How to build your own automatic and programmable 6S discharger

[redbird300]

We already covered some lipo charging gear in the past, but never really talked about discharging. Let’s see what we could do about that. Making one myself sounded like a nice challenge, you'll understand why further on.

Warning: you’ll read that I encountered a heat related problem using the firstly proposed hardware, but I deliberately choose to keep that info in this thread, as it might be useful for others to avoid the same problem. I also mentioned how to solve that problem. Consider this as well as a learning thread instead of only a building thread.

This will be a post consisting of six parts, at least in first instance, covering the first working prototype:

1. The general idea
2. The circuit
3. The components aka shopping list
4. The build
5. The testing and first use
6. What’s next ?

[redbird300]

1. The general idea:

OK, so some of you might need to discharge lipos from time to time. You charged a bunch of lipos, and rain starts falling down during the whole day for example. Or you can’t go to the field anyway, as the wife suddenly decided that you need to mow the lawn. Bad luck, but we need to deal with these conditions.

Best solution is to avoid the need to discharge at all times, by only fast-charging what you really need at that moment and nothing more, but this scenario is not likely for everyone, or in all circumstances. So let’s see what we could do .

The most common practice must be to power up your lipo charger, put it in discharge mode and push a lot of buttons to pick the discharge rate and end voltage, hook up the pack, and wait until the discharging is finished, which takes a lot of time with most chargers. The discharging capabilities are not very powerful, even some high end chargers allow only for 60-80 Watts discharge capacity. Many smaller chargers only allow for 20 Watts, which takes ages to discharge a large pack: a fully charged 6S/5000mAh pack would need around 3 hours to lose half of it’s charge for example ! I am not talking about regenerative discharge into a car battery here, as this involves even more work to hook it all up. It allows for high currents, but I looked for another solution.

Some people use automotive headlight bulbs to discharge lipos, but even when combined with a timer this involves some risks. There is no safe and 100% sure way to discharge to an exactly predetermined and controlled voltage, and the individual cells are not actively monitored.

Wouldn’t it be nice to have a simple device, not needing to be connected to a power outlet, where you hook up the pack and which discharges fast, automatic, safe and easily ? Well, that’s the general idea. Problem is that I didn’t find these in any shop, so I decided to try to make one myself. Note that the idea is not entirely mine, I simply read about a similar project somewhere, and just made my own version.

What did we want exactly ?

- Something fairly easy to build yourself, given some basic electric knowledge, without complicated self made electronic circuits. This involves a little tric of course, which consists in using a specific lipo monitor with an external alarm port (the Hobbyking “Cell-Log 8”), as you’ll see further down this thread.
- Must be able to reach a decent discharge current, I started with the idea of a 80 Watt prototype, but it will become clear that making a more powerful one is very easy with this design. In fact, just changing the value of one component will be enough.
- Not excessively expensive. Yeah, we need that money for that new heli, lol.
- Independent from any external power supply or electrical outlet. We already have cables enough as it is, and this will allow for field use.
- No active cooling if possible, to avoid noise and extra power problems, a drawback is that we’ll need a very large heatsink for decent passive cooling !
- Programmable settings like end voltages, as well for the whole pack as each individual cell, a safety timer, etc…
- We didn’t pay attention for this first demo version to the casing yet, it’s meant for indoor use only, but anyone wanting to use it outdoors could make a nice case around it.
- Neither did we try to make the device usable for any packs other than 6S lipos, the most popular kind nowadays, right after the 3S packs. As the 3S packs have less capacity, the problem seemed less urgent to me. You could still discharge these with any normal charger in a reasonable amount of time, or even build a second discharger suitable for 3S packs, by using a 12Volt relay, etc…

Impossible ? No, not really, but there are some problems to solve, read on in part 2..

[redbird300]

2. The circuit:

Let’s have a look at the electrical diagram of the first prototype:




How does it work ?

In short:

The lipo pack is simply connected to the circuit by the deans connector and the balance plug, and all the rest starts automatically, the circuit will stop discharging and switch off the pack once it is discharged to the programmed alarm level in the Cell-Log monitor.

In detail:

The connected Cell-Log monitors all voltages during the discharge, we program it for a “normally closed” (NC) setting on the alarm output port, which means this output will be closed as long the Cell-Log judges that the actual voltages are above the alarm level, like on a fully charged pack. This makes the (normally open) relays contacts close, and by that the pack will be connected in series with the high-power discharging resistors. So now the pack is discharging, at a rate determined by the resistors value. Once the voltages have dropped enough, the Cell-Log will activate the alarm port, which will open, and the relay contacts will subsequently open as well, so no more discharging, the pack is not connected to the resistors anymore. Note that the pack voltage will rise a bit again without the discharge resistors attached, and the circuit will restart discharging again for a short time. This will happen a lot of times before the pack reaches a low enough voltage to make this stop. It’s not really a big problem, as you could stop the discharging yourself at that moment, but it is hard on the relay contacts. That’s the price to pay for such a simple setup. The blinking of the connected light and the sound of the relay contacts opening and closing gives an extra indication that the discharging process is (almost) finished. BTW, we’re already thinking about a different version where this won’t happen anymore, but you’ll need to give me a few more weeks before we’re there. There are several possible solutions, but I want the most simple and reliable one.

Note: if the relay coil would fail and refuse to close, there’s no harm done with this kind of setup, the lipo will simply not start discharging.

The manual of the Cell-Log clearly indicates how the alarm port works. It is a transistor driven port, capable of switching up to 50V/500mA, which is largely enough for a relay coil. FYI, the coil of my very ordinary relay needs 21 mA at 24 Volt. I attached a copy of the Cell-Log manual, it explains all about the alarm features and settings. On page 9 you’ll find this typical application, and that’s also how we’re going to use it:


The diode connected over the relay coil suppresses the back EMF, don’t worry too much about it. Note though that it has to be mounted the “wrong” way: the negative pole is connected to the positive pole of the pack !

The small 24V light bulb is just an extra optical indication that the pack is discharging, but it’s not really necessary, the Cell-Log’s screen will also tell you that.

The relay and the light bulb will take (small amounts of) power from your pack while discharging, and will work on slightly lowering voltages during the discharge, but that causes no problems.

The high power resistor(s) now: I picked a value of 7,8 ohm in this case, because that will give me a (for a 6S pack): 25V/7.8ohm = 3.2A current, which will make for 3.2A x 25V = 80 Watts of dissipation power. I found two resistors of 3.9 ohm in series to be convenient, as these are cheap in a 50 Watt version, and that way they will each only need to cope with 80W/2 = 40 W. If you find a single 7.8 ohm/100Watt power resistor, that’s fine also, and even easier. I also found some high power resistors up to 250 Watts each, but I considered these too expensive.

[redbird300]

3. The components, aka shopping list:

- A Cell-Log 8M, you don’t need the advanced 28$ version with USB logging, the 14$ version also has the alarm output connector and the wire, and will do fine: http://www.hobbycity.com/hobbycity/store/uh_viewItem.asp?idProduct=10952

- A high power resistor, look at the calculations above. I took two 3.9 ohm / 50 Watt pieces in series, to obtain 7.8 ohms / 100 Watts. These have a fixed heatsink around the resistor, but these are not enough to work at the max specs ! They mainly serve to guide the heat to a mandatory external heatsink. Price around 5$ each.

- A male deans connector, or any other connector that fits your packs.

- A single pole relay, with a 24 Volt coil, and contacts capable of switching over 3 amps, I took a 30VDC/16Amp version to be on the (very) safe side. 6$ in my local shop. This one uses 21 mA when it’s working on 24 Volt, not that this matters much.

- A general purpose diode: a simple 1N4148 type (75V/0.2A) might do, but I took a 1N4007 (1000V/1A) out of my spare parts box. You won’t have change small enough in your pocket to pay for one of these.





- Some large alu heatsink(s) and thermal compound. These are 100 x 75 x 65 mm each, with a thermal resistance of 1.6 Kelvin/Watt, together that makes 0.8 K/W. This isn’t really enough in fact, theoretically you need twice that amount of heatsink, see below for the calculations. I expect temperatures to rise rather high. Anyway, one large heatsink is also OK. Think big, it is like the heat that a 80 Watt light bulb produces that you need to dissipate. Real heatsinks like these are rather expensive, 25$ a piece at my local shop, but you might find some in a broken amplifier or large PSU, or even use some thick scrap metal. Painting it black with heat resistant paint also helps. Metal fins pointing upwards allow for a smaller heatsink.





Note 1:
You could use several automotive 24V light bulbs in parallel instead of the resistors, like three 25 Watt types, eliminating the need for these heat sinks, but I don’t like all the intense light they produce, and they will get hot also.

Note 2:
Heatsinks are (roughly) calculated as follows, neglecting thermal resistance between the resistor and heatsink, and neglecting the fact that heatsinks are usually rated for vertical mounting:
Pick the highest ambient temperature at which you would ever use the device, let’s say 30 degrees Celsius, that’s a hot summer day for us down here. Now pick the highest temperature that the heatsink should ever reach when in use, let’s say 60 degrees Celsius for the sake of the example. We also know that we need to dissipate 80 Watts. Thus, the thermal resistance should be (60-30)/80 = 0.375 °C(elsius)/W or K(elvin)/W. Any heatsink spec will list this thermal resistance number, lower values are OK, not higher ! Google “heatsinks”, and you’ll see what I mean. That will also give you an idea of the size you’ll need.

My combined heatsinks only have a thermal resistance = 0.8 K/W, so I expect temperatures to rise up to around 100°C, if the ambient temperature would be 30°C and with 25 Volt applied with the given resistor value. We’ll see what happens later in this post.


Rather optional, depending on how you assemble the project:

- A support plate, I used some 5 mm plexi.

- A piece of circuit board.

- A holder to plug in the relay, but you might solder straight to the relay pins if you wish. If it ever fails, it will be harder to replace though, involving some soldering.

- Some small 24V light bulb, or a LED with a suitable resistor in series (anywhere in between 2000 to 3000 ohm will do fine, at least 0.5 Watt, but take a little larger like 1-3 Watts, to avoid heat here).

- Some wiring, we’re only talking about 3 amps now, it doesn’t need to be so thick.





Not in the pics: some small hardware of course, like bolts and nuts, you’ll see those during the build stage.

[copterboy]

I'm reading with great interest Raf. That happened to me on Thusday, all packs charged ready to go. Get a phone call about work, and I'm no longer able to go, and the rain is now pouring down for the next few days, and I'm working most of next week.

A simple/quick solution to having packs lying around fully charged would be great

[redbird300]

4. The build:

I did cut a piece of plexi, rounded the edges, and attached the two heatsinks with M3 bolts, nuts and distance holders, to allow for a nice airflow all around the heatsinks. Mark the holes to be drilled in the heatsink for the resistor(s) now.




Apply a thin and equal layer of thermal compound between the bottom of the resistor and the heatsink.
This is important, don’t forget ! Like here on the right one:



Fix the resistors on the heatsinks firmly with bolts, nuts and spring washers (threadlock doesn’t like heat).




Wait a minute now, stop the build, and let’s test the supposedly kinda limited heatsinks right away: let’s connect the resistors to an adjustable PSU, and measure the temperatures with an infrared thermometer while allowing more and more current to pass through them, up to the expected max current of about 3 amp. We’re starting with a humble current of 1 amp:




Here are all the testresults, with an ambient temp = 23°C. I applied different currents during several minutes, up to the moment that the temp of the heatsinks didn’t rise much anymore:

8V/1 amp = 8 Watt: no noticeable rise in temperature, 3 or 4 degrees at most.
12V/1.5 amp = 18 Watt: heatsinks around 38°C, resistors around 42°C
16V/2 amp = 32 Watt: heatsinks around 48°C, resistors around 57°C
20V/2.5 amp = 50 Watt: heatsinks around 64°C, resistors around 78°C
24V/3 amp = 72 Watt: heatsinks around 79°C, resistors around 100°C (!!!)
BTW, as a reference for the Fahrenheit lovers: pure water boils at 100°Celsius (at normal pressure).

Conclusions: The relationship between power and temperature is linear, and the prototype does get very hot. And it gets about as hot as calculated above. Note that it will only get that hot if it would have been discharging a long time without a break, which is unlikely in real conditions. The whole test took way over an hour.

There are several ways to deal with this heat problem, just to name a few:

1) Change nothing, and take care not to get burned. Bad idea for a final version, IMHO. The circuit wouldn’t suffer from it though, as these resistors support way over 200°C in fact.
2) Use a protective, heat resistant cover over the heatsinks, but allowing for a very good airflow.
3) Use active cooling, like two silent 12V computerfans in series, or one 24V fan, and connect these to the 24 Volt circuit like the resistors are connected, avoiding the need for a separate power supply. This is definitely the way to go to build a very powerful discharger. If you use two 12V fans, take care that the power consumption and thus the internal resistance is equal, it’s best to take two identical ones.
4) Use larger heatsinks, about the double in size to get temperature down to a reasonable level.
5) Use resistors with a higher value, which will make a less powerful but cooler discharger. For example, two resistors of 6.8 ohm in series will make for 13.6 ohm. That means 1.77 amps at 24 Volt, or 42 Watt of dissipation power.
6) Use cutting wire as a discharging resistor, to get rid of the resistors and heatsinks. You need to find a heat resistant isolating object to act as a support though. Independent of the type and the resistance/meter of the wire, you want to aim for 8 ohm total resistance, that will determine the length of the wire you need. And you’ll need to check that it doesn’t get too hot at the calculated length.
7) A combination of the above.
.
Because this is only a first and temporary prototype I choose to do nothing for now, which allowed me to finish the device quickly without any changes. It’s a test, making a cooler or even larger one will be easy once we know exactly what we’re dealing with, and this experience helps. Note: I intend to use larger heatsinks in a second and more final version.

Now after this delay, let’s continue the build: put the relay, diode and light bulb on the circuit board, fix the Cell-Log with some double sided adhesive tape or so, and solder all the connections, wiring and connector per the diagram above. Mount the circuit board on the support plate. You’re basically done, unless you want a nice case around it. Do allow for enough airflow though in that case.

[ragge1]

thats nice
a suggestion , why not move to Norway , Bergen so we can be neighbors Raf , can help you on your projects

[redbird300]

5. The testing and first use:

Double check your circuit for bad wiring first, and set the desired alarm values in the Cell-Log by connecting a lipo to it, but with the alarm cable of the Cell-Log and the main leads of the lipo pack NOT connected yet. Now it’s ready for use, hook up your pack and the device will start discharging. Now there is nothing left to do but to wait until the pack is disconnected by the circuit, and the alarm sounds.






This worked fine with the prototype, and ended with the above mentioned blinking of the light and switching on/off of the relay, like once every second. This will slow down after a while, until the moment the pack or one of the cells doesn’t reach the set voltage anymore. Or until the timer in the Cell-Log reaches the programmed setting. I already started experimenting to achieve this, but made a wrong estimation in some electronic calculations and blew up the alarm port of the Cell-Log. The rest of the device still works, only the alarm port is cooked. Well, stuff like that happens when experimenting, I need to get another Cell-Log before proceeding, and that will take some time. I opened it up, but these surface mounted circuit board don’t allow for much DIY repairs.




Anyway, I hope that with all this info anyone could start his own project now if he would wish to do so. It would be nice if any questions about it are posted here, instead of PM’ing me, they might be useful for others.

[redbird300]

6. What’s next ?

A second version will be coming up, with some improvements:

- Much more cooling capacity, allowing for larger discharge currents without excessive heat.

- Solve the problem of the repeated switching on/off at the end of the discharge. We’re already working on that, and my thanks to Erwin for some nice ideas about that.

- Build the device in a more final form, were the circuit board will be protected inside a housing, etc…

All that could take some time though, but the first prototype did work and did the job well, and that’s a good start.

As that other guy said “I’ll be back…”

[redbird300]

Here's the diagram of the second (and final) version, with two changes:

1) A lower resistor value with higher power rating. This would make for around 170-180 Watts of discharging power, still working on appropriate cooling for that, but it should not be a big problem.

2) We use a relay with a double row of contacts now, instead of one with a single contact. This will make for no repeated on/off switching of the device anymore at the end of the discharge cycle. When we hook up a full pack, nothing will happen now, as the contact "B" is still open. Push the "start discharge" button for a brief moment, and the relay coil will get the necessary current to close the contacts A and B. When releasing the push button, the relay will stay active, as the contact B is closed now, and it will stay that way. At the end of the discharge, the alarm port will cut off the relay coil, and even when the alarm port activates again, it won't be able to reactivate the relay, as contact B is open again. Done.

[jperkosk]

Excellent, Raf, as usual
So that's what you're up when you go MIA!
A suggestion for another option if I may (there are always multiple options), from my industrial applications: If you change from relay switching to solid state switching then you don't care about blinking, it could blink away forever gradually slowing down and your version #1 would be just fine. You can either use a commercial SSR (solid state relay) or rig one from 1 transistor and 1 resistor, for 3A current a transisor working in a saturated state wouldn't need to be big . Of course your version #2 solves all that!

[Greatsteaks]

Doesn't a good charger have discharge built in? My icharger 208b does. I'm not by any means bashing your build, I'm just wondering if I'm doing something wrong.

[redbird300]

Originally Posted by Greatsteaks

Doesn't a good charger have discharge built in? My icharger 208b does. I'm not by any means bashing your build, I'm just wondering if I'm doing something wrong.

A quote from my first chapter:

The most common practice must be to power up your lipo charger, put it in discharge mode and push a lot of buttons to pick the discharge rate and end voltage, hook up the pack, and wait until the discharging is finished, which takes a lot of time with most chargers. The discharging capabilities are not very powerful, even some high end chargers allow only for 60-80 Watts discharge capacity. Many smaller chargers only allow for 20 Watts, which takes ages to discharge a large pack: a fully charged 6S/5000mAh pack would need around 3 hours to lose half of it’s charge for example ! I am not talking about regenerative discharge into a car battery here, as this involves even more work to hook it all up. It allows for high currents, but I looked for another solution.

And furthermore, it's fun to build it yourself, lol.

I have two 3010b iChargers myself, discharging is slow, limited to 80 Watts at most, and I need to hook it up to a power supply before I can even think about discharging.

[redbird300]

Originally Posted by jperkosk

Excellent, Raf, as usual
So that's what you're up when you go MIA!
A suggestion for another option if I may (there are always multiple options), from my industrial applications: If you change from relay switching to solid state switching then you don't care about blinking, it could blink away forever gradually slowing down and your version #1 would be just fine. You can either use a commercial SSR (solid state relay) or rig one from 1 transistor and 1 resistor, for 3A current a transisor working in a saturated state wouldn't need to be big . Of course your version #2 solves all that!

+1 Jerry, I even prefer that solution myself, but I wanted something easily reproducable for now, without any electronics in it. Probably I'll use a SSR in the end anyway, in the device that I'll use myself for real.

[Greatsteaks]

Got it.

[chansen1953]

hey Raf, exacttly was is the problem with leaving Lipo's fully charged ??
and how long is too long ?? . Is one week ok...
I charge all by batts up and usually use them over the next 2-3 days, but sometimes it may be a whole week before I get to use them.. depending on the weather etc.
It takes me many hours to charge them all ( 10 x 6s packs) , so I cannot really do this on the day I want to go flying ??

[Tjomp]

What can I say but ?

