HobbyKing.com New Products Flash Sale
Reply
Thread Tools
Old Sep 06, 2009, 02:57 PM
Registered User
United States, AL
Joined Apr 2007
821 Posts
Mini-HowTo
How to build a constant-current discharger

My desire is that building this will allow more people to post quality voltage-time discharge graphs of batteries. This will allow more batteries to be tested in a standardized way to help determine the quality of various packs(esp. the ever-changing chinese packs).

The basic device we are going to build can be seen at this link: http://www.webx.dk/oz2cpu/radios/dc-load.htm We will make a few small changes to make it fit our use better. I've uploaded the schematic from that page here to save it for posterity. Our goal will be to discharge a 2200mAh 3s LiPo pack at 50a for under $50. I'm no electronics wiz, but I know what the basic circuit symbols are and understand v=ir and p=iv. If you learn those simple things you can understand how this thing works and make small changes to fit your needs.

First, a look at the schematic and our minor changes:
The device labelled "7809" is a 9v voltage regulator which provides a stable reference voltage to the circuit. It is critical and can't be changed. The two capacitors on it are per the spec sheet from any manufacturer of the voltage regulator and ensure smooth operation.
The big triangle is an op-amp. It measures the battery discharge current and controls the device to keep the current at a constant level. It is a tiny 8 pin chip with two op-amps on it, we only use one of them.
The "sense resistor" is a 0.001 ohm shunt that provides a small voltage drop that the op-amp reads as battery current.
The transistor(s) are mosfets that regulate the battery current. A mosfet is sort of like pinching a garden hose to control how much water can flow through it. The mosfet's resistance is controlled by the op-amp to keep the current flow constant. The 100 and 0.1 ohm resistors are if you use more than one mosfet - the 0.1 ohm resistance helps balance current flow through the mosfets and the 100 ohm resistors dampen changes in mosfet resistance to prevent oscillation due to manufacturing tolerances in the mosfets.
The potentiometer varies the reference voltage to the op-amp, ie. it sets how much current flow the op-amp will keep across the sense resistor.
The left switch is simple an on-off switch. The right switch switches between constant current and constant resistance mode. We will delete this switch and hard-wire the circuit for constant current.
The other small components are as called for on the mosfet/op-amp spec sheets and simply help the circuit perform ideally.

How many mosfets one needs depends on how much power you want to dissipate. A 3s LiPo at 50a is around 500w of power. You can look up the specs for different mosfets, but what it comes down to is how well you can cool the mosfets. 500w is a lot of heat and it's hard to get so much out of a tiny electrical part the size of a BB. Devices such as the cc400 can dissipate 400w of power but cost around $400 - a bargain as other devices cost much more.

So our plan of attack is this: Build the control circuit and use it to test single mosfets to see how much power we can reliably put into them. Our goal is 75w per mosfet(7a * 11v = 77w) and seven total mosfets to get 50a. For testing purposes we will replace the 180k ohm resistor with an 820k ohm resistor. As pictured the circuit will provide 50a of current adjustment. With the resistor change the adjustment range will be reduced to 11a and provide finer control for testing.

At mouser.com you can buy the parts for the control circuit and 3 mosfets for about $14(plus shipping). If you omit the switch it's only $12! Three mosfets because we will surely burn some up testing our cooling ideas. We got our parts today and will build the thing this weekend. The big hurdle will be figuring out a way to cool the mosfets. Commercial devices use fans and large(ie. expensive) heat sinks to dissipate the heat, we'll try out some no-frills cooling options which is how we will make this thing so inexpensive. Next we'll also review changes to make if you want to use this for a significantly different current/voltage range(ie. 6s packs, more/less current, etc).
biskit is offline Find More Posts by biskit
Reply With Quote
Sign up now
to remove ads between posts
Old Sep 06, 2009, 02:59 PM
Registered User
United States, AL
Joined Apr 2007
821 Posts
Pictured below is the "prototype". The mosfet is clamped to the heat sink(pliers) just to test the circuit. The tiny blue lipo is the power supply and the 3s A123 pack is being discharged at 1a. The yellow blob is the shunt. Everything fit easily on an old radio shack generic board I had laying around. Perhaps somebody smart can design a PCB to fit everything neatly and easily and sell them for a couple dollars?

The 820k resistor to change the current range from 50 to 11 amps gave almost exactly a 10a adjustment range. I left the power switch off for testing and didn't include the constant resistance ability either. It worked the first time too with no problems.

