Modified LM324 Based 150W 72V 10A Electronic Load [COMPLETED]

[VEGETA]

I meant, instead of having one big heatsink for 4 mosfets... how about putting one per mosfet? One of those famous to-220 heatsinks. Will it work?

If your device is about 100 watts that means 25 watts per mosfet, so my question will such heatsink be able to tolerate these 15-25 watts? with or without fan cooling.

You need to specify which heatsink you are thinking, a pict would do.

Usually they are kind of small and good for LM137 when you need a bit more umpf out of it without them.

JS

I meant one of these:

https://www.aliexpress.com/item/5pcs-lot-38x34x12-8mm-TO220-TO-220-heatsink-heat-sink-radiator-for-IC-triode-7805/32622932747.html

and these are more quality ones (branded): https://eu.mouser.com/ProductDetail/Ohmite/FA-T220-64E?qs=sGAEpiMZZMttgyDkZ5WiumlCfl50RTwzVA%252bY4U4BtvA%3d

And in the end of the case (inside) one of these small fans: https://www.aliexpress.com/item/High-quality-3010s-30MM-30-x-30-x-10MM-12V-2Pin-DC-Cooler-Small-Cooling/32603431500.html

so 4 heatsinks close to each other and one fan sucking air from them and toss it outside.

[JS]

I meant, instead of having one big heatsink for 4 mosfets... how about putting one per mosfet? One of those famous to-220 heatsinks. Will it work?

If your device is about 100 watts that means 25 watts per mosfet, so my question will such heatsink be able to tolerate these 15-25 watts? with or without fan cooling.

You need to specify which heatsink you are thinking, a pict would do.

Usually they are kind of small and good for LM137 when you need a bit more umpf out of it without them.

JS

I meant one of these:

https://www.aliexpress.com/item/5pcs-lot-38x34x12-8mm-TO220-TO-220-heatsink-heat-sink-radiator-for-IC-triode-7805/32622932747.html

and these are more quality ones (branded): https://eu.mouser.com/ProductDetail/Ohmite/FA-T220-64E?qs=sGAEpiMZZMttgyDkZ5WiumlCfl50RTwzVA%252bY4U4BtvA%3d

And in the end of the case (inside) one of these small fans: https://www.aliexpress.com/item/High-quality-3010s-30MM-30-x-30-x-10MM-12V-2Pin-DC-Cooler-Small-Cooling/32603431500.html

so 4 heatsinks close to each other and one fan sucking air from them and toss it outside.

The first ones are definetly too small, the second ones go down to 3°C/W to air the biggest. That, plus 2°C/W from junction to heatsink, with 25W is 125°C rise over ambient, way too much for it to work. With a fan might get to more reasonable numbers.

Having them in different supplies makes no difference in theory, thermal circuit is equivalent given the 4 HtoA resistance are equivalent to the big one for the 4. Now, the one dissipating more will get warmer, but the heatsink also and more heat will be dissipating making difference smalls. The major difference might be mechanical or electrical if using the heatsink as the conductor to tie the tabs.

JS

Enviado desde mi LG-M250 mediante Tapatalk

[VEGETA]

I tried searching for one big heatsink but couldn't find a standard one. So this project, what heatsink comes with it? or should you put a heatsink of your own?

Say I want to make such a project which will be sold as a commercial open source one... okay? now how can I source heatsink if there are no standard ones? especially a cheap one with a fan to make it affordable.

So you think one heatsink is better than 4 smaller ones? perhaps you are correct since it could dissipate all the heat better but I don't really know.

[JS]

I tried searching for one big heatsink but couldn't find a standard one. So this project, what heatsink comes with it? or should you put a heatsink of your own?

Say I want to make such a project which will be sold as a commercial open source one... okay? now how can I source heatsink if there are no standard ones? especially a cheap one with a fan to make it affordable.

So you think one heatsink is better than 4 smaller ones? perhaps you are correct since it could dissipate all the heat better but I don't really know.

