Author Topic: Precision Power Supply  (Read 5064 times)

0 Members and 1 Guest are viewing this topic.

Offline szymonm2

  • Contributor
  • Posts: 25
  • Country: ie
Precision Power Supply
« on: June 22, 2016, 07:00:09 pm »
Hello, Im trying to make a precision power supply with constant current and constant voltage. I already built a power supply using the ltc3780 buck converter but I need something for precision work such as lasers and leds.

Im needing a schematic where the given power supply can set the constant current before plugging in the load.
24W would be the max power of the power supply, I dont mind building even a 20w one as long as it is precise.

Thanks.
 

Offline Signal32

  • Regular Contributor
  • *
  • Posts: 234
  • Country: us
Re: Precision Power Supply
« Reply #1 on: June 22, 2016, 10:35:11 pm »
Im needing a schematic where the given power supply can set the constant current before plugging in the load.
Do you need a power supply that will display the current limit before you plug in the load ? If yes, maybe your ltc3780 supply can be modified, post the schematic.
Or do you need a power supply that when hooking up the load, it will regulate the current to the set limit very fast ?
 

Offline szymonm2

  • Contributor
  • Posts: 25
  • Country: ie
Re: Precision Power Supply
« Reply #2 on: June 22, 2016, 10:49:00 pm »
Im needing a schematic where the given power supply can set the constant current before plugging in the load.
Do you need a power supply that will display the current limit before you plug in the load ? If yes, maybe your ltc3780 supply can be modified, post the schematic.
Or do you need a power supply that when hooking up the load, it will regulate the current to the set limit very fast ?


the ltc3780 can limit the current to max 200mA which is way to much for precision work.
 

Offline Signal32

  • Regular Contributor
  • *
  • Posts: 234
  • Country: us
Re: Precision Power Supply
« Reply #3 on: June 22, 2016, 11:08:40 pm »
the ltc3780 can limit the current to max 200mA which is way to much for precision work.
By default you may be right, but there is no reason you can't modify it to impose better current limiting, if you're happy with the design you have.
You would do this by sensing the current yourself and controlling the voltage on the feedback pin.
If you want a new design, I think you need to give more details what you're looking for for people to be able to come up with ussefful suggestions : Ex: linear, buck, boost, switching, lcd display, bluetooth, etc
 

Offline szymonm2

  • Contributor
  • Posts: 25
  • Country: ie
Re: Precision Power Supply
« Reply #4 on: June 22, 2016, 11:21:06 pm »
the ltc3780 can limit the current to max 200mA which is way to much for precision work.
By default you may be right, but there is no reason you can't modify it to impose better current limiting, if you're happy with the design you have.
You would do this by sensing the current yourself and controlling the voltage on the feedback pin.
If you want a new design, I think you need to give more details what you're looking for for people to be able to come up with ussefful suggestions : Ex: linear, buck, boost, switching, lcd display, bluetooth, etc

Ok, So I dont really have much suggestions, as this project is really flexible. I would rather if it was a switching power supply but I do not really mind if its linear. LCD 16x2 screen I think is a must have in order to show the constant current set before applying a load to it. I also want it to have X.XXX Volt resolution and X.XXX A amp resolution. As I mentioned the power supply is a precision one and I intend using max load of about 20W on it. Anything more than that I have the ltc3780 powersupply.

Thanks
 

Offline MosherIV

  • Super Contributor
  • ***
  • Posts: 1336
  • Country: gb
Re: Precision Power Supply
« Reply #5 on: June 23, 2016, 12:30:53 pm »
Hi

Your definition is too contradictory:
"Hello, Im trying to make a precision power supply with constant current and constant voltage."

Please define "precision power supply" - are you after absolute voltage or precise current?
What makes it precision? - are you after thermal stability or very fine (microVolt/amp) control

You cannot have both constant current AND constant voltage - it must be one or the other.
Most bench PSUs are constant Voltage with a current limit.
Constant voltage supplies vary the voltage, with a max current limit, they drop the voltage or shut down on current limit.
Constant current supplies vary the voltage to maintain a fixed current.

