Author Topic: Exploring the potential of a basic CC Load  (Read 3551 times)

0 Members and 1 Guest are viewing this topic.

Offline enut11Topic starter

  • Frequent Contributor
  • **
  • Posts: 933
  • Country: au
  • Love building/modifying/restoring test equipment
Exploring the potential of a basic CC Load
« on: June 20, 2018, 06:58:57 am »
I had previously built 2 Constant Current loads based on a Chinese 4-MOSFET PCB and they work well.
https://www.eevblog.com/forum/projects/upgraded-lm324-based-150w-72v-10a-electronic-load/

Commercial Instruments offer other functions such as Constant Voltage, Constant Resistance or Constant Power, to name a few. I want to explore some of these functions using a PCB sourced on eBay that uses a single MOSFET.
https://www.ebay.com.au/itm/DC-1V-100V-Constant-Current-Source-Electronic-Load-Board-75W-0-10A-Power-Tester/322376631212?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2057872.m2749.l2648

enut11
an electronics nut from wayback...
 

Offline enut11Topic starter

  • Frequent Contributor
  • **
  • Posts: 933
  • Country: au
  • Love building/modifying/restoring test equipment
Re: Exploring the potential of a basic CC Load
« Reply #1 on: June 20, 2018, 07:14:48 am »
This article on the Keysight website offers a simplified explanation of some of these modes of operation.
https://community.keysight.com/community/keysight-blogs/general-electronics-measurement/blog/2016/09/16/how-does-an-electronic-load-regulate-it-s-input-voltage-current-and-resistance

If you compare the Constant Current operation (Fig1 in the article) to the Constant Resistance operation (Fig 5 in the article), you will see some similarities. Ie, by simply changing the OpAmp positive input from a Vref (CC) to a sampled input voltage (CR) is all that is needed to change between these modes.

enut11
« Last Edit: June 21, 2018, 09:08:49 am by enut11 »
an electronics nut from wayback...
 

Offline JS

  • Frequent Contributor
  • **
  • Posts: 947
  • Country: ar
Re: Exploring the potential of a basic CC Load
« Reply #2 on: June 20, 2018, 07:19:47 am »
Hi, I see you keep looking into this... Seems like a DC dummy bug got into your system. I've done and you saw a CV/CC one. To do CR might be tricky to get stable, but feeding the sense voltage to the set current should do. To get true CP you need a multiplier or a divider inside the loop, hard to do precisely and stable in analog, I guess one approach for this would be two ADCs, a DAC and a CPLD and run the control loop on the digital domain. I've seen first order approximations to limit the power where I+V was controlled but not close enough to I×V for this application, unless a very tight working area is defined and parameters adjusted for that point.

How are you planning to archive the different control variables?

JS

If I don't know how it works, I prefer not to turn it on.
 
The following users thanked this post: enut11

Offline enut11Topic starter

  • Frequent Contributor
  • **
  • Posts: 933
  • Country: au
  • Love building/modifying/restoring test equipment
Re: Exploring the potential of a basic CC Load
« Reply #3 on: June 20, 2018, 07:39:22 am »
Hi @JS. Yes, I think I got a Load bug. I had so much fun with the two CC Loads built previously that I wanted to go further with this theme. In fact your CC/CV posting was the inspiration for me to look at other uses for these circuits. After all, 90% of the effort goes into the chassis and hardware, so why not a multifunction instrument?

Baby steps at this stage. I come from a long line of analog experimenters so will take the simple approach where possible.

The first step that I have taken here is to nut out the circuit diagram for the XHDZ-FZ-2g Load. It looks a bit more complicated than previous designs but all the basics are there: TL431 Vref, LM258 linear opamp, MOSFET, 10mOhm shunt. The second 1/2 of the LM259 opamp appears to be a Source voltage amplifier, presumably because of the low value shunt resistor.

Not sure of the purpose of the LM393 Comparator. May be some form of protection??
enut11
« Last Edit: June 20, 2018, 07:53:22 am by enut11 »
an electronics nut from wayback...
 

Offline enut11Topic starter

  • Frequent Contributor
  • **
  • Posts: 933
  • Country: au
  • Love building/modifying/restoring test equipment
Re: Exploring the potential of a basic CC Load
« Reply #4 on: June 20, 2018, 07:43:14 am »
The PCB pix
an electronics nut from wayback...
 

Offline JS

  • Frequent Contributor
  • **
  • Posts: 947
  • Country: ar
Re: Exploring the potential of a basic CC Load
« Reply #5 on: June 20, 2018, 07:56:04 am »
The 393 input doesn't look quite right, but it could be doing some over voltage protection or limiting the current as voltage gets higher. My design did it the other way around, as it works in CV till the set current as maximum, this is more sage I guess but mine also had thermal shut down so it should hold itself.

If you are going fully analog leave the CP mode for last and run some sims so you get it closer without blowing so much stuff...

JS

If I don't know how it works, I prefer not to turn it on.
 

Offline enut11Topic starter

  • Frequent Contributor
  • **
  • Posts: 933
  • Country: au
  • Love building/modifying/restoring test equipment
Re: Exploring the potential of a basic CC Load
« Reply #6 on: June 20, 2018, 09:29:06 pm »
@JS, yes the LM393 usage seems odd but should not affect 'normal' operation. Maybe reverse polarity protection??

Constant Resistance Mode
If we now look at Fig 5 from the Keysight Notes above, to change from CC to CR modes requires a simple mod to the PCB. That is, the 10K pot must be disconnected from the TL431 Vref and connected to the load DUT input via a resistor. The value of this resistor will determine the Constant Resistance that can be dialed up using the pot. What we have to keep in mind is the watts generated so we don't blow up the MOSFET!
enut11
an electronics nut from wayback...
 

Offline JS

  • Frequent Contributor
  • **
  • Posts: 947
  • Country: ar
Re: Exploring the potential of a basic CC Load
« Reply #7 on: June 20, 2018, 09:40:10 pm »
CR seems simple but with two feedback path working together it can get unstable quite easily... Experimenting with compensations in the sim is quite easier than in real life

I dont remember how inputs are named in the 393, I assumed closed connection on the open collector when + in was higher than the other, but now I'm thinking you could bebright and that's thebother way aroubd but still doesn't make much sense as internal diode will kick in anyway.

JS

If I don't know how it works, I prefer not to turn it on.
 

Offline enut11Topic starter

  • Frequent Contributor
  • **
  • Posts: 933
  • Country: au
  • Love building/modifying/restoring test equipment
Re: Exploring the potential of a basic CC Load
« Reply #8 on: June 20, 2018, 10:46:43 pm »
CR seems simple but with two feedback path working together it can get unstable quite easily... Experimenting with compensations in the sim is quite easier than in real life

JS

Suggestions on how you sim a circuit welcome.

What range of Constant Resistance?
If we look at the specs of a commercial 150W load, eg the Maynuo 9812, the CR mode has a wide range or 0.03-10K ohms. This is only available through range switching so to start I decided to go with <100 ohms and <50 watts.
an electronics nut from wayback...
 

Offline enut11Topic starter

  • Frequent Contributor
  • **
  • Posts: 933
  • Country: au
  • Love building/modifying/restoring test equipment
Re: Exploring the potential of a basic CC Load
« Reply #9 on: June 20, 2018, 11:05:59 pm »
The attached spreadsheet pic shows one way of achieving variable resistance by using the standard 10K pot on the CC Load. By using an input resistor of 22K, the load resistance can be varied from 1.1 ohm to about 108 ohm. Resistance vs pot rotation is non-linear but should still be usable.

At an input voltage of 10v, the 50 watt limit is reached at mid pot position. This is where a panel meter that displays Watts is very handy as the power can rise rapidly with input voltage. I will probably be using this panel meter in the final version of the Load. See Reply #12 in the following link on how to modify it to read down to zero volts.
https://www.eevblog.com/forum/projects/smart-panel-meter-can-it-be-modified-to-read-down-to-zero-volts/msg1540907/#msg1540907
enut11
« Last Edit: June 21, 2018, 08:03:42 am by enut11 »
an electronics nut from wayback...
 

Offline JS

  • Frequent Contributor
  • **
  • Posts: 947
  • Country: ar
Re: Exploring the potential of a basic CC Load
« Reply #10 on: June 20, 2018, 11:09:43 pm »
I use multisim for most of my simulations, multisim blue is free from mouser and you can run the sim as your bench with virtual instruments.

JS

JS

If I don't know how it works, I prefer not to turn it on.
 
The following users thanked this post: enut11

Offline enut11Topic starter

  • Frequent Contributor
  • **
  • Posts: 933
  • Country: au
  • Love building/modifying/restoring test equipment
Re: Exploring the potential of a basic CC Load
« Reply #11 on: June 21, 2018, 01:49:22 am »
@JS, you mentioned possible instability in modifying this Load for CR. Can you elaborate on that please?
enut11
an electronics nut from wayback...
 

