Author Topic: EEVblog #102 – DIY Constant Current Dummy Load for Power Supply and Battery Test  (Read 33225 times)

0 Members and 1 Guest are viewing this topic.

Offline DavidTopic starter

  • Frequent Contributor
  • **
  • Posts: 279
  • Country: gb
It's great to see another technical blog (no multimeter reviews!  :P ).  Dave, any plans to post a blog on the latest power supply you designed? (I seem to remember you saying you were working on one...) There seems to be a lot of talk about SMPS/DIY bench supplies on the forum and it would be good to see how you went about the design.

Keep up the good work,
Dave
David
(United Kingdom)
 

Offline ddrl46

  • Regular Contributor
  • *
  • Posts: 95
  • Country: nl
That power supply might be his secret project ;).
 

Offline EEVblog

  • Administrator
  • *****
  • Posts: 37740
  • Country: au
    • EEVblog
Yep, all will be revealed in time...

There will probably be a follow-up to that blog, actually going through designing a switchmode supply.

Dave.
 

Offline DavidTopic starter

  • Frequent Contributor
  • **
  • Posts: 279
  • Country: gb
Yep, all will be revealed in time...

There will probably be a follow-up to that blog, actually going through designing a switchmode supply.

Dave.

Brilliant!  :)
David
(United Kingdom)
 

Offline orbiter

  • Frequent Contributor
  • **
  • Posts: 619
  • Country: gb
  • -0 Resistance is Futile
Is that a scratch on your Fluke Dave? :(

 

Offline ddrl46

  • Regular Contributor
  • *
  • Posts: 95
  • Country: nl
Awesome, just in time!
I'm working on a power supply project using linear regulators (aka LM317 / LM350) but i'll wait and try to make a switch mode now  ;D.
 

Offline NiHaoMike

  • Super Contributor
  • ***
  • Posts: 9018
  • Country: us
  • "Don't turn it on - Take it apart!"
    • Facebook Page
For something bigger, wouldn't it be better to PWM the MOSFET in order to keep power dissipation in the resistors? High power resistors are pretty cheap in the form of heating elements and light bulbs.
Cryptocurrency has taught me to love math and at the same time be baffled by it.

Cryptocurrency lesson 0: Altcoins and Bitcoin are not the same thing.
 

Offline RoelA

  • Contributor
  • Posts: 10
Awesome, just in time!
I'm working on a power supply project using linear regulators (aka LM317 / LM350) but i'll wait and try to make a switch mode now  ;D.
You should also watch episode #90:
Quote
Just how different are linear, LDO and switching buck voltage regulators? I'm glad you asked!
You might be surprised...
It's tutorial time.

it on youtube.
 

Offline ddrl46

  • Regular Contributor
  • *
  • Posts: 95
  • Country: nl
I've already seen that, i've seen all the EEVBlog episodes so far. I've already been experimenting with my own switch mode regulators but it hasn't been working out that well so far. I just need a little bit more explanation.
 

Offline DavidDLC

  • Frequent Contributor
  • **
  • Posts: 755
  • Country: us
Dave, can you post the complete schematic with values ? It does not matter it is done in DaveCAD.

 8)

I hope you can to it.
 

Offline ddrl46

  • Regular Contributor
  • *
  • Posts: 95
  • Country: nl
He drew the schematics already in the video did he? It has all the values and stuff.
 

Offline EEVblog

  • Administrator
  • *****
  • Posts: 37740
  • Country: au
    • EEVblog
He drew the schematics already in the video did he? It has all the values and stuff.

Yes, all the values are in the video. It's just one opamp, one load resistor, one MOSFET, and one POT!

Dave.
 

Offline ddrl46

  • Regular Contributor
  • *
  • Posts: 95
  • Country: nl
Just bought all the parts, just gotta build this :).
 

Offline smpowell

  • Contributor
  • Posts: 10
Re: EEVblog #100, #101, #102 – can't transfer video
« Reply #13 on: August 01, 2010, 10:34:34 pm »
Hi Dave,

I've been watching EEVblog on my TV by downloading the Youtube file and transcoding onto my Tivo. But the latest ones can't be transferred.

I'm assuming something changed to set a copy protection flag on the new videos.

Could you change things back to the way they were?  It's nicer to watch on the "big" screen.
 

Offline DavidDLC

  • Frequent Contributor
  • **
  • Posts: 755
  • Country: us
He drew the schematics already in the video did he? It has all the values and stuff.

Yes, all the values are in the video. It's just one opamp, one load resistor, one MOSFET, and one POT!

Dave.

Yes, but I want to save it to my computer, so a picture or pdf will do the work.

If not, I will copy from the video, just want to know if Dave has something already he can post.
 

Offline EEVblog

  • Administrator
  • *****
  • Posts: 37740
  • Country: au
    • EEVblog
Re: EEVblog #100, #101, #102 – can't transfer video
« Reply #15 on: August 02, 2010, 02:19:27 am »
Hi Dave,

I've been watching EEVblog on my TV by downloading the Youtube file and transcoding onto my Tivo. But the latest ones can't be transferred.

I'm assuming something changed to set a copy protection flag on the new videos.

Could you change things back to the way they were?  It's nicer to watch on the "big" screen.

No change at my end.
I have no idea, sorry. Anyone else?

Dave.
 

Offline ThunderSqueak

  • Supporter
  • ****
  • Posts: 167
  • Country: us
  • Dont be a freak... dont be a freak... dont be a ..
    • ThunderSqueak!
I usually watch your blogs on my big screen too, via itunes podcast.  I have it set to download it automatically (checks once a day) and then when I get home and sit down my ps3 or xbox360 tells me "you have a new video".  I watch most of my podcasts this way.    I had no issues with the last video at all.   

Only videos missing are the first 13 episodes as I do not think Dave uploaded them to itunes.

I did try playing the youtube videos on my ps3, xbox360, and my bluray player (all 3 do youtube) and had no issues at all.  Perhaps your tivo is just being stubborn? The format appears to be the same old standard format youtube uses. 
Currently working with non-binary computing, no reason for it... just doing because I can ^^
 

Offline Simon

  • Global Moderator
  • *****
  • Posts: 17816
  • Country: gb
  • Did that just blow up? No? might work after all !!
    • Simon's Electronics
well as he did it in DaveCAD it looks like there is nothing better and it was probably only drawn (literally) because he was doing a video on it. So your homeowrk is to go back to the video, pause it and copy the schematic and values to your usualy schematic drafting program (of which there are many free) and then you can print that to a PDF making printprogram (and yes there are plenty of those for free too)
 

Offline EEVblog

  • Administrator
  • *****
  • Posts: 37740
  • Country: au
    • EEVblog
I did try playing the youtube videos on my ps3, xbox360, and my bluray player (all 3 do youtube) and had no issues at all.  Perhaps your tivo is just being stubborn? The format appears to be the same old standard format youtube uses. 

There is now an XBMC add-on available to browse and watch the blog!



Dave.
 

Offline EEVblog

  • Administrator
  • *****
  • Posts: 37740
  • Country: au
    • EEVblog
well as he did it in DaveCAD it looks like there is nothing better and it was probably only drawn (literally) because he was doing a video on it.

Correctamundo!

Dave.
 

Offline Simon

  • Global Moderator
  • *****
  • Posts: 17816
  • Country: gb
  • Did that just blow up? No? might work after all !!
    • Simon's Electronics
well it's that simple a circuit I'm sure you would have built it out of your head
 

Offline toli

  • Frequent Contributor
  • **
  • Posts: 313
  • Country: il
Sure is a great little thing to have. Especially now that I need something to check a couple of transformers, and than the PS I'm working on (once I find the right case for it).

I only have other op-amp's and some PMOS (IRF9Z34N). I've drawn something that should work with these, and will appreciate it if someone more experienced would just take a look and let me know if its ok. It runs just fine on PSpice. R1 will be the pot. I might change the input and output pins on R1 since the current goes up with the value of R1, so that way I would still need to turn it clockwise to tune the current up.

The op-amp might actually be the OPA604 as I have a few spare of these as well (they work about the same on the spice model).
My DIY blog (mostly electronics/stereo related):
http://tolisdiy.com/
 

Offline smpowell

  • Contributor
  • Posts: 10
Re: EEVblog #100, #101, #102 – can't transfer video
« Reply #22 on: August 08, 2010, 08:02:59 pm »
Hi Dave,

I've been watching EEVblog on my TV by downloading the Youtube file and transcoding onto my Tivo. But the latest ones can't be transferred.

I'm assuming something changed to set a copy protection flag on the new videos.

Could you change things back to the way they were?  It's nicer to watch on the "big" screen.

No change at my end.
I have no idea, sorry. Anyone else?

Dave.

Updating all the software (pyTivo/ffmpeg/python) fixed the problem.  I'm guessing that the older version of the transcoder (ffmpeg) was choking on the video files.

Stephen
 

Online jimmc

  • Frequent Contributor
  • **
  • Posts: 304
  • Country: gb
...I only have other op-amp's and some PMOS (IRF9Z34N). I've drawn something that should work with these, and will appreciate it if someone more experienced would just take a look and let me know if its ok. It runs just fine on PSpice. R1 will be the pot. I might change the input and output pins on R1 since the current goes up with the value of R1, so that way I would still need to turn it clockwise to tune the current up.

The op-amp might actually be the OPA604 as I have a few spare of these as well (they work about the same on the spice model).

Sorry neither the NE5532 nor the OPA604 will work in your circuit if you want to set less than about 2A.
For low currents (R1~0) both inputs will be close to the +ve supply of the OP-AMP.
Neither of these types will work under these conditions.
The parameter you need to look for is the 'Common Mode Input Voltage Range' which needs to include the positive supply for your circuit to work down to zero current.
Simplest to look for an amplifier with 'Rail to Rail' input spec.
(There are a few devices that typically include the positive supply but not the negative eg TL071 or LM301.)

(Interesting to note that the LM324 in Dave's original circuit is only spec'd from 0 to (V+ -1.5v) at room temp and so 3 turns of the 10 turn pot are not usable.)
 

avrfreaks

  • Guest
Dave, it there a schematic or some kind of layout of this Dummy Load PSU as you show it off.

Cheers. :-\
 

Offline EEVblog

  • Administrator
  • *****
  • Posts: 37740
  • Country: au
    • EEVblog
Dave, it there a schematic or some kind of layout of this Dummy Load PSU as you show it off.

The schematic is in the video. Nothing on paper, sorry.

Dave.
 

Offline ThunderSqueak

  • Supporter
  • ****
  • Posts: 167
  • Country: us
  • Dont be a freak... dont be a freak... dont be a ..
    • ThunderSqueak!
because people keep asking, snapshot of the schematic from the video
Currently working with non-binary computing, no reason for it... just doing because I can ^^
 

Offline Simon

  • Global Moderator
  • *****
  • Posts: 17816
  • Country: gb
  • Did that just blow up? No? might work after all !!
    • Simon's Electronics
hm did you check the terms and conditions of DaveCAD to see if reproduction of generated material is legal ?  ;D
 

Offline ThunderSqueak

  • Supporter
  • ****
  • Posts: 167
  • Country: us
  • Dont be a freak... dont be a freak... dont be a ..
    • ThunderSqueak!
hm did you check the terms and conditions of DaveCAD to see if reproduction of generated material is legal ?  ;D

no..


hmmm, that might explain the men in black who showed up at my door this morning


 :o
Currently working with non-binary computing, no reason for it... just doing because I can ^^
 

Offline Simon

  • Global Moderator
  • *****
  • Posts: 17816
  • Country: gb
  • Did that just blow up? No? might work after all !!
    • Simon's Electronics
no that must be because you have aliens in your loft but don't know it. it's ok the men in black are friend with the aliens.
 