I would suggest you look at some simply linear PSU designs for how to control voltage and be able to set a current limit.

http://www.microsyl.com/index.php/2010/03/31/bench-power-supply-0-25v-0-5amp/


« Last Edit: June 23, 2016, 12:33:03 pm by MosherIV »
 

Offline MarkF

  • Super Contributor
  • ***
  • Posts: 1642
  • Country: us
Re: Precision Power Supply
« Reply #6 on: June 23, 2016, 12:46:05 pm »
Why not something like a HP 6216A 0-25V/0-0.4A Variable DC Power Supply.  I have 3 of these and they work very well.  You set the current limit by shorting the outputs and adjusting to your desired limit.
 

Offline ebclr

  • Super Contributor
  • ***
  • Posts: 2045
  • Country: 00
Re: Precision Power Supply
« Reply #7 on: June 23, 2016, 12:46:18 pm »
 

Offline jeroen79

  • Frequent Contributor
  • **
  • Posts: 515
Re: Precision Power Supply
« Reply #8 on: June 23, 2016, 04:30:01 pm »
Often the currentlimit/constantcurrent mode will measure the current using the voltage dropped over a shuntresistor and then compare this to a reference voltage.
If you measure the reference voltage you will know the currentlimit without having to short the terminals.
 

Offline szymonm2

  • Contributor
  • Posts: 25
  • Country: ie
Re: Precision Power Supply
« Reply #9 on: June 23, 2016, 09:35:26 pm »
Please define "precision power supply" - are you after absolute voltage or precise current?
What makes it precision? - are you after thermal stability or very fine (microVolt/amp) control

Im trying to achieve precise control of the volts and current. At least 1 miliamp and milivolt of resolution is needed..
 

Offline Marco

  • Super Contributor
  • ***
  • Posts: 4700
  • Country: nl
Re: Precision Power Supply
« Reply #10 on: June 23, 2016, 09:37:54 pm »
Unless the power supply is designed for constant current mode it will probably have a charged capacitor ready to blow your laser diode behind the output enable switch, the current limit will kick in a little too late.
 

Offline MosherIV

  • Super Contributor
  • ***
  • Posts: 1336
  • Country: gb
Re: Precision Power Supply
« Reply #11 on: June 23, 2016, 10:06:12 pm »
Hi

"Im trying to achieve precise control of the volts and current. At least 1 miliamp and milivolt of resolution is needed."

I would recommend op-amp driven circuits. They are easy to understand and can offer very precise cotrol.
The link I posted earlier I find the easiest to understand,
http://www.microsyl.com/index.php/2010/03/31/bench-power-supply-0-25v-0-5amp/
download the schematic and read the text.
Ask questions about it.

You have not specified whether you want
a) Constant Voltage (you set the voltage and let the current vary)
OR
b) Constant Current ( you set the current and let the voltage vary)

It is impossible to do both, see Ohms law.
 

Offline szymonm2

  • Contributor
  • Posts: 25
  • Country: ie
Re: Precision Power Supply
« Reply #12 on: June 23, 2016, 10:27:36 pm »
You have not specified whether you want
a) Constant Voltage (you set the voltage and let the current vary)
OR
b) Constant Current ( you set the current and let the voltage vary)

It is impossible to do both, see Ohms law.

Then how do some powersupplies come with a function of switching the modes between constant current and constant current such as a cheap power supply from china called "YIHUA PS-3010D"
 

Offline Marco

  • Super Contributor
  • ***
  • Posts: 4700
  • Country: nl
Re: Precision Power Supply
« Reply #13 on: June 23, 2016, 10:36:56 pm »
He just meant you can't do both at the same time.

A power supply can do both at different times, but response time and overshoot for current regulation is likely going to be pretty poor on most generic power supplies.
 

Offline Signal32

  • Regular Contributor
  • *
  • Posts: 234
  • Country: us
Re: Precision Power Supply
« Reply #14 on: June 23, 2016, 10:39:35 pm »
It is impossible to do both, see Ohms law.

