Dump Load Controller Mk2 boards

[commanda]

These are the second version pcb files. All the led's & switches are mounted directly on the boards. This will tidy up the wiring, at the expense of more pcb real estate, a few more wire links, and a little more restrictive on panel layout and choice of switches.

Board 1 (LM3914)




Board 3 (Timer)




Board 5 (Alarm)




This one hasn't previously been published. It physically mounts 4 fets on one board with a HEF4104 level translator chip, with 7805 & 7812 regulators. With this you can use regular fets rather than the logic-level units I used originally. Layer the Drain & Source tracks with solder to reinforce the pcb tracks.




Board 6 (Fet Interface)




These boards should mix'n'match with the original artworks.

Caveat Emptor; I haven't actually built these. I have double & triple checked them, but I'm only human after all.

Amanda

[dinges]

And in case anyone wonders: point-to-point wiring also works fine. But next time, I think I'll get me a wire-wrap set first...

So if you're desperate enough to build one, there's no need to etch PCBs. Vero board works fine too, and is more flexible. Besides, all my projects are basically prototypes anyway.

How's work coming along on your controller? (I mean, the actual building of it).

[phil b]

Amanda, what format are the files in? My new Windoz system thinks it's Express PCB. Is that correct? Can you post in another format?

Thanks, Phil

[zapmk]

Amanda,

Iam in the same boat as Phil b, (I've always use PCB Express)

would love to build the boards.

But total clueless on how to load or use AutoTrax on a windows system.

 -Zapmk

[commanda]

Dinges,

You gotta learn how to make pcb's. After you do you'll never use veroboard again.

Next time I make a board I'll take some piccies and write a howto.

Autotrax is available for free download. Oztules posted the link recently. It is a dos program. On windows, right click on traxedit.exe and select properties. Set it to full screen mode. I also posted a while back on where to get the vesa drivers so you can run it at greater than vga (640x480). I've used it for more years than I probably care to admit. Good thing is the drivers used by the pcb manufacturers are also free, so any shop that makes pcb's can handle the format. Unlike expressPCB where only one shop can make the boards.

Phil & Zap;

I downloaded expressPCB; can't see any way to convert formats, unfortunately. If all you want to do is make the boards as I've laid them out, I can convert them to pdf format. If that's the case, let me know and I'll convert the lot to pdf and make a zip file of it.

Amanda

[zapmk]

Amanda,

If it would not be to much trouble, PDF format would be great.

 -Zapmk

 

[dinges]

Thanks Amanda,

But I have made PCBs in the past. I'm not good at it, but I could if I put my mind to it. It's something I hate doing though.

Secondly, with pre-designed PCBs, you're stuck with what the designer put in them. If I have components that are slightly different, it's hard to use them, without messing about. Want to use a different potmeter? 10 turn vs. single turn? Or vertical instead of horizontal? No problem with vero-board.

Since most of my projects are prototypes (one-off's), I love vero-board. For its flexibility. It can be pretty quick to build too. You just have to stay alert, and everytime you install a component or a wire link, mark it on the the schematic. Otherwise, things WILL go wrong.

Once I learned working with vero-board, I switched over from etched PCBs and have never looked back!

I'm pretty experienced with this mode of construction and love it. It can be mechanically very sound too. In fact, even better than with PCBs. Plus, before I build, I usually give the vero-board the shape of the house it will be in. That way, I can fill every nook and cranny of the case, should I want to. And, you can build VERY small with vero-board. Combination of SMD and through-hole mounting can yield very tiny products!

As far as a description of etching/creating PCBs goes; you could write another explanation yourself, but there's plenty of descriptions to be found on the web, for anyone interested...

BTW, good to see there's quite a bit of interest in that dumpload controller.

[Dave B]

Just to add a little more confidence to a dumpload circuit using the LM3914. I designed and posted my dumpload controller using the LM3914 nearly 2 years ago and it is still working fine. I trigger solid state relays off of the output to any selected LED(s) as using the LM3914 as a graphic volt meter. Very simple, very effective and has built in hysterisis which is adjustable that delays triggering on or off of the relays which can be more efficient in gusty conditions. I am using 2  LM3914's to drive 15 LEDS so 15 trigger levels can be selected and with a pot to adjust the voltage range this gives an unlimited selection of voltage levels to trigger from. Anyone interested in more detail just let me know.  Dave B.

[commanda]

Dave,

I presume this is the article you're talking about?

http://www.fieldlines.com/story/2005/1/3/222114/8823

What happened to your photos?

I know I went completely nuts and over-designed it to the max, but whatever. I have my reasons (other non-tech members of the house and the fact I may not always be there to babysit the system being very high on my design criteria list).

Solid State Relays (SSR's) are a great invention. Pity they cost so much.

I presume your alternator is 3 phase, and you're rectifying to dc to feed the heating elements?  Any problems with electrolysis?

One problem I do see with your circuit is running the SSR's in parallel with the leds. Different color leds have different forward voltages; some leds (red in particular) might not have enough forward voltage drop to trigger the SSR.

Hysteresis; I solved that problem by using a clocked latch (74HC374).

Good to hear yours is still running fine after 2 years. Hope I can say the same 20 years from now.

Amanda

[commanda]

If it would not be to much trouble, PDF format would be great.

Not too much trouble. Couple of days; when I get back to work on Monday. Will add external dimensions so you can scale your printer if need be.

Amanda

[commanda]

Peter,

Wire-wrap; been there, done that. Beyond a certain level of complexity, becomes impossible. A first generation Z80 with all its support chips, PIO's, UART's, memory controllers, glue logic, etc, on a 6" x 4" STD bus card is impossible. Well, almost impossible. I know, first hand experience. Even more fun when you walk into the project half-completed, and are given the responsibility of making someone elses circuit work. Like I said, been there, done that. Wish we had digital cameras and the internet back in the '70's and early '80's.

As for my controller; was finished several weeks ago. Currently waiting on the front & rear panels to come back from the engravers. Then I've got to find some heating elements, cut them down, and mount in a frame that will sit in some air-conditioning duct. I've found the address of my local heating element specialist, just got to get around there one lunchtime.

I find it hard to believe you mount smd components on veroboard. Maybe you're just crazy :-). Or at least, crazier than I am. All the smd stuff I use is 50 thou pin spacing, half of veroboards 0.1 inch track spacing. Or worse, metric pin spacing (CS5534). Last week I made a board to adapt a surface mount 8 pin chip to a board designed for a through hole part. The boards final size was 1/2 inch by 1/2 inch. Made a prototype using the same technique I use to make all my boards.

It probably is redundant, but I made some boards today, and took some photos of the process. I've been asked often enough about it now that I think it could be useful for some on here for me to actually detail how I make boards. Once you get it down pat, from artwork on computer to physical pcb ready to drill & populate is measured in tens of minutes. No darkroom, no yucky chemicals except for the etchant.

Meanwhile, work continues on the Grand Unified Dc-Dc Converter (GUDDC) Trademark & Patent Pending; just kidding. Don't believe in Patents. Don't believe in sweeping all-encompassing statements either, which makes the previous statement either redundant or irrelevant. Patents do have their place. On the other hand, "there is nothing new under the sun" is also true.

Amanda

[dinges]

LOL.

Never used wire-wrap myself, but I understand that's how many people built their first computer. Actually, I've read stories about the Z80 & wirewrapping, so... Maybe it has to do with the operator, not the process BTW, checked the prices of simple wirewrap tools. Amazing. They must be made of gold, with a diamond tip. And these are the manual ones, mentioned as 'budget' stuff (with excellent price/value ratio). Better keep on looking, maybe the 2nd hand electronics websites have one, or e-bay.

By veroboard, I mean print with .1" spaced holes with little copper islands. SMD resistors and capacitors can be mounted very nicely on them! ICs not, of course. I find this method ideal for adding small pull-up/down resistors, and the powersupply decoupling caps. Will make a few pictures of the finished boards, should clarify my construction method a bit more. Top layer wires all going in one direction, bottom (solder) layer perpendicular to that. If space is really at a premium, plain point-to-point wiring.

I assume you make PCBs by means of the photocopier & iron method? Have heard a lot about it, tried it a few times, with mediocre succes (though the PCB was usable, but barely). But especially in the manufacture of RF prints, I'm still looking for a good solution. Amaze me, please.

Peter.

[Dave B]

I really don't know what happened to the photos. I know enough about electronics to be dangerous and can usually design simple circuits. Although maybe not the most efficient I like to keep it simple. Solid state relays are expensive but again this makes for a more simple circuit and the proof is in the reliabilty so far (it would be interesting to know how many times these have switched.) I am rectifying a tap off the wild AC for my DC input to the LM3914, it is here that a filtering cap can be adjusted for the hysterisis. This was a nice accidental discovery which I didn't plan for. I am driving the heating elements with the AC. I have 5 green, 5 yellow and five red leds indicating the voltage ouput steps in the bargraph. I believe you are correct in that when selecting a red led to trigger off of I need to open the connection to the led in order to trigger the relay. Not a big deal but just does not light at the same time as the others would. I really need to design the power fail circuit which would automatically connect a load if the grid fails. Free wheeling is a scary thing. Any ideas on this ? Maybe just a normally closed AC relay to start with, I'll figure something and it will be simple. I think I'll try etching boards in the future, it sounds like fun. Thanks for sharing your circuits although like others I cannot open your files. Dave B.

