Automated Load Tester

[ronv]

They are pretty pricey to have made. I did the layout in Express PCB. Their price is 2 (minimum qty.) for $87 including shipping. I'm sure they can be found for less, but we would need to have someone here that knows Eagle or some other layout package that supports gerber files redo the layout. The do it yourself method can be painful, but is least expensive. You would need raw pcb, etchant, and drills. Although if you don't have a drill press or dremel drill press it's not much fun either. Total cost this way maybe $20 for 1 plus $6 shipping for the drills. The factory made ones are really nice if you don't mind the cost.

 

[tkc100]

OK
Ron
I will make my own PCB
I really don't mind doing so I just thought it might be easier to have it made.
Now on to the next question
In an earlier post you brought up the idea of protection from the unit being hook up backwards.
At the time I wasn't concerned.
Earlier this week I during a moment of mental lapse attempted to connect a battery bank up incorrectly.
Don't know where my mind was but the battery lost a post and I'm a bit gun shy.
Did you give any further consideration to this?
Is it possible or would it require too much modification.

 

[ronv]

This one is 1/2 protected.
Diode D1 protects the circuitry from reversal, However the Fet switches will conduct thru their body diodes so you will get a 25 amp spark. Probably not enough to hurt anything but enough to let you know it is backwards. We could add a big diode on a heatsink to resolve this if you like. It would lower the load current to about 22 amps but we could probably find another combination of parallel resistors to get it back to the 25 amp range.

 

[ronv]

Sorry, there was a typo in the last post. It would lower the current to about 22 amps, but we could use 10 2.2 ohm (standard value) 25 watt resistors in parallel and get back to 25 amps with the diode. The current will be a little less accurate but not much.

 

[OlPhart]

Expanding voltage range trick

That ~.90 volt range can be expanded. A trick I use is a negative voltage regulator run backwards: ground is the most positive terminal, use it as input. The input is the most negative terminal, put it to ground. The output is between them, so is still the output.

What this does is subtract the regulators' rating in a precise way: your range of 6.25-5.35 becomes 1.25-.35 with a (very) low dropout 5 volt regulator. This can now be amplified to suit an ADC or whatever. Good Hunting... <<<)))

 

[tkc100]

OlPhart
I am afraid that one goes right over my head.
What do you think ron?

Now as to the matter of circuit protection. Even though I was dazzled by what a whole lot a amperage can do I'm not concerned with a 25 amp spark. Especially if it complicates the design.
If sometime in the future I should suffer another mental lapse I just want to be certain it would not damage this circuit.

 

[ronv]

That was my thought as well. Kind of like hooking up the jumper cables.
I bet a dead short on those big batteries is spectacular. Lucky it didn't explode, I guess they can do that when shorted.
We should probably talk about the box this will go in and how the heat sink for the resistors will work. I'm not sure if you have a look in mind. Maybe we can both rummage. Right now the board is about 3X5 so we can make it larger to fit the box. The big resistors should be outside the box. Maybe shielded so they can't short out to anything if it tips or falls.
I think olphart was offering a way to scale the difference in battery voltage so it could be amplified. I think we have about .004 volt resolution now so probably not a problem.

 

[ronv]

Take a look at this one. I think with some minor changes to the PCBA so the leds could come out the front panel. Front has the switches and leds the back has the 50 watt resistors. Might have to put a brick in it to hold it down.

https://www.mouser.com/catalog/specsheets/EPD-200290.pdf

 

[Mosaic]

Just noticed this thread.
I did some designs last year based around desulphating with pulsed power thru a bank of High pulse current FETs etc. basically pulse zapping the batts with around 36V of very LOW ESR caps (around 30,000uf)

Part of the design dealt with automated discharge and pulsed recharge to derive Peukert Indices and profile each battery (bank) to maintain a historical record of each bank.

It was to be MCu driven so that unattended operation would not only profile the battery condition but also determine whether the pulse charging made a worthwhile difference to reversing any sulphating.

Along the way i picked up some interesting links which may be useful here:

**broken link removed**
http://www.marine-electronics.net/techarticle/battery_faq/b_faq.htm
**broken link removed**

BTW, I have all that parts for my design based around a 16F886 but no time to build the unit yet. It uses a rewound multitapped microwave transformer to get the voltages/currents. Hopefully soon.

 

[tkc100]

Well I haven't drop off the edge or given up on this project.
Just had to triage my life this past week and this project wasn't a top priority.
I am happy with the design and have the funds to go ahead
ronv I will do my best to look over the parts list tomorrow. As far as the box goes I have in mind more of a hand held look to it. Perhaps two shallow boxes separated by some spacers. One for all the electronics and the other to house the load resistors. I will do some poking around and see what I can find.

Let's discuss the PCB a bit. I have made several PCBs before but using photosensitive materials.
I don't currently have access to all the equipment I used then. Not that it was complicated or anything but it does require some stuff I don't have at hand right now.
I have done a little bit of research and came up with these methods of using a laser printer for the PCB layout. It looks all too simple is it possible it could work?
**broken link removed**
**broken link removed**
One is pretty much the same as the other it is just that the second method claims to have been able to use simple magazine paper in place of the photo copy paper.
Anyone had any experience I might learn from?
Remember for now this is a one off work and I wouldn't have any use for left over materials.

 

[ronv]

Load Box

Just as well you were gone for a few days. During that time I found what Rick Perry would call a "Euoops" It wouldn't come on right. While I was at it I redid the PCBA to get the LEDs close to the edge and to get the high current portion on the other side of the PCBA to make it easier to wire. I have updated the parts list.
Attached is the new schematic and a view of the layout (a little smaller so it should fit better on standard size boards.)
Make sure to download Express PCB so the artwork will be the correct size. I'll have to send those files private message they don't want to download here.
I have used the toner transfer method, but always with the press and peel. But I think the magazine paper would actually work better. I go to my local post office to use their laser printer and putting strange paper in their copier makes them real nervous. I'll attach all the files. Maybe someone will give them a check. I will also give them another look tonight.

 

[tkc100]

ronv
This thread is getting pretty long and we may have exhausted others interest in this project but I thank you for your perseverance.
I though I would have time today to review all that has been done up to now and also look over the parts list. I didn't!
Tomorrows will try again.
I download the PCB software and will install it tomorrow also.
There are of course some additional part I will have to add PCB stuff, clips, wire, box, and I think I would prefer a three position slide switch to the rotary one but I will take care of all this.
Just as a side bar... I noticed on another site where an individual was using the same laser transfer method to label the instrument cabinet. Looked good in pictures may give it a try.
All this design stuff is really a bit over my head and the learning curve is straight up. I will be glad when all the parts are in front of me on the workbench.

 

[Reloadron]

I don't know?

Good threads on good topics & projects are supposed to get long. I still return to it and read the post, I won't overly add anything unless I feel I have a worthwhile suggestion but I still return to the thread.

Ron

 

[Mosaic]

I use toner transfer using quality magazine paper (eg. taken from astronomy mag or style magazine) . Good enough accuracy to do DFN chips and 603 smt devices on double sided boards with 10 mil traces. I etch with nitric acid, mixed to etch 1 oz CU boards in 5 mins or less.
Slow etching is bad for small detail work. Doing regular DIP work with 16 mil traces is a no brainer. I find using 0.8mm PCB is easy to cut with a decent scissors, no sawing.