[redbird300]

Originally Posted by chansen1953

hey Raf, exacttly was is the problem with leaving Lipo's fully charged ??
and how long is too long ?? . Is one week ok...
I charge all by batts up and usually use them over the next 2-3 days, but sometimes it may be a whole week before I get to use them.. depending on the weather etc.
It takes me many hours to charge them all ( 10 x 6s packs) , so I cannot really do this on the day I want to go flying ??

Chris, I don't wanna burn myself in such a discussion, as I am not a qualified chemical specialist, but I'm just trying to follow some well known manufacturers guidelines about not leaving lipos fully charged, every hour they are charged is supposed to deteriorate the packs slightly, I'm just thinking "better be safe than sorry".

[dauste]

Originally Posted by chansen1953

hey Raf, exacttly was is the problem with leaving Lipo's fully charged ??
and how long is too long ?? . Is one week ok...
I charge all by batts up and usually use them over the next 2-3 days, but sometimes it may be a whole week before I get to use them.. depending on the weather etc.
It takes me many hours to charge them all ( 10 x 6s packs) , so I cannot really do this on the day I want to go flying ??

FYI: https://www.helifreak.com/showthread.php?t=250524

[sutty]

Interesting project Raf, particularly with regard to the amount of energy that needs to be dissipated. I realise that as soon as you do the maths it is pretty easy to understand, but without giving it too much thought I hadn't realised. I guess you only have to consider how it can bat those blades around for up to 10 minutes, depending on settings, and it is pretty obvious really.

Anyway, now that I have seen yours, and the calculations, I can understand why there isn't a simple, cheap, small, plug in device to do just this. This having been said, if there were a cheap, low current device, with overdischarge protection, I would probably buy a few, dependant on price, as I would be happy to just plug them in and wait.

Are there any devices that can emit energy, in large amounts, that is not heat? I haven't looked at this, but I remember hearing that modern LEDs are much more efficient than say light bulbs, and waste much less energy as heat. Might there not be some high power ones, of which you could get several, and light your room to about 1 billion candle power, so you were blinded as wou went in the room, lol, but without it getting very hot? Haven't done the maths here either, but I was just thinking aloud.

Thanks for that post Dave. Interesting read about the possible cumulative effect of leaving them charged up. I had been trying to keep mine charged up for no more than about half a week, thinking that this would be fine, but when I read that it is likely to be cumulative, I really should only be charging them if I am certain to fly. At the moment I am effectively charging them every week, whether I get to fly or not, and then discharging them, 3 or 4 days later, if I don't get chance to fly. So in about 10 weekends of non flight, it could be possible for me to clock up about a month of sitting there fully charged, which I would have always previously considered to be really bad form.

Much better it would seem to be able to charge them really quickly, like with Raf's high power portable charging station, virtually as you make the deicsion to fly. In Raf's case, when he is actually at the club. The problem with that is that even with all the home made cost savings, that is still a very expensive approach.

In your situation Chris, what can you do? It is fairly well described by the manufacturers and in threads like the one Dave posted, that it is detrimental for your lipos to be left fully charged. This is why they are delivered at a storage value of 3.8 - 3.85V per cell. At this voltage, particularly in a cool place, they are good for extended periods of none use, but you have no choice when it takes you so long to charge them up. I'd be in a similar sitaution to you if I were regularly having to deal with batteries for my 600, which I suppose I'd like to be in the not too distant future, but for the 450 I can get at least I can get all the 450 batteries charged, from storage, in about an hour and a half. If I charged at 2C, then I guess I could get it down to just under an hour, which is manageable.


I wonder if it would be better for me to regularly charge at 2C, instead of 1C, so I could charge only if I am almost certain to fly, 1Hr before I go out. It might be better than being gentle on them at 1C, preparing them the night before, and frequently ending up needing to discharge them, 3 or 4 days later, because the weather never came my way? Quite an interesting dilema I think.


Always charge last minute at 2 or even 3C, when the weather is already good.

Always have batteries ready from Fri to Mon inclusive, then discharge if not used.


Which is worse for my batteries I wonder?



Cheers


Sut