[commanda]

Theres probably a simple fix for the red led problem.

Put a resistor in series with the led. Connect the SSR across the resistor/led combination. Measure the current in the led, and select a resistor based on ohms law to drop a few volts across the resistor.

R=E/I

so if it's drawing 5 mA, a 470 ohm resistor will drop approx 2.5 volts. Added to the led voltage should be enough to turn on the SSR.

Of course, looking up the datasheet on your particular SSR's so you know definitively how many volts they need to turn on reliably and working back from there is the way to go.

Technically more elegant would be to run all the leds with series resistors and parallel the SSR straight off the resistor. Size the resistor to exceed the minimum turn-on voltage of the SSR. Then it wouldn't matter what color leds were used where.

Amanda

[commanda]

Here's a zip file of pdf's of all the artworks for the dump load controller.

Each artwork has 2 pdf's; one composite view and one track view only.

Those which have 2 varieties (Mk1 & Mk2); the MK1 is as I built mine. The MK2 versions have leds & switches mounted on the board to reduce wiring. I haven't built any of the MK2 boards so caveat emptor. Obviously, where a board has both MK1 & MK2 versions, you only need to build one of them.

Zip Filesize = 390,916 bytes

Usual commercial rights reserved.

Amanda

[dinges]

Or simply put 2 red LEDs in series; the original one, which is in the front (I assume), and a 2nd one somewhere on the pcb, which isn't seen from the outside. Its only purpose is to drop 1.7V extra.

Even one or two 1N4148 diodes in series with the red LEDs would work.

BTW, Amanda, I like overdesigned And if it's 20+ year reliability you are looking for, eliminate every electrolytic from the design. Only tantalum. The more I learn about electrolytics, the more I begin to dislike them...

Off-topic: the ESR-meter I've built has really paid off. And tought me a thing or two about elcos.

[commanda]

Or simply put 2 red LEDs in series; the original one, which is in the front (I assume), and a 2nd one somewhere on the pcb, which isn't seen from the outside. Its only purpose is to drop 1.7V extra.

Even one or two 1N4148 diodes in series with the red LEDs would work.

NOT technically elegant. Will it work, yes.

I tend to be pedantic, a perfectionist, and have a severe dislike of things assymetric.

Not always in that order.

In the last 12 months I have overhauled 2 computer motherboards which were faulty. Replace every electrolytic capacitor. In the last 15 years I've lost count of how much stuff I've repaired by judicious replacement of suspect electro's. I only ever buy the best quality high temperature low ESR electros.

Tantalums. Watch the polarity. I've seen them fail 5 years after being installed backwards. And thousands of microfarads and hundreds of volts is real difficult to get in tantalum.

I can save you the cost of that ESR meter. Give the device in question the sink test. Get a boat, go out to the middle of Sydney Harbour (or nearest geographical equivalent) and toss the suspect device overboard. If it sinks to the bottom; it WAS OK. If it floats, it's no good :-)

Think about it. It's a joke.

Amanda

[ghurd]

A decent red LED at 5ma will be well under 1.7Vf, maybe 1.2V?

G-

[dinges]

Ghurd Ghurd Ghurd.

What am I ever going to do with you... <shaking head>.

Just measured with my LED tester (yes, I've actually built one, in its case; a 9V battery with various resistors to give various currents, from 1mA to 30mA:

1. mA : Vf=1.8   V
   
2. mA : Vf=1.85  V
   
5. mA : Vf=1.9  V
   
10. mA: Vf=1.95 V
    
15. mA: Vf=2.0  V
    
20. mA: Vf=2.05 V
    
30. mA: Vf=2.15 V
    
    

Now, if your defence is going to be that my test-LED wasn't a 'decent' one, I'll bang my head to the wall Please don't make me take a larger sample of LEDs...

[dinges]

You don't consider it a technically elegant solution? Because I did. Not kidding here. Actually, I thought it was more elegant than using a resistor, because with a diode or LED the extra voltage drop is nearly constant with varying current. Guess my '2 engineer'-quote applies here

"Think about it. It's a joke."

Still thinking about it. I don't get it. What's so funny?

Will give my dumpload controller the float test soon. I hope it sinks.

[commanda]

Peter,

With an LM3914 the led current is constant. The current is set by the resistors from reference out to ground.

Sink test for electro's. Do you know how deep Sydney Harbor is? Want to try & get your sink-tested electro back from the bottom? That's the joke. If it sinks to the bottom it was OK (past tense). Moral of the story; if it's suspect chuck it out & replace it anyway.

For any given circuit, it can be extremely difficult to define exactly where the borderline for good/bad will-work/won't-work is.

Amanda

[commanda]

Yes but.....

I think you'll find the SSR's need at least 3 volts to turn on reliably. The Opto22 modules we use at work do.

