meter to calibrate this.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf