Tinkering with the Naos Raptor - 20A, 120W dc/dc converter...

[mamu]

With not using the control pin, I pressed the button several times with the batts in backwards and no issues at all - batts not even warm to the touch.

[sphearion]

While I don't like the idea of firing the raptor under load like the easy okr, what is the downside of doing it this way?  If its nothing to worry about and the chip is technically designed to do this, seems the easiest remedy in the short term is not using the control pin.

[CraigHB]

One time I had a board fail on me with a soft short, board got hot, battery got hot, but the fuses didn't trip.  Not to say your board is bad, just that it may be drawing enough current to get hot, but not enough to trip the fuses.  I could be way off on that being a factor in your case, just something to consider.

You don't ~have~ to rely on a fuse for reverse polarity protection.  You can add a MOSFET and that will eliminate your problem.  The only real benefit in relying on a fuse is to reduce the part count and reduce the size.  There's nothing wrong with using both.  Even with a MOSFET, it's still better to have short circuit protection in the event the board fails to cover that fault.

[sphearion]

the Fet I used robbed enough power that you could not use the chip at its higher settings.  While it did stop the reverse voltage, losing 1 - 1.5v of output on the chip makes it no better than an OKR-t10 IMO

[CraigHB]

Just use a good transistor, won't be an issue.  The main power switch in my own mods uses a P-Channel MOSFET with only 2 mOhms on resistance.  At 20A, that's a voltage drop of only 40mV and a power loss less than a Watt.  Sounds considerable, but it isn't, only a percent considering you're inputting about 80W at those currents.  For the currents you would typically see powering an atomizer, it's nothing, at 5 Amps input, it's 10mV drop and 50mW power loss.

http://www.vishay.com/docs/62860/si7157dp.pdf

[mamu]

While I don't like the idea of firing the raptor under load like the easy okr, what is the downside of doing it this way?  If its nothing to worry about and the chip is technically designed to do this, seems the easiest remedy in the short term is not using the control pin.

With not using the control pin, I had issues with getting up to 6v output - it would only go to 5.1v.  Never did figure out why, but once I hooked up the control pin, I got 6v output.  Also, to get an easy low voltage protection is wiring a zener to the on/off control.

[sphearion]

I am using a 6.2v zener between control and switch for low volt protection. and that works great.   I dislike not being able to get the full 120watts.   Not that i ever use it that high. just being able to

[mamu]

Just use a good transistor, won't be an issue.  The main power switch in my own mods uses a P-Channel MOSFET with only 2 mOhms on resistance.  At 20A, that's a voltage drop of only 40mV and a power loss less than a Watt.  Sounds considerable, but it isn't, only a percent considering you're inputting about 80W at those currents.  For the currents you would typically see powering an atomizer, it's nothing, at 5 Amps input, it's 10mV drop and 50mW power loss.

http://www.vishay.com/docs/62860/si7157dp.pdf

Thanks Craig.  I ordered a few and will give it a go.

[sphearion]

mamu, what about isolating the control pin from ground not only with the pull down resistor. but also with a schottky diode. maybe there is some weird leakage there that is causing the battery heat, (im a noob just trying to figure out why the control pin being connected without the button pressed its only connected to that ground)

[whoi]

Thanks Craig.  I ordered a few and will give it a go.

Very interested to see the final schematic.

[mamu]

mamu, what about isolating the control pin from ground not only with the pull down resistor. but also with a schottky diode. maybe there is some weird leakage there that is causing the battery heat, (im a noob just trying to figure out why the control pin being connected without the button pressed its only connected to that ground)

Tried it.  When I reversed the batts, the diode got so hot it melted the insulation from the ground wires - also burnt holes in the breadboard where those wires were and that reaction was fast - like a second or two.    but the batts didn't get hot.

So took the diode out and wired the fuse to the on/off control pin instead of Vin.  If you look at my diagram in post #16, I just switched the fuse from Vin to in/out at the fire switch.

Surprisingly, it didn't make a bit of difference for input/output function as the board functions just the same as if the fuse was wired to Vin.  But I had to also wire ALL the ground wires to the center ground pin and not the outer ground pin for it to work, not sure why that is, but when I put the batts in backwards, the zener diode caught on fire when ground wires were wired to the outer ground pin. 

I re-read the datasheet and googled info on this Raptor board, but cannot find why there are two common ground pins and what the traces are for each.  I wonder if one is just for the Turn-Table loop and PGood, but have no idea.  The Trim ground pin is independent and not common with the other two ground pins like it is with the OKR.