Then how do some powersupplies come with a function of switching the modes between constant current and constant current such as a cheap power supply from china called "YIHUA PS-3010D"
Have you not noticed all the fake / counterfeit products that come out of China. Chinese manufacturers seem to have no problems breaking many laws just to sell their products.  The Chinese power supply you mention probably already violates a few safety laws so they probably decided it might as well violate Ohm's law.
The power supply works basically by choosing by itself weather to be in constant current or constant voltage, depending on the load.

The 1mV/mA goal is very ambitious and you're unlikely to achieve it cheaper in a one-off design than buying a used quality product.
You may also want to read this thread: https://www.eevblog.com/forum/testgear/power-supply-with-fast-current-limit-regulation/
Would recommand buying something like this : http://www.ebay.com/itm/Keysight-Agilent-6612C-20-Volt-2-Amp-GPIB-Power-Supply-/301786728901?hash=item4643e409c5:g:vMsAAOSw5ZBWN5P8
or http://www.ebay.com/itm/HP-66311B-Mobile-Communications-DC-Source-0-15V-0-3A-5A-Peak-/252420584811?hash=item3ac570756b:g:ZbYAAOSwuhhXWv~f
« Last Edit: June 23, 2016, 10:41:21 pm by Signal32 »
 

Offline MosherIV

  • Super Contributor
  • ***
  • Posts: 1336
  • Country: gb
Re: Precision Power Supply
« Reply #15 on: June 23, 2016, 11:04:36 pm »
"Then how do some powersupplies come with a function of switching the modes between constant current and constant current such as a cheap power supply from china called "YIHUA PS-3010D""

"He just meant you can't do both at the same time."

That is what I meant. PSUs are normally designed to work in 1 or the other mode.
As I said before, a bench PSU is normally Constant Voltage but when you reach the set current limit, they switch over into Constant Current mode.
With a Constant Current PSU, when it reachs the max available voltage, it is then stuck in constant voltage mode.

Personally, I found Constant Voltage easier to understand (and probe ) as a beginner.

"The 1mV/mA goal is very ambitious and you're unlikely to achieve it cheaper in a one-off design"
I do not agree with that. As I said, go for an Op-Amp driven circuit, they have the closed loop bandwith to achieve the control you are after. (also may cause you problems with response time and overshoot but do not worry about that for now)

Stay away from switching PSU circuits, the inherent switching noise will make you goal of mV/mA resolution very difficult to acheive.
Go for linear PSU.
Over rate the transformer, you asked for 24W - go for 35W (look for 35VA) , this keeps you well out of the ripple/magnetic saturation limits of the magnetics.
 

Offline Signal32

  • Regular Contributor
  • *
  • Posts: 234
  • Country: us
Re: Precision Power Supply
« Reply #16 on: June 23, 2016, 11:24:47 pm »
"The 1mV/mA goal is very ambitious and you're unlikely to achieve it cheaper in a one-off design"
I do not agree with that. As I said, go for an Op-Amp driven circuit, they have the closed loop bandwith to achieve the control you are after.
There's a reason why ~$100 and even ~$200 power supplies have only 2 decimal digits. Here are some considerations:
For 20.000v you'll need a 16 bit DAC which + good PCB layout (x2 for current control) These can be somewhat expensive in one-off quantities.
0.001v is 0.005% of 20V, which gives you an idea of what tolerances / required thermal stability you'll be working with.
If you're not using a DAC to set the voltage / current, but a pot, when the temperature of the pot will change you'll see a noticable change in the output voltage because of the pot's thermal drift.
And that's only the start of it.
 

Offline Marco

  • Super Contributor
  • ***
  • Posts: 4700
  • Country: nl
Re: Precision Power Supply
« Reply #17 on: June 24, 2016, 02:02:02 am »
Even fast current regulation is relative ... they are generally not designed to be turned on into a low impedance load (like a diode) and not exceed the current limit.
 

Offline System Error Message

  • Frequent Contributor
  • **
  • Posts: 469
  • Country: gb
Re: Precision Power Supply
« Reply #18 on: June 24, 2016, 07:00:28 am »
is it normal for a bench PSU of constant voltage to drop by .5V under load? such as with 2A load?
 

