-
You can do that, even with riskier stuff as long that's hidden well within the case and marked.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
If you intend to leave it outside, I wouldn't recommend it, as it might very well end shorting with simmering in your desk.
Regards!
Enviado desde mi LG-M250 mediante Tapatalk
-
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.
Hmm, interesting data. I ran one again tonight at ~50 watts (4.4A at 12v) for about 10 min. This time I used in IR thermometer and measure 60C on top of the MOSFET case. Junction temperature will be higher but no way to measure it. Heat sink was about 10C cooler. Ambient here is 18C.
Four MOSFETs pumping out 150W into the new heatsink should be interesting!
enut11 -
You can do that, even with riskier stuff as long that's hidden well within the case and marked.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
If you intend to leave it outside, I wouldn't recommend it, as it might very well end shorting with simmering in your desk.
Regards!
Enviado desde mi LG-M250 mediante Tapatalk
No room inside. The heat-sink and fans will be outside the box with a big warning sticker
-
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.
Hmm, interesting data. I ran one again tonight at ~50 watts (4.4A at 12v) for about 10 min. This time I used in IR thermometer and measure 60C on top of the case. Junction temperature will be higher but no way to measure it. Heat sink was about 10C cooler. Ambient here is 18C.
Four MOSFETs pumping out 150W into the new heatsink should be interesting!
enut11
You can estimate the junction temperature from the case temp, power and junction to case thermal resistance, usually to the back of the case, where the heatsink goes. Often a hole is drilled to measure there with a thermocouple.
50W
60°C at the case
50°C at the heatsink (0.2 °C/W case to heatsink thermal resistance)
20°C ambient (0.6 °C/W heatsink to ambient thermal resistance)
Look datasheet to find juncture temps.
JS
Enviado desde mi LG-M250 mediante Tapatalk
-
No room inside. The heat-sink and fans will be outside the box with a big warning sticker
[emoji85]
The thing is that it isn't just at 30V, it is at whatever your DUT is at, and that can be much more.
JS
Enviado desde mi LG-M250 mediante Tapatalk
-
True @JS. That is why I have programmed the LED display to flash on and off at over 33v to remind me that is the limit for this CC Load.
If the heat-sink is up to the task of dissipating 150W and the transistors are at a reasonable temperature I could experiment with mica washers at a later date.
enut11 -
Trial fitting of the 3mm copper plate and power MOSFETs to the aluminium heat-sink for the CC Load. Hoping to sink 150W here. Heat-sink measures 150x69x36 and will have twin fans.
Note to myself: 'check component clearances before drilling holes!' The left MOSFET Gate and right MOSFET Source are very close to the bolt heads. There is just enough clearance to use heat-shrink on the MOSFETs and bolts.
enut11
-
Are you sure that plate is making good contact with the heatsink, I'd use a few screws more! [emoji3]
Mechanicals are a bitch, they are my project stoppers ans I deal with that sort of thing at work every day, just heavier stuff. When I'm with electronics I want to be soldering, not tapping holes!
JS
Enviado desde mi LG-M250 mediante Tapatalk
-
@JS. The reason there are not more screws is that I ran out of them! Agreed, the mechanicals, not the electrics, form the bulk of a project. That is why some of my projects end up being used without a box.
Pictures showing CC Load twin fans mounted and the heat-sink screwed to the box using L-brackets. Next is the final wiring and testing.
enut11
-
That's nice dude!
For the insulation, you could do it between the cooper and aluminum, such surface has very little thermal resistance, insulating it you will be adding much lower resistance than doing so in the transistor cases. You need a big piece of insulator, but maybe a thin sheet of maylar would do, I don't know the thermal characteristics of it, but a few micron would do for electrical insulation, can take the heat no problem and it's cheap.
JS
Enviado desde mi LG-M250 mediante Tapatalk
-
Wiring all done. Added labels to front panel. Decided to leave some open space at the back to let the MOSFETs breath. I added a heat shield to isolate the PCB, etc from the heat generating components. It is important to keep the TL431 voltage reference as stable as possible.
Had time for a quick test to see that it works, and it does. Final testing up to 150W soon
enut11
-
Nicely done!
good idea about the heat shield, it looks siny, I hope is not metal, but... you forgot the resistors inside! They are working pretty cool but still 2.5W will warm a bit.
You kind of cheated with the front panel display I guess
JS -
Thanks @JS. Yes, the Source resistors will create some heat at high amps. I will run a long term test sometime and measure the box internal temperature. I will monitor the 2.5v Ref and if it proves to be a problem I could relocate the heat shield further towards the front. Btw, the shield is hard cardboard.
enut11 -
150W-10A-33V Electronic Load test results with open case.
Blue temp sensor monitors MOSFET (at the notch) and was in contact with the metal tab.
Green temp sensor monitors heat-sink copper.
Black temp sensor monitors 0.1ohm Source resistor.
I also used an IR Thermometer on the MOSFETs and it proved to be more reliable. Even so, it too may have given a slightly low reading given the small surface area of a MOSFET.
Temperatures were taken for currents up to 10A/118W. Max MOSFET temperature was less than 50C. The copper heat spreader was 5C lower. I did not measure the Al heat-sink.
Temperature difference between the 4 MOSFETs was less than 2C. This seems reasonable to me. I took the trouble to match the MOSFETs for Vgs vs Load and also matched the 4 Source resistors within 1%.
I also took the load up to 150W/13A for about 10sec but did not take any temperatures.
@JS, thanks for your tip on Source resistor heat. The results show a maximum of 35C at 10A. I will now look at shielding the TL431 Ref from this potential heat source.
Thanks to all who contributed to this project. Collective knowledge helps to get to the best solution faster.
That's about it for this project. I will test temps with a closed case and report any significant results.
enut11
-
30ºC rise at 118W, math says 0.25ºC/W from case to ambient, doesn't look so real to me, 0.21ºC/W from heatsink to ambient. That's at maximum load, with lower loads even higher.
This means 0.04ºC/W from mosfet to heatsink, or for each one 0.16ºC/W.
Datasheet for your mosfets says 0.24ºC/W from case to heatsink, flat surface grease, did you use insulation at the end? I will guess not for the next calcs.
https://www.infineon.com/dgdl/irfp4668pbf.pdf?fileId=5546d462533600a40153562c8528201d
You did said the case probe was touching the heatsink, so I'll say it's half way and you acctually has 0.24ºC/W the datasheet suggests from case to heatsink. This means the case is actually at 54ºC, just add 2ºC to the cases, we are still at a safe 50ºC. Now, from case to junction, that's 58ºC at the junction, or 41ºC raise at 118W, 0.35ºC/W J-A. For your rated 150W you could say you'd only have 52 ºC temp rise, at a more confty room for my girl (I'm fine at 17º) that's 75ºC at the junction. This things are rated for 175ºC at the junction, running them at 100ºC at J is safe enough for me, you should cover the 150W on your sticker with a 220W even if your measurements weren't soooo reliable
you still get a good margin for junctions to get to 175ºC.
How long did you run the tests? this is a big thermal mass you need to heat up before taking reliable measurements. You could even find the thermal capacity of your heatsink in the process, if you measure the heatsink temp over time you should see an exponential grow in temps, and is just a RC circuit. Even if you don't let it settle to the final temp you could get a better approx on your final values. Few samples, set 100W to do the math easy, read 15s for the first 3 intervals, then 30s intervals, take maybe 10 samples, quit early if you see the temps are converging. Taking the heatsink temp as the cooper plate seems fine, then you just take from there to ambient as a single number (settling time would be a bit longer considering the full scheme)
I can't wait for my 121GW to do the work for me, I just would need to check my phone that temps aren't going outside safe ranges and come back to process data with the PC so comeonnnnn Dave, ship fast!!!
JS -
@JS, to reduce thermal resistance I did not use any insulation between MOSFETs and copper or between copper and heat-sink. For the time being I can live with a live heatsink.
I will have to think about your thermal analysis. I take it that you are satisfied the 150W label is warranted?
I am now happy with the performance of this CC Load. When I purchased the original kit I thought that it was going to be an easy project! In the end I am glad it was a challenge. I got to learn about the behaviour of MOSFETs, a device designed for rapid switching, and how to force them to behave in a linear fashion. Also, the headline numbers on the spec sheet, Rds, Ids max and Vds max are relatively useless in linear mode. For DC operation I found you have to rely almost entirely on the SOA diagram and even downrate that! An interesting journey.
enut11 -
@JS, to reduce thermal resistance I did not use any insulation between MOSFETs and copper or between copper and heat-sink. For the time being I can live with a live heatsink.
I will have to think about your thermal analysis. I take it that you are satisfied the 150W label is warranted?
I am now happy with the performance of this CC Load. When I purchased the original kit I thought that it was going to be an easy project! In the end I am glad it was a challenge. I got to learn about the behaviour of MOSFETs, a device designed for rapid switching, and how to force them to behave in a linear fashion. Also, the headline numbers on the spec sheet, Rds, Ids max and Vds max are relatively useless in linear mode. For DC operation I found you have to rely almost entirely on the SOA diagram and even downrate that! An interesting journey.
enut11
Yes, seems fine for me, I think you do have some margin to insulate the heatsink if you wanted to do so. At least for the thermal analysis.
I only watched the thermal specs on the datasheet, I guess you are well within SOA for VA rates.
JS
-
To answer your question on how long I let the CC Load run before recording temps, it was generally several minutes, until the temp change was insignificant. I started with low amps and finished on the highest setting.
enut11 -
In that case, anf if you are limited by SOA, you can ease up on thermals and add the insulators.
I'm not saying you are going yo die because of the live heatsink, but a falling screwdriver can ruin your day... I'd also add a grill around so nothin can fall between the transistors. It seems pretty open space between project case and heatsink.
Be aware, cases always stall my projects, I think your solution is great! Just some comments to be on the reliability side.
JS
-
Thanks JS. Another solution would be an isolated guard over the heat sink. I do believe that direct mounting the MOSFETs (ie no insulating washers) keeps them cooler at max amps and that can only be good for component life.
In one test I forgot to switch on the fan and the heat-sink got hot very quickly. However, temps came down fast with a bit of air moving through the fins so another reason to be happy with the thermal performance of this heat-sink.
In the final setup I found that the pot allowed the amps to be set well over the 10A limit that I needed. I inserted a 15K resistor after R22 and in series with the pot. I have produced a label that now sits on top of the plastic enclosure and it shows the expected amps for each rotation of the 10-turn-pot.
enut11
-
A final reply (I think) on this project. After several more tests I am happy to report that this 'Take #2' DC Electronic Load does work reliably all the way up to 10A and 150 Watts max. I have limited my version to 33v (by programming this into the panel meter) but the MOSFETs are rated to 200v so it would probably work up to 100v.
Bear in mind that my heat-sink is 'live' to improve thermal performance, hence my self imposed voltage limit. Also, whether the copper heat-spreader I used was necessary, I really cannot say except that it works and I am happy with that.
Thanks again for your suggestions and constructive feedback Forum members.
Here is a pic of the CC Load display taken during testing of a 32v/5A (160W) Gophert CPS-3205 power supply.
-
Can you not do a quick summary of the parts that differ from the original load?,ie resistors changed and fets etc with the values?.
-
Hi @tekguy. Built that 4 years ago! A very handy instrument.
LM324 150W 10A Constant Current Electronic Load
Changes from standard kit:
x4 FETS changed to higher rated IRFP4468 130A 200v (preferably matched - see Reply #33. Other FETs are suggested by Forum members in the thread.
x4 FET Source resistors from 0.22R to 0.10R (5watt).
Added 15A automotive fuse and holder to input.
Added V/A digital panel meter with its own 7.4v 2S Li supply. Reply #15.
Added LM317 manual fan speed controller. A better design would be a heat sensing circuit to automatically switch on fan at a particular temperature.
Added 8.2R + 10uF snubber across input to suppress unwanted oscillation.
Added 7812 voltage regulator to stabilise TL431 and IC supply.
Added SPDT switch for local/remote input voltage sensing.
Added an additional 15K resistor in series with the helipot. The actual value depends on the final maximum current required.
Most of the parts are available on eBay and easily searched. Direct references are not given as they go out of date quickly as sellers change their listings.
Listing in Reply #35 may or may not work.
enut11
PS
Parts/kits seem to have changed since I built my electronic load 4 years ago. There are other designs, with fully built circuit boards and heatsink/fan already attached, available on eBay.
I did see the original kit that I built available from this supplier:
https://tinyurl.com/42ejr993
The photo below shows the general layout including two heat-shields, made of thick cardboard, to isolate the voltage reference.
-
where did the 7812 go in the schematic you talk of,can you draw it?.
-
@tekguy
If you look at the circuit in Reply #7, the 7812 12v regulator can be inserted between C7 and R7.
You will need 12vAC or 15vDC at the power input jack for proper regulation.
The 7812 is really a belt and braces solution. A well regulated DC input will also work.
enut11