Help with PSU (Temp control fan, load bank, & PWM circuit)

Hey guys, I've been thinking about building a load similar to the one earlier in this thread. Is there anything you would do different if you were starting from scratch?

Hey guys, I've been thinking about building a load similar to the one earlier in this thread. Is there anything you would do different if you were starting from scratch?

Click to expand...

The exact circuit topology used is highly dependent on the power supply you are trying to load up.

The one described here is good for one single very high current rail that is expected to be reasonably well regulated. It is a constant current load using N-FET's as the main heat dissipating element. The massive amount of watts being dissipated means massive cooling is needed. For this we decided to go with soldering the tabs directly to a copper pipe and run water through it to remove waste heat efficiently. This works well *IF* you only have one rail you need t load up.

Now... If say you want to make a load for a PC computer power supply with it multiple rails, you will have problems with the above design. That's because in the above design, using N-Channel FET's, we have to make it a common positive system, not a common ground system. This makes it impossible to connect two different voltages of a common ground supply to the same load because all the positives are inter connected. For this reason, I am working (slacking off actually) on a P-Channel FET version that is designed to have all the tabs connected to a single common ground. This circuit is very similar to the first, just the feedback circuitry is inverted. It can also be used to load up a single high power rail just as the first version can. So it gets the best of both worlds. The only drawbacks are dollar-per-watt the N-channel version is cheaper, and it is easier to conceptualize.

Edit: For say... a battery discharger or similar, the first unit with under voltage shut down would be optimal. Also things like fire box, battery thermal monitoring and shut down, and power meter/fuel gauge would be ideal. Though Li-chem battery's are best stored with like 70% charge if I remember correctly.

As to what he would do differently... I would say nothing, but I have no say since I didn't build his. I personally think it went rather well though considering. I have seen much more convoluted and ugly designs that left the builder completely bewildered and unable to make the device work even a little bit, let alone be able to finish it. That being said, I think that ronv did a nice job taking the key of the circuit and making it into a fully featured working "product" design. It has what it needs and it works, nuff said.

I need to stop procrastinating and make mine about the same already...

-()b

I am VERY happy with the dummy load as is. I know ()blivion mentioned a couple times that possibly we shouldn't have burned off as much energy in the current sense resistors because they do get quite hot, but then again, they are rated up to 275C, so it's probably fine.

https://www.mouser.com/Search/Produ...00virtualkey59420000virtualkey594-AC05W0R130J

There is the issue with the voltage and current jumping around slightly, I am not sure what could be done to fix that, or even if it is enough of an issue to be concerned with. I wonder if some capacitors in the right places would help...don't ask me what the right places would be though, just a thought from a novice. I do have caps on the power inputs of each op-amp, but I wonder if caps on the feedback inputs and or output of the op-amps would help noise issues too...I don't know, we need the experts to weigh in on that. I am just making wild guesses here based on very limited experience with these.

()blivion said:

The exact circuit topology used is highly dependent on the power supply you are trying to load up.

Click to expand...

It would primarily be used to discharge Lipoly batteries and occasionally to test single rail power supplies. I'm thinking something in the 500-600 watt range. Do you have any experience with the linear "Extended FBSOA" power mosfets that IXYS produces? They're expensive but they appear to be purpose built just for this kind of application. I'd probably use an Arduino or something similar to control it.

Dean

Boy they do make some brutes!
I have no experience with them, but the few I looked at were kind of low current/high voltage types.
They do make some really high wattage lower voltage types like:
https://www.electro-tech-online.com/custompdfs/2012/10/99176.pdf
To your first question:
I wish we had come up with a way to measure the amps without the X10 multiplyer.
Also wish we could scope the PSU to make sure the variation in voltage is not due to the load box.
Other than that it seems pretty solid.
Were you planning on water cooling (I was amazed at how well this works) or air?

Jer, Could you post a final picture?

I will post a final picture probably tomorrow if that's ok, I am working on writing up an a classified add tonight to sell an in-line power switch that I made.

I am not familiar with "Extended FBSOA" from IXYS. I will probably get crap for saying it, but a transistor is a transistor IMO. Safe operating area can be roughly approximated by the package wattage. The transistors you mention have great wattage, thus very high/large SOA. But as ronv pointed out, something with a lower voltage and higher current might be a better fit. Any kind of sub 24 volt system is certainly going to get all it's watts from current. So it's best to have overhead in that aspect of your part.

Li-Po charging/discharging is tricky business because (as I'm sure you know), Li-X batteries have a tendency to blow up if you charge/discharge them faster than their C rating. They also will be permanently damaged if you discharge them past a certain point. The best approach (as you would like to do anyway), is to use an MCU system to carefully watch the charge-levels/current/temp during the whole process. Now one *CAN* go with Arduino to do this. There is certainly nothing wrong with Arduino based systems. But one could minimize cost and size and create a more customized task specific system if one uses a bare PIC/AVR/$MCU on a proto PCB. The downside is there is more development time necessary to get a finished product. But that is not a large problem as this site happens to be great for such things. (For example, I can develop PIC based systems).

Before you do anything though, might want to search the net for a Li-X discharger schematic that has already been built. If you can't find something, THEN might be time to make a new thread.

ronv said:

Boy they do make some brutes!
I have no experience with them, but the few I looked at were kind of low current/high voltage types.
They do make some really high wattage lower voltage types like:
https://www.electro-tech-online.com/custompdfs/2012/10/99176.pdf
To your first question:
I wish we had come up with a way to measure the amps without the X10 multiplyer.
Also wish we could scope the PSU to make sure the variation in voltage is not due to the load box.
Other than that it seems pretty solid.
Were you planning on water cooling (I was amazed at how well this works) or air?

Jer, Could you post a final picture?

Click to expand...

Yeah, most are 500v & up but they do make a few 100v & 200v parts. One of the 100v parts is good for 200A. Check out this puppy: IXTK-TX200N10L2 (scroll down to the bottom to see the SOA graph). Too bad they're so spendy.

Dusey52 said:

It would primarily be used to discharge Lipoly batteries...Dean

Click to expand...

Interesting, I didn't even think of that. I guess it could be used to discharge lipos for storage, or could it, hm, I don't know. I can discharge them with my PL8, so it's not really needed, but an interesting idea anyway.

Made my first sale...whoohoo!!! (just a y-connector with an in-line on/off switch, haven't started selling the PSUs yet).

jocanon said:

Interesting, I didn't even think of that. I guess it could be used to discharge lipos for storage, or could it, hm, I don't know. I can discharge them with my PL8, so it's not really needed, but an interesting idea anyway.

Click to expand...

The discharge capability of most chargers is pretty wimpy. With your load, 'as is', you'd have to watch it like a hawk without an auto-cutoff circuit else you'd destroy the battery.

It shpuld be pretty easy to discharge them at 5 or 6 C rate with a low voltage cutoff at 3.3 volts or so per cell.
May be time to start a new thread to get more intrest and details.

Agreed. I didn't intend to hijack this thread. I was just curious if you'd do anything different if you were to to it again. Sounds like you're happy with it.

Dusey52 said:

Agreed. I didn't intend to hijack this thread. I was just curious if you'd do anything different if you were to to it again. Sounds like you're happy with it.

Click to expand...

No problem Dean, let me know if/when you start a new thread, not because I think I can contribute to it much, but more because I would be interested to follow it. As to the discharge rate, I think we can go up to the rating of the lipo, correct? These are high discharge lipos, for instance, my 4 cell 2500mah GensAce lipos are 25C discharge rate, in theory, could I safely discharge them at that rate with my dummy load as is, but I would just have to watch it like a hawk to be sure it doesn't go below 3.3 volts or so?

In theory you could disharge them up to their 'C' rating. In practice, I wouldn't go that far. In any case you'd want to be very carefull not to over discharge them. Imagine discharging your favorite battery at a 30 or 40 amp rate, the phone rings, you get distracted, the next thing you know your battery is toast. A low voltage cut-off would prevent that from happening.

Final Product!

Here are a bunch of pictures from different angles. The basic frame work to support the PCB is a 1X2 board I screwed into the sides of the CPU case (this really helped make the case more sturdy) then I attached 2 1X2s stacked on top of each other on each side perpendicular to the one going across the width on which I mounted the copper pipe and the PCB so eveyrthing is quite sturdy in there.

I can't thank you guys enough ronv and ()blivion for helping me build this! It's amazing, exactly what I envisioned when I set out on this project and you guys really made it happen! When I am all done with my power supplies (artwork and all), I will send you each one as a token of thanks if you would like.

I will send you each one as a token of thanks if you would like.

Click to expand...

Sure, I would take one if you're offering. Can't think of what I will do with so much power. Weld I guess... LOL. I'll take any broken ones too if you don't mine. I'm sure to need spare parts for when I melt the good one.

I PM'ed you my shipping address.

In theory you could disharge them up to their 'C' rating. In practice, I wouldn't go that far.

Click to expand...

Yup, really bad idea. Li-X cells are particularly bad when it comes to wear and tear. And even though they may be rated up to a certain level, they still wear down faster if you discharge them quicker. Especially if that discharge rate is close to the max allowed.

I personally wouldn't make a battery discharging system unless I had very very compelling reason to do so. With most all secondary battery technology, you can only cycle them so many times before they accumulate permanent damage and become rather useless. It's a little different with Li-X since they already have a ticking timer on them from the moment they are manufactured. But still, you can always shorten that timer by being rough with them.

()blivion said:

Yup, really bad idea. Li-X cells are particularly bad when it comes to wear and tear. And even though they may be rated up to a certain level, they still wear down faster if you discharge them quicker. Especially if that discharge rate is close to the max allowed.

I personally wouldn't make a battery discharging system unless I had very very compelling reason to do so. With most all secondary battery technology, you can only cycle them so many times before they accumulate permanent damage and become rather useless. It's a little different with Li-X since they already have a ticking timer on them from the moment they are manufactured. But still, you can always shorten that timer by being rough with them.

Click to expand...

All good points. The two uses that I'm most interested in are:

1) Discharging from full to storage charge in a reasonalable amount of time. As you know, LiPoly batteries don't like being stored at full charge and the discharger feature in most hobby chargers is pretty low power (slow).

2) Testing/verifying manufacturers wild claims about their battery specs. The quality of these hobby batteries is all over the place and it would be nice to be able to bench test them at the loads they'll see in actual use. Voltage sag under load, actual capacity, etc...

Those are good ideas, I really like the idea of testing the Lipos...you could post up the data you collect in rcg. I bet a lot of people would appreciate that, maybe not some manufacturers

I worked for a while on a similar project for lead acid batteries (solsr systems). Very similar to the load box with a constant current load, low voltage cut off and a timer. But pretty expensive just to find out which batteries are the best. You could do the same thing with the load box, a voltmeter and a watch. It would just take a lot of time.