badinfluence

Somethings are noy yet set in stone. My goal is to be under what the BRZ Electric kit is supposed to be. (2000 to 2700)


Right now the most expensive parts are

Custom M90 Supercharger (Got this for a deal)
Lipo Batteries (300 to 500)
Supercaps (58F)x6 (500)
Motor Controller (200)
Arduino controller with Bluetooth(50)
Tuning? (Not sure yet)
HHO or WMI (200 max)
PWM Charging for Battery to Supercap (50) [Instead of a diode]
Shunt/Amp meter setup (20)
Welding/Fabrication for the flanges *unknown* Might just by my own welder for cheaper.
Power Cables 2x Rockford 0/1 AWG+ (100)
Power Cables 4x Rockford 4 AWG+ (100)

The idea is I want to do it first, and then give an idea of how it could be done better. So far I have

>Don't go with a massive motor like I did.
>Use 72-90 volt motors (That is what I used)
>Stay with root or screw type superchargers just get a smaller version than the M90 The centrifugal (turbo kind) are a complete rip off for 2 grand. Get a K04 from a Mazdaspeed 3 used and make the motor assembly yourself with a Castle RC motor and Andymark transmission with custom gears if you want to go centrifugal. Paying a grand for someone elses' isn't worth it.
>Not recommended on the Series II because of the tuning difficulties currently.

I hope to keep this going so others can eventually have an open source solution. I also plan on sharing my "boost" code and schematics once I get it working and improved.

I have a collapsed lung right now and have for some time so that ate up some serious cash and time.

AAaF

Have you seen this thread?
https://www.rx8club.com/series-i-maj...ussion-256515/

This is a cetrifugal compressor, but when you are independent of engine RPM, this should not be a problem. I know these are supposed to be efficient, but I do not know how they compare to yours, do you?

EDIT: I can see that you have linked to the product earlier in this thread. Nevertheless, the high RPM motors together with a centrifugal SC looks like a attractive option. Have no clue what these motors cost though, neither what voltage they run at. And I guess you would need a more complicated AC drive to run them, but I'm not sure.

AAaF

And good recovery regarding your collapsed lung. Hope it heals as fast as possible.

badinfluence

This setup is a prime example of great engineering and providing a good product. However, On the high flowing RX8, I think it would just be more of a pain to keep it up and running just to not restrict airflow. I have had more chances to think since I talked about it last, and it would be kind of an annoyance to have all the time. It would be VERY easy to just add a bypass and then it would be a much better product on the RX8.

AC motors are more difficult to control, but they are AMAZING when controlled properly. You can even regain lost power with them and regenerate energy. You can with a DC motor too, but the circuit is VERY complicated, large, and unreliable.

AC motors essentially take a very little amount of power to regenerate, but that is a given since they operate like a transistor in the circuit when regenerating. It is very complicated, but the best way to research the regenerative drive is to look at Industrial Machine tools designed to run on 3Phase or converted to run on single phase. It is all the same principle.

Sadly, the company I am working for is doing quite poorly and may collapse at any moment. I was able to acquire 8 12 7ah batterys for testing, but I need to built a proper charger, get the motor controller, and then get the Supercaps so I don't fry the batteries with way to much current draw. Prob about 500 to 1000 dollars.

Lately I have been working on something really cool with electronics. A laser CNC machine built from scratch. I am building it to solve a cutting problem I am having with plastic, but everything I have learned will work towards cutting and milling parts for the bypass and cooling units on the ESC.

I will jump in on the forum when I have a bit more time, but I don't want to create an argument in someone else's thread so I need to think of my responses and make them factual and not argumentative. Everyone sounds like a dick on the internet, wither they mean to or not sadly.

Legot

iTrader: (1)

That doesn't use an ac motor.

badinfluence

Do you have some data on what it is?

It looks like it has 3 wires going to the motor, but has a ESC that could just be a DC PWM controller. I just haven't seen one of the Phantoms in person to tell.

There is debate on how ESCs work and if all RC motors are AC or DC depending if they have a commutator. Especially with the brush-less models. The real answer is in the ESC.

A lot of the homebrew versions of these kits are a 1/8 Scale RC motor, ESC, and a Lipo setup. (Brushless motor) Mamba is one of the ones I have seen used.