Offline Signal32

  • Regular Contributor
  • *
  • Posts: 234
  • Country: us
Re: Precision Power Supply
« Reply #19 on: June 24, 2016, 07:22:31 am »
is it normal for a bench PSU of constant voltage to drop by .5V under load? such as with 2A load?
No, it is not normal.
If you're measuring directly at the output terminals, then there might be a problem with your supply. A tiny drop might be expected (0.02V)
If you're measuring at the end of the cables that you plug into your supply, yes it's normal, the voltage drop is across the cable.
Does the power supply voltage display actually say that the voltage drops ?
 

Offline System Error Message

  • Frequent Contributor
  • **
  • Posts: 469
  • Country: gb
Re: Precision Power Supply
« Reply #20 on: June 24, 2016, 08:51:04 am »
Im measuring from my router, its one of those meant to use in a datacenter.

The output voltage on the bench PSU says 24.5V. The router reads 24.0V and is trying around 2A. Both the router's and PSU's amp display are the same.

During lower loads around 1.3A the voltage drop is only about .2 or .3V. When i measure with a multimeter with no load the voltage is accurate to about .1V variance.

edit: I just measured the output with wattmeter with my router running. Output reports 24.5V, multimeter says 24.3V at output and router says 24V. The voltage isnt a big issue since the router is designed to operate from 13V to 36V but i thought i would test the supply under load.
« Last Edit: June 24, 2016, 08:57:56 am by System Error Message »
 

Offline Signal32

  • Regular Contributor
  • *
  • Posts: 234
  • Country: us
Re: Precision Power Supply
« Reply #21 on: June 24, 2016, 09:20:01 am »
Im measuring from my router, its one of those meant to use in a datacenter.

The output voltage on the bench PSU says 24.5V. The router reads 24.0V and is trying around 2A. Both the router's and PSU's amp display are the same.

During lower loads around 1.3A the voltage drop is only about .2 or .3V. When i measure with a multimeter with no load the voltage is accurate to about .1V variance.

edit: I just measured the output with wattmeter with my router running. Output reports 24.5V, multimeter says 24.3V at output and router says 24V. The voltage isnt a big issue since the router is designed to operate from 13V to 36V but i thought i would test the supply under load.
What type of bench PSU is it ? I'm understanding that it's the standard adjustable voltage / current kind. Or is it fixed-output 24V ? The obvious question would be why not use a more efficient fixed voltage PSU ? If it is a fixed voltage PSU, the voltage drop is not that surprising.
 

Offline System Error Message

  • Frequent Contributor
  • **
  • Posts: 469
  • Country: gb
Re: Precision Power Supply
« Reply #22 on: June 24, 2016, 09:29:11 am »
Its a PSU that allows you to vary the output voltage and current limit. Its a big, heavy and bulky PSU with 2 adjustable outputs that can also operate in series/parallel/independent, and a constant 5V 4A output.

Im using the PSU because im trying to test it. My router just so happens to have such a big voltage operating range and if i want to adjust the load all i have to do is run a throughput test within the CPU itself and perhaps let it heat up too.
 

Offline MosherIV

  • Super Contributor
  • ***
  • Posts: 1336
  • Country: gb
Re: Precision Power Supply
« Reply #23 on: June 24, 2016, 12:01:57 pm »
"For 20.000v you'll need a 16 bit DAC"
The OP has not specified what type of control yet. Yes, I agree it will need a high order DAC, probably a 24bit if you want to work around DAC non-linearity if he wants digital control. Personally, I would advise againt digital control, digital adds noise into a system and at mV/mA you need to avoid noise at all costs.

"here's a reason why ~$100 and even ~$200 power supplies have only 2 decimal digits."
"0.001v is 0.005% of 20V, which gives you an idea of what tolerances / required thermal stability you'll be working with."
The OP has asked for help in designing/building a PSU to mV/mA - so let try to help him.
It is good in pointing out the issues, how do we over come them?

"when the temperature of the pot will change you'll see a noticable change in the output voltage because of the pot's thermal drift. "
The OP has not said that thermal drift is important. I agree that this need to be considered when working to mV/mA resolution.
So we need to make sure the design includes a high precision V reference with low thermal drift AND that the pots used also have low thermal drift.

(Still cannot believe UK voted out  |O )
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf