AC-DC Power Supply and Load Bank Build

Great! <...we...> need to hire a qualified local Electrical Engineer, then the project will really get cracking along.

Seriously?

So we are currently considering a transformer and phase converter setup to move to 240 3-phase and our current service will easily handle a branch circuit up to 120A (after efficiency calculations) so theoretically power isn't necessarily a concern.
Here is the problem in a nutshell: frankly we don't need to purchase a power supply that has incorporated technology of anything other than providing the power. We would build all of the controls and optimize outputs at the software level. Power supply build isn't our main concern truly (we would just be wasting 3-5 thousand dollars if we purchased a power supply capable of outputting 5kW as they integrated systems are not necessary to pay for others engineering because we can handle that)
In reality the bulk of our issue is finding an acceptable and safe solution for load banking. Just using thin film or wire wound resistors is not the real answer as we should probably have high power n-channel mosfets, a sensing resistor to ground, and a dual op-amp setup to provide loading...but man the design is killing us.
We know we can build one, control it (again at the software level) and come in much less that the 3-5 grand you would pay for a 5kW dummy load. Therein lies the problem. On the software side, we have unlimited free resources, we have plenty of empty enclosures awaiting whatever we would want to build, so essentially our only cost is the component BOM

Rent an appropriate sized welding machine for your DC power, then buy a bunch of these for your loadbank. It shouldn't take more than an hour to size the components.

But I sense you want sophistication without the cost; unfortunately that takes more money/time, you get what you pay for.

Not necessarily sophistication without the cost, providing the sophistication internally.
We had looked at thin film or wire-wound resistors but having a sensing and driving circuit would be best so we can shunt, and control.
As far as the welder goes, im going to go out on a limb and say that while an arc welder would certainly provide high current output, welders are now mostly equipped with ground fault protection and thermal resets. Similarly, they wouldn't be as nice as we would like. Anything we would be powering would have multiple sets of redundant wireless communicating master/slave dc-dc converters for power conditioning, with multiple levels of redundant circuit protection but I still don't think i would trust the output from an arc welder to be the input into one of our very sophisticated power distribution units.
thoughts?

DC is DC, and capacitors and inductors can provide filtering, but you have have more on your plate than "just" a 5kW power supply/loadbank. You want things like tightly controlled, zero ripple, DC output into a computer controlled, infinitely variable loadbank.

Once you tell the forum what your design parameters/goals are, you might get better answers. Having done MW-sized projects like yours, I can empathize with your desire to get this right the first time. That's the purpose of properly written specifications, and what I meant by "sophistication".

Just a remark. Sizing of wirewound resistors for long term service rarely leads to expected results.

Has anyone had any experience that disagrees with this statement?

You are exactly right and therein lies one of the problems with wirewound.
Thin film give much better longevity, precision, and accuracy
Lets face it though, I could easily have multiple different thin film resistors for each output on the distro but you are by no means looking at a continuous output if you are just running through one resistor as a load. It sounds good out loud but is terrible on paper.

That depends upon your design goals, whether you're designing to cost, functionality, and/or longevity. No experienced power engineer would use a single value or series string in a mission critical application, nor would they choose a power rating that does not have an adequate safety factor. Instead they would choose a series-parallel combination of resistors with a power rating to ensure that the loss of any single path does not materially affect the EUT (Equipment Under Test).

If you like thin film because they're compact and non-inductive, then remember to derate them to accommodate their temperature time/current overload/heating characteristics. Dissipating 5kW continuously in a compact space is a non-trivial pursuit, especially if failure is not an option. TANSTAAFL.

At 5kw I would use multiple 500 Watt wire wound resistors in parallel at about 50% of their power rating. In my experience vitreous coated wire wound properly sized resistors will last for >40 years. Some of them have a very low temperature coefficient.

I don't see film resistors being practical at 5kw dissipation.

Again, as others mentioned, you have given no real design specifications.

A Resistive load bank, the most common type, proves equivalent loading of both generator and prime mover. That is, for each kilowatt (or horsepower) of load applied to the generator by the load bank, an equal amount of load is applied to the prime mover by the generator.

A simple Google search found this:

https://accusrc.com/product-Agilent-Keysight-6683AJ01-9145?gclid=EAIaIQobChMIt9nHpJ_72QIVgWp-Ch26nAWKEAQYASABEgJsM_D_BwE

I seriously doubt that you can design one cheaper.

Could certainly design cheaper in terms of parts, maybe not labor, but am HOPING for single phase so we don't need to install a static phase converter (though it isnt out of the question)
Thanks though

What are the specs for required ripple and regulation?

Note that the unit shown requires 208V 3-phase power. That is common in the U.S.

You do realize the whole AC versus DC power distribution debate that happened over a century ago also considered is it more cost effective to distribute high voltage or high current locally. Silicon (semi-conductor) is far more abundant in the earth's crust than copper. Copper is so expensive today they don't even make US coins out of it. I would nominally use multiple power supplies distributed with one at each load instead of one large power supply and a lot of copper distribution bus work to minimize voltage drops.

Your power distribution ideas are impractical, to say the least.

Then again I've found that most physicists think they can ignore the laws of physics.

I think the difference between a physicist and an engineer is more along the lines of what is theoretically possible and what is practically possible.

Engineering is the art of applied science. Nuts and bolts approach.

Physicists are more theoretically inclined and are more likely to be star gazing.

Except that I have a BS in Applied Physics, Masters in Quantam Electric Field Theory and BS in Electronics Engineering with a concentration in digital signal processing...oh yeah and 8 years experience in naval ac/dc electronics.

So saying that this ‘physicist’ is a star-gazer is quite misguided. I see your point and raise your three degrees and nearly 20 years experience.

we could certainly use multiple bussed low power supplies combined for load sharing but are looking at designing something with ease of processor level control. LabView is my baby and I will have multiple component constructed solid state relays all bussed together with each output having processor level control.

Again I say, power supply is not the main concern, per sé. It’s more of a: this is the supply I plan to build, but 5kW variable load banks are expensive to purchase. Thanks for your criticism though ��

Oh puhleeze!! I'll see you with a BSEE(Power Engineering), MSCS, and PE, then raise you with twice as many years of hands-on experience designing, testing, and commissioning high power AC/DC generating/control equipment, then re-raise you with 8 years as an Adjunct Professor developing/teaching Masters-level Power Engineering courses and Digital Lab Techniques. So check the alphabet soup after your name along with your ego, at the entrance to this forum.

Many posters have called you out for your lack of the critical specifications that are necessary to properly design your loadbank. Most of the bench engineers I've known could spec/design your circuitry on the back of a napkin during their coffee break. Either impress us and provide the data, or fold.

Well then, I bow to you kind sir with what seems like 6+ years of school, 40 years of experience and 8 years teaching. So at 71 you must be quite froggy.

You see the response was referencing someone saying that as a physicist I stargaze and can’t apply knowledge, so I provided credentials (which didn’t include my time as an adjunct physics lecturer but since you threw it in I will as well)

As far as the loadbank is concerned, what specs are you looking for? No worry for ripple since power will be conditioned out of the distro.

so make it easy: I need a total load of 175A continuous at 28VDC.

I have 28 outputs from my pdu ranging from 1-35A, so I need a loadbank that can load them all individually and combined. Now technically the entire system will never exceed 40A but when combining the specs of everything being powered, the total need is 175A and each output is protected with multiple redundancies accordingly. Each output is also controlled via touchscreen with embedded VIs using a National Instruments SbRIO. Each solid state circuit I have created is engaged via 3.3V signal, dual reverse polarity redundant protection, 3 outputs are 12VDC and one is 5VDC and utilize master slave wireless load sharing dc-dc converters.

Particular enough for you or should I continue?

and if you can sketch on a napkin a load Bank capable of being controlled with LabView, and varied throughout the entire spectrum of need with that many inputs...well excellent job old chap

Let me summarize your comments. You claim multiple advanced degrees in Science and Electrical Engineering. You also claim decades of employed experience using this schooling. You come to this open forum claiming to design low voltage high current DC power distribution asking how to build and modify test loads and the supplies themselves.

This should be a trivial task for somebody with your claimed schooling and experience. Obviously either your schooling was not very good, you forgot all you were taught or they are just false claims. For your sake I hope the latter is true. Liars can change for the better.

You know, you’re right, I’m probably lying about what I build...or I am like every other engineer on the world and know that I will always have more to learn and look to my industry for assistance. Perhaps I shouldn’t have considered reaching out only to be ridiculed. Forgive me for utilizing an engineering forum for assistance in developing my thoughts into action.

<...Instrument Physicist...> is a role with a different skill set to that needed, which falls within the domain of the Electrical Engineer.

Stephen Hawking was a well respected physicist and "star gazer".

I am sorry you take offence at being lumped into the same category.

Look up at the stars and not down at your feet. Try to make sense of what you see, and wonder about what makes the universe exist. Be curious. Stephen Hawking

Read more at: https://www.brainyquote.com/quotes/stephen_hawking_627123

In my family we have engineers and a physicist, whats interesting is that some of the engineers are nuts and bolts guys/gals and some are star gets.

But that is cool, let them be what they want to be. They are all making a contribution to our country. Our physicist had an office he shared with another teacher at his university, he showed me on his blackboard how the beauty of math.

High Current DC Power Supply - TS Series: 5 kW, 10 kW, 15 kW, 20 ...

and

Load Banks

Thank you. Easy to buy, not quite as easy to build to specs. DO you have any insight perhaps? Thanks in advance

I like to know if there is any way to diy a load bank , for example , simulating a motor load ...