The issue is getting the bitch to turn fast enough to spin that wheel. I personally think a CVT between the motor and the turbine wheel would fix it's limits, but that wouldn't be fun to figure out. (Well it would, just don't have time.)

lastphaseofthis

ac induction motors are in the tesla, but DC brushless motors are in hobby RC. they funtion very simular. they're both fed by 3phase. however the outside stationary coil is powered by A/c on the tesla, that powers the inside coil thru INDUCTION. The dc motor uses a neodblah blah magnet that while super powerful on the rc scale, would be inadequate compared to the current you can induce on an A/c motor.

so the deisgn is almost the same, but the behavior is far different. Ever heard of super conductors?
i would like to see a super conductor a/c induction motor built. one the size of your alternator would theoretically be around 2000+hp. kinda pointless to run a turbine at all then.( however superconductors must be kept around the temp of liquid nitrogen. bah ha.) waiting on room temperature super conductors and batteries in the megawatts instead of kw. then we can have flying cars.

the only reason i go so far into science is because that what this is, a science experiment. for a practical upgrade to make more power on an rx8, this is highly non-cost effective . at over $2700, you only have 20-30 hp to show? whose got buy that?
now that we can forget about meeting product qualifications, we can focus on making it work with disregard to cost, in the name of science. also i've seen electric superchargers on drag cars with over 1000 hp. you tube them. some of them don't even take advantage of brushless motors!

badinfluence

I think you merged two things together. My electric motor uses about 20HP, as in HPe+ to make 0-21HPm to drive the M90. No power gain has been recorded because it isn't finished. I am trying to overcome some power control issues that most setups have with golf carts and DC motor controllers. So far I have about 1200 or so in it not 2700. I don't intend to blow money, and this isn't a production. The BRZ turbine/RC motor is in production, mine will always be a one off..... It was never designed to be produced, just a science experiment like you said. I have learned a TON about transistor controls, mosfets, and drive strategies so as far as investments go, it has paid for itself. I am using free batteries I got recycled, a free charger, and I charger I made for 10 bucks. So really if you tried to duplicate it, it isn't worth the money. For my investment, it still will continue to pay off.

My project is mostly to generate more torque, not HP. I think the rX8 is almost perfect for the HP range it offers, just the torque curve could be a bit lower on the range.

If you are describing the BRZ supercharger I absolutely think it is overpriced and not a good deal. Weight vs Turbo, with some tweaking and supply as demand it could be worth it. Right now it is just to much cheddar considering the awesome mod scene that car has. I see a widely supported turbo coming soon if one doesn't already exist. For weight with LIPO batteries, it could be worth it for around 1500....

Legot

iTrader: (1)

You haven't started yet?

badinfluence

Little update on the project. I have been learning some methods of isolated control technologies for Motor drives. What I have is an IGBT, 2 to 4 to be exact, with 2 gates and 3 poles that will allow for some regeneration and braking of the motor and a simple, yet clever bypass valve that will require no blow off. Coming up with the power to just test the supercharger has been annoying since I need to be able to control a total of 13kw and at normal 12 volts, that is way to much for most gates to take. So I have been working on coming up with 48 volts to test with, but I don't want to whack my batteries to the point I wipe them out, but I don't want a useless test since I have to find the sweet spot in the power grid and attempt to generation some boost via compressor. (closed instead of open like an air tank)

I am thinking 48 volts is enough, but my options are virgin lead batteries which are heavy as ****, or LiPos with a clever limiting system and super capacitors with a pretty indepth system of balancing them so they don't overcharge getting whacked with 2kw.

The thought of a 13REW swap has always been fresh on my mind here, but the whole project was a success,even if I pack it in because it got me into electronics in a big way. This supercharger is going on something, but not until the power problem is solved properly.

I got a little more specific on the microcontroller, using a BLE arduino, but I am wondering if a Spark Photon or Core (if my damn photon ever ships....****), is a better option. I have several options, but I want the best tune-able environment and I like the idea of a cloud IDE for the tuning computer. It will have some modules and extensions like it's own kit, so it could be used as an ECU helper for anyone looking at doing something like this project.

