24V Load Ballast

[T3sl4co1l]

Setting up to do some higher power load testing.  A long time ago I built a 0-20V +/-120A sync buck unit, working with a 24V (lead acid) reservoir, with CC/CV/programmable behavior.  Well, that's great if I'm powering loads big or small, but strong sources can easily overwhelm the battery (I have a pair of 12V 122Ah deep cycle batteries hooked to it), particularly if it's discharged a bit (between float cycles) so the ESR is high.  (The converter unit is protected against overvoltage, it simply turns off if its input rises too high -- OVLO.  But that's obviously a non-starter for sustained operation.)

So I'm making a load dump to wire in parallel with it:

https://www.seventransistorlabs.com/Images/ShuntBallast.pdf





Simply a shunt regulator, with switches into fixed resistors, so, hysteretic control.  Quite straightforward.

The relatively large capacitors guarantee that the cycle time is no faster than a few hundred microseconds; and they need low ESR (ca. 17mΩ ea.) so that the zero in the transfer function doesn't increase frequency a dangerous amount.

Here's typical operation with a 10 ohm load resistor and a moderately soft rectified supply:



Switching is quite well behaved, given the modest drive strength.  (An emitter follower is needed, to save on pulldown resistance, but otherwise things are quite relaxed.)



Note the overshoot due to the inductive load resistor, clamped by the TVS.  I've actually got SMCJs in here (handy), but SMAJ is more than enough.

Now to build the frame to support the resistors, and, well, the rest about the power supply I'm going to test with it...

Tim

[m k]

It seems that those multi layer modules are not always as complex as one could think.

[T3sl4co1l]

?

Tim

[m k]

Just that it can be totally functional without fancy curves, maybe even better and not overheating, what seems to be pretty common behavior these days.

Last of overheating ones I've seen is controller of electric blanket.
One could think that heat is something they have had in mind.

BTW,
some may think those other layers are a bit narrow but that's their problem.

[T3sl4co1l]

Sure, well it's hysteretic like any other of those controls.  And it would be equally disastrous for this thing to latch on, so I've at least used a fairly reliable switch (solid state), but there's no redundancy or protection here so if it does fail, it's going to get quite toasty, and sink the battery rather quickly (hopefully, perhaps blowing the fuse?).

Still not sure what "layers" is referring to, something lost in translation...

Tim

[T3sl4co1l]

And here's the mechanical build for it, mostly:



And now my hands are all pins and needles (well, mostly needles) from the fiberglass...

Probably want to add screens/guards around it eventually, but at least it's SELV so, as long as it doesn't fall over, right.  Need to make up some cables to wire up the resistors and board, and, probably just double-stick tape the board to the frame, don't really care to drill more holes and whatever.  The lugs are solid enough on the board (huge solder joint), even with the fairly inflexible 10AWG wire I have, they should last a long time without much work.

Tim

[m k]

Dang, I thought the circuit board was much bigger.
Oh, wait...

Nothing important with layers, just a sort of a comic relief and including a presumable definition of jump wire.
Since around here we are sometimes a bit picky.

BTW,
I'm quite sure that many have no real sense what kind of an energy hog is general heating and what it means with batteries.
My prediction is also that electric vehicles are not going to change that.
Same with dimensions and some young electricians with wire thicknesses compared to glass tube fuse internals.

BTW2,
I've crawled in the insulation just enough that doing more is not very tempting.
One fiber, possibly glass, was also inside of a side of a palm making skin effect for years, ripped out with pliers and what not but no difference, seems to be gone now without a notice.
Stamped hay also as a kid but around here a swimming water is never very far so it's compensated.

[SeanB]

Nice Sfernice adjustable resistors you have there.

[T3sl4co1l]

Nay, local brand name -- Milwaukee (now Vishay).

Tim

[SeanB]

Sfernice was also Borged into Vishay, they seem to buy up all the smaller competitors over time.

[T3sl4co1l]

Yep. They hold quite a number of traditional brands now...

Tim

[T3sl4co1l]

And, wired:



Added a stiffener across the bottom; G10 (~ FR4), roughed on the face, JB-Weld'd in place.  Shouldn't be coming loose in a while.

All the pink (polyester fiberglass) stuff is neatly sanded by the way, smooth to the touch, and also easy to glue if I should want to.  (Tempted to glue some of the frame joints, but having them removable via bolted connections is probably better for now.)

The wire is some crazy MIL spec, nickel-plated, mineral-loaded PTFE stuff I picked up surplus, years ago; as you might guess from the plating, it's a right bitch to solder (possible, but man does it take forever, hence I've not used it much in the past), but perfectly suited to crimped lugs as here.  The two wild wires on the sides, I probably should've cut to (direct line) length instead, but, eh, they look neat enough, despite the lack of wire tie?

Some final updates to the supply/converter it's attached to, and I'll be ready to test at power...

Tim

[TimFox]

Construction question:  how do you cut those nice gaps between rectangular pads on the circuit board?

[T3sl4co1l]

Utility blade -- one straight cut to score the copper, then a few cuts with the blade at an angle to make a V-score into it.  Clean up with a scraper (to flatten the top surface, removing burrs; or at an angle to smooth the sidewalls of the V).

I often plan layouts before cutting them, ...as you can see on the background paper.

That's a 2oz board BTW.  A bit more effort to cut this way, it's a big harder to score all the way through the copper.  1oz is probably ideal for this sort of technique, still thick enough to peel off when you need larger areas cleared.

Tim

[m k]

Have you thought Dremel style router?
Seems to be a top heavy yes.

I bought a set from China once, it was incomplete.

Some small vinyl cutters can possibly have enough pressure also but they need a flat bed.
I didn't know they can be as small as they actually are.

[T3sl4co1l]

I don't make nearly enough protos to justify a mill for myself -- but that's a good way to do it, when you already have one or when you do have that need.

Tim