Author Topic: Dump Load Controller (Read 7051 times)

commanda · « **on:** June 01, 2006, 08:45:28 AM »

This project has been on the back-burner for quite a while now, so I thought it about time I made some progress on it.

It is based on the LM3914, with a few bells & whistles. It shows the correct way to get a voltage output from the LM3914. The output is latched (74HC374) so you don't get dump loads cycling on/off rapidly. The '374 is clocked by a '555, probably at 30 second intervals. There are 8 dump loads switched in sequentially as the battery voltage rises. The first led is green, and the circuit is adjusted such that this led comes on at the float voltage. The next 8 leds are amber (these operate the dump loads). The last led is red, and also connects to the alarm circuit.

The individual dump loads are fused, and there is fault monitoring for each (74HC86). If the load goes open circuit, or the fuse blows, or the fet goes short-circuit, will raise an alarm immediately (whether the dump load is active at the time or not). An open circuit fet (unlikely) will not raise an alarm until that dump load becomes active.

There is also a mute switch for the audible alarm, which also activates a flashing red led whilst muted.

There is also a switch on the sense input, to switch between float and boost charge. Not shown on the circuit, but it also turns on a flashing led whilst in boost mode.

Physically, it is being built into a plastic 1RU rack case.

Photos to follow as it goes together.

Amanda

WXYZCIENCE · « **Reply #1 on:** June 01, 2006, 10:20:08 AM »

Commanda, what type of fet are you using for you load control. Are you driving it with 5v from the 374? Joe.

commanda · « **Reply #2 on:** June 01, 2006, 03:11:44 PM »

Fets I have are IRF540. The '374 is cmos, and runs off 12 volts.

My symbol in Autotrax for the '374 doesn't have supply pins.

Amanda.

commanda · « **Reply #3 on:** June 01, 2006, 04:58:29 PM »

A couple of (very) quick photos.

I've split the circuit over 2 pcb's. First pcb almost assembled contains the LM3914 (under the ribbon cable), transistors, and ribbon cable connection for the leds.

Second pcb, of which I've done the layout but not yet etched the board, contains the rest of the logic. The 2 boards connect together with ribbon cable.

Across the front of the rack are the 8 x 10 amp DC circuit breakers.

Amanda

willib · « **Reply #4 on:** June 01, 2006, 11:58:31 PM »

nice work amanda !

have you tested it ?

what method did you use to make your boards?

i wish another person had started his post with a schematic * cough * oztules * cough cough *

commanda · « **Reply #5 on:** June 02, 2006, 02:08:42 AM »

The LM3914 part has been prototyped & tested using a kit from Jaycar of the circuit published in May 2006 Silicon Chip magazine. It's set with a range of 1.0 (one point zero) volts from led 1 to led 10, for a 24 volt system that's 27.0 to 28.0 float, and 28.0 to 29.0 in boost (equalize) mode. These are flooded lead acid batteries, and I'm open to suggestions on exact voltages to use.

The rest of the logic hasn't been prototyped & tested, but it's nothing I haven't done a zillion times before, so don't expect any surprises.

The first board with the LM3914, which you can see in the photos, was done with the "join the dots with a texta" method. I powered it up this afternoon with no surprises.

The second larger logic board will be done with transfer film. Print the artwork onto transfer film with a laser printer, then run the film & blank copper clad board through a laminator to transfer the ink to the copper.

Amanda

oztules · « **Reply #6 on:** June 02, 2006, 06:49:56 AM »

Good too see your "to do" list is turning over to your wind projects (as opposed to the transport projects)

Shaping up to be a very nice comprehensive dump controller. I can't think what else you could add to it.

I had scratched out a circuit using 2 lm393's over a simple 9 stage voltage divider and go from there, but after seeing this......maybe I will change that into a peak current meter for measuring current surges in equipment to get a feel for the peak draw on startup (fridges, etc) as opposed to what I can measure with normal meters....... and await your projects completion.

You have obviously given this some thought. Sooo.... Why did you not go for a PWM solution? I'm not smart enough to know why, so I will build both and see what advantages each system provides.

I know Hugh and others have all used the stepped approach, but i haven't figured out why.

very well done thus far Commanda

............. oztules

amiklic1 · « **Reply #7 on:** June 02, 2006, 07:04:52 AM »

Thanks for the piuctures. Could you upload the PCB pictures to make it easier to work with this? I would like to build this one and to take some experiments.

commanda · « **Reply #8 on:** June 02, 2006, 04:09:44 PM »

I can't think what else you could add to it.

I can. An R/S flip flip in place of the boost switch. Push-button turns it on (Set) and starts a timer. Timer end Resets the flip flop. This might become an add-on at a later date. Can be done with a single 4093 for shorter time constants, or a 4093 and a 4020/40/60 for longer time constants.

Why did you not go for a PWM solution?

Ultimately, part of my dump load might be an underground reticulation cooling system. The stepped control will make this sort of thing easier, as against a pwm.

The fault monitoring also makes it practical to use low-reliability loads, like light bulbs.

Amanda

commanda · « **Reply #9 on:** June 02, 2006, 04:13:53 PM »

I will upload the Autotrax pcb files, once I get the second board built & tested, fix any last minute tracking errors &/or changes, put values on all the components, and make the component numbering match up between the circuit and the pcb's.

Amanda

BruceDownunder · « **Reply #10 on:** June 03, 2006, 05:51:23 PM »

Great work Commanda,as usual.

lovely to see some one actually "doing it"

Bruce

domwild · « **Reply #11 on:** June 04, 2006, 12:44:44 AM »

Interesting work!

commanda · « **Reply #12 on:** June 09, 2006, 10:15:56 PM »

It's official; Amanda has finally cracked and lost the plot.

You can't run 74HCxxx off 12 volts.

Can't believe I actually did that. I work with this stuff every day. Maybe familiarity does breed contempt.

Anyway, no harm done. Thank goodness for current limited bench power supplies. The circuit actually works just fine off 5 volts anyway. Except I'll have to get some fets with logic level gates.

Amanda

oztules · « **Reply #13 on:** June 10, 2006, 02:01:38 AM »

Must be the season for it. I designed my l4990 pwm booster, and couldn't for the life of me work out why it i couldn't even get the oscillator to run, checked the data sheets..... ooops the threshold voltage is only 8v for the l4990a........ but smarty here is using l4990..... identical except 4990 is 16v 4990a is 8v...bum.. thats the only difference.

So now it's a 24v system booster.

Some days it pays for me to read what it actually says, not what i think it says.

Your not alone..............oztules

commanda · « **Reply #14 on:** July 07, 2006, 09:09:53 PM »

Final couple of photos before I send the front & back panels to the engravers for some labelling.

Overall, it's a bit of a rats nest, and a single unit rack enclosure is a bit of a tight squeeze. I have tried playing around with the pcb layouts, to get it a bit more tidy. I'm trying to get all the leds and switches mounted directly on the pcbs. Whilst it would certainly decrease the amount of wiring, it also cuts down on the options of exactly how to build it. I have a new board, which mounts 4 of the fets and an optional logic level translater chip, with screw terminal terminations for the heavy drain & source wires. Haven't built it yet. I will eventually upload that artwork for anyone that wants it.

I used 2 pin microphone connectors for the 24 volt dc input, the temperature sensor input, and the temperature meter itself. The cable bundle has a 12 way automotive connector on the end. 8 red wires go to the dump loads, and 4 black wires go to battery negative.

The 2 little 50mm square by 10mm deep fans in the lid still need a series dropping resistor to slow them down a little, and their own switching fet.

Amanda

dinges · « **Reply #15 on:** July 07, 2006, 09:18:23 PM »

Funny. Was wondering today how things were going with your controller.

Looks pretty nice, despite the rat's nest. But there's definitely little space left. A bit of lacing wire and bundling of cables will make it look much better.

Am definitely interested in the new board with logic-shifter (was thinking myself, wouldn't a resistive divider be enough?) and in-built FETs. Will build one such board for the 19" rack, so I have a stand-alone dumpload unit. But should I ever have to dump much more energy, a different, stand-alone power dissipating unit will have to be built. Am already toying with the idea, but it's all still very vague.

As for construction of mine, am still waiting for the perf. board to arrive so I can start constructing the main board. Hopefully main board will be started & finished next weekend. Then wiring up the backside of the 19" rack and off we go...

Peter.

(PS: it's very silent w.r.t. your controller from the rest of the board; am I the only one who is or will be building one?)

commanda · « **Reply #16 on:** July 07, 2006, 11:24:51 PM »

Peter,

You probably are the only other person building one right now. There are several issues here as I see it.

One is the electronics ability.

Two is the realisation that this level of technology is what it takes to actually live off-grid.

In my case, what will probably happen is the land will be paid off, the house will be built, and my partner and the two kids will move there, while I'm still in the big smoke for a few more years. Without me there 24/7 to babysit the system, it has to "just work".

Whilst I admire adherents of the KISS principle, and scrounging and making do with what one can acquire, the other side of that coin is that in the long run you can't beat purpose designed & built equipment. And whilst some bits can be bought off-the-shelf, what do you do in X number of years from now when it breaks down. Maybe the manufacturer no longer supports that particular model. Or its got to be sent away for months whilst some "agent" tries to fix it. Or if you build something based on some scrounged bit of gear, how do you replace it if it breaks down.

I'll upload the artwork for the fet board shortly. Haven't done a circuit diagram as such because it's so simple. The logic level translater chip is an HEF4104. If you can't use Protel Autotrax format, contact me directly and I'll do a pdf version for you.

Amanda

oztules · « **Reply #17 on:** July 08, 2006, 03:24:46 AM »

It looks the part Amanda.

I feel that over time it will become THE last word in DYI dump controllers. As for builders, they will come. Many more people seem to build the Dans mills, than advertise that fact here. People turn up here having built one years ago, and I'm sure many will never make themselves known.

Even if the whole unit is not replicated, I'm sure different bits will be borrowed and put into different designs, that suit other RE folks.

It's also nice to see Peter has made a tidy fist of it so far.... my turn is coming too.

..........oztules

ghurd · « **Reply #18 on:** July 08, 2006, 10:11:05 AM »

Amanda,

What voltage are the steps between dump loads?

Glen

dinges · « **Reply #19 on:** July 08, 2006, 01:56:13 PM »

Ghurd, I think the answer is in post nr. 5; there's a 1V window (between 27.0V and 28.0V), so I guess steps of .1V. (in boost/equalization mode it's 28.0-29.0V).

The good thing is if one dumpload fails, battery voltage will rise till the next dumpload kicks in. Seems like an inherently safe design.

Peter.

commanda · « **Reply #20 on:** July 08, 2006, 05:15:47 PM »

Peter is correct. I've calibrated mine to have 1 volt full scale deflection.

The two adjustment pots on the LM3914 are connected in parallel (not in series like most circuits). This means you can get the full scale deflection extremely small.

It is also perfectly reasonable to calibrate it any way you like. For instance, you could set it up to be a battery meter, say 20-30 volts, and only hang dump loads off the top 3 outputs.

Amanda

News:

Author Topic: Dump Load Controller (Read 7051 times)