So I replaced the zener and re-wired all the ground wires to the center ground pin.  Put the batts in backwards and aok nothing gets hot let it set for a few minutes and aok BUT when I press the fire button the batts, batt springs, and batt wiring gets hot.

So I've got reverse polarity protection as long as I don't press the fire button. 

But of course, this is still not a solution.

So with all the ground wires now wired to the center ground pin, I re-wired the fuse back to Vin, put the batts in backwards and no go - batts and springs and batt wiring gets hot immediately.  So the fuse needs to be wired to the on/off pin when it's enabled. 

I'm at a loss at what else to do, so will work with the P-FET Craig linked to when it gets here.

I just wonder if this is the same with reverse polarity with the OKR when using the on/off pin.  I've got an OKR-T10 on hand to tinker with, so may wire that up and give it a go when I have time.

[Breaktru]

This is really weird mamu. Technically a fuse in the battery positive should do the trick.

I wonder if you put the fuse in the Negative of the batteries if that would help? Or perhaps also in the positive and negative of the battery. 


Update:
I have been thinking about the 2x 9A fuses = 18A with trip current of 2x 15.3 = 30.6A
Wouldn't you think that before the trip current is reached that things would get super hot?

[mamu]

Thanks for the suggestion, breaktru.

Tried a fuse on positive and with a fuse on negative - still no go.

By no go I mean that I have reverse polarity protection as long as I don't press the fire button.  The batt and batt spring wired to ground get the hottest the fastest.

I think I'll try a fuse at the junction wire connecting the batts in series.

I had wondered about that trip current being high like it is, and I think you're right, breaktru - that is a high trip current so things are going to get hot before it trips at 30A.  I feel the fuses during this testing to see if they are getting hot, and they're not getting hot.

When I tested the DNA with a 6A hold (2x 3A in parallel per batt), 12A trip per batt, the fuses got hot really fast when I put the batts in backwards and the batt and batt contact didn't even get warm to the touch so they were protected with the fuses.  And that was with independent fuses per batt.  With a common fuse, the batts are not protected.

I wonder exactly how much current is being dumped from the batts when in series when put in backwards.  I'm using a pair of the 30A Sony 18650 batts for testing.

All this time with all this testing the board has been protected, but it does have some overcurrent protection.  It does take a minute or so for it to start working again after a fault though.

From the datasheet:

Overcurrent Protection
To provide protection in a fault (output overload) condition, the unit is equipped with internal current-limiting circuitry and can endure current limiting continuously. At the point of current-limit inception, the unit enters hiccup mode. The unit operates normally once the output current is brought back into its specified range.  The typical average output current during hiccup is 10% of Io,max.

[CraigHB]

Thought about this a bit more and might have some insight for you.  The reason a fuse works in providing reverse polarity protection for a converter is because of the current path formed by the switching MOSFETs.  One of the characteristics of a MOSFET is it can only block current in one direction.  When reverse voltage is appliced to the drain and source, it behaves like a diode.  The formal term for that is a MOSFET's body diode.

Specifically, the reason MOSFETs behave this way is because the gate and source substrates are tied together with a common connection.  You see that in the schematic symbol.  Otherwise MOSFETs can latch on under certain operating conditions.  There are acutally 4 wire MOSFETs available that don't exhibit the body diode characteristic, but you have to know what you're doing with those since they can latch on under certain conditions.  They're often used in SCR type devices that utilize this characteristic.

So, under reverse polarity, the power supply sees two diodes in series formed by the converter's switches which is basically a hard short.  This allows enough current flow to trip the fuses.

I don't know why, but for some reason, use of the enable pin is limiting current flow enough for the fuses not to trip under a reverse polarity condition.  It probably has something to do with the switches the Raptor uses and the reverse voltage applied through the enable pin.  It may be causing one of the transistors to become active somehow and limit current.

The only solution I can think of for this situation is to use a transistor for reverse polarity protection instead of relying on a fuse.

[Iamthebadass]

I want to try and use a P or N channel fet for reverse polarity, but every time I try to search I get confused. I do not exactly know what I need as far as drain source and gate source. It is really confusing me. Craig, could you maybe explain to me what each number means for one of these pups? And how it would relate to this circuit? (Ie. 30A, 8.4v)

[mamu]

That makes sense, Craig. 

The MOSFET at the on/off control also allows us to use a low rated switch, even a tact switch if we want.  Whereas, if we don't use the on/off control, we need a rated switch that can handle the load.

The fuse in this situation with using the on/off control isn't protecting the batts and you're right, an external MOSFET is required to protect the batts.

I should get the P-FET you linked later on in the week and will update.

I think first I'm going to hook Source to Vin just to see.  If that doesn't work, I'll hook it to the on/off control.

[CraigHB]

I want to try and use a P or N channel fet for reverse polarity, but every time I try to search I get confused. I do not exactly know what I need as far as drain source and gate source. It is really confusing me. Craig, could you maybe explain to me what each number means for one of these pups? And how it would relate to this circuit? (Ie. 30A, 8.4v)

There's some old threads here where we've discussed MOSFET specifications in depth.  Maybe someone can be so kind as to link to one of them.  You might be able to find them using the forum search, that's almost a cliche, but it really can be difficult to find threads through the sea of results using the forum's search function.

There's also a thread that discusses reverse polarity protection using a P-channel MOSFET, again same thing to say about finding that one.

[Iamthebadass]

Thanks craig, I think i googled my way into oblivion, but i think I may have it.

And then I realized you linked a fet, so I checked it out and ordered some to try.

I swear I can never leave digikey without going "oh well I could use 100 more resistors... and capacitors... and diodes" Sheesh.

needless to say I am good on those for quite some time.

[CraigHB]

I understand that, I have a whole drawer full of components, but fortunately most of the chips are free samples.  I'll get an idea for something and order some samples, then decide to use something different.  Sometimes I buy a good of number of something only for it to end up in the collection due to a change of heart.  I seriously have about 20 different power FETs on-hand, just got a few more today from TI as a matter of fact, really amazing ones.  MOSFET specs get me excited, check out the ones on these babies;

http://www.ti.com/lit/ds/symlink/csd16321q5.pdf

http://www.ti.com/lit/ds/symlink/csd25402q3a.pdf

It's not just the low on resistance, but the insanely low gate charge, these are lightning fast MOSFETs with virtually no appreciable power loss.  Best ones I've ever seen, though the speed is only a factor for high speed switching like in a power converter, not a factor for reverse polarity protection or user on/off switching.

[Iamthebadass]

how do you get free samples? I want in on this!

[CraigHB]

Just ask for them really.  Go to a site like TI and create a user account.  Go to the samples area and order what you want.  They'll usually give you a quantity from 2 to 5 depending on the maker.  Some have given me quantities as high as 10 for free.  The one caveat is that samples are intended to be offered to professionals looking to produce products using the parts.  They will usually ignore requests if the account was created with an email from Yahoo or Google or other public provider.  Best to have a unique or business like domain on your email, but often an ISP domain will do.

[Iamthebadass]

OK craig, I got those FETs in...

And they are smaller than hell. How am I supposed to solder onto these with no board for them to sit in???

[CraigHB]

How am I supposed to solder onto these with no board for them to sit in???

Just make a host board for them.  It's really not a big deal. 

You can get copper clad board cheap on ebay; link.  Any copper clad is fine, the thinner stuff is easier to work with and you can use 1oz or 2oz copper.

It's a simple pad layout, just three pads, gate source and drain.  You can probably use a Dremel tool to score out the pattern or you can etch the board using masking tape as an etching solution mask.

Once you have a board, just reflow solder the part on there.  All you need is some solder paste, not expensive; link.  You can reflow solder using anything that gets the board hot enough to melt the solder, heat gun, toaster over, frying pan, whatever.

Once you have the part mounted, hand solder wires to the pads and cover the whole assembly with heat shrink tubing for insulation and protection.

Careful handling MOSFETs, they're highly vulnerable to ESD damage.  You should use a grounding strap when handling them until they're installed in a circuit.  You can get cheap tape sytle ones that you can just stick onto something with a ground from utility power like the bottom of a PC chassis; link.

[Iamthebadass]

Thanks for the info.

I'm just gonna "borrow" and ESD strap from work.

[mamu]

Reverse polarity protection for the Raptor... tested and good to go with protecting the batts if put in backwards and with no adverse effect on the circuit - voltage drop under load is the same with and without the P-FET.

Wiring in a P-FET:
Part # SI7157DP (many thanks to Craig for recommending the appropriate P-FET to use for this board - the P-FET I had used for reverse polarity protection for the DNA was no go with this board and I'd prob still be searching for the right one  many many thanks for sharing, Craig!!)

S: Vin
D: Batt +
G: Batt -



1.  This is a teeny tiny smd part and needs mounted.  I used a 3x4 piece of veroboard and removed areas of  copper as shown below...


