Game Changer Page 3 with video of 1700amps!!!!

[speedmd]

Making it round as an extension of a motor case would be very economical compact and clean looking. Even if it were separated by some small venting (fan) section it could share the same extrusion tooling as the motors outer case. Like it. Nobel design goal.

 

[Arlo1]

when I said "hard sell" it is difficult to determine physical size from pictures. It looks like this controller is about 10" in diameter. I may be waaaay off and that's totally cool, but that was my impression. What are the actual expected dimensions? Anyway, for most EV solutions on here, a 10" diameter controller is probably biggish.

This is Alpha beta version 1.0
the next revision is 1/2 the diameter. When finally ready for market it will take up a lot less space then something that is square or rectangular of similar current.
Don't worry about the size of Alpha beta 1.0 it has a huge heat sink and it is just to prove the design works and find bugs I want to change before ordering the boards to make it smaller.

Size will depend on the current you want to flow though the motor.

 

[mrass]

Hi Arlo I like what you are doing here! But why not use a resistor between gate and drain?

 

[Arlo1]

Hi Arlo I like what you are doing here! But why not use a resistor between gate and drain?

What do you hope to accomplish with that?

 

[mrass]

A faster but foremost more reliable turn-off of the mosfet, it's pretty common practice. About 2k2 ohms (locally!) per mosfet should be ballpark.
Not that I doubt your skills or would interfere with your design, just being curious!

 

[Arlo1]

A faster but foremost more reliable turn-off of the mosfet, it's pretty common practice. About 2k2 ohms (locally!) per mosfet should be ballpark.
Not that I doubt your skills or would interfere with your design, just being curious!

Where have you seen this done?

Your telling me you want me to tie the 118v battery supply to the gate.... thats what the drain to source difference will become during switching this will try to turn the gate on when its off.
What you are suggesting will be instant death of a controller with anything over 20v on the battery voltage. As you need to understand this will slowly apply more then 20v to the gate. In fact it will slowly charge the gate to full battery voltage! The gates are typically rated for 20v max!

As well. It will slow the turn off down not speed it up. When you turn the fet off the drain voltage will raise if you tie that to the gate it will try to turn the gate back on not off.

Show me examples of this Gate tied to Drain....

Or do you meen Gate tied to Source which is a pull down resistor and will slightly help speed up turn off slightly. But a pull down is actually there just to keep the fet off in a situation when the driver does not have control of it.

 

[mrass]

Whoooops, a typo indeed... did mean gate to source.

 

[Arlo1]

Yup and for that I use a 4k7 pull down!

I have a better sch I will post later but the pdf that is there does show the pull down Its just 1 for all fets and I do realize they have their own gate resistors but those will be a DNP part or if needed <1 ohm each. They are just there in case.

 

[mrass]

Yup and for that I use a 4k7 pull down!

I have a better sch I will post later but the pdf that is there does show the pull down Its just 1 for all fets and I do realize they have their own gate resistors but those will be a DNP part or if needed <1 ohm each. They are just there in case.

Good man! wil make me sleep better tonight (was trying to think of a reason why not to put a pull-down resistor in, also was reading up on your gate driving issues on the first 25 pages of your other thread..).
A 4k7 alone would drain (pun intended) a 5nf (typical for bigger mosfet?) gate capacity from 10v to 1v in about 54 micro seconds.
Unless you`re actively driving the gate low it would be to slow for decent switching speeds.
Keep up the good work!

 

[ElectricGod]

Am I reading the schematic correctly? 16 mosfets per phase? Yes...you will be able to power a D9 bulldozer with this controller!

 

[Ohbse]

Am I reading the schematic correctly? 16 mosfets per phase? Yes...you will be able to power a D9 bulldozer with this controller!

You don't even need to refer to the schematic, just look at the photos of the assembled phase. Check out Arlos other projects, he's been turning it up to 11 (and blowing it up) since way back.

I'm guessing end state is enough reliable controller to get the Leaf drivetrain to saturation for 10 seconds?

 

[Arlo1]

Am I reading the schematic correctly? 16 mosfets per phase? Yes...you will be able to power a D9 bulldozer with this controller!

EDIT:

I have a design I am working on for a 96 fet controller. That's down the road. You did not see that. I will change the SCH it is more or less the right one but I have changed it to just show 1/2 the phase because that's what each board is.

I made a mistake of linking the wrong SCH

This is a 48 mosfet controller for now (8 parallel)

 

[ElectricGod]

Arlo...cool stuff! I wish I was as smart as you.

 

[Arlo1]

Arlo...cool stuff! I wish I was as smart as you.

Dedication
Hard work
And a lot of time on the computer at least 50% on this forum is how I got here.
So if you want its totally possible for you as well.

 

[ElectricGod]

Arlo...cool stuff! I wish I was as smart as you.

Dedication
Hard work
And a lot of time on the computer at least 50% on this forum is how I got here.
So if you want its totally possible for you as well.

I went to school for an EE, but never used it for a job. I'm pretty rusty now, but yes...little by little I'm "relearning" what is useful to me for EV projects.

 

[oatnet]

Really cool, Arlo!

-JD

 

[ElectricGod]

Will you mass produce them? Any idea how much per controller?

 

[Arlo1]

Will you mass produce them? Any idea how much per controller?

Maybe.... And something between $10 and $100,000 its really hard to guess how much power you want. I have some plans for this but its big power not intended for ebikes. Part of the goal is to get the power density as high as possible and keep the system voltage as low as possible.

 

[InterestedinEVs]

Can't wait to see this one playing out! Gimme a hollar when you're ready!

 

[okashira]

Very cool. Great work. So you are going to use one of sbe's ring caps? I wonder if they can go lower voltage for better esr, smaller size, cost. All their standard stuff is HV

Is there a way to also put the gate driver in the center also? This way you can make all the gate drive legs/traces exact same inductance and impedance as well.

 

[toolbag]

Those SBE caps are indeed sweet. They even come with a 25mm thru hole if you want to get fancy putting one around an output shaft.

One issue there is that you still need some bulk capacitance in addition to the awesome film cap. I'm using one if their 225uF 1200V parts on a very high frequency SiC inverter, and we saw some scary oscillations on the DC bus on the order of 50% of the bus voltage (this was at ~200V, 100A DC). We'll get away with just 225uF on the vehicle with much shorter DC bus leads than on the test setup, but for a normal mosfet based inverter I'd go with a thousand extra uF at least.

 

[Arlo1]

Start reading the posts from HighHopes on this thread to properly calculate capacitance https://endless-sphere.com/forums/viewtopic.php?f=30&t=31804&hilit=Calculating+cap&start=25


I got mine. I actually went for the bigger one for now as the one I wanted was out of stock for 2 months. But they are working with me to custom design their cap for this controller.

 

[Arlo1]

One more good day or evening and I should be ready to put some power to it.

The plan is to tuck the current sensors under the cap. And in the future the guys at the Cap store will custom make me caps with a number of posts divisible by 3 because as this one is I need to cut 1 of the 8 inner tabs off to make clearance. But don't work it wont mess with geometry.

This cap is overkill for this but they were out of stock for a couple months while they made more of the smaller ones.
I will likely get something custom done in the future to get the size down.

But first I prove the whole thing then I make it better.

Oh to dream of 1 day getting or building a CNC....

 

[SjwNz]

CNC plasma cutter

 

[toolbag]

Have you looked at the loop inductance from that standoff layout? I started with something like that, but then I saw that the loop inductance from my loop (which was much smaller in area than what I'm seeing here) was an order of magnitude more than the ESL of the CAP .