You'll also find that green & yellow leds have a higher Vf than red leds.

Amanda

[dinges]

Thanks for explaining the joke... :/

[dinges]

Forgot to mention I tested a red LED. BTW, white ones require even greater voltage (about 3.6V).

The opto22 modules require, from memory, 3V at 1mA (not sure about the current). They can be driven with voltage of up to 32V. When driven by higher voltages, a resistor should be used to limit current to the 1mA level.

[ghurd]

I'll call that a very old, or strange, or possibly damaged, red LED.

Most (I would have said 'ALL' yesterday) red LEDs will run with a low AA, and most will be overdriven with a fresh alkaline cell (1.65V) and without a resistor.

Sorry.  Larger sample. Just 2 more should do it.

I don't see how it would make any difference, as long as they are all the same and you have enough.  But those numbers look more like a yellow LED.

And if the white LEDs are taking 3.6V@20ma, get some new ones!

For lighting, I would not use a white LED with a 20ma Vf above 3.25V. Those few go straight to the landfill.

Are you using English or Metric volts?  

Maybe try another meter?  Really, some read higher V when the battery is low.  You knew that. If I got the above readings as an average, first thing I'd do is grab another meter...

G-

[dinges]

<lots of expletives deleted>

"But those numbers look more like a yellow LED."

Next thing you know I will be accused of being colour blind...

The voltages correspond perfectly to what I've learned about red, yellow and green LEDs. Must be a metric thing, no doubt. You are no doubt using Imperial LEDs.

Oh yeah, the LEDs are old. Can't remember having bought them, all mine are 'ex-equipment'.

Some more measurements:

1. 67V (5mm ; 10mA; red)
   
1. 95V (3mm ; 10mA; red)
   
1. 95V (3mme; 10mA; red)
   
   

Tested 2 white LEDs (same manuf. batch):

3. 10V (10mA)
   
3. 25V (20mA)
   
   

Now, I must admit. The meter is not a calibrated one... It's an old Philips analog one (PM2412) that I fixed. But it's accurate. Don't make me get out the Fluke 8000A. Or the AVO8...

So, I hereby challenge you to make the same measurements (let's just stick to red LEDs, for the moment). Seriously! If red LEDs exist that operate of 1.2V, I'd be utterly amazed. And like I said, all my LEDs have a few years of life behind them. I like to call them 'tested', as opposed to 'used'.

[dinges]

Oh yeah. That original red LED I tested; it must be from before '90. It had long leads (was still 'new', rather: unused). Had chosen it because it was easy to measure for me, with its long leads.

How I got the LED? From my highschool days, one of a few I was given by the physics teacher for some of my experiments. Probably in 1988 or 1989. Now, that was when I was given them. They probably had them for a while there too...

Hope this satisfies you. I'm going to swallow a happy pill. I've got a headache; you think banging my head to the wall could have anything to do with it?

[dinges]

Nope, one happy pill isn't enough. I still find myself looking for datasheets of LEDs.

A link to someone with more time than common sense, dedicated to the humble LED:

http://ledmuseum.home.att.net

http://en.wikipedia.org/wiki/LED

And now, for the killer:

http://www.oksolar.com/led/led_color_chart.htm

There. Now what do you have to say... I've read it on the internet, so it must be true!

[ghurd]

Dang!  I have used red LEDs with a single cell many times, but you win.

ETG-5TS630-30, 5mm red LED. First quality, 2 samples, same lot, 2004 date.

2. 3ma 1.81/1.82V (that is 2.3ma)
   
5. ma 1.88/1.89V
   
10. ma 1.98/1.98V
    
20. ma 2.09/2.07V (data sheet typ = 1.94Vf@20ma)
    
30. ma 2.23/2.20V (just for fun)
    
50. ma 2.36/2.31V
    
    

The 2.3ma is because of a flat spot in the power supply pot.

So the idea could have merit.

G-

[dinges]

Could it be because fresh cells are a bit above 1.5V, say, 1.6-1.7V? You won't be getting many mA's through them that way, but enough to light them.

I remember that, back in '88, they had a very nice IC (LM3909) whose only function was to make a LED flash. The major benefit of it was that, with the LM3909, you could make a signalling LED work off 1.5V.

It actually was the very first circuit I've ever built... (on vero-board, BTW) One IC, one cap, one LED. It blinked. Guess it's a sign of the times that my very first project used an IC...

The IC is obsolete now. Just found a website which shows the internals of a LM3909, rebuilt with discrete components:

http://home.cogeco.ca/~rpaisley4/LM3909.html

[ghurd]

LM3909 is/was again being made.  NTE876  US$11.36 each.

Conflicting information, production is delayed or discontinued, but they are still shipping the part.