2.  Solder the source (green wire), drain (red wire), and gate (black wire) on the veroboard by inserting the bare wire end from the underside of the veroboard in the appropriate edge hole and sliding the wire though to the top, apply flux to the bare wire end, then apply solder to the iron and swipe across the wire including the copper traces connected to that wire...


3.  Position the P-FET in place, apply flux to all pins, then solder to the board - line up the gate pin (orient toward the black wire) and solder it first, then apply solder to the tip of the iron and swipe it across the 3 source pins (toward the green wire) making sure you don't solder the gate pin to the source pins, then apply solder to the iron and swipe it across the 4 drain pins (toward the red wire)... make sure the 3 source pins and the 4 drain pins have connecting solder and that the gate pin remains isolated from the source pins...


4.  Insert shrink tube over the P-FET and apply heat...


Updated wiring for reverse polarity protection...

[CraigHB]

(many thanks to Craig for recommending the appropriate P-FET to use for this board...

Welcome, glad to be of service.  That particular FET has very low on resistance, less than 2 mOhms with the gate-source voltage of two series cells.  You should see very little voltage drop with that one, less than 20mV at 10 Amps.  In fact, the wires connecting it probably have more resistance than the transistor does.

[mamu]

It's perfect, Craig, doesn't appear to be affecting circuit resistance at all, not that I could measure anyway.

And it's doing a terrific job - absolutely no heat at all from any component, including the batts when they are put in backwards.  I left them in backwards for several minutes and nothing, not even warm to the touch and the P-FET, fuses, etc not even warm to the touch.

[CraigHB]

The FET actually shuts off when reverse polarity is applied so no current flows at all expect for a tiny immeasurable leakage.  You could leave the batteries in backwards indefinitely and not a thing would happen, not even the batteries would drain.  At least that should be the case.  You can verify that by checking voltage between the gate and source reads zero when the batteries are in backwards.  It will ready battery voltage when the batteries are installed correctly.

[Breaktru]

Nice guys. Finally a solution for the Raptor.
So let's some it up. For battery reverse protection the FET will only be necessary for the Raptor. And not needed for the OKR or 08100w as the PTC will cover reverse protection for those two converters. Would you agree?

[Mandro]

Could I ask a (probably simple) question ?
Is this reverse protection only needed to avoid user error ?

[mamu]

Nice guys. Finally a solution for the Raptor.
So let's some it up. For battery reverse protection the FET will only be necessary for the Raptor. And not needed for the OKR or 08100w as the PTC will cover reverse protection for those two converters. Would you agree?

The thing is, with this Raptor I tested reverse polarity without the on/off control wired in and was aok for reverse polarity - same as is for the OKR and 08100w without using the on/off control pin. 

So, I'm thinking with the OKR and TI board, both will need the FET if using the on/off control pin if wanting reverse polarity.  The Raptor, OKR, and 08100w are the same as for current control when using that enable pin.

I wish someone who has either of those boards in a mod and using the on/off control and with just fuses would test it and chime in about reverse polarity.  I just haven't had time yet to work with the OKR-T10 else I would do it.  If I get around to it though, I'll test it as I will be using the on/off control for undervoltage protection.

Could I ask a (probably simple) question ?
Is this reverse protection only needed to avoid user error ?

Yes - all this trouble and extra wiring and P-FET for having the stoopids and not paying attention.   

Of course, if a modder is using lipoly batts with these boards, no worries about reverse polarity and no need for the FET.

[Mandro]

Yes - all this trouble and extra wiring and P-FET for having the stoopids and not paying attention.   

Of course, if a modder is using lipoly batts with these boards, no worries about reverse polarity and no need for the FET.

Thanks mamu, I'll leave it out then. I'm building a raptor mod ATM for myself and as usual, I'm running out of space lol

[sphearion]

I should be getting some of these 781-SUP75P03-07-E3 today from mouser,   I will update if these work as well because they could easily be added directly between + and - on the battery sled, giving you a new positive on the sled taking minimal space and without needing a host board.

Here is to efficient fets!

[Visus]

Thanks Mamu/Craig just ordered one to play around with myself.

Now it will take me another year to order the diode, fuse, and fet lol..   
Still waiting on parts for MCU mod 

[Breaktru]

I probably won't be adding the FET since I only build mods for myself, don't use the enable pin and mostly use soldered in Li-Pos.
But if the need arises I know where to look. Thanks Craig and Mamu

[CraigHB]

Could I ask a (probably simple) question ? Is this reverse protection only needed to avoid user error ?