Offline btfdev

  • Contributor
  • Posts: 15
Just re-watched the video and got a quick question.

Noted at 2:10, Dave mentioned a LM317 could be used a constant current load when you ground the output.

So *WHY* MOSFET solution you need a power resistor going to ground but LM317 does not need one?

Any idea?

Thanks, BTFDev.
 

Offline SeanB

  • Super Contributor
  • ***
  • Posts: 16284
  • Country: za
With the LM317 direct to ground you are just relying on the built in current limiter, at around 1.6A or so, to provide constant current sinking provided you just ensure the voltage across the device is above about 3V and it has a big heatsink. If you use a resistor across the output to ground then it will keep the voltage across the resistor constant at 1.25V over a wide input voltage range, from about 5V to 30V in most cases. The resistor allows programming a lower current, you can go from under 1mA ( but with a slight error from adjust pin current and quiescent current of the device) to around 1A and have a pretty stable current. It does need a heatsink that will keep the device under 60C for best stability though,  especially at the higher range.

 

Offline mianchen

  • Regular Contributor
  • *
  • Posts: 225
  • Country: gb
Two questions:

1. What are the benefits of using a mosfet compared to a bjt transistor? I have a few 2N3055, is it a bad idea to use them? I've tested it on a bread board, and it seemed ok and worked as 'expected'. I don't have a MTP3055, but have a few P36NF06L and I managed to kill one when the set current is >2 A with a 3c/W heatsink.

2. How to calculate the heat dissipation for the transistor? The output voltage of the PSU to be tested (Vout) - Voltage  drop across the resistor (Vr) x the set current going through the resistor?

Sorry for the noob questions

Thanks
 

Offline ModernRonin

  • Contributor
  • Posts: 44
1. What are the benefits of using a mosfet compared to a bjt transistor?
In this particular case, the major advantage is that a voltage on the gate of a MOSFET controls the current through the drain-source. With an NPN transistor, it's a current through the base that controls the current from collector to emitter. Voltage control is usually easier than current control.

In the general case, the advantage of MOSFETs is that when used as a switch, they have a lower resistance than an NPN. They are more power efficient, when used as a switch. But in this case we're not using the MOSFET as a switch. We're using it as a variable resistor. So that advantage is not useful to us here.

Also, NPNs can go into thermal runaway. NPNs conduct better the hotter they get. With MOSFETs it's the opposite, the hotter they get the worse they work. So MOSFETs can't go into thermal runaway nearly as easily.

Quote
> I have a few 2N3055, is it a bad idea to use them?
First, you need to know that there is a minimum voltage drop across an NPN transistor operating in linear mode of about 0.7 volts. This will put a limit on the maximum current that the dummy load can sink. However, the limit is pretty high. Something in the neighborhood of: (1.5 * input voltage) amps.

Second, NPN transistors can go into thermal runaway as mentioned before. This probably won't be a problem with the circuit Dave used, because the opamp will automatically throttle back the NPN if it goes over-current. But just be aware that thermal runaway can happen.

Third, it depends on how you're driving the base of the 2n3055. If there's X amount of current flowing through the collector of an NPN, then you need (X / beta) flowing through the base. The 2n3055 has a beta anywhere between 20 and 70. So in the very worst case, you will need to supply X / 20 amps of current into the base. If the opamp you use to drive the base of the 2n3055 can supply that much current, then no problem. But if it can't supply that much current...

Assuming a "typical" opamp with 20mA of output current capability, that would mean your dummy load would be able to handle a maximum of 20mA * 20 = 4000 mA = 4 amps. This is a lot less than the 15A maximum that the 2n3055 can handle.

To ensure the full 15A limit can be reached, you would need an opamp that can output 15A / 20 = 750 mA. There are opamps that can supply that much current, but they're not very common and may not be cheap.

