Schottky Battery Paralell setup... How?

RLT

10 kW
Joined
Oct 2, 2007
Messages
683
Location
Ruidoso, NM; USA
Well, I thought I knew how to do it, but apparently not. :oops:

How would I wire these big Schottky diodes to make a circuit to parallel two battery packs?? :?:


View attachment sch.jpg
Thanks!!!

Datasheet is here:
 

Attachments

  • 249nq135.pdf
    119.4 KB · Views: 171
first you need 2 of them. each of these is a single diode.

next you need a plate of copper or aluminium to bolt them to. this plate will work both as a heatsink and as a conductor for the Cathode end of the diodes. i would recommend a chunk of copper because it conducts better. drill and tap 4 X 6-32 holes in it so you can mount 2 of these diode blocks on it. also dril and tap a 1/4-20 hole that can be used for the output lead.

now when you screw the 2 diodes down to the coper plate you have created a common cathode dual diode which is what we need. connect one of the battery + wires to each of the diodes using 1/4-20 hardware. then lead off a single B+ output wire from the 1/4-20 hole you drilled in the coper plate.

the copper plate is live and must be isolated and or insulated from the surface to which you are mounting it.

rick
 
HMMM...

That's what I did.

With only one battery pack attached, I only got an about a 7 volt drop in voltage (40.X to 33.x volts)
PBA.gif
Tried it with 4 different diodes with the same results, so I figured I was doing something wrong.
 
I think this is what you need to do:

PBA-2.gif

There will be some voltage drop ... but it shouldn't be that big.

You are using the diodes to isolate a pair of batteries in parallel, right?
 
RLT said:
With only one battery pack attached, I only got an about a 7 volt drop in voltage (40.X to 33.x volts)

WTF? That's huge drop in voltage. Normal diodes only drop 0.7V, and these are schottkys; there should only be around 0.2-0.3V drop. Mebbe more at high current levels, but those are rated for 150V 240A. I wouldn't worry about them getting hot :D.
 
I just measured a V drop from 48.00V straight from a power supply to 47.98V through one of these diodes (no load).
ie. a drop of 0.02V
 
OK, Here's what has me confused...

That first diagram I drew was just for initial testing... I wanted to make sure that it was set up properly before adding the second battery pack, as in kbarretts's diagram (thanks, K).

But shouldn't the volts where I am getting 33 volts be 0 (or at least near 0) volts?..

I know the forward voltage drop is supposed to be around 0.7V or less on schottkys, but shouldn't the reverse voltage drop be a heck of a lot more than 7.0V???

Or is it that there is just enough current leaking through that the meter picks up a fair amount of voltage, but at such a low current that it would be insignificant when both batteries are in circuit?

Yeah, That's probably it.
D'OH!
HomerDoh.jpg
 
Sometimes you get a tiny current passing through the diode, which will show some volts, even though it will block any substantial load. Try measuring resistance instead of volts. It should show complete discontinuity in one direction and a few hundred ohms in the other.
 
obviously my english isn't so good as to be so misunderstood.

but thankfully both K Barret and Fechter understood what i meant and provided a nice sketch. thanks guys.

rick
 
PJD said:
Beginner question, but why do you have to isolate the parallel connected packs with these diodes?

Usually it's the BMS ... if one cuts out, the other gets fried. This can happen on both parallel and series rigs. You can also use these diodes to isolate packs in series ... but the wiring is going to be slightly different.

There is also an issue with some nickel chemistry batteries self destructing if charged while connected in parallel.

I't basically just a safety thing. A cheaper solution is to just swap packs when one gets low, either manually, or with a simple switch.
 
it also keeps the parallel strings isolated from each other so that a pack with a slightly higher voltage will not discharge and attempt to charge a weaker pack.

it is most noticable on SLA and NiMh packs where each parallel string has it's own self discharge speed. the pack the string that discharges the fastest will drag down the other string.

same effect occurs with parallel strings where one sags more than the other under load.

or any other situation where one string has a higher voltrage than the other.

rick
 
It also keeps the parallel strings isolated from each other so that a pack with a slightly higher voltage will not discharge and attempt to charge a weaker pack.

it is most noticable on SLA and NiMh packs where each parallel string has it's own self discharge speed. the pack the string that discharges the fastest will drag down the other string.

Same effect occurs with parallel strings where one sags more than the other under load.

or any other situation where one string has a higher voltrage than the other.

I thought that one string, or battery balancing the other was the whole purpose of putting the sub-packs in parallel. Admittedly, in the case of SLA's it is usually better to use "buddy pairs" than parallel strings.

And, how do you charge the pack with the diode in place?

Also, how does one BMS cutting out (opening the circuit) damage the other BMS?

Thanks.
 
PJD said:
Also, how does one BMS cutting out (opening the circuit) damage the other BMS?

Thanks.

Current spike on the remaining BMS.

This is the reason Ping will void your warranty if you parallel or series his packs ... one pack's BMS trips, and the other one gets spiked and fries before it can shut itself down.
 
Obviously you can parallel cells up, and everybody does it. But you then have to keep them that way and charge and discharge them together.

I'm a little less happy with connecting strings of cells in parallel, because of the possibility of an imbalance on one string, or of a cell failing or going negative at deep discharge.

If you connect two or more packs in parallel with isolating schottkies, you avoid these problems. The packs can be treated independently, they can be charged separately and used at different states of charge. You can even mix packs of different chemistries and cell sizes. The advantage over running one pack down and then manually switching is that when the packs are paralleled up, they will share the load. This reduces the peak current seen by an individual pack - which can be important for battery life.

I have a 4 into 1 Schottky combiner on my bike, so I can plug in a set of battery packs. The initial reason for doing it was to cut the peak current per pack. I had a Li-Ion main pack with some small NiMH packs to act as boosters. It made a huge difference when hill climbing as the NiMH would add the extra current for the short term when the volts on the main battery sagged.

Nick
 
Ok. I see. I am one of the relatively rare, multi-K watt electric scooter riders here, and the battery pack is a 120 pound permanently installed (until they go bad) feature of the bike. So the issue of paralleling removable battery packs just doesn't come up.
 
Does anybody have a good source for buying schottkey diodes? I was just going to get a couple of the 40A isolators from ebike.ca but he doesn't have any in stock. I figured I'd build my own, but have no idea as to where to buy some.

Any suggestions?
 
So, I'm finding more of the three leg style Schottky diodes online. I am assuming that these can also be used? Is this how one would wire one?

schottkeydiodepic.jpg


schottky.png
 
That is correct.

What you have there is a common-cathode dual diode T0-220 package. A couple of things to watch out for are:

Other dual diode packages may be connected as common-anode or as series.
If there is a metal tab at the end of the package or a bit of metal showing on the back, it is probably connected to the middle pin, so it will be live.
Think about the current and the heat dissipation. Suppose you are passing 30A and the voltage drop is 0.66V, that is 20W in the package so its a good idea to bolt it to a heat sink of some sort.

Nick
 
That means the heat sink needs to be isolated from the ground (frame) right??????????
otherDoc
 
You will need to either isolate the heatsink (electrically but not thermally) or to isolate the diode from the heatsink.

The latter is easy enough. There are standard ways of doing it, with silicone pads that go between the diode and the metal, and plastic bushes that isolate the screw from the diode. Sorry don't have a picture or a link to show you, but they should be available from the same suppliers as the diode.

If the diode package is all plastic, there isn't a problem. But if it has a metal part, then it is likely that the metal part is live.

Nick
 
Right. Keep it insulated.
You can use insulators between the diode and the heat sink so that the sink can be attached directly to the frame or just insulate the whole heat sink.

The leads going into the diode are really prone to breaking off if they get bent more than once too. It's good to securely attach the wires to the heat sink with zip ties, glue, whatever, so that pulling on the wires does not place any stress on the diode legs.

You can use the fact that the anode is connected to the heat sink tab to your advantage. Just snip off the middle leg and make your wire connection to the tab. You can use a bolted ring terminal or solder directly to the tab. The tab is not prone to breaking off and it gives you more room to attach to the other legs.

You can put a jumper between the two outside legs and use both diodes in parallel as a single diode. This doubles the current rating. Since both diodes are on the same die, they will always be at the same temperature so current sharing will be pretty even.
 
So, does this Schottky look like a good option? I'm currently running 36V 40A total, and this one would allow me to upgrade to a higher voltage.

http://digikey.com/scripts/DkSearch/dksus.dll?Detail?name=497-3220-5-ND

How hot do these things get? I'm thinking of replicating what ebikes.ca assembled (which sounds like they will no longer carry/produce).

So I'm thinking of a potting box like this:

http://digikey.com/scripts/DkSearch/dksus.dll?Detail?name=377-1174-ND

Using these for heat sinks:

http://www.frozencpu.com/products/7325/vid-110/Thermalright_8800GT_PWN_Heatsinks_NVIDIA_8800GT_ATI_3850_3870.html?tl=g40c21

And embedding the diode and wires (with heatsink fins exposed) in epoxy.

http://search.digikey.com/scripts/DkSearch/dksus.dll?pname?site=us;lang=en;name=3M6435-ND

Sound like a good plan?
 
Back
Top