Legot

iTrader: (1)

These are not difficult problems to solve.

How have you not started yet?

badinfluence

Why do you think I have not started?



If they aren't difficult, what are your ideas?

Legot

iTrader: (1)

Forgo all of the regen stuff because you won't be able to get it to work effectively. Turn the pump on and off with relays and an H bridge to get braking (you shouldn't need braking though). I'd go MSP430 rather than Atmega328, but that's just a personal preference.

Done.

badinfluence

Ok so your the second person to solve those problems. For the record it was just a post update to start with, not a problem. the 3 pole igbts make an h bridge, and have an extra pole good for 210 amps. So I solved that 2 months ago.

The regen isn't something I am trying to do, just a side effect of the 3rd pole I added to remove the excess voltage without blowing the drive and the motor.

(H bridges always have the possibility of delay and drive burn out because one leg doesn't switch)

I have the atmegas and the spark. The spark isn't an 328, I think it is 90mhz.

Legot

iTrader: (1)

No, use the bridge to drive the relay's. You're not trying to do anything anywhere near the speeds that requires you to use transistors.

I would use they Atmega's, they tend to be pretty robust (and you don't need the kind of power the Spark has).

badinfluence

Can you explain this a bit more? I am not sure what you mean by relays.

I am using the IGBTs because nothing I have else can handle 84 volts 200 amps.

total electrical power can touch 13KW, so even at 10% DC it is quite a bit of holding back. That would blow most transistors, fets, and relays.

The Spark is more for the Cloud IDE and ability to adjust without having an ISP or anything that would have to be removed or tapped into. I also like the fact that I can use a LiPo to directly power it. Not that it is going to be using much power anyway with everything being isolated.....

I want to be able to control the amount of Power the supercharger puts out, and that is why I went big. It can handle 20psi, but I only intend on putting out around 5psi, so like 30% power is needed to do so. I can't find much else that can control a perm magnet with DC.

So far I have just about everything to build this monster's electrical system, except batteries, supercapacitors, and/or a way to come up with 13KW.

http://www.farnell.com/datasheets/101164.pdf
That the is the Half Bridge I am using. (IGBT) Right now I have 4 of them.

badinfluence

Moved a bit more forward on this one. I have found that the Supercaps I was originally looking at are at a insane pricepoint (HCAP 3500Farads 2.7v). So I opt'ed to go with the 350 Farad at 2.5 volts instead. I got them at $7.50 each instead of the Digikey price of 9-11 dollars. These specifically are 2.5v not the 2.7v with the round terminations at the top. The ones I have are 2 tabs on the top and bottom just like a D battery. This makes it so much easier to solder/weld them vs the round ports + & - being centimeters from each other and hard to fabricate a tab between.

I have 54 total, so I am trying to figure out the best way to convert the power. Here is something I have in my favor with a brushed motor.
-I CAN regenerate with very minimal effort since the brushes interrupt when they rotate.
-The motor has a ceiling of 90 volts, but runs anywhere between 90 and 5 volts. The problem is at to low of a voltage resistances and current goes up and could nuke the brush holders and brushes.
-While the motor is disconnected for a split second, I could use the momentary disconnect condition to charge the caps, then as it reconnects pull the charger, making a forever 50/50 duty cycle based on the RPMs. I need to figure out how to very the frequency instead of the Duty cycle, but I think it holds some promise.
-Also in addition to this, if I know the RPM & When the brushes are touching I can record it, put it to a serial connection/gauge/lcd, & use it to make a accurate table. It also would allow me to know when to ramp up or down the charger based on the RPM & charging requirements of the DC to DC converters.


Right now I am thinking for the initial testing, I am going to use some 10/12/15A DROK "600W" DC to DC boost converters and a target voltage of between 48 and 55 volts. (22 CAPs x 2 for captivity. So 30 Farads at 55 volts.) The idea being a fully charged 12 volt battery is 12.8ish, so 4 fully charged 12v batteries in series gets me very close to my max voltage/max captivity for the caps.

I also have a TripLite UPS that runs at 48 volts that had the batteries go bad. So I have the inverter & battery charger since it is a double conversion unit.