Using an Al enclosure could be a solution, you might want to electrically isolate the transistors fromr it for safety, but would end in a cheaper product as you get two for one. You still need an enclosure. Then if you want to add a fan it can just circulate air from outside to the inside and the inside of the case would have the extra airflow.

The bigger ones from the second link are probably fine, you will be depending on airflow and having well distributed airflow can be tricky.

I've seen catalogs for big heat sinks, you need sub °C/W here, 3/4 would be equivalent to the 4 of the link. I can't mention one brand or a supplier as have plenty of big heatsinks from a friend who took it from broken phone transmision lines powe supplies, so I didn't come in need for them.

You would be amazed how the ones I have match the PCB I've done for the μSupply without planning for it.

JS

Enviado desde mi LG-M250 mediante Tapatalk

[enut11]

The heat-sink, fans and power MOSFETs have arrived.
I purchased ten IRFP4668 130A 200v MOSFETs on eBay with a view to selecting 4 matched ones.
I will publish how I tested and matched them shortly.
The heat-sink measures 150x69x36 also purchased on eBay. I do not have any thermal specs for it.
The fans are 70x15mm 12v units from eBay
enut11

[JS]

  Sorry I join a bit late, wouldn't be possible to have all the mosfets for the higher range with 4x0.1Ω resistors and switch a series resistor in with all of them for the lower ranges? Then you just turn one mosfet in or leave them do whatever as it doesent matter if they are unbalances. Possibly using a different output low connector so you dont have to switch. Then you just need a switch to switch in or out the second connector but it  just has to handle 1A.

  I'm building a dummy load right now, small, just one IRF540 but I added CV control up to the CC limit and thermal shutdown when needed till it cools down a bit. After it I'd probably scale up things to get more umf, but I don't usually need to deal with high power stuff so who knows when.

JS

Enviado desde mi LG-M250 mediante Tapatalk

[enut11]

Hi @JS. The first CC Load that I built (see above) had a switched range with 100mA or 2.5A available and it works well. I switched the Drain input to one of the MOSFETs using a 2.2ohm Source resistor for the lower range. Switching the Source resistor for different ranges may also work.

For this version (Take #2 Electronic Load) I am aiming for the full 150W/10A advertised for the kit.
Choice of MOSFETs is often dictated by economics. I decided that 4 'robust' switching MOSFETs would do the job. Will know soon.

Would you share how you implemented thermal shutdown?
enut11

[VEGETA]

The heat-sink, fans and power MOSFETs have arrived.
I purchased ten IRFP4668 130A 200v MOSFETs on eBay with a view to selecting 4 matched ones.
I will publish how I tested and matched them shortly.
The heat-sink measures 150x69x36 also purchased on eBay. I do not have any thermal specs for it.
The fans are 70x15mm 12v units from eBay
enut11

Can you please provide the links and prices? shipping?

Also, I would be interested to see what is the performance of heatsink alone at say 5A compared to with fan. I hope you can do that, you would benefit us.

One way you can determine heatsink temperature coefficient, it is by recording its temperature at no current (= ambient) then put a known power like 1W and measure its temperature, then repeat with 10W... finally you will get to know the coefficient.

[enut11]

Proposed MOSFETs location on heat-sink. The cardboard represents a template for a copper heat spreader sheet between MOSFETs and heat-sink. The idea comes from Kerry Wong who demonstrated a useful temperature reduction in his Youtube video (see Reply #12 above).

Setup I used to test and match the MOSFETs. Flying leads for Gate, Drain ad Source from the PCB go to screw terminals allowing quick change of MOSFET.

I monitored the MOSFET Vgs at 1mA, 10mA, 100mA, 1A and 2A Drain currents while keeping an eye on the device temperature.
enut11

[JS]

  Here is the full scheme in mine, it simulated ok, I should solder the components tonight and tomorrow pick up my new scope and test it.

  Nothing fancy, LM34 (the BJT in the schematic as I didn't had the part) comparator to a diode with some hysteresis, it should cut down at about 70ºC at the sensor and go back up at 50ºC. I could probably add a hand switched fan if I see the need, I have an old CPU heatsink to start the tests, the PCB allows direct mounting to the heatsink of the mosfet, LM34 and sense resistor if needed, 0.22Ω 2W, would be 1.4W at 2.5A I'm thinking as absolute max so it should be fine, I don't think to use this puppy up there, I'd probably use a log pot for the current so the control in the lower side is more smooth, there are a few tweaks to be made to the circuit, but the current PCB can accept external control (or set sensing) and output sensing with the available pin headers.

  The CV and CC can be switched by the pin header between the opamps and the mosfet, I think CC should be left always on, I'm running some simulations now for doing so, replacing the D2 with a resistor and eliminating R14.

  The circuit has a few tweaks to be made, I think I can do better for the step response but it already sim better than the basic config, as the voltage control makes it smoother but much slower, I'd like faster response with less overshoot (don't we all). I did tried the snubber on the output, step was much better seen from the dummy load but not from the DUT side, it could be wired externally and switched in and out.

  Switching mosfets are optimized for a different task, be aware, they are usually good for high short peaks of power in them but not optimized for long sustained power applications. Have an eye on them and tell us how it went.

JS

[enut11]

The heat-sink, fans and power MOSFETs have arrived.
I purchased ten IRFP4668 130A 200v MOSFETs on eBay with a view to selecting 4 matched ones.
I will publish how I tested and matched them shortly.
The heat-sink measures 150x69x36 also purchased on eBay. I do not have any thermal specs for it.
The fans are 70x15mm 12v units from eBay
enut11

Can you please provide the links and prices? shipping?

Also, I would be interested to see what is the performance of heatsink alone at say 5A compared to with fan. I hope you can do that, you would benefit us.

One way you can determine heatsink temperature coefficient, it is by recording its temperature at no current (= ambient) then put a known power like 1W and measure its temperature, then repeat with 10W... finally you will get to know the coefficient.

MOSFETs here:
https://www.ebay.com.au/itm/IRFP4668PBF-TRANSISTOR-MOSFET-N-CH-200V-130A-TO-247AC/123131585110?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2057872.m2749.l2649

Heat-sink here:
https://www.ebay.com.au/itm/Aluminum-Heat-Radiator-Heatsink-Cooling-Fin-150x69x37mm-Silver-Tone-F6/282280422071?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2057872.m2749.l2649

Fans here:
https://www.ebay.com.au/itm/Hot-Quiet-7cm-70mm-70x70x15mm-12V-Computer-PC-CPU-Silent-Cooling-Case-Fan/112935110852?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2057872.m2749.l2649

Fan covers:
https://www.ebay.com.au/itm/70mm-Iron-Net-Fan-Cover-CPU-Fan-Grill-Computer-Cooling-Fan-70x70mm-For-AC-DC/282789089373?ssPageName=STRK%3AMEBIDX%3AIT&var=582088181424&_trksid=p2057872.m2749.l2649

Heat-sink needs drilling and tapping for MOSFET bolts.
 
"Also, I would be interested to see what is the performance of heatsink alone at say 5A compared to with fan"
This is not enough info. A heat-sink is an energy dissipator. Tell me how many 'watts' you need tested and I will see if it is possible.
enut11

[enut11]

Hi @JS
Interesting circuit. Have you Posted it separately on the Forum?  Should raise some useful discussion.

Temp control on mine is manual. Awaiting for these controllers to arrive for testing.
https://www.ebay.com.au/itm/Temperature-Speed-Controler-DC-12V-Denoised-Speed-Controller-for-PC-Fan-Alarm/132595582255?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2057872.m2749.l2649

I have yet to think about step response. That will come later.

I like the ability to switch CV/CC.
enut11

[VEGETA]

Try to test it at say 30v/5A = 150 Watts. Or make a spreadsheet with 1W, 5W, 10W, 50W, 100W, 150W, and so on. Record the mosfet's temperature and heatsink temperature. This could be a nice experiment as well as a reference for such heatsinks since they don't have actual data. You could build the accurate data yourself by recording it, then deduce the temperature coefficient from it. Next, we can all use that coefficient with such heatsinks as an approximate reference value.

I found some nice heatsinks as well as cheap like 5$ delivered! this is one example:

https://www.ebay.com/itm/Large-Big-Aluminum-Heatsink-Heat-sink-radiator-for-Led-High-Power-Amplifier-CN3/253326048354?epid=1974571145&hash=item3afb68bc62:g:xIcAAOSwiHpaPdH9

it would be so nice if these can operate properly at 150 watts with or without a fan.

Your experiment can provide reliable data about when do we need a fan too.

[JS]

Hi @JS
Interesting circuit. Have you Posted it separately on the Forum?  Should raise some useful discussion.

Temp control on mine is manual. Awaiting for these controllers to arrive for testing.
https://www.ebay.com.au/itm/Temperature-Speed-Controler-DC-12V-Denoised-Speed-Controller-for-PC-Fan-Alarm/132595582255?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2057872.m2749.l2649

I have yet to think about step response. That will come later.

I like the ability to switch CV/CC.
enut11

  I haven't yet, I have this and a µSupply on the bench right now, the µSupply is already populated, only missing the LT3080, a 7805 for supplying the µC and a few caps I should get tomorrow. The dummy only has 3 unsoldered resistors but I plan to start soldering right now.
  I want to build them and be able to give some real world data on the circuits before posting, I guess two posts are on the go right now, maybe a single one, I have a few projects as I'm rigging up the lab and maybe would be good to have them in one place. I built a mΩmeter (measures consistently down to a single mΩ and has 2 digits to spare) and planning a differential probe a bit different to the projects I've seen out there. The idea for this would for measuring small signals floating highish (measure components floating inside circuits and things like that) not too much bandwidth, say audio and bit else, lowish noise and easy to get components.

JS

PS: 1h later PCB fully populated, well, it's missing one cap (voltage control compensation) would go in tomorrow.

[JS]

  That value looks useful for transient tolerance but not much for a DC dummy load as it's going to be used in a steady state a lot. No doubt a higher number will translate in higher reliability, tolerance to transient, but junction-heatsink thermal resistance is also quite important. Both will go up together but power switching transistors might very well have higher P2t relative to the thermal resistance. Big transient response should be managed as it not only can be bad for the transistors but for the DUT as well and we don't want to harm the DUT.

Where do you get the values from?

JS

[enut11]

Using the setup from Reply #33 I tested the 10 IRFP4668 MOSFETs for Vgs at various Drain currents at Vds = 5v. Based on this data I have selected #3, #4, #8 and #9 for my CC Load.
enut11

[JS]

Hey, here is my tiny one! Builded, tested* and posted 

I do like the constant voltage mode to be able to test constant current sources...

https://www.eevblog.com/forum/projects/cccv-dummy-load/

JS

*barely tested, just low power under a 9V battery.

[enut11]

Having set aside 4 reasonably matched MOSFETs from the 10 IRFP4668 purchased it was time to see how far I could push one of the remaining ones.

Aim for this CC Load project is 150W/10A/30v with 4 power devices, ie 37.5w/2.5A per transistor if they all share the current load evenly.

1) My power supply is good for 28v/1A/28W so I tested with that first using the setup in Reply #33. MOSFET case temperature taken at the transistor notch was around 30C (ambient 18C).

2) Next I tested using an IBM Server 12v/69A PS. At ~12v/4.4A (CC Load pot limit) the MOSFET was now creating ~50w heat and the case temperature was around 49C (ambient 19C).

So, with one device able to handle up to 50W this project looks to be on track for 150W total with 4 devices.

In these tests the MOSFET was bolted to a fan cooled Al heat-sink using thermal paste. Even with direct mounting, ie no insulating washer, I noticed that the heat-sink was much cooler than the transistor in all tests. This tells me that the real challenge will be the rate of heat transfer through the transistor case-to-heatsink junction.

Next step is the heat-sink and this will be the biggest effort with lots of drilling and tapping.

[JS]

Yes, case-to-heatsink thermal resistance is a bitch, with 4 transistors they are kind of in parallel, so you expect to see the same thermal difference in that juncture if using the same mounting technique. Anodized heatsinks from a respectable vendor might have enough anadization to be considered enough insulation, if you need them to be isolated.

Other improvement that could be made is to use a better thermal paste, any decent metal based PC grade is an improvement over regular silicon based ones.

As mentioned cooper plates might be helpful, as it's softer than aluminum might get better contact for the same pressure and then you have a greater area from the cooper to the aluminum and the cooper softness helping here again.

JS

Enviado desde mi LG-M250 mediante Tapatalk

[VEGETA]

How about this panel meter that you use?

I read that it doesn't go to lower voltages, is that correct? can it be mounted high-side and low-side?

[sorin]

According to SOA on the datasheet IRFP4668 can handle:

500W @  5V (not a realistic value, limited by Thermal Resistance)
30W   @10V
10W   @20V

So I don't think that is suitable for your application.

[VEGETA]

According to SOA on the datasheet IRFP4668 can handle:

500W @  5V (not a realistic value, limited by Thermal Resistance)
30W   @10V
10W   @20V

So I don't think that is suitable for your application.

Where did you get such info?

I used IRILZ44N for my design for 30v/2A specs, it could go much more. It is in an isolated package which is a plus.

[sorin]

IRILZ44N don't exist maybe you mean IRFZ44N or IRLIZ44N? 

IRFZ44N is not specified for DC Load, just for pulse Load.

Its not about how large the transistor is but who the transistor is constructed internally.
High current, low voltage transistors usually are composed from many (10-100 or 1000) small current transistors inside the die.
If you use this transistor in linear mode then the current is not shared equally through them.
On the forum are several discussions explaining this topic.

Where did you get such info?

On the IRFP4668PbF Datasheet Fig 8, page 4.

I used IRILZ44N for my design for 30v/2A specs, it could go much more. It is in an isolated package which is a plus.

And what about the junction temperature???
Have you measured that?
Isolation is not a good thing on out application because  it have a large thermal resistance Junction-to-Case 3.3 °C/W

Its all about long term reliability...

[enut11]

Yes, case-to-heatsink thermal resistance is a bitch, with 4 transistors they are kind of in parallel, so you expect to see the same thermal difference in that juncture if using the same mounting technique. Anodized heatsinks from a respectable vendor might have enough anadization to be considered enough insulation, if you need them to be isolated.

Other improvement that could be made is to use a better thermal paste, any decent metal based PC grade is an improvement over regular silicon based ones.

As mentioned cooper plates might be helpful, as it's softer than aluminum might get better contact for the same pressure and then you have a greater area from the cooper to the aluminum and the cooper softness helping here again.

JS

Enviado desde mi LG-M250 mediante Tapatalk

As for CC Load #1 above, I have decided to mount the MOSFETs directly to the heat-sink, ie no mica washers. This means that the heat-sink will be 'live' but we are only dealing with low voltages <30v. I will be using thermal paste from Jaycar (Australia) intended for CPUs.

I have finished drilling and tapping the heat-sink and will post pix soon.
enut11

[enut11]

How about this panel meter that you use?

I read that it doesn't go to lower voltages, is that correct? can it be mounted high-side and low-side?

@VEGETA
See this posting

https://www.eevblog.com/forum/projects/smart-panel-meter-can-it-be-modified-to-read-down-to-zero-volts/

for info on how to modify it. It can only be mounted low-side due to the common ground.
enut11