Offline JS

  • Frequent Contributor
  • **
  • Posts: 947
  • Country: ar
Re: Exploring the potential of a basic CC Load
« Reply #12 on: June 21, 2018, 02:39:21 am »
You have an outer voltage to current feedbac loop and an inner current feedback loop. When voltage gets higher current gets higher, the source could sag down if it has a highish output impedance, then voltage gets lower which would set a lower current, then current gets lower, voltage gets higher and so on. Certain loads will induce some oscillations.

To deal with this I'd try to make the voltage to current stage much slower than the current loop so settling is slower but current reacts fast in relation to the sensed voltage, and folows its response. That would behave like an inductive load, current responding slow to the input voltage change. But there's always some input condition where it gets tricky to stabilize. Dealing and testing with it is easier in a simulator as you can generate any source you like, you can test different input conditions and find a feedback configuration that works nice. Look my design of the CV/CC load where I did so and I came up with a reasonable fast and stable config, I've only run simulations on the final circuit, the one I've build was a previous version.

JS

If I don't know how it works, I prefer not to turn it on.
 
The following users thanked this post: enut11

Offline enut11Topic starter

  • Frequent Contributor
  • **
  • Posts: 933
  • Country: au
  • Love building/modifying/restoring test equipment
Re: Exploring the potential of a basic CC Load
« Reply #13 on: June 21, 2018, 02:49:51 am »
OK @JS. Good explanation. Will have a good look at your design soon.
How does a few uF across the 10K pot sound. Will this reduce unwanted oscillations in the Load?
enut11
an electronics nut from wayback...
 

Offline schmitt trigger

  • Super Contributor
  • ***
  • Posts: 2205
  • Country: mx
Re: Exploring the potential of a basic CC Load
« Reply #14 on: June 21, 2018, 03:33:08 am »
Looking forward  for what you come up with for the CR Mode.
I once attempted to design one but failed, as others have mentioned it became unstable.
:-(

Unfortunately I became distracted with other things and never got around to completely debug the design.

 

Offline enut11Topic starter

  • Frequent Contributor
  • **
  • Posts: 933
  • Country: au
  • Love building/modifying/restoring test equipment
Re: Exploring the potential of a basic CC Load
« Reply #15 on: June 21, 2018, 05:00:30 am »
Initial experiment with CR Mode was successful.

I cut the PCB track from the TL431 Vref and soldered a 22K resistor between pot and Load input (Drain). I also soldered a 1uF cap across the pot to dampen the input changes to the LM258 Opamp.

Using a variable power supply as input, I was able to set the Load pot for a safe current draw. Then, varying the voltage on the power supply produced a linear change in current, just like a resistor!  :).

Also, the resistance numbers from the spreadsheet above proved to be close to reality, a nice bonus. The practical span was 1-100 ohms using the 22k input. A switching arrangement of input resistances would proved a wider span of loads.
enut11

All tests were under 1A. Next step is to test it at higher currents.
enut11
« Last Edit: June 21, 2018, 09:11:45 am by enut11 »
an electronics nut from wayback...
 

Offline enut11Topic starter

  • Frequent Contributor
  • **
  • Posts: 933
  • Country: au
  • Love building/modifying/restoring test equipment
Re: Exploring the potential of a basic CC Load
« Reply #16 on: June 21, 2018, 07:58:29 am »
Well, the Load passed the 50 watt stress test in CR Mode, 7.3A/7v, so all appears to be ok. Pot was set to max (~1 ohm across the MOSFET).

With a 22k input resistor and a 10K pot, the effective range of resistance load that can be dialed up is between 1 ohm and 100 ohms.

Increasing Rin will increase the top to over 100 ohm loads. Reducing Rin will decrease the bottom end and allow testing with <1 ohm loads. A practical value for Rin is probably between 10K and 100k.
enut11
an electronics nut from wayback...
 

Offline JS

  • Frequent Contributor
  • **
  • Posts: 947
  • Country: ar
Re: Exploring the potential of a basic CC Load
« Reply #17 on: June 21, 2018, 01:43:25 pm »
Nice, you are hands on with this, you do have a cap at the output of the pot, right? That would do the kind of response I was talking.

You could limit the low level with a resistor underneeth the pot, so you have one range from <1Ω to say 50Ω and a secons range from 50Ω to 3k or something, giving better control overall.

Other thing you could do is to add a CC limit to the CR mode, fixed with a diode limiting the voltage at the wiper or variable using an opamp with a diode in the loop for clamping at the set current. ~1Ω can go to a pretty high current very fast if voltage moves arround, and if variable it could protect the DUT as you could imagine.

JS

If I don't know how it works, I prefer not to turn it on.
 

Offline JS

  • Frequent Contributor
  • **
  • Posts: 947
  • Country: ar
Re: Exploring the potential of a basic CC Load
« Reply #18 on: June 21, 2018, 02:52:32 pm »
I read more posts here but I can't see them now. Sometimes you want a smaller cap plus a bigger cap with a series resistor, to make two compensation networks, but it depends on how slow and how much overshoot you can live with. I don't have my computer now as I'm on the road for a week now, and doing the math for this is quite extensive as ypu want to test with different DUTs.

If you are only interested in steady state tests you can do slow and safe, a few μF after the pot could do but the working condition will change with the setting as the output impedance of the pot changes, adding a resistor inseries with it will reduce the problem, other way would be to put the cap in parallel with the pot and maybe change it when you switch ranges, so it's always under control, you could choose a different compensation for the different ranges.

JS

If I don't know how it works, I prefer not to turn it on.
 

Offline enut11Topic starter

  • Frequent Contributor
  • **
  • Posts: 933
  • Country: au
  • Love building/modifying/restoring test equipment
Re: Exploring the potential of a basic CC Load
« Reply #19 on: June 23, 2018, 06:12:59 am »
Nice, you are hands on with this, you do have a cap at the output of the pot, right? That would do the kind of response I was talking.

You could limit the low level with a resistor underneeth the pot, so you have one range from <1Ω to say 50Ω and a secons range from 50Ω to 3k or something, giving better control overall.

Other thing you could do is to add a CC limit to the CR mode, fixed with a diode limiting the voltage at the wiper or variable using an opamp with a diode in the loop for clamping at the set current. ~1Ω can go to a pretty high current very fast if voltage moves arround, and if variable it could protect the DUT as you could imagine.

JS

Thanks @JS. The 1uF cap is across the input to the pot and seems to work for the CR Mode but I have not looked at it on a scope yet.

Good idea to jack up the earthy side of the pot to add CR ranges and improve setability.

BTW, I had another look at your CC/CV circuit but without an explanation it is a bit difficult to understand. Will try again.

Next: Constant Voltage Mod
I am currently working on the Constant Voltage mod for this PCB but it is proving more difficult than anticipated.
Constant Voltage is used to test current sources by clamping the voltage at a set level, regardless of current. The Keysight notes from Reply #1 show a basic schematic in Fig 3. By implementing this into my CC Load, the circuit went into a low frequency oscillation. Still working on it.
enut11
an electronics nut from wayback...
 

Offline JS

  • Frequent Contributor
  • **
  • Posts: 947
  • Country: ar
Re: Exploring the potential of a basic CC Load
« Reply #20 on: June 23, 2018, 02:40:45 pm »
Ask the questions in the topic ans I'll explain further...

JS

If I don't know how it works, I prefer not to turn it on.
 

Offline void_error

  • Frequent Contributor
  • **
  • Posts: 673
  • Country: ro
  • I can transistor...
Re: Exploring the potential of a basic CC Load
« Reply #21 on: June 25, 2018, 07:01:42 pm »
Had a look at the Keysight article and I'm wondering how would one digitally adjust the resistance in constant resistance mode? A common digital potentiometer usually has 256 steps (not a lot) and they're quite inaccurate... but there's also the possibility to calibrate the whole thing when using a microcontroller.
Trust me, I'm NOT an engineer.
 

Offline JS

  • Frequent Contributor
  • **
  • Posts: 947
  • Country: ar
Re: Exploring the potential of a basic CC Load
« Reply #22 on: June 25, 2018, 07:45:11 pm »
I left an explanation of the circuit earlier in my topic, feel free to ask for details.

Forgot a little detail, my approach used a CV with a CC intervention, kind of backwards to what you are trying to archive here so not much useful stuff there to your application working on an existing CC load. My first attempt wasbckoser as it used a CC and a CV with an or gate, but you would need to do some mods to your board anyway to add the diode/resistor at the gate.

The approach you might want is to do something like I did in my last approach but backwards, setting a reference current to keep the voltage at the desired value.

JS

If I don't know how it works, I prefer not to turn it on.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf