Dummy Load II

[jocanon]

I am thinking if its just a matter if convience then let's just do whatever is easiest. I know that it works with the separate supply, so I would say just keep it the way it is (except for the oscillation of course).

 

[Mr RB]

As for powering the dummy load logic off the load, I was thinking if we try it we should do what Mr RB was saying and install a heavy duty switch in series down stream from where the logic gets it's power.
...

Sorry I did not explain that clearly enough. My switch is in the control circuit, it's just a small current switch that switches the biasing of the power transistors off. No need for a heavy duty switch.

...
So since that problem was gone I tried mr RB's method. It seems pretty good except there is a lot of ringing (osc.) when there is ripple on the 24 volts.

With a heap of integration after the opamp (before the FET gates) there should be no ringing in response to target PSU ripple. The voltage on the FET gates should be very constant due to the (relatively) big cap there.

The control amp will modulate its output up and down to respond to the fluctuating current sense voltage, but that is what it is supposed to do. The main thing is the FETs don't turn on an off in response to the target PSU ripple, which would be very very bad. The opamp turning on and off is perfect.

 

[ronv]

**broken link removed**Here are some pics of different time constants on the FET gates during power up of the 24 volt supply. The ringing bothers me - maybe it shouldn't. What do you all think. As you can see it decreases as the time constant gets shorter until it is back to what we have now.
J, I added the voltage regulator for 2 reasons-well 2 1/2.
1- the present circuit is sensitive to noise on the 12 volts. So as everyone has suggested I added the regulator.
2- When the 12 volts is already on when you power up the 24 volt supply the current goes very high with Mr. RB's idea. Adding the regulator running off of the DUT assures the 12 volts comes on after the 24 and has less noise on it than perhaps your 12 volt supply. Note: I also placed a cap on the pot to futher filter it and to make sure it turns on slow.
1/2 - we can then use the 12 volt supply for the ammeter.
So what do you guys think?

PS. Did you try the cap across the pot to see if that helped stabilize the readings?

 

[jocanon]

...PS. Did you try the cap across the pot to see if that helped stabilize the readings?

ronv, not yet. I have been crazy busy, have company coming over for dinner tonight. I will do it first thing when I get off work tomorrow and let you know the result.

Jeremy

 

[Mr RB]

Why have the integrator time constant so short? I use something more like 2k7 to 10uF or 47uF. Ideally it should be significantly slower than cyclic events on the target PSU like 100/120Hz ripple. That way the FET gates remain at a constant DC bias even if the target PSU has a 120Hz ripple component.

 

[ronv]

I must have misunderstood this response.

https://www.electro-tech-online.com/threads/dummy-load-ii.132379/#post1104888

 

[Dusey52]

ronv, did you decide against swapping out the op amp for one with lower Vio?

Another possiblilty for powering the ammeter is an isolated DC to DC converter. There are many of them on Ebay for < $10, some as cheap as $3.

 

[()blivion]

I must have misunderstood this response.

Dummy Load II

He must have changed his mind. Also, your link is dead, mine works though it's the bottom of *THIS* post, right?

J, I added the voltage regulator for 2 reasons-well 2 1/2.
1- the present circuit is sensitive to noise on the 12 volts. So as everyone has suggested I added the regulator.
2- When the 12 volts is already on when you power up the 24 volt supply the current goes very high with Mr. RB's idea. Adding the regulator running off of the DUT assures the 12 volts comes on after the 24 and has less noise on it than perhaps your 12 volt supply. Note: I also placed a cap on the pot to further filter it and to make sure it turns on slow.
1/2 - we can then use the 12 volt supply for the ammeter.
So what do you guys think?

I think it's looking good. But I would like you to test some things for me if you can.

(1) Try and set up the DUT/supply with a high(ish) series resistance. This is to check and see what might happen if the DUT can't supply current enough to raise it's output voltage. I want to know what the dummy load control circuit will do when handed these lemons.

(2) Can you put a significant delay on the dummy load control circuits power. To test what happens if the DUT starts up fully before our circuit has power. This was more or less the problem with my attempt to power the circuit from the load.

 

[ronv]

Dusey, I have put the opa234 in the latest. It looks really good.
I'll take a look at ebay. J, has 12 volt supplies I think, but I'm curious.
() Good idea, I'll try it. I went back to a zener cause the 317 was a little flakey. Might be my model. So with the zener the logic will always be a little behind the DUT. I think with this one no 12 volts is no FET on. I bet you had PFETs where when they were grounded they were on?
I'll try the resistor now.

 

[ronv]

Tried the resistor. Pretty interesting. Initially the 24 volts goes to 24 volts. Then the 12 volts comes up and it starts to draw current from the 24 volts and the voltage drops. At some point the 12 and the 24 are the same voltage because it's below the zener voltage. So what happens is the gate voltage drops, so the current drops until the voltage doesn't reach the gate threshold voltage. I ran it with just the 2 fets from 5.5 amps down to a few ma. I was suprised it behaved.

 

[ronv]

Dusey, Those are cool. Didn't know they made such a thing!

 

[()blivion]

I bet you had PFETs where when they were grounded they were on?

Oddly enough, no.

I was having problems running off the DUT with the very first N-FET version. But, to be fair, I wasn't using any regulation at all. My supply was just 12V from the DUT. It might have been my Op-Amps too. The thought crossed my mind that when under-powered they might hold their outputs high for some strange reason, or just because they are running out of the min voltage spec. I should bread board a test circuit with the exact part number I was using, and a fet others as a control, and see what happens. With no load on the output pin, even a little leakage current could pull the outputs to high enough voltage that it makes the FETs short. Threshold of the FETs was around 2.5v-3V. I imagine it would be quite easy for the 12v powered Op-Amps to hit that by accident, especially when not powered correctly. Not having regulated power probably didn't help

While I'm on the subject, the 3-terminal regulator will likely overheat if it has to drop 12 volts off of the 24 supply without a heat sink. As I understand linear regulator operation, the farther the input voltage is from the target voltage for a given current, the more watts it will dissipate because of V[SUB]drop[/SUB] times current. I recommend a resistor before the regulator. Or something else to try and share some of the extra power. That is... if we are still going to run with a 3-term regulator, which it looks like we're not.

We really don't need that much current. Just enough to power the Op-Amps + the reference/control. Maybe... 100mA~200mA max? A simple high power zener+resistor+filter cap will work fine I'm sure.

Could also use a fixed regulator model if you have one. Or make your own one transistor regulator. Like this...

 

[Mr RB]

I must have misunderstood this response.
...

No, chalk it up to my casually mentioning a few points at the same time, which is a bit fuzzy.

Re what I said at the bottom;
"Also the integrator does not need to be THAT big, it just needs to have a response time significantly slower than the control loop response and slower than any regular noise ripple etc on the feedback signal."

The important bit is the integrator needs to be big enough so the FETs remain at a constant bias even when the target PSU has ripple, the last thing you want is the FETs trying to switch themsleves on and off to respond to PSU ripple and mains ripple.

Normally I would have the time constant on the output integrator slow enough so it won't be responding to 100Hz or 120Hz mains ripple on the target PSU. The opamp itself will respond to the mains ripple and vary up and down at 120Hz (maintaining an average FET bias) but when that is happening the voltage on the FET gates should ideally be DC, so they act like a massive power resistor of a fixed value, giving the correct average current even though some ripple exists on the target PSU.


If you wanted to add that load on/off switch, it can be added as part of the integrator so it cuts the voltage to the FETs base. So switching the switch on will ramp the current gently up to max, switching it off will ramp the current down to zero. (Assuming both target PSU and opamp PSU are running).

 

[jocanon]

hey guys, I am going to have to put off testing the dummy load for one more day...sorry, i had a few orders come flooding in and I need to put together the packages so I can get them in the mail in the morning. If you can believe it, I actually made more at solidhobby today than at the CPA firm I work at! Still not ready to quit my day job just yet though I promise I will do the tests just as soon as I have a spare minute...thanks everyone for the help and input! Oh, and I would have had a little more time to probably be able to do the tests tonight anyway, but when I got home the duplex connected to my unit (which is vacant) had a broken pipe and flooded the entire duplex! I had to deal with that for a while.

 

[ronv]

All that fun and tax season soon.

Once upon a time you talked about a 1700 watt load box. Can you elaborate on what supplies you want to test?
I'm thinking we would need to come up with a way to limit the current at the high voltage. If it is just one supply we could do it digitally but if it is a whole range of supplies it would need to be something that limits the current based on the supply voltage. Right now the pot just won't go higher than about 60 amps. I don't think you want to design it so it can do 60 amps at 50 volts - thats 3000 watts.

 

[jocanon]

Don't remind me (tax season)...

Correct you are, I don't think I will be testing 3000 watts, unless I hook it up to a 240v dryer socket to test for someone overseas...still I don't see myself doing that, so we don't even need to go there...there is not 1 specific power supply I am looking at yet, there are a few. Basically, the thing is, the PSU I am currently using is quickly becoming obsolete and not available in bulk anymore. I still have a large number of them in stock so I am good to go for many months, but soon I will need to re-tool and find another model. Some of the candidates I am looking at go up to 60+ amps.

edit:
If I run tests on 50v supplies it will most likely be at a lower amperage because a regular household circuit if I am not mistaken will only go up to about 15 amps before it trips the breaker. I am already pulling about that to get 1300 watts DC.

 

[ronv]

Maybe you mean 60 amps, not 60 watts?

Yes I understand the lower current at 50 volts, but right now there is nothing to protect you if you forget and leave the setting at 50 or 60 amps. (constant current)
It is easy to make it 60 amps at 24 or 12 volts - harder to make the current automatically go down when the voltage goes up past 24 volts.

 

[jocanon]

Maybe you mean 60 amps, not 60 watts?

Yes I understand the lower current at 50 volts, but right now there is nothing to protect you if you forget and leave the setting at 50 or 60 amps. (constant current)
It is easy to make it 60 amps at 24 or 12 volts - harder to make the current automatically go down when the voltage goes up past 24 volts.

Yes, corrected the watts to amps...I see what you are saying. We can just design it assuming I will not test 50v supplies. If I do end up testing them, I will just do so at my own risk and be very careful. I don't have any immediate plans to test 50v supplies. I could always test them as two separate 24v supplies too.

 

[ronv]

OK, so say 26 volts at 60 amps maximum?

 

[jocanon]

OK, so say 26 volts at 60 amps maximum?

Yep, sounds good.