I am blown away. A really great exposition. Simple, practical and I like the verification bits. I learnt more about thermal resistance of heatsinks here than all the books and articles and datasheets I have been reading.
Excellent video. So good to see electronics theory applied in a practical way. I studied electronics 20 years ago at college as part of my apprenticeship at a telecoms company, and although I came away with useful qualifications I didn't really learn much that I could apply: the course was 95% theory and 5% practical!
You know, I really wish I had found Dave's videos last. After watching these vids, it makes it really really hard to watch the videos from other people, even though they contain good information as well. His teaching style is just so in tune with how I learn and his explanations are in depth, but don't treat the listener like an idiot. This series is just brilliant.
+Joseph Nicholas You're equating age with knowledge of electronics? Odd. So an 82 year old retired CEO of a Fortune 500 company with no experience in electronics is a better candidate for this channel, and a 26 year old with a master's in EE is better for Lorton? Hmmm - how's your logic circuit doing? Swinging toward the arrogant side a bit maybe?
M Miller Let's be honest. Electronics and computers in general are not the most exciting topics. I thoroughly enjoy Dave's ability to turn a mundane subject into something pleasant to watch. His personality is what brings me back again and again to his videos.
Had a vision of factories full of engineers waiting impatiently for Dave's latest video to come out, designing away as he talks, then quick into production, the first on eBay wins!!!
5:19 absolutely important thing no explained at all. This is a video for truly advanced people. I suggest to search Rail to Rail Op amp to understand what wasn't said here.
The other part of the heat dissipation calculation is that the MTP3055 has a thermal resistance between junction and case of 3.13 degrees C per Watt. That means with your 11 W example, that the mosfet junction will be another 3 x 11 = 33 degrees above the case/heatsink - a little over 100 degC with a 20 degC ambient. That's within the maximum operating temperature of 175 degC for the device, but a little toastier than I like. The corollary is that your heatsink/mosfet combination has an absolute maximum power dissipation of around (175-25)/(4.5+3) = 150/7.5 = 20 W with a 25 degC ambient. You might want a better cooling solution if you start looking at higher voltages and currents - 12 V and 2 A would well exceed the capabilities of that heatsink. And to get past 50 W, you'll need a mosfet in a different package.
Notice that the op-amp feedback controls the DC current depending on the PWM input (because the voltage across the 1R is proportional to the current), NOT resistance. For power, P=V*I, which is the same as CC iff the supply voltage is constant (not necessarily true! Think batteries as they discharge). Constant-resistance would be to vary current according to supply voltage according to Ohm's law, i.e. you have to actively maintain the constant resistance since the OPA/NMOS system is CC.
Dale, doing it in that way creates a constant resistance load, so the current will change when the power supply voltage under test changes. Wicth Dave's circuit the current sucked from the power supply is constant.
@cborrero2000 The opamp output gives whatever Gate voltage is required to keep the constant voltage across the load resistor. It's a standard building block circuit.
Dave, you're GOOD! Keep the video's coming. This video is already 7 years old when I saw it this Dec 2917. So you are now 7 years a little bit older since you made this video. But keep up the good work.I LEARNED A LOT FROM IT AND SO DO OTHERS. ATTABOY! Thanks.
I have a remark about the diagram at 16:40. In constant resistance mode, to justify Ohms law, you'll need to know the input voltage and adjust the current accordingly. Since the microprocessor sets the FET current with PWM, it also needs to monitor this input voltage. The inputvoltage (=Voltage across FET and current shunt resistor) can be monitored using a resistive divider over the input connectors. The divided voltage is then connected to a second channel of the microprocessor ADC (or use an analog multiplexer). For constant power, same story.
Agreed, the battery voltage is useful for data logging and as you note required if you want to do constant resistance (or constant power) actively controlled discharging.
Opamp is required to maintain the constant current. With a pot + manual intervention, you could vary the current, but the current will vary the moment supply voltage varies. Using opamp in closed loop will make sure that the current is same irrespective of any input voltage.
Very well explained and executed. ...If you'd only gone one step more and shown how to relate the allowable maximum temperature of the MOSFET to the power dissipation of the heatsink to verify that the heatsink would adequately protect the MOSFET for the maximum current you were allowing the device to draw.
Thanks for the video Dave..very useful...Please give some tips on building higher currents sources like 7 Amps constant current source. Thanks Regards.
Great Idea Dave! I want to build one! A request: Please add 17:17 to the next "how to entertain a geek Highlights! Especially the part about doing it with "intelligent control" ;)
Finding a voltmeter in the right range that works in a grounded configuration has proven very tough. The cx101 appears to unobtanium these days. And I want a digital meter, don't care for analog for this application.
One doesn't have to keep the voltmeter connected all the time, so set the test unit using a regular DMM and then do the testing. It may be more convenient to have the meter built-in...but it isn't required. Or...one could set up fixed resistance value points using carbon or precision resistors as desired with a rotary switch to then set the current to some common values one might want (e.g. 100mA, 250mA, 500mA, 1000mA etc). I find in most cases I don't need an infinitely variable current load to test things.
@brikotube No real reason, I just had a suitable one. Same circuit works with BJT's too. If you battery power the thing then the extra base current requirement of BJT's might be a problem.
Nice explain on the load. One bloke made one of these, but his oscillates.. and he thinks it's supposed to do that! Aaah! I knew that's not right.. your circuit verified that. His has no bypass caps.. thus oscillation. Also, excellent hack on the meter! I had done the same hack to my LED panel meter. Cheerz
Pretty awesome little project. I want to make one but I think I'd like to "crank it up a notch" and have a microcontroller involved that varies the current, takes measurements and spits them out back to a computer. Press a button, capture data and graph :-). The original use of your board seems to do something along those lines where the uC controls the current, unless I misunderstood.
Awesome as always I would like to see a circuit for constant current and voltage both adjustable to hook to a server power supply or two supplies for 24v up too say 100a DC thanks keep them coming
24v at 100a is 2.4kw thats a lot of poer to be dissapaiting in semiconductors and a heatsink. I woud use some enameld copper wire in a tub of water, not so convinient but a lot cheaper
It occurs to me that what I said earlier on transient is not strictly true. The op-amp isn't purely a proportional controller (ideally it is, in reality it has an s-dependent transfer function). You could probably add a capacitor somewhere to move a pole and increase the damping. I'd have to sit down and look at it. (Also, that's assuming you mean transient response to a control voltage change, not a change of the supply voltage-that's a different problem I've been thinking about.)
Excellent video, and with the knowledge gained I am going to build a project I have been putting off: an intelligent battery charger! I have some I2C DACs from Microchip which should be very suitable for setting the opamp input voltage.
I always look at the worst case scenario whenever I re-design a circuit to suit my needs and in this case I would have soldered a 1M resistor from the wiper to ground. Similar to input conditioning when I do programming. It avoid unforeseen bugs and aids troubleshooting. I also make a point of establishing and highlighting test-points within the circuit since nobody I know troubleshoots at a leisurely pace.
Great video, quick question though, with the 1 ohm resistor in place can you not just connect the pot to the mosfet and change the gate voltage to increase/decrease the current? This works in my simulations so I'm struggling to understand why the Op-Amp is required. Cheers!
Hi Dave. Would it be possible to use an op amp and a logic mosfet to give a variable constant current source rather than a load? I have built a similar circuit with the LM317 and fixed resistors selected by a rotary switch to give me 10mA, 20mA, 30mA, 40mA etc. But it would be nice to have a completely variable current source from 1mA up to 1A perhaps? Cheers, John.
I dunno if this will help but when i use pc power supply as a voltage source i noticed you dont get much out of it unless you have a load on the 5v side(maybe 12v side, it's been a while) so just hit up radio shack and grab a 5w load resistor, strap it to some old cpu heatsink w/fan and load that puppy up. =D hope that helps you get your supply into the optimal range your lookin for.
Something like this should be easy to scale up by increasing the number of MOSFETs fitted and by using fan-cooled heatsinks. I use a CPU heatsink from an Xbox 360, although it's not fan-cooled, but I don't need to dissipate 150W like you do! You may have to get creative when clamping a CPU heatsink, though. Mounting back-to-back might make it easier (I have a spare, unused MOSFET under mine, so each TO220 epoxy block faces the other one's metal tab, with legs sticking out opposite ends).
Great video as always. Several years old now but... hoping someone can point me to the video he mentions at the end about doing the power supply curves using the load.
Can this also work as an adjustable current circuit for a power supply ? Say you drop that MOSFET before the Constant Voltage (CV) side of the power supply, what would happen ?
I HAVE A QUESTION... How can I know what kohm potentiometer value should I choose to fully take advantage of the 10 turns???? I have seen videos where people buy the expensive 10 turn pot and only with a few turn they reach the max voltage.
Steady-state error is probably due to a low op-amp open-loop gain (or your current sense resistor isn't precise enough). For transient, do you know any control theory? This is a pretty simple proportional controller with very high gain, so depending on the opamp/NMOS transient can be horrid; you could try a PD controller or a PI controller with a smaller proportional gain (the integral eliminates steady-state error, smaller gain improves transient). It requires several circuit changes, though.
Since MTP3055 is an N channel mosfet, shouldn't you put the load resistor on the drain side? Or does it not matter in this case? I'm trying to replicate something similar with constant power mode as well
If I understand this correctly most of the power is dissipated by the mosfet? Is there a design where the power would be dissipated by resistors? I'm thinking a big ass cap, a pwm controlled mosfet and big ass power resistors. I feel something like that would be easier to make programmable with a uC and scaled up in terms of maximum load.
It would be 2.5A with a rail-to-rail opamp. The lm324 goes up to max VCC - 1.5V so the full swing for a VCC of 5V is 0 to 3.5V, divided by 2 results in the ~1.75V. Why 1.3A instead of 1.75A? I'm sure the answer is buried in the datasheet of the opamp or the tolerance of the resistors, etc
@@catalinvasile9081 You have the right idea. The LM324 isn't going to swing above about 3.5V with a 5V supply. Since the gate threshold voltage for an MTP3055 is specified as somewhere between 2V and 4V, it's pot luck whether the circuit can actually turn on the mosfet. In the worst case scenario needed to sink 2.5A, there will be 2.5V across the current sense resistor, so the source will be at 2.5V and the gate could need as much as 4V above that just to turn on. Then as the mosfet has a minimum forward transconductance of 4A/V, it may require an additional gate voltage of 2.5/4 = 625 mV to supply 2.5 A. That means you have to design for a voltage of at least 2.5 + 4.0 + 0.625 = 7.125 V at the gate. Finally, the LM324 needs another 1.5V "headroom" which shows that you need a 9V power supply for the opamp to accommodate the worst case with that mosfet if you want to sink 2.5A. If the actual circuit limits to 1.3A, that would mean that the LM324 output and the mosfet gate are at about 3.5V and the mosfet source is at about 1.3V, leaving 2.2V as gate-source voltage, which is a bit better than what you might expect as it has a typical gate threshold voltage of 2.7 V according to the data sheet.
Will this circuit work as a chopper for a stepper motor driver that pulls too much current through its h-bridge or do I have to go with something more complex? Adding a heatsink to the h-bridge is not an option.
Hi Dave, Thanks a lot for this video. Do you think that we also could develop an AC version of an electronic load. I know this would be a complete new design. Please let me know your opinion. Best regards from Switzerland, Rene
Hi! I need to build a dummy load for a PC power supply that can draw up to 100A and another unit that can draw up to 30A that are adjustable in steps of 10A and 5A respectively. I know my basic electronics but I would have absolutely no idea on where to start with this. I have heard MOSFETs are very scalable and can achieve this with adequate cooling, but I need a circuit diagram or something so I can make it. Please help!
I need to test an IPSU with 9 different power rails ranging from 1.25V to 5V and current less than 1A would this design be recommend for my case? Thanks.
Hi , I would like to test the discharge procedure of a 24vdc (7 days autonomy )Nicad battery bank . Please advice, which is the best device to use for that application? and how can i buy the same?
Hi Dave I'm trying to use the CX101b LCD but it's a pain in the ass to configure, can you tell me what resistor values you use to get it to read 1999???? I know that your schematic values are not the same as what you use. Thanks.
Will something like this work for testing amperage and voltage at same time on a solar panel with a watt meter? If not could you please make a video of one that would do so for A47 volt panel or higher?
Hi Dave, I put together one of these on a breadboard using similar components, but the transient response seems to be pretty bad. Also, there is steady state error. Any ideas why?
Hello, since i have stumbled upon your blog i found very useful stuff here and learned something. I'm trying to build a small test rig for testing vacuum tubes and transistors under working conditions. So far i have used resistors are constant current load but as usual resistor stack is never big enough and you need to use some high power ones. So i thought why not to use variable current source instead of resistor so i'm wondering with some modifications could this circuit work in a way i imagined it should. Ofc limiting current for supply voltage is only one part of that test rig, other is how to drive gate/grid/base but that's another story and not connected to this video.
Hi - thanks for the video, I actually learnt a lot from this ! Can you give some pointers for research into scaling this up to, say, 60A ? Is the approach to parallel a few of these circuits or use a bigger MOSFET ? Any help would be appreciated.
is it going to work if i power the op-amp from a different supply. did the gate of the mosfit will work. I was thinking to combine the two grounds? Help pls
I could have sworn you had a smaller and simpler dummy load in a previous episode, but I can't find anything. Am I just imagining that or did I miss it when looking through older episodes?
Considering he says, "it's about time I probably built one," and holds up a dummy load at least twice as big as what I remembered, or thought I remembered, that implies he hasn't made one before and my memory is way off. That's why I'm confused.
@ElvisAviator I've found out on how to get the LCD cx101b to work reading millivolts even though it's one millivolt offset, I think is due to the lcd +-1%.
Hi Dave, I need to variably load my small wind turbine while testing it in a wind tunnel. My electronics skills are not that good, I am a simple aerodynamics engineer. Could this little device be scaled up to dissipate about 150W?
Dave I'm having trouble with "drift" with the OpAmp voltage follower. I used a 1ohm, 10% power resistor, an LM324, and a IRF540 Power Mosfet. I am using a 9V power supply for the LM324 (measuring it, it actually shows about 10V) I notice that if I use this to test battery discharging I get "drift". I'll start out at say, a setting for 500ma, and as the battery voltage drops, the current *increases* , and i'm not sure why. I thought perhaps its something to do with the MOSFET changing its conductance or something? I always end up having to adjust the input pot setting over time to lower the current back to 500ma It doesn't make sense to me becasue the OpAmp should technically keep the voltage /current constant even if the resistor drifts, wouldn't it? It almost seems like my OpAmp is drifting. Also, my MOSFET doesn't get hot at all, granted I'm only running 1.5 V at 500ma , and the resistor is .5V of that (I set the OpAmp + input to .5V so I get .5A on the 1ohm resistor), so ya it's only having to handle what, 500mw, not that much I guess.
I am blown away. A really great exposition. Simple, practical and I like the verification bits. I learnt more about thermal resistance of heatsinks here than all the books and articles and datasheets I have been reading.
That dave CAD sure is some piece of sophisticated software you got. :)
Excellent video. So good to see electronics theory applied in a practical way. I studied electronics 20 years ago at college as part of my apprenticeship at a telecoms company, and although I came away with useful qualifications I didn't really learn much that I could apply: the course was 95% theory and 5% practical!
You know, I really wish I had found Dave's videos last. After watching these vids, it makes it really really hard to watch the videos from other people, even though they contain good information as well. His teaching style is just so in tune with how I learn and his explanations are in depth, but don't treat the listener like an idiot. This series is just brilliant.
+Joseph Nicholas You're equating age with knowledge of electronics? Odd. So an 82 year old retired CEO of a Fortune 500 company with no experience in electronics is a better candidate for this channel, and a 26 year old with a master's in EE is better for Lorton? Hmmm - how's your logic circuit doing? Swinging toward the arrogant side a bit maybe?
+M Miller I think you are reading wayyyy too much into what Joseph said come on lol
M Miller Let's be honest. Electronics and computers in general are not the most exciting topics. I thoroughly enjoy Dave's ability to turn a mundane subject into something pleasant to watch. His personality is what brings me back again and again to his videos.
Had a vision of factories full of engineers waiting impatiently for Dave's latest video to come out, designing away as he talks, then quick into production, the first on eBay wins!!!
5:19 absolutely important thing no explained at all. This is a video for truly advanced people. I suggest to search Rail to Rail Op amp to understand what wasn't said here.
Umm no its not that advanced
The other part of the heat dissipation calculation is that the MTP3055 has a thermal resistance between junction and case of 3.13 degrees C per Watt. That means with your 11 W example, that the mosfet junction will be another 3 x 11 = 33 degrees above the case/heatsink - a little over 100 degC with a 20 degC ambient. That's within the maximum operating temperature of 175 degC for the device, but a little toastier than I like.
The corollary is that your heatsink/mosfet combination has an absolute maximum power dissipation of around (175-25)/(4.5+3) = 150/7.5 = 20 W with a 25 degC ambient. You might want a better cooling solution if you start looking at higher voltages and currents - 12 V and 2 A would well exceed the capabilities of that heatsink. And to get past 50 W, you'll need a mosfet in a different package.
Notice that the op-amp feedback controls the DC current depending on the PWM input (because the voltage across the 1R is proportional to the current), NOT resistance. For power, P=V*I, which is the same as CC iff the supply voltage is constant (not necessarily true! Think batteries as they discharge). Constant-resistance would be to vary current according to supply voltage according to Ohm's law, i.e. you have to actively maintain the constant resistance since the OPA/NMOS system is CC.
Dale, doing it in that way creates a constant resistance load, so the current will change when the power supply voltage under test changes. Wicth Dave's circuit the current sucked from the power supply is constant.
Excellent tutorial Dave , Hats off to you .Like your Dave CAD .tells almost everything on this CKT
@cborrero2000 The opamp output gives whatever Gate voltage is required to keep the constant voltage across the load resistor. It's a standard building block circuit.
Dave, you're GOOD! Keep the video's coming. This video is already 7 years old when I saw it this Dec 2917. So you are now 7 years a little bit older since you made this video. But keep up the good work.I LEARNED A LOT FROM IT AND SO DO OTHERS. ATTABOY! Thanks.
2917!? You're from the future!? What's it like there? lol
I have a remark about the diagram at 16:40. In constant resistance mode, to justify Ohms law, you'll need to know the input voltage and adjust the current accordingly. Since the microprocessor sets the FET current with PWM, it also needs to monitor this input voltage. The inputvoltage (=Voltage across FET and current shunt resistor) can be monitored using a resistive divider over the input connectors. The divided voltage is then connected to a second channel of the microprocessor ADC (or use an analog multiplexer). For constant power, same story.
Agreed, the battery voltage is useful for data logging and as you note required if you want to do constant resistance (or constant power) actively controlled discharging.
Opamp is required to maintain the constant current. With a pot + manual intervention, you could vary the current, but the current will vary the moment supply voltage varies. Using opamp in closed loop will make sure that the current is same irrespective of any input voltage.
Awesome little circuit! I had thought about trying to make a constant-current load before but wasn't sure which route to take. New project I think!
@thenaimis Correct, that's what the original board does. Micro with constant current/resistance/power and logging capability.
Very well explained and executed.
...If you'd only gone one step more and shown how to relate the allowable maximum temperature of the MOSFET to the power dissipation of the heatsink to verify that the heatsink would adequately protect the MOSFET for the maximum current you were allowing the device to draw.
Thanks for the video Dave..very useful...Please give some tips on building higher currents sources like 7 Amps constant current source.
Thanks Regards.
Great Idea Dave!
I want to build one!
A request:
Please add 17:17 to the next "how to entertain a geek Highlights!
Especially the part about doing it with "intelligent control"
;)
Excellent...might have to put one of these together.
"from a couple of junk box parts". Like CX101 meters. My junk doesn't resemble that at all :-(
+Juanjo Aparicio Perhaps a 0-1 volt analog panel meter is in your junkbox? It doesn't have to be Digital.
Finding a voltmeter in the right range that works in a grounded configuration has proven very tough. The cx101 appears to unobtanium these days.
And I want a digital meter, don't care for analog for this application.
One doesn't have to keep the voltmeter connected all the time, so set the test unit using a regular DMM and then do the testing. It may be more convenient to have the meter built-in...but it isn't required.
Or...one could set up fixed resistance value points using carbon or precision resistors as desired with a rotary switch to then set the current to some common values one might want (e.g. 100mA, 250mA, 500mA, 1000mA etc). I find in most cases I don't need an infinitely variable current load to test things.
Excellent video. I love these little design projects.
@brikotube
No real reason, I just had a suitable one. Same circuit works with BJT's too. If you battery power the thing then the extra base current requirement of BJT's might be a problem.
Nice explain on the load. One bloke made one of these, but his oscillates.. and he thinks it's supposed to do that! Aaah! I knew that's not right.. your circuit verified that. His has no bypass caps.. thus oscillation. Also, excellent hack on the meter! I had done the same hack to my LED panel meter. Cheerz
Pretty awesome little project. I want to make one but I think I'd like to "crank it up a notch" and have a microcontroller involved that varies the current, takes measurements and spits them out back to a computer. Press a button, capture data and graph :-). The original use of your board seems to do something along those lines where the uC controls the current, unless I misunderstood.
@Nermash Yeah, the Gossen is a damn fine meter!
@SuperRnine PSU positive voltage is connected to the MOSFET gate and the negative voltage to the ground.
Useful, I think I'll build one of these.
I'm surprised not to find anything like this on eBay.
@eeadata Why not? The Opamp will adjust the gatevoltage as well. You need to know the Thresholdvoltage, the amp, the max DS-Voltage and P-dead.
You are better than most of university professors
I like those analog teaching videos you made back then.
Admit it, you just love that Gossen Metrawat and german craftmanship:)
Awesome as always I would like to see a circuit for constant current and voltage both adjustable to hook to a server power supply or two supplies for 24v up too say 100a DC thanks keep them coming
24v at 100a is 2.4kw thats a lot of poer to be dissapaiting in semiconductors and a heatsink. I woud use some enameld copper wire in a tub of water, not so convinient but a lot cheaper
Absolute fantastic... every question I had was answered :-)
I really like your stuff
Greetings from Germany
TOM
It occurs to me that what I said earlier on transient is not strictly true. The op-amp isn't purely a proportional controller (ideally it is, in reality it has an s-dependent transfer function). You could probably add a capacitor somewhere to move a pole and increase the damping. I'd have to sit down and look at it. (Also, that's assuming you mean transient response to a control voltage change, not a change of the supply voltage-that's a different problem I've been thinking about.)
Btw if you change the shunt resistor value to 0.5 ohms, you will have a wider current range up to 2.6 amps.
excellent dave, I love your videos
Excellent video, and with the knowledge gained I am going to build a project I have been putting off: an intelligent battery charger!
I have some I2C DACs from Microchip which should be very suitable for setting the opamp input voltage.
I always look at the worst case scenario whenever I re-design a circuit to suit my needs and in this case I would have soldered a 1M resistor from the wiper to ground. Similar to input conditioning when I do programming. It avoid unforeseen bugs and aids troubleshooting. I also make a point of establishing and highlighting test-points within the circuit since nobody I know troubleshoots at a leisurely pace.
Great video, quick question though, with the 1 ohm resistor in place can you not just connect the pot to the mosfet and change the gate voltage to increase/decrease the current? This works in my simulations so I'm struggling to understand why the Op-Amp is required. Cheers!
@Films4You Sanyo Xacti HD1010,
Thanks a lot Dave.
What is the use of the extra op-amp?
Why does the MOSFET need to be logic-level?
Hi Dave.
Would it be possible to use an op amp and a logic mosfet to give a variable constant current source rather than a load? I have built a similar circuit with the LM317 and fixed resistors selected by a rotary switch to give me 10mA, 20mA, 30mA, 40mA etc. But it would be nice to have a completely variable current source from 1mA up to 1A perhaps?
Cheers, John.
I dunno if this will help but when i use pc power supply as a voltage source i noticed you dont get much out of it unless you have a load on the 5v side(maybe 12v side, it's been a while) so just hit up radio shack and grab a 5w load resistor, strap it to some old cpu heatsink w/fan and load that puppy up. =D hope that helps you get your supply into the optimal range your lookin for.
Great video. Can you suggest equivalent of cx101?
Something like this should be easy to scale up by increasing the number of MOSFETs fitted and by using fan-cooled heatsinks. I use a CPU heatsink from an Xbox 360, although it's not fan-cooled, but I don't need to dissipate 150W like you do! You may have to get creative when clamping a CPU heatsink, though. Mounting back-to-back might make it easier (I have a spare, unused MOSFET under mine, so each TO220 epoxy block faces the other one's metal tab, with legs sticking out opposite ends).
The intro is so 2009 😅
Nice on tho, exactly what i was looking for
If this would be used in lab power supply as a adjustable current source where should the load be connected ?
I was going to build me one of these, but I was going to use a 2N3055 power transistor as the sink element.
It is interesting. what will happen if connect two constant current load in series. Will they go crazy?
Аляксей Вясельеў One of them will win, and control the situation.
Thanx for sharing you knowledge, you're amazing!
It goes all the way up to 1337. Gnarly!
Hello - can you suggest how I may use this to test multiple outlets on an industrial PDU.
Great video as always. Several years old now but... hoping someone can point me to the video he mentions at the end about doing the power supply curves using the load.
DaveCAD :D Where can i download it?
Can this also work as an adjustable current circuit for a power supply ? Say you drop that MOSFET before the Constant Voltage (CV) side of the power supply, what would happen ?
Your junk box must be a lot larger than mine!
Thanks for the very interesting vlog. :)
HEY DEAVE, WHAT A BOUT A NEW VIDEO LIKE THIS, WHAT'S NEW IN POWER SUPLY TESTING!
I HAVE A QUESTION... How can I know what kohm potentiometer value should I choose to fully take advantage of the 10 turns???? I have seen videos where people buy the expensive 10 turn pot and only with a few turn they reach the max voltage.
Can I get the PCB design and BOM to implement it in a holiday DIY project?, Thank you.
Steady-state error is probably due to a low op-amp open-loop gain (or your current sense resistor isn't precise enough). For transient, do you know any control theory? This is a pretty simple proportional controller with very high gain, so depending on the opamp/NMOS transient can be horrid; you could try a PD controller or a PI controller with a smaller proportional gain (the integral eliminates steady-state error, smaller gain improves transient). It requires several circuit changes, though.
I have been wonder if it is possible to using 1000 watt MOSFET with MIC1337 timer (the perfect 50% cylce) while testing power supply at 12V 60A?
I'd love to make one with a microcontroller!
Since MTP3055 is an N channel mosfet, shouldn't you put the load resistor on the drain side? Or does it not matter in this case? I'm trying to replicate something similar with constant power mode as well
Hello Dave. Excuse my ignorance but, why you used a mosfet for the sink element instead of a BJT?. Thanks
If I understand this correctly most of the power is dissipated by the mosfet? Is there a design where the power would be dissipated by resistors? I'm thinking a big ass cap, a pwm controlled mosfet and big ass power resistors. I feel something like that would be easier to make programmable with a uC and scaled up in terms of maximum load.
Why did you calculate a max of 2.5A (given the 1/2 voltage divider between the opamps) and then you got only 1.3A in the actual circuit???
BIll Geo I think it's because of wire resistance. (They were thin).
It would be 2.5A with a rail-to-rail opamp. The lm324 goes up to max VCC - 1.5V so the full swing for a VCC of 5V is 0 to 3.5V, divided by 2 results in the ~1.75V. Why 1.3A instead of 1.75A? I'm sure the answer is buried in the datasheet of the opamp or the tolerance of the resistors, etc
@@catalinvasile9081 You have the right idea. The LM324 isn't going to swing above about 3.5V with a 5V supply. Since the gate threshold voltage for an MTP3055 is specified as somewhere between 2V and 4V, it's pot luck whether the circuit can actually turn on the mosfet.
In the worst case scenario needed to sink 2.5A, there will be 2.5V across the current sense resistor, so the source will be at 2.5V and the gate could need as much as 4V above that just to turn on. Then as the mosfet has a minimum forward transconductance of 4A/V, it may require an additional gate voltage of 2.5/4 = 625 mV to supply 2.5 A.
That means you have to design for a voltage of at least 2.5 + 4.0 + 0.625 = 7.125 V at the gate. Finally, the LM324 needs another 1.5V "headroom" which shows that you need a 9V power supply for the opamp to accommodate the worst case with that mosfet if you want to sink 2.5A.
If the actual circuit limits to 1.3A, that would mean that the LM324 output and the mosfet gate are at about 3.5V and the mosfet source is at about 1.3V, leaving 2.2V as gate-source voltage, which is a bit better than what you might expect as it has a typical gate threshold voltage of 2.7 V according to the data sheet.
Looks like R35 needs a touch more solder on the top side.
Don't you just love the versatility of Op Amps?
Will this circuit work as a chopper for a stepper motor driver that pulls too much current through its h-bridge or do I have to go with something more complex? Adding a heatsink to the h-bridge is not an option.
How to determine the require gate voltage in order to obtain 1v source voltage?
is there any source available?
why did u not measure the temp on the transistor tab?
Interesting. Keep up the good work Dave!
excelente video.... aca un nuevo suscriptor.... saludos desde Yaritagua Venezuela
Hi Dave,
Thanks a lot for this video.
Do you think that we also could develop an AC version of an electronic load. I know this would be a complete new design. Please let me know your opinion. Best regards from Switzerland, Rene
Hi! I need to build a dummy load for a PC power supply that can draw up to 100A and another unit that can draw up to 30A that are adjustable in steps of 10A and 5A respectively. I know my basic electronics but I would have absolutely no idea on where to start with this. I have heard MOSFETs are very scalable and can achieve this with adequate cooling, but I need a circuit diagram or something so I can make it. Please help!
is there a better opamp ? with a low output impedance? thx
I need to test an IPSU with 9 different power rails ranging from 1.25V to 5V and current less than 1A would this design be recommend for my case? Thanks.
Hello Daniel, when can i find app obdtools of obdlink please ?
Hi , I would like to test the discharge procedure of a 24vdc (7 days autonomy )Nicad battery bank . Please advice, which is the best device to use for that application? and how can i buy the same?
Hi Dave I'm trying to use the CX101b LCD but it's a pain in the ass to configure, can you tell me what resistor values you use to get it to read 1999???? I know that your schematic values are not the same as what you use.
Thanks.
thank you so much. you are the best!
Luckily in 2018 you can get a 250W boost converter for $4 and limit 100W leds without heatsink
Will something like this work for testing amperage and voltage at same time on a solar panel with a watt meter? If not could you please make a video of one that would do so for A47 volt panel or higher?
9:30 if you put 10 resistors in parallel your tolerance goes down to 0.3%.
Hi Dave, I put together one of these on a breadboard using similar components, but the transient response seems to be pretty bad. Also, there is steady state error. Any ideas why?
Excuse me, but what determines the gate voltage?
Hello, since i have stumbled upon your blog i found very useful stuff here and learned something. I'm trying to build a small test rig for testing vacuum tubes and transistors under working conditions. So far i have used resistors are constant current load but as usual resistor stack is never big enough and you need to use some high power ones. So i thought why not to use variable current source instead of resistor so i'm wondering with some modifications could this circuit work in a way i imagined it should.
Ofc limiting current for supply voltage is only one part of that test rig, other is how to drive gate/grid/base but that's another story and not connected to this video.
Hi - thanks for the video, I actually learnt a lot from this ! Can you give some pointers for research into scaling this up to, say, 60A ? Is the approach to parallel a few of these circuits or use a bigger MOSFET ? Any help would be appreciated.
is it going to work if i power the op-amp from a different supply. did the gate of the mosfit will work. I was thinking to combine the two grounds? Help pls
Where did you get the PCB?
i'm gonna quit killing LEDs in series in case of faults and shorts with that)), beauty!!!
I could have sworn you had a smaller and simpler dummy load in a previous episode, but I can't find anything. Am I just imagining that or did I miss it when looking through older episodes?
Maybe LM317 he explains at start of the video?
Considering he says, "it's about time I probably built one," and holds up a dummy load at least twice as big as what I remembered, or thought I remembered, that implies he hasn't made one before and my memory is way off. That's why I'm confused.
You are drawing that mosfet wrong. Is it an enhancement mode or a depletion mode device?
@ElvisAviator I've found out on how to get the LCD cx101b to work reading millivolts even though it's one millivolt offset, I think is due to the lcd +-1%.
dave,
did you get a new 87V in this video? it looks newer than the one you had before...
did your previous 87V broke if it is indeed different?
thanks
What's the bypass cap for on the LM324?
Hi Dave,
I need to variably load my small wind turbine while testing it in a wind tunnel. My electronics skills are not that good, I am a simple aerodynamics engineer. Could this little device be scaled up to dissipate about 150W?
could you show this in 0-100 amp model
Dave I'm having trouble with "drift" with the OpAmp voltage follower. I used a 1ohm, 10% power resistor, an LM324, and a IRF540 Power Mosfet.
I am using a 9V power supply for the LM324 (measuring it, it actually shows about 10V)
I notice that if I use this to test battery discharging I get "drift". I'll start out at say, a setting for 500ma, and as the battery voltage drops, the current *increases* , and i'm not sure why. I thought perhaps its something to do with the MOSFET changing its conductance or something? I always end up having to adjust the input pot setting over time to lower the current back to 500ma
It doesn't make sense to me becasue the OpAmp should technically keep the voltage /current constant even if the resistor drifts, wouldn't it? It almost seems like my OpAmp is drifting.
Also, my MOSFET doesn't get hot at all, granted I'm only running 1.5 V at 500ma , and the resistor is .5V of that (I set the OpAmp + input to .5V so I get .5A on the 1ohm resistor), so ya it's only having to handle what, 500mw, not that much I guess.
could I use a 3404ad dual oppamp instead of a quad amp and use a IRL520n fet?