I also bought a Canbus triple to help give me some new ideas for tuning and ability to vary everything I am planning.

badinfluence

Ok I made this thread way to soon, and I will fess up to that. However, the entire project has reassured me that I need to move into Integrated Electronics, and given me some confidence in playing with AC and DC voltages.


Here is the update.

The setup will now be controlled by several watchdogs made up of ATTiny85s or ATTiny84s that will send a warning bit, and a FUCKFUCKFUCK!!!! bit to the main processor.
Here are the parameters being fed into the Main processor and the watchdogs respectively
MAP (Analog and I2C)
Airflow Rate (Analog and I2C if I can find an I2C version)
Motor RPM
Motor Current A
Motor Current B
Motor Status
PWM Rate (Feedback)
PWM Frequency (Feedback)
Current Forward and Backward (+amps/watts or -Amps/watts)
DC Volts A
DC Volts B
DC Volts C
Resistance A
Resistance B
Bus A Undervolt
Bus B Undervolt
Bus C Active
Regen

So 3 Buses for the system
Bus A
Capacitor Bus 75V balanced individually through passive recharging and high frequency PWM.

Bus B

Battery Bus from Boost Converters. A combination of Batteries in separate banks that boost power to the capacitors before they charge the motor to spin


Bus C

Motor Bus with regeneration via P Channel, N channel, and SSR setups.


Motor Frequency will vary between 750hz and 7500hz and will more than likely run around 1500hz when spinning after the soft starting. (5KHZ, 1KHZ, 750hz, 500hz, 250hz intervals like a binary setup)

Capacitors will charge at 48khz at a Duty Cycle that is the opposite state of the motor frequency so the motor cannot draw massive currents and blow up the boost converters or batteries. (I think the only smart way to do this is to alternate a blocking diode on a power mosfet protected from reverse protection and alternate potential current from the boost converters.

badinfluence

Worked a little on the brains of the show a bit....



Pictures: Pismo Labs Hogh Power Wireless 802.11an/ng Acess point
Netgear GS108T-200NAS smart switch
Raspberry Pi 2 without extra junk on board (for carputer and communication with CANBUS, ESC Photon, watchdogs, ATMEGA32U4, and internets
RPI Model B running PiMusicBox to replace 3.5mm audio jack with AirPlay, Pandora, Spotify, and Internet Radio (headless)
Synchronis Buck/Boost converter for power supply 2
Particle Photon for Control via I2C and Optocoupler board not shown.
USB Hub with onboard 802.11n and BT4.0
Watchdog board with Zeners for ensuring power spikes and voltage changes do not effect any PCM changes or spikes.


So the stuff not in the photo is
200 Farad Combination 15.0Volt Supercap with specialized charging and balancing circuits I developed for the ability to remove the battery and keep the PCM voltage constant and mostly happier than stock or how my car is stock. (bulk dealer screwed something up with the regulation)
Enclosure for the Router Combo(which is removable, portable, and can be powered via most AC adapters for anything 5 to 30 volts, solar powered, powered by the car socket, or Lithium batteries.
Cradle point Router for VPN access to the "homebase" aka "Walter" or "CLOUD V.P.S.." Whichever has the lowest latency
LCD for HDMI control Ina H.U.D. Configuration on the Windshield
Custom Bluetooth Controller for operating everything without a Keyboard/Mouse/input device. (So it works while your driving and does not take your eyes off the road)
Intel portable Windows PC for MazdaEdit and Tactix adapter
CANBUS Triple for logging and PID interface

everything on this post EXSISTs, but is not all just for this project. The router is basically a portable network in a box complete with VPN, VLANs, WiFi, and brainpower from the RPI2.
Same thing with the Bluetooth controller. It was designed for a kick *** retro gaming controller for SNES and PC gaming and functions as a human interface device, a logic device, and a physical security key for locking computers and passwords.


So now you know how insane I am, but I think this was known before.

On another note, all car progress is halted till I replace the Midpipe and get the can running correctly. (Thought it was the coils and plugs, but the cat is intermediately blocked, but only just for now) This is stopping me from getting base calibrations for voltages on the cars sensors and PID behavior.




Brains! Happy Halloween and Now I am ready to go an hour back in time

badinfluence

OK here is an update!


