Quark ESC Power Board Comparisons/Mods
 

I am in the infant stages of challenging myself to build a custom power board that will use higher amp MOSFET's but be controlled by the Quark Brain Board and Software.

The MOSFET's that I plan on using are:
75545P
N236
A1
The absolute max ratings are 75A, 80V, 0.010 Ohm, N-Channel, UltraFET Power MOSFET. This style has compatible on/off speeds with the FET's used in the Quarks.

I noticed in looking up information on the 75545P that since it's a N-channel MOSFET, it is neither a PNP nor a NPN FET. It doesn't have a preference to which polarity of DC current controls it's gate.

I've been digging around in my 33Amp ESC doing some research on it's FET's. This ESC uses two types of FET's. The TPC8107, Silicon P-Channel MOS Type(PNP type), and the IORP550C NPN Style FET. These have max voltage limitations of 30 volts. The data sheets don't show a max amp rating, but it can't be much.

I was wondering how to adapt the circuits to the N-channel MOSFET's knowing it was going to be a big problem until I began looking at my Quark 80 amp ESC.

The Quark 80 Amp uses 24 of the the 22N03S MOSFETs! These are N-Channel also. The Data sheet on this MOSFET shows absolute ratings of a 30 volt limitation with a 50 amp current limitation. Much Better!

I can only see the front 12 MOSFET's, and I'm not tearing my Quark 80 completely down just to see. Does anyone know if the MOSFET's on both sides of the board are the same?

I'm wondering also, What are the numbers on the ones used in the Quark 125 amp model? I looked around at some of the other posts pics, and I can't make it out. They appear to be the same, plus the 125 amp model uses 36 of them.

I would also like some FET input on the Quark 65 amp model if anyone has access to that.

I know that adapting a larger MOSFET in place of a smaller MOSFET will be much easier to do than using a brain board that currently controls a NPN, PNP Fet arrangement.

I'm going to look at this in two steps.

Goal #1. Use a MOSFET controller like the 80 amp or 125 amp to control a custom built powerboard capable of 200 or 300 amps continuous. Same software, same voltage limitations.

Goal #2. Is it possible to regulate the voltage to the existing logic controller in such a way, that it will control a power board that is at a much higher voltage? Regulate incoming power to the logic board, and also incoming feedback from the motor? I'm a long way from answering this one, but there must be a way.

Any help, or input would be greatly appreciated.

Wow! First of all, let me say you are tackling a complex project here!

When you say PNP/NPN, I assume you are familiar with bipolar transistors. Transistors are "current" driven, have a current gain (beta), and have a set collector to emitter voltage drop even at saturation. MOSFETs are voltage driven (they don't load down driving circuits) and their drain to source v drop is determined by the Rdson value and current flow. Think of them more like a faucet where the source is the water tank, the drain is the output, and the gate is the valve. Rough analogy, but sufficient. There are other differences, but you see what I mean - it's like comparing apples to oranges.

Aside from being N or P channel, are they enhancement or depletion types? I would think you would NEED enhancement types, but I dunno.

Sorry I can't help you any further on this as I've never really dealt too much with FETs at the design level.

The Quark 125 uses the same FET parts, just more of them. IIRC, they are the same on both sides.

Thanks for the reply Brian. You're very knowledgeable and I was hoping you would chime in on this.

I like your analogy. It's a good one since DC Current only flows one way. I've often thought like that too.
I find myself switching back and forth between the two sets of data sheets comparing figures and I find no reference on either as being enhancement or depletion types.

I think I'm in the ball park on compatibility. The donor drive unit that gave up these bigger MOSFET's served as a controller for a 36 volt drive motor. The motor is brushless and is a 6.6kw unit. This was not a sensorless setup by any means. It pushed around a 9500 pound Industrial truck at 9mph.
The Controller these came out of has an extensive brain board that uses an encoder bearing in the motor to calculate timing. Based on encoder information, the controller knows motor speed and armature postion. Much like the Novak sensor based setups.

I want a high Amperage Car Controller that will do multipole motors! A HV setup that will do 36 volt reliably would be a dream come true.:yes:

I'm sorry but I do not think that this is a good idea, based on your questions...

P.S. But if you willing to try, then shoot me PM and I will try to help you out with design.

Hi Griffin,

Thanks for volunteering to lend a hand. I appreciate that. I have followed some of your work and respect you as being one of the most knowledgeable people around here on this.