Yes, but the thing is, I'm the biggest idiot of all when it comes to something being idiot proof.  There is no doubt at all that with removable cells I will put them in backwards at some point.  With these high energy cells, it will definitely destroy the mod if that happens even if it's not matter of personal safety.

For non-removable cells, reverse polarity protection is not a requirement since it takes the idiot out of the equation, but in some cases you still need the short circuit protection.  Buck regulators can limit current and protect from short circuits, but boosters can not.  It's just how it works out with the difference in the electrical topology between the two converter types.  Also, even for a buck regulator, it's never a bad idea to have backup short circuit protection in the event the board fails to handle it.  That's not an unreasonable contingency.

[Mandro]

Thanks Craig, looking at mamu's photo, it doesn't look like the fet will take up too much space in this configuration so maybe I will include them from now on.

[mamu]

I like these challenges for the learning experience.

I tell ya, when I was testing and that zener caught on fire it totally freaked me out - there were flames coming off the zener and melting it and the breadboard.  And testing with the diode and it melting the insulation of all the ground wires made me uneasy lol.

All for a good cause though.   

Think I'll wire up an OKR-T10 this evening and see what's what with that enable pin without a FET.
 

I should be getting some of these 781-SUP75P03-07-E3 today from mouser,   I will update if these work as well because they could easily be added directly between + and - on the battery sled, giving you a new positive on the sled taking minimal space and without needing a host board.

Here is to efficient fets!

Please do keep us updated, sphearion, as I am curious too.

[Breaktru]

Think I'll wire up an OKR-T10 this evening and see what's what with that enable pin without a FET.

Will you be wearing your fire suite?
Be careful

[Iamthebadass]

I just added one to my raptor build. works like a charm. I am using a 1590G case, so it was really tight.

I soldered it directly to the fuses to save space. and its sitting to the left of the battery sled.

I literally have no space left in this mod, maybe 2x2mm left, but I seriously connot believe how little room is left.

It works like a champ.

Thanks again Mamu, craig, and Breaktru.

[sphearion]

781-SUP75P03-07-E3 today from mouser do in fact allow the chip to fire at 6 volts.   the fact that you can solder them flat against the front of your battery sled using by folding the legs to the left and right and the gate and drain legs are just long enough to reach the contacts on the battery sled, then bend the source leg toward the face of the chip enough for some heatshrink tube, and solder your + to it. and  it takes up very little space.

[Iamthebadass]

Thanks spheroin, I may try some of these out as they are easier to use than SMD fets.

[CraigHB]

That one's not too bad.  Looks like it has 8 to 9 mOhms at the voltages you see with series cells.  For the currents you would typically see powering an atomizer you'd be looking at about 25mV drop and a 75mW loss, about a half percent, certainly tolerable.  It's a bit ugly for higher currents, at 15A input, losses would be about 120mV and 2W which is about 2%.

[mamu]

That's great news, sphearion.

Did you test input/output loaded voltage difference to see if there is any difference with using the FET vs without the FET?

[sphearion]

I did not test before adding the fet, honestly i was kinda excited to see if it worked at all,   The last fets i used if i turned the pot up to 6v the raptor would just humm.  this time, i turned it up, put my .8ohm atomizer on top and hit the switch while measuring voltage with my multimeter and saw 6.001 volts ouput loaded.

[mamu]

I totally understand the excited YAY it works. 

[sphearion]

is it possible that a surge of voltage through a P-Fet could cause it to stop working, IE Inserting batteries? 

[whoi]

Anyone have an updated schematic and part list (or just a labeled schematic)?

Nevermind, I didn't see the updates :0)

[Mandro]

I'm still working on my raptor mod and was going to use a fet for the reverse protection but decided to use the fet on the switch instead of using the enable pin.

[Visus]

I'm still working on my raptor mod and was going to use a fet for the reverse protection but decided to use the fet on the switch instead of using the enable pin.

I am not sure if this was sorted but here's Mamu's quotes on that..

if anyone decides to use this module in a mod, I would highly recommend using the control on/off pin even though it adds a bit of a mA drain (similar to the OKR module).  Without it, voltage drop at higher output volts is significant

With not using the control pin, I had issues with getting up to 6v output - it would only go to 5.1v.  Never did figure out why, but once I hooked up the control pin, I got 6v output.  Also, to get an easy low voltage protection is wiring a zener to the on/off control.

.