The mosfet is on about 6 inches of wire so that I can easily test different cooling methods to see how much power each one can dissipate. Without the "power" switch - with the potentiometer turned all the way down there is no measurable current and with a battery plugged in without power applied to the circuit there is also no measurable current flow. So I'll probably leave the switch out completely and it'll save $1.50 on the parts!

Next: test cooling devices to find out how many mosfets we need and the value of the ballast resistors.
biskit is offline Find More Posts by biskit
Last edited by biskit; Sep 06, 2009 at 03:13 PM.
Reply With Quote
Old Sep 06, 2009, 02:59 PM
Registered User
United States, AL
Joined Apr 2007
821 Posts
and reserved
biskit is offline Find More Posts by biskit
Reply With Quote
Old Sep 06, 2009, 11:34 PM
CamLight Systems
New York City, USA
Joined Oct 2003
1,172 Posts
Quote:
John,

I am using 4 x Thermaltake A1744 coolers with 4 x IRF2907 Fets per cooler.
I measured the cooler and came up with 0.151C/W at 350W.
Fet J-C = 0.45 C/W
Fet C-HS = 0.5C/W
Source resistors are 0.15ohm so @ 25A, 16V (400W) resistor power is 23.4W and power in Fets is about 90W as Schottky losses are about 15W.
Thermal calculation gives a temp rise of 141C, which allows an ambient of 34C before we get 175C.
Whilst I agree with your conservative approach to junction temperatures when designing commercially, now it is just a hobby, so if a Fet fails I just swear and have a cup of tea/coffee. In truth I have not yet had a failure and my experience of powerfets is that they will take a lot of power or current abuse. Overvolts is the killer. I also put auto fuses in series with each cooler to protect the pack but have not yet been needed.
Sorry I got the resistor values wrong in original post - age and memory!
Wayne
Wayne,

Ahh, your source resistors are a lot larger than mine...great for taking some power away from the FETs. I needed to support up to 100A discharges at 0.9V with only 4 FETs so my source resistors are low, 0.005ohm.

And with a 2C/W thermal resistance at 70W for that heat sink (Thermaltake's rating), I can easily see it dropping to 0.151C/W at the much higher temp levels you're running them at. Very nicely done, you have a great setup!

I completely agree with rating a hobby setup at 175C. I do that for my commercial units that I keep for inhouse use. I sell them rated at 400W (at 140C junction) but take out the power limiting firmware and run them at almost 500W (at 175C junction) for myself. I have to assume at least one of my clients is going to do a discharge with the heat sink fully dust-clogged and outdoors on an August Arizona afternoon, hence the conservative 400W rating. Personally, I've only popped one set of FETs over a several year period and that's when I hooked up a 24V pack but had the current (and total power level) set for a 12V pack. I was picking tiny balls of copper out my shirt for a qute a while there. And it sounded like a 22LR going off when the MOSFEts blew.
JohnMuchow is offline Find More Posts by JohnMuchow
Reply With Quote
Old Sep 06, 2009, 11:38 PM
CamLight Systems
New York City, USA
Joined Oct 2003
1,172 Posts
Quote:
Originally Posted by biskit
Pictured below is the "prototype".
<snip>
Next: test cooling devices to find out how many mosfets we need and the value of the ballast resistors.
Congrats on the first test!
If you can, have the wire from the board to the MOSFET's gate as short as possible, especially if the capacitor and resistor are on the circuit board end. It helps keep the MOSFET from oscillating and going boom.

You may be just fine as is. This is a just-in-case, best practices, kind of thing.
JohnMuchow is offline Find More Posts by JohnMuchow
Reply With Quote
Old Sep 07, 2009, 03:52 AM
Registered User
Rugby, UK
Joined Feb 2007
812 Posts
John,

I have looked at your website and see that you have to cope with single Nimh cells, hence the 5 milliohms. Even with the 0.15 ohm resistors there is a fair amount of unbalance but I just prayed and it seems OK!
Impressed with your range of equipment. Tried to send an e-mail but cannot extract your address as my computer will not connect via your auto system. Can you send me a PM with e-mail address pl;ease.

Wayne
Wayne Giles is online now Find More Posts by Wayne Giles
Reply With Quote
Old Sep 07, 2009, 04:22 AM
CamLight Systems
New York City, USA
Joined Oct 2003
1,172 Posts
Hi Wayne,
Sorry about that. The e-mail address is encrypted to prevent spam. PM on the way.
JohnMuchow is offline Find More Posts by JohnMuchow
Reply With Quote
Old Sep 07, 2009, 08:23 AM
Dieselized User
gkamysz's Avatar
Chicagoland
Joined Feb 2000
7,150 Posts
I've built that circuit and it works fine. I haven't wired up all the fets yet, but I have an 8Lb heat sink they are going to be mounted to. I'm planning on dumping upwards of 2kW so I'll probably still need a fan.

Greg
gkamysz is offline Find More Posts by gkamysz
Reply With Quote
Old Sep 07, 2009, 09:33 PM
Registered User
United States, AL
Joined Apr 2007
821 Posts
So what to do with all the heat? To discharge a typical 3s 2200mAh size LiPo pack we have to get rid of about 25kcal of heat, or about 105kJ(about the energy you get from eating two Nilla Wafers). This usually means large heat sinks and fans, but that would blow our budget out of the water.

Speaking of water... water can absorb a huge amount of heat. Even more if it's frozen and has to melt. I'll save you the math, but our 105kJ will turn 1/3 L of ice at 0c into water at 0c. 1/3 liter is less than a can of coke holds. Most people who have RC toys will have running water also, so why not just use a tiny amount of water from the tap?

With a single mosfet clamped to a tiny piece of aluminum and held under barely running tap water the mosfet leads and case were barely warm to the touch at 90w. The aluminum plate was cold. The 3s a123 pack was almost dead and with the 10a current limit that was the max I could get from it. I'll try a 4s pack, but it'll break 100w easily. Maybe 150w from a single mosfet?

So the plan: a water cooled heat sink! I'm picturing an aluminum block(maybe copper if I can find something suitable and cheap, copper is expensive!) about the size of a deck of cards with 4 or 5 mosfets on it and a few passageways drilled with water flowing through it. Maybe a cheap electric pump to recirculate ice water from a small container around 1L or so. Or just hook it up to the sink and use tap water.
biskit is offline Find More Posts by biskit
Reply With Quote
Old Sep 07, 2009, 10:40 PM
Registered User
United States, AL
Joined Apr 2007
821 Posts
Discharged 2Ah from a 3s lipo at 10a non-stop. Power started at 130w and was around 110w after 2Ah. Wires were a bit warmer but still nowhere near hot. Charging a 4s a123 now to see if it can sustain 150w! Don't even care about the ballast resistors, we can easily get 500w from just four mosfets! Maybe we can get by with three?

This little piece of aluminum is very thin and the clamp blocks water from the surface directly opposite the mosfet. A deck-of-cards size block of aluminum would have neither of those problems.

Questions for electronic geeks:
1. how complicated would it be to add a thermistor to cut power if it gets too hot?
2. how complicated would it be to add an automatic cutoff so you don't accidentally kill your batteries?
3. is there an ideal ballast resistance now that we don't have to rely on them for a significant amount of power dissipation?
biskit is offline Find More Posts by biskit
Reply With Quote
Old Sep 07, 2009, 11:09 PM
Registered User
United States, AL
Joined Apr 2007
821 Posts
This is getting boring! 4s a123 at 150w into a single mosfet. I noticed that if I touched the potentiometer while holding the plate the current went up a bit - seems like the case of the mosfet is electrically connected to the ground or something? Anyhow, I could get it up to 170w and that was no problem either. Had to up the water flow from the initial trickle though, but everything was still only warm to the touch. I turned the water off and the drops on the plate sizzled instantly - without moving water it would self-destruct within 5 seconds easily.

So we can meet our goal with just three mosfets. I'll probably use four since that'll make the cooling block symmetrical and leave lots of safety/growth room. Total parts cost with 4 mosfets and 4 ballast resistors is about $25. That leaves $25 for the cooling block!
biskit is offline Find More Posts by biskit
Reply With Quote
Old Sep 08, 2009, 01:45 AM
Uh Oh Jungo
ParkJeff's Avatar
Australia, WA, Perth
Joined Feb 2003
770 Posts
biskit,

I'm interested in building an electronic load too as the setup I am currently using (halogen lamps) does not present a constant current load.

Could you show us a few more details of your prototype such as the type of MOSFETs you're using and what you're using for the 1 milliohm shunt?

Thanks.

Jeff
ParkJeff is online now Find More Posts by ParkJeff
Reply With Quote
Old Sep 08, 2009, 03:59 AM
CamLight Systems
New York City, USA
Joined Oct 2003
1,172 Posts
Quote:
Originally Posted by biskit
Questions for electronic geeks:
1. how complicated would it be to add a thermistor to cut power if it gets too hot?
2. how complicated would it be to add an automatic cutoff so you don't accidentally kill your batteries?
3. is there an ideal ballast resistance now that we don't have to rely on them for a significant amount of power dissipation?
Great testing!