I look at myself in this, like a student, that has alot to learn. But I like to challenge myself, sometimes things work out for the better, and sometimes they don't. But in the process, along the way, knowledge is gained.

I have spent several hours tonight just reading about principles of operation. There's alot to learn, but this is how we grow. :yipi:

Cool, I was planning to start writing in Brian's thread about "ESC's".
Creating threads on modifying ESC's my goal was to help hobbyists to improve their controllers without knowing too deep how it works. And for any good electrical engineer all my tricks are quite obvious, but to design good ESC will require more then just EE...experience helps allot! :)

Ok, according to the data sheets, here's the resistence values:

Big MOSFETS - RDS(ON)= 0.010 ohms max with 0.0082 ohms typical
No Gate Resistence value listed

Quark 80 amp MOSFETS - RDS(ON)= 2.2 megaohms
Gate Resistence = .6 ohms

The Big Mosfet data sheet doesn't list a Gate resistence value.

I wonder how much current the Gate lead circuit in the Quark 80 Brain board can support?

They should have Rg mentioned somewhere, maybe on the Graph...
This parameter responsible for FET's frequency response, noise and etc when you run simulation and important in design stage.
"Big" fets are usually slow, bulky and rated for high power plus in most cases have lower RDS(On) then "smaller" ones.

You need to check for fet driver part number in Quark 80 brain and study the circuit around it. To calculate required Voltage and Current check for switching frequency and total gate charge. Then do the same calc with different loads and temps.

Here's the Data sheet on the Big Mosfets that I intend to use for the new powerboard.

http://www.fairchildsemi.com/ds/HU%2FHUF75545P3.pdf

Here's the datasheet for the Quark 80 Power Mosfets.

http://www.ortodoxism.ro/datasheets/...3S_Rev1.11.pdf

Griffin, could you take a look at these and tell me if they are capatible with each other?

I'm going to research the driver fet's in the Quark 80 now.
Thanks.

Keep in mind the higher you go in voltage the higher the resistance will be. You can find 40V MOSFETs that have extreme high AMP ratings, but they will be limited to silicon and/or die ratings. Browse www.irf.com or www.infineon.com and you can find a great selection of MOSFETs. There are also various other brands out there.

Well, it is hard to give you straight answer, because more info requires about application, like in your case what are the current ratings of fet's driver current and power ratings of system providing power to fets' driver IC's? Do they use bootstrap technique in Quark 80 if yes then what capacity bootstrap cap has and what are diode ratings?
Then what exactly you would like to accomplish, stock configuration or higher voltage/current one. How many Fets you would like to use...

As far Rg for "Big" Fet - check on page 6 inside PSPICE model RGATE value 0.87

And, I tend to agree with lutach there are plenty fets, what so special about this one?
I mean both MGM, Quark, Castle... are using probably the best fets for the price and ease of assembly, when you shooting to do something custom you want to get the best (regardless of price and assembly pain) and deal with tuning and tweaking until "green light".

These "Big" Mosfets are ones that I have on hand. Just wanted to do some comparisons and learn more about compatibility.

Correct me if I'm wrong, but Mosfets with a lower on resistence value translates into more voltage at the motor leads with less heat buildup in the ESC.

I took my Quark 80 back out of the case to investigate the driver fets and their ratings. They are sandwiched between the power board and the brain board. I can't get a good look at them. I'm hoping to catch Zeropointbug and ask him for a closeup view of the brain board that he has already disassembled.

"I mean both MGM, Quark, Castle... are using probably the best fets for the price and ease of assembly, when you shooting to do something custom you want to get the best (regardless of price and assembly pain) and deal with tuning and tweaking until "green light"."

In Business, I'm sure they are balancing the most profitable (cheapest) way to go and live up to the ratings of their ESC's.

On the Custom work, I can't aggree with you more. I'm not worried about the cheapest route.

Is it possible to use a Quark brain board from an ESC that's rated for 2 to 6 Lipo cells and adapt it to a larger capacity power board for a 10 or 12 lipo cell rating?

I see.
Rg and RDS(On) are not the same. RDS(On) directly related to voltage drop across FET.
You cannot tell how good FET is based on RDS(On) you need to check dynamic parameters the ones with "ns" units. Transition time from off/on will determine how efficient switching happen, less time better. Later we can cover why too fast is also not good.

Unsoldering, unsoldering and picture might not be enough...probing and digging layout.

Over 25V opto-coupling or other technique will require to isolate from high-voltage switching problems. So, probably won't work.

It would be the high voltage back EMF sensing that would cause the problems in a HV setup?