In September of 2015, I got laid off so that halted production of anything that costs money. I have been working with some separate ideas for how to control the motor drive and I believe I have identified the PIDs/Data Points I would create.

I am not hopeful that anyone here might be able to help, (only because I dont think many watch the thread), but I am on a mission for a precision Motor drive controller for a Brushed DC motor between 6KW and 24KW. Switching would be done by a IGBT in a Half Bridge or Bridges, so the controller would just be a controller.... Not a direct driver. Everything will be isolated from the controller including drive signals and feedback.

(Hopefully someone has an idea or a direction for this one, and can help me. I would love any input I can get on the controller. I do want to build from scratch or keep it workable from an existing unit. Controllers like what I am looking at get expensive real fast.....)

Unknown Controller Chip (T.I. Prefered)
-1A-250A/300A/400A/500A Max. (20KW Max, but no plans to run full tilt)
-Closed Loop and Open Loop Operation if Possible
-Current Control & Monitoring if possible
-RPM & Timing Prefered with Variable Frequency Driver
-Ability to drive motor between 100V to 0V
-Chip Voltage Prefered either 1-3.3V, 3.3V, 5.0V, 12.0V(10-16v)
-I2C, UART, SPI, or NON-SMBUS protocol preferred.

So I have a lifetime of knowledge about electronics, Computers, VFD, and ICs, but I have never tried to build a drive quite like this. Just because I am saying I have that, doesn't mean I am trying to show off or brag, just letting everyone know the scale of what I am looking for, and that I have not been able to do so, so far.

Right now the idea is to control with a more powerful controller, Possibly the New Ardunio Zero's chip, ESP8266, Spark/Particle Photon, Spark/Particle Electron, or a combination of controllers including the Raspberry Pi Zero and Raspberry Pi 2. (I got the Zero the first week it was on Adafruit! Love the price points for this stuff now! Buy a burger or buy a SOC Microcontroller!)

This isn't the Exact IGBT I am using, but it is close.
BSM200GB120DN2 Infineon Technologies | Mouser

Somewhere Around 200A-250A is the model I have. (I have 4-8 of them)

Looking at Lithium Iron Phosphate and Sodium-Ion batteries to provide the umph the setup will need. Possibly upgrading the alternator to a more beastly unit following the power curve, ignoring the idle(the opposite of car audio builds)

So the only restriction with the Motor Controller is my fabrication skills and tools. I do not have the ability to cut my own PCBs, so a DIP, SOIC, or TSSOP is preferred. The final product will be produced and cut into a PCB process, and I may give them away or sell them if I have to do a minimum order. Depends on the Company or if I get my Laser Cutter up and running I am building from scratch.

Power generation may change a little bit as I am moving and may end up with the ability to finally fabricate an intake for the system, removing a serious roadblock I have right now. At this point, Houston has taken it toll on my baby.... 2 people have fucked it's side up, and its value has gone down I think. The good news is, that it is going to just be mine at some point so it isn't all bad! When I get more paid off a Turbo, rebuild, and complete fab is planned!

For now I am thinking of making bolt ons for everything without touching the frame or welding anything. I have been watching Mad Max and playing the game a bit to much, so hopefully that fever goes down.

So I have to pack up my lab to move out of Texas, so hopefully that doesn't kill the project off..... Sad as it is to do, I hate packing up all my progress and hoping it stays where I can see it and start it back up again. Moving always sucks.....

badinfluence

Ok, found some traction on the project and a hell of a big problem fixed!


I am looking for traces of my idea on the internet and checking the Patton office to make sure I am not infringing, but if it works I am filing for a patent!

The basic idea is a big *** Boost Converter that can Boost,Buck, Flyback, Store, Forward, Drain, Charge, and redirect to multiple devices, except to the tune of 100 amps+. On the market I have seen maybe 3 of those together, but not in one IC package.

So basically I can't find a good motordrive controller, so I am making one big unit that will run the entire show off the car's stock voltage without nuking anything. It is all about the "control" of the circuit and monitoring how big of an impact to the VCS your making.

A good next step is to study the PCM's Duty Cycle commands to the Phase Alternator and see if I can interrupt them, add to them, and redirect the current source away from the PCM to prevent damage or massive EMI!