 

[tkc100]

Well 54 some post later and it looks as if it is about time to do some business with Mouser.
I just finished reading all the posts and I would like to summarize the project to make certain everything is on task.
The proposed project is an automated load tester for 6 volt deep cycle batteries.
When the tester is first hooked up to the battery a green LED will come on to indicated the battery is fully charged OCV greater than or equal to 6.35 volts. This actual voltage may require some initial calibration to compensate for temperature but this will be a trial an error process after the tester is assembled.
Following the reserve capacity specifications provided by the manufacture and the prescribed testing procedure established by BCI the tester will load the battery with a 25 load and then measure the time necessary for the voltage to drop to one of three preset voltages:
50% SOC = 6.10 volts
80% SOC = 5.99 volts (only 20% SOC left in battery)
Complete discharge = 5.25 volts
During the entire test cycle a yellow LED will remain lit indicating the test is in progress.
When anyone of the three preset trigger voltages are hit the test is complete, the yellow LED will go out and the timer will show the elapse time.
ronv thanks so much for mothering this thing along this far and a thank you everyone else that has participated in all the brain work.
The final schematic and PCB are shown in post #51
This is a last chance to review, comment or change anything. Can anyone see a need to do so?

 

[ronv]

I think you are ok. The only thing is the "on" led in the parts list is red. Make sure to use rev. 1 artwork and I think you are good to go.
I'll make up a little test procedure to get the adjustments close and make sure all is well before 25 amps is used.

 

[tkc100]

Sounds good!
I was just going over the parts list.
There are of course some parts I will have to add. Stuff you probably just have laying around.
PC board, enchant, wire, grommets, Clips and of course the enclosure.
I thought that would be simple. I envisioned just a simple aluminum box. After looking at the timer it needs to be about 2" deep and then of course it has to house the PCB. I thought I would get another matching box (length and width and hopefully shallower) for the load resistors. Then sandwich the two together with some spacer in between. I thought I could ventilate the smaller box with some holes and the box itself would be a decent heat sink.
Straight forward enough but when I was at Mouser I felt like the county boy's first trip to a shopping mall. The selection is vast and I haven't master the art of searching just yet. I will keep at it.

 

[tkc100]

OK so much for the idea of two matching boxes. Easy enough to get two identical boxes but not so easy to find two with the same width and length with differing depths.

How much heat do I need to anticipate these four load resistor producing?
One test cycle could run as long as 6 hours.

What do you think about this idea.

An aluminum enclosure properly sized for the PCB and timer and then mounting an aluminum plate on the back the same length and width as the box. Mount the resistor on the plate and separate the plate from the box with some spacers. The plate would act as a heat sink and there would be a lot of ventilation around all four sides.

 

[ronv]

I'm glad you ask that question! The 50 watt resistors are really only 20 watt without a really big heat sink.
Take a look at these guys. They can do the 160 watts though they will be hot so will have to be shielded from touch, but still with air flow. Like the others it will take 4 1 ohm in parallel. They present another problem in that they are inductive, but let me check that out.

 

[tkc100]

ronv
I actually just stumbled upon this problem when I was working on how to mount the resistor and make it work with an enclosure. Since that time I discovered as you did that the 50 watt resistors require 129 sq inches of heat sink per resistor or as you stated in your post they must be derated.

I was doing a little figuring and poking around and was wondering what if we used one .25 ohm 100 watt resistor. It requires 294 sq inches of heat sink but derating it by 50% would only require 20% of that surface area or 60 sq inches. That's still a lot of flat surface for a had held tester. I looked at a number of heat sinks but the problem I ran up against is the size of the resistor ≈6.5" x 3" x 2" Mouser Part #: 71-RH100R2500FJ01 $68 If I'm reading the data sheet right that's one big resistor. I found a number of heat sinks that looked good and had adequate surface area but would not accommodate that size of the resistor. It might be possible to take two and with a little fabrication make one big enough.
Or does 60 square inches mean the retire box it's mounted on/in. In which case a simple cheap aluminum enclosure just big enough to house the thing would provide more than enough surface area. That would take us back to the idea of two boxes back to back with some spacer in between.

This is kind of a side note but how in the world can a resistor maintain + or - 1% with temperature fluctuation? As I was reading the data sheet it indicates the current carrying capacity goes down as the temperature rises but it seems to me in my limited understanding of electrical theory the resistance would go up.

The devil is in the details!!!!!