That is minor and must be solved at the same time when you figure out how to generate supply voltage for your ESC microprocessor and subsystems (It is complicated going from above 35V down to 5).
You still need to sample supply voltage, if you plan to monitor battery voltage to use LiPo cutoff.

-
As far high voltage problems, it is a complex to describe, but think of your ESC as a hammer. When it is small you can use it and neighbors won't even hear you, while if you start using pneumatic one then entire neighborhood will know that you mean business. :)

One thing I have to add. As you all know there are some delays with m ESC, but the engineer that is building it told me that it would be easier to make a 80 or 100 volts controller. I already told him I will do a few after I receive the one I'm waiting for. Just te thought of having a ESC capable of 18S and 22S lipos just makes me happy LOL. I think I like HV stuff mainly because I got electrocuted twice. First time was in Brasil with 220 volts, I was 7yrs old. The other time was in HS in Florida. It was from a nasty Hi Amp Audio Grade battery that knocked me a few feet back LOL.

If you like that, you should check out a TV flyback transformer sometime! Don't know if it would produce enough current to kill, but the voltage sure hurt!

I have sold a few thousand units custom made for Philips Brasil. Thye had something like 12Kv. I'm not sure if it was 12000 volts.

Yeah, that's 12,000v. Usually, since the voltage is stepped up, the current is stepped down so there isn't much current available. But, it only takes ~15mA to kill (that's 0.015A)...

Good thing for fuses and circuit breakers LOL.

I decided to take a few pictures of this HV power board before I completely tear down the original configuration and perhaps learn a few things from analyzing what they've done.

I notice that they have all the gate leads tied together through a series on resistors inline. I'm not sure what effect this will have, but I think it governs the amperage the gate leads are allowed to draw.

The Blue Component marked "Wima" is like a small Capacitor, it will charge up under an Ohm meter. This is connected between the source and the drain. A Cushion? I would think they need to charge up first, before full output power is seen.

http://i160.photobucket.com/albums/t...m/DSC06623.jpg

Wait second! Are those more Low ESR Capacitors I see? (*Joke*)
http://i160.photobucket.com/albums/t...m/DSC06621.jpg

I'm thinking about using this board for testing. Strip the resistors, and use the existing pathways for the Mosfets. The Caps are already in place too.

This thing is huge though. It's 8" x 3" x 2". If I can make it work and put it through some testing, I'll reconfigure the entire thing to make it compact.

Absolute ratings for this board is 375 amps. Perhaps it will do 300 amps continous with the proper heat sinking. (which I have also)

Come to think of it, How am I going to test this thing? I know a Data Logger is in order, but to seriously load it? Hmmmmm...........aah well, one thing at a time.

I have a AXI 4130/20 that I could bolt a 20" inch prop onto. It's rated for 8 Lipo cells. Maybe open up the doors and blow the house out. LOL

Please keep us updated Lutach. You sure have my interest in that. But please, be careful around all those flowing electrons. LOL

I will keep everyone informed, but I might wait until I have one unit here for testing. Now I'm really liking your controller. Size wise, yes it looks big, but the power you might be capable of producing, the size of the controller will not matter. Keep me informed on how good your monster (In a good way) performs.

Thanks Lutach

This powerboard has 30 of the larger 80v 75amp Mosfets. This will be a hefty gate load for the drivers to carry. This is why I need to study Brain Board Components alot closer to calculate driver load carrying ability.

That is nice. You might be able to use 18S lipos with your controller. Good for speed runs LOL. Make sure all the components are rust free.

Yesterday, my Quark 80 amp died. It still blinks the red "heart beat" at the led's. It also lights different led's with different throttle positions on the controller. No power is being sent to the motor.

I suspect the problem is on the power board. I hope it is.

Can I adapt the brain board to work with this larger power board?

That's what I'm thinking about doing. It would be a 6S ESC that would kick butt.

The larger power card will do 300amp at 36 volts. That would be 450 amps on 6S.

When you guys have the ESCs ready.. I'll be sh*ttin' my pants to test it out.. I've gotta make use of all my lipos...

hehe, that low voltage Brain! :wink:

Come stop by and I will show you my 'Tesla Coil' running at ~700,000volts AC... very load, and looks VERY deadly, however you can hold you hand in the arcs and you only feel very minor shocks... that's because it's running at very high frequencies ~500khz resonant freq.

I have heard of ppl using those flyback transformers in small Tesla Coils for the primary circuit.