Right now I have 56 Supercaps, but I am going to need more than that for testing and scaling this around. I am still sticking with the 2.5v 350F (max 3v) because the 2.75V 350s have the same 3.0 v limit and have those stupid fork designs that are just annoying to work with.







So as it stands here is the tested numbers for the car running on caps

Hot Weather: 50 to 200 Farads depending on how long you leave it. Assume 3A parasitic draw and add 4x 7W Solar Panels and a 15V boost converter capped at 1A max. (With or without battery)

Cold Weather: Assume you will need 200 Farads here and they need to be in the engine bay! Pulling from 10ft of 0awg wire works, but it makes the PCM mad. Really the PCM needs its own power source for consistency even with the stock design of everything. (Constant Current, Constant Voltage, + Decoupling cap + Filters)


All Weather (The answer to the need a battery problem)
So here is the hybrid design finally I came up with
4S LIPO 16.8VDC 50C 5A (recommend 3-15A at least 50C)
IGBT Half Bridge Module 50-200A
Buck/Boost Isolated 5v Module
5v USB Microcontroller such as ATMEGA32U4
10-50A Diode
58 Farad Supercap
PC817 Optoisolators
Starter Solenoid

Basically this turns into a hybrid charger when the car detects the keyfob (or if you dont have one, the ignition switch or brake lights)
It charges the supercap to 15V (DO NOT EXCEED 15V under ignition for more than 5 seconds.... [All parts of VCS are rated for 16V, except power parts rated for 20-24V pulse only, PCM can't take 15v for very long])
Once it is ready the PC817 tells the Micro-controller to light an LED and slow the PWM charging.
Once the vehicle enters changing mode, the IGBT opens so the doesn't pull from the LIPO battery or destroy anything.

So basically it is a buck converter with some logic with Isolation. Sounds complicated, but really it is about 8 parts and can be built on a Saturday. I want to make sure it is ready and I dont get sued before I post the schematic of it.......

The circuit includes a variable charger for the LIPO battery, but at default it is set to a 1A charge max with 250MA standby. THe circuit protects from Running, Charging, and Standby so none can happen at the same time and blow **** up.


Side note, it is better to disconnect the battery for colder weather starts. It is a draw off the cap and has enough resistance to take voltage from the cap to the Starter Motor for about 1 seconds till the caps are the lower voltage. With it isolated, the car starts faster and the PCM doesn't reset.

badinfluence

This has not been forgotten. I recently got the ability to use some Specialized lab power supplies with precision voltage and current limiting, so I can see how the supercharger behaves at very low power ranges. Here is the testing schedule I have come up with.

25V

5A
10A
15A
20A

32.5V

5A
10A
15A
20A

40V (Max the supply with do)

5A
10A
15A
20A

20A is as high as it will do sadly.... But that is about 800 watts, so that is realistic to draw right off the alternator (1440W), so it will give me a good idea of how much voltage I have to store to get that above 1440W for X amount of time. (1440 is Watt Hour, and I need roughly 5000 Watt Seconds for at the max 5 seconds)

I also have a running MIG setup for assembling the flanges I have had for 2 years, so after getting the AEM intake pipe that holds the MAF housing, and building the supercharger cage and bypass, I will be able to mount the bastard, even if it is disconnected and not working.


Side note: I ran my battery relocation through the driver side, and it does not fit the 4AWG cable I have, so I have slightly exposed wire. It will NOT fit in the kickpanels, so for anyone doing that mod I recommend having 4 wires of 8AWG which is much cheaper and more flexible.

EarlQHan

I commend you for trying to do this but your project seems to be unfolding at a much slower pace than the available technology. ESC's are now in production vehicles, as will 48V. In my opinion, the rotary is an excellent candidate for a mild-hybrid ESC, so best of luck with your endeavors.

9krpmrx8

iTrader: (46)

What production vehicles has an electric supercharger?

EarlQHan

Audi SQ7. But a lot of OEMs have tested them. The benefit is power on demand, with little-to-no parasitic loss and no backpressure, allowing for downsized/downsped engines. The downside is the cost and associated complexity