1. Not complicated electronically but there may be some physical implementation problems. The thermistor would have to be protected from the water flow, which would cool it and give you false readings. You could drill a small hole behind a MOSFET, thru the heat sink, and epoxy the thermistor in the hole, touching the rear of the MOSFET (the best place to measure temps).

It may not be worth it though as you'll need a thermistor, op-amp (comparator), some diodes, and some resistors and capacitors to create the circuit. The thermistor would vary a voltage to the comparator who's output would go high above a certain temperature. The output of the comparator goes through a diode (to isolate it) and on to the current sensing pin of the load's op-amp. When the comparator goes high, the load's op-amp will think that a huge amount of current is flowing through the sense resistor and it will reduce its output to almost zero, turning off the MOSFETs.

2. An auto-cutoff uses another comparator and a reference voltage which is a known fraction of the pack voltage at the cutoff point. Have the same fraction of the pack voltage go to the comparator along with the reference voltage and when the pack voltage drops too low, the comparator goes high. Do the same thing you did for temp cutoff (the diode to the load op-amp) and the MOSFETs will now turn off when the voltage drops too low. You'll need to add a latch to keep the MOSFETs from turning back on when the pack voltage rises back up though. There are some great latching voltage monitoring chips out there. I use the TL7700 a lot for up to 40V but you can divide the voltages down easily with a pair of resistors.

3. No ideal ballast value but the larger the value is, the more stable the paralleled MOSFETs will be and the more power they won't have to handle. Though, the water cooling seems to make the power handling an unnecessary worry.

Regarding the change in current levels when touching the MOSFET case and the pot. The case is connected to the MOSFET's source pin and is at the battery positive voltage (when used in your load). It can cause instant destruction of your load if you're touching it and touch the wrong part of the circuit. There's a shock hazard for you too at higher pack voltages. The plate will have to be insulated from human touch or you'll have to place an insulator between each MOSFET and the plate. This will significantly reduce the power handling of the MOSFET though due to the insulator making it much harder for heat to flow out of the MOSFET. Another option is to hardcoat anodize the plate, turning the entire surface into an insulator. Don't use standard anodizing!!!
JohnMuchow is offline Find More Posts by JohnMuchow
Reply With Quote
Old Sep 08, 2009, 08:10 AM
Registered User
United States, AL
Joined Apr 2007
821 Posts
Parkjeff: mosfet is "IPP048N06L G" from mouser. The shunt is just a 1 mohm resistor, but it has to be able to dissipate 2.5w at 50a. I used a "OARSXPR001FLF" from mouser because it was the cheapest I could find. Once I finalize everything I'll make a project list on mouser and you can just click "buy" and get all the parts you need.

John: that sounds like charlie brown's teacher to me! I found a little chip that monitors a thermistor and switches an output. Only two resistors to set the hi/low temperatures to switch at it seems like. Not worried about getting the temperature exact, just curious if it'd be simple to make it to where it wouldn't self-destruct if you used it without water running. I'll save the extra fancy stuff for after it's done I think, otherwise I'd never finish!

So who can make me a nice little aluminum block with a few holes in it?

I also noticed it'd be easy to make a simple 1-sided PCB without needing jumpers or anything. How much does it cost to have some made?
biskit is offline Find More Posts by biskit
Reply With Quote
Old Sep 08, 2009, 08:23 AM
Uh Oh Jungo
ParkJeff's Avatar
Australia, WA, Perth
Joined Feb 2003
770 Posts
biskit,

Thanks for the reply and part numbers. Looking forward to seeing your completed discharger. It's really an interesting and worthwhile project.

Cheers,
Jeff
ParkJeff is online now Find More Posts by ParkJeff
Reply With Quote
Reply


Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Discussion PIC PWM and FETs as constant-current discharger? meteor DIY Electronics 21 Jan 09, 2010 10:17 AM
Mini-HowTo How to build a constant current battery discharger biskit Batteries and Chargers 25 Sep 06, 2009 07:53 PM
how to build a vaccum forming box? rayc Scale Kit/Scratch Built 14 Feb 28, 2008 10:46 AM
Discussion DIY analog constant-current discharger? meteor DIY Electronics 7 Jan 06, 2006 10:09 AM
Anbody know how to build a prop? Pork Tornado Parkflyers 4 Aug 03, 2001 04:10 PM