BTW, you can go back and do all of the above math with 70 instead of 20 to see what kind of base current you would need in the "perfect world" case. But I don't recommend you count on the world being perfect.

(There's a relatively simple base driving circuit that you can make with a resistor and small MOSFET that will greatly amplify the output current of your opamp. But you're such a n00b I don't want to confuse you with it just now. It looks basically like this, except with the transistors swapped: http://www.neufeld.newton.ks.us/files/electronics/mosfet-motor/MOSFET-motor-driver.png )

Quote
> 2. How to calculate the heat dissipation for the transistor? The output voltage of the PSU to be tested (Vout) - Voltage  drop across the resistor (Vr) x the set current going through the resistor?
The usual answer is "P = IV", but in this case we should probably use the equivalent but alternate form:

P = I^2*R
or
Watts = (Current through component)^2 * Resistance of component

Now, the current you know. You just decide you want 1.3 A constant current, and that's that. The constant-current load will adjust its own resistance until that constant current is reached.

The resistance you will have to calculate, based on the input voltage and constant current. Example: You have 12V and set a constant current of 400 mA (= 0.4 A). Use V = IR to calculate the equivalent resistance:

V = IR
V/I = R
12/0.4 = R
R = 30 ohms

Now use P = I^2*R:

P = 0.4^2 * 12
P = 1.92W

This is much less than the ~100 watts than the 2n3055 is rated for.

But don't get a false sense of security. Remember that the equation depends on current squared. As the current goes up, wattage also goes up very quickly. Another example:

Constant current = 4 A; Vin = 24 volts; R = V / I so R = 6; P = I^2*R so P = 16 * 6 = 96 watts! You're getting close to the power dissipation limit of your 2n3055.
 

Offline btfdev

  • Contributor
  • Posts: 15
Hey folks

In the winter holidays we did our own dummy load based on Dave's circuit.
http://www.binarytaskforce.com/Weblog/2013/02/08/dummy-load/]
[url]http://www.binarytaskforce.com/Weblog/2013/02/08/dummy-load/

[/url]

Sure it is a great tool!

Enjoy

Team BTF
 

Offline rodcastler

  • Regular Contributor
  • *
  • Posts: 98
  • Country: cl
Dave says he adds the second OpAmp because he had the dual package lying around but what is exactly the benefit of the second opamp? Apologies if this is a newbie question but I wish he had explained it a little deeper in the video.
« Last Edit: June 23, 2014, 11:45:26 am by rodcastler »
 

Offline Lee697

  • Regular Contributor
  • *
  • Posts: 82
  • Country: au
Dave says he adds the second OpAmp because he had the dual package lying around but what is exactly the benefit of the second opamp? Apologies if this is a newbie question but I wish he had explained it a little deeper in the video.

Curious about that question myself - I breadboarded the dummy load yesterday, and just used the pot as Dave indicated in the video, and it works well....
 

Offline JayJones

  • Newbie
  • Posts: 5
Reviving an old thread...

Can anyone tell what I might need to modify to run this from a 9VDC adapter, and any way to use a standard N-Channel Power mosfet?
I blew up 4 FQP50N06 mosfets already trying to figure out which parameter I'm exceeding in the datasheet :)

Thanks!
 

Offline rqsall

  • Regular Contributor
  • *
  • Posts: 87
  • Country: nl
At least you will need to reduce the source resistors (1 ohm in Dave's version) to something less than 0.5 ohm or so. You need more Vgs with a non-logiclevel mosfet to turn it on, and if the 1 ohm source resistor is already eating 2 volts at 2 amps, and you have 5V supply on your opamps, then there isn't much left for Vgs.

If you blew them up, however, I suspect you ran too much current through it, perhaps the gate voltage was too high due to faulty wiring?
 

Offline brutester

  • Contributor
  • Posts: 17
Attached is even simplier way of doing it. Just use a power NPN. The one I've used was removed from old computer PSU. It is not as good as Dave,s approach, but hey - it is powered from the device and does the job :)
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf