badinfluence

Umm, can you specify what you meant by this? I am not sure if you are agreeing with me or giving me advice? I ended up with the same bath, SLA in the trunk, dumping the LIFEPO4/LI-ion in a conventional sense.

The Plan is to put a Lithium battery on a slow PWm duty cycle to charge the rear cap/battery only, not charge the lithium. The lithium would charge off a Ignition wire and isolate when Ignition is not on, battery voltage is high enough, and when charging the lithium or charging the capacitor. This will NOT be a pre-built Lithium, since I have seen pictures of them melting down on Amazon. If the PCB was correctly setup, it should not have been able to melt, it should have self isolated. This is also IF a PCB/BMS even existed to start with.

Example:
Car voltage = 14.1, IGN=On
-Isolate Lithium upkeep charger, charge lithium from IGN source.

Car Voltage=12.7, IGN=Off
-Isolate Lithium upkeep charger, isolate lithium from IGN source

Car Voltage=11.9, IGN=Off, Lithium Battery = 12.7v
-Enable Lithium Upkeep charger=DC30%, Isolate Lithium from IGN source

Car Voltage=10.7, IGN=Off, Lithium Battery = 12.7v
-Enable Lithium upkeep charger=DC85%(MAX), Isolate Lithium from IGN source

This is at least 2 weeks away and I am stuck between using a 4S LIPO or a 3S Lipo. Current WILL be limited at 1C rate. So longer life and the possibility of using Panasonic 18650s instead, and the charger controller prevents overdischarge, and overcharging. I don't want to try and make my own lithium controller since they seem to be well engineered already and I already overengineered a simple thing here.

9krpmrx8

iTrader: (46)

Thanks, I like to keep it clean. Yes I am turbo and it adds heat but really only in the vicinity of the turbo and I have a ton of heat shielding on the firewall, manifold, and a turbo blanket.

The marine fuse box feeds power to my gauge modules, boost controller solenoid, and the Bosch water pump and pump fail safe for the turbo cooling circuit.

Harlan

iTrader: (3)

Just some friendly advice.

I did have another idea which I may use eventually. Using a couple 18volt tool batteries with car chargers and buck converters you can have an easy to build setup with existing charge control and battery management. It also has the added benefit of giving you cordless tools in your tool box.

badinfluence

Do you use cordless tools that much in your 8?

Wait....this isn't like a website I found one time right????? Those are supposed to be for screwing screws not .....

Harlan

iTrader: (3)

Once you change a tire with a cordless impact driver you will never want to go back again.

hoss -05

iTrader: (4)

He is running a low mount turbo, it heats things up a bit but not as bad as a top mount.

His marine fusebox is used for auxiliary electrical systems such as his coolant pump, sensors and boost control equipment. Marine stuff is great~! I have used blue sea systems equipment in the harshest environment I can think of and it never failed me.

badinfluence

Ok off topic, but what model do you have? I bought a cig lighter one and a big *** breaker bar recently. On the cordless ones, I always had the stupid adapater snap on me. (The part that takes the part that screws in the driver to the 1/4/3/8/1.2 head.)

badinfluence

Yes, Blue Sea kicks ***. I saw it and recognized it immediately. The only thing I think is a bit overpriced of theirs is the non-surface mount breakers considering most are just busman rebranded. (Some are waterproof though)

Harlan

iTrader: (3)

I'm using a Ryobi one from home depot. It's cheap but it gets the job done, and takes of lug nuts unless they've been over torqued.

badinfluence

Ok made some major strides here and found a few snags.

The biggest ****** making it hard to swap to a cap is the Radiator fans. When they run after the car is turned off they rape that battery pretty damn good. Disabling them would be a foolish move. If anything I would want them to run more often.

So that bastard aside, here is what I got

I found that the 16.2v caps are not as efficient as the 15v caps I made myself. The 16.2v module has active balancing which is great, but the caps are 16.2v max/rms @ 2.7-2.75v per cap, whereas my original caps are 2.5v. The way these Double Layer caps work is you need to get closer to the max voltage to get the full charge, otherwise you have to scale the Farads down. IE 16.2/14.5v = % (58 F * % = actual storage) So I am loosing at least 2 volts, if not 4 and reducing the ability to store more of a punch to start the car and keep those Ignition coils primed.

So my plan is to keep my boat anchor in the trunk and leave the 16.2v be until I redesign my 15v setup with better balancing. I am going to go with 12 caps so ~100F captivity, and use a 4S Lithium Ion BMS setup and a buck+constant current charger to keep the voltage at 14.5. (this does not in any way fix or solve the radiator fan problem). It would only be operational when the battery is above 25% charge, the Cap is below 14.0 volts, and the Ignition is turned off. (ACC would run for 5 minutes then something would have to happen)

The BMS setup's captivity is 75 watts, so it will not do miracles, just enough to get the benefits of a faster start and more accurate first firing at startup.


So far with the "boat anchor" in the trunk and the cap at the front things got so much better with starts, and especially hot starts. I believe just relocating the coils and battery alone helped (heat), but adding the cap is a cherry on top because it holds the higher voltage and when it drops below the battery voltage, the battery rushes current to it and keeps the resistance low between the cap and starter.

Legot

iTrader: (1)

You could just cut power to the fans when the ignition is off. They're not at all necessary when it's not running.

Harlan

iTrader: (3)

Even if you prevent the fans from running with the ignition switch off the Cap bank still doesn't have enough power for more than a couple seconds of cranking. Add to that the high parasitic draw which can make a regular battery go dead in about a month, and you have an option that will only last a couple of days without an aux battery or a charger.

Capacitor only is not a realistic option.

badinfluence

Have you tried it?

.09ohm resistance vs 12ohm for a linear output makes a battery look like a joke. My car starts on the 2nd crank with the cap without the battery and very very similar, but slower with the battery backing it up. (Cap still has the lower resistance, but it has to charge the battery with it's lower resistance)

At 77% Max charge@45Farads at 12.6. 84% Maxcharge 52 Farads it cranked that much faster and settled at 13.2 volts with passive balancing with the 2.5v instead of the 2.75.

Batteries have increased resistance below the 12.0 mark vs having the same resistance (like a motor wants), between 14 to 6 volts. The motor will continue to crank till the PCM or relays cannot stay on. (12-5v PCM, 12v relays@~<8v) You have 15 volts available instead of 2 volts at different resistances.


The radiator fan runs because Mazda thinks it is needs to run. Some things OEMs are just dumb, but if Mazda needs it to run, then I will make it work somehow. I have not gotten a chance to time and measure the drop of the radiator fan, but from the logic it looks like it only runs IF the voltage is above a threshold. I never said the Cap could not do it, I said they rape the battery. That is why I am trying to prevent.

Harlan

iTrader: (3)

Yes I have tried it. Had a silicone coupling blow off and it only gave me a couple of attempts before it was less than 10 volts and wouldn't crank. That was with a bigger cap bank too. It runs down below 10 volts in three days without a charger. I now have a 35ah deep cycle in the trunk.

badinfluence

Series 1 or 2? Advanced Keyless Entry or Regular key?

What is the silicon coupling?

What size bank? I was going to go with 120 or 180@15v instead of 58. I tested with 58@16.2 which put it out of reach for dipping down. Also did you have a balancing circuit, with what color LEDs?

I have been studying the engine draw and I think I have a solution for the draining below 10v. My starter cranks at 8, so I wonder if that is just a change in series or something else.

The solution is parallel charging with a high frequency pulse. This allows the caps to charge correctly rather than linearly in series. They also do not like to follow rules for C+C+C+C=V, so they don't charge correctly because of the low resistance of all the units. (IE, the last 3 caps act more like a wire than a storage device which is a problem for the way they charge)

I have a 100W 16.8 volt Lithium battery with BMS circuit I am using to run the show. I am running 3 isolated DC-DC converters at 5v each that run from 30v to 8v (the lowest I have tested them). They are isolated so they can be used in parallel without shorting the car out by connecting a +5 to a -5/VSS of another converter.

The science is more important to me than the results, so even if it isn't practical I want to find a way to make it work well since I spent the money on it. I think the Hybrid solution of both seems to work best for an adverage joe, maybe a solar trickle charger and a higher voltage AGM/SLA 12v battery (13ish) so it doesn't pull the cap down.

Harlan

iTrader: (3)

Series 1 with a turbo, the coupling was on the charge pipe. It was a 500F cap bank with 6 2.7V 3000F capacitors, the balancing circuitry came with the caps. Below about 10v the starter solenoid would not en-gauge. Below around 8volts the key-less entry wouldn't work either. This is just what I remember offhand, I didn't take notes.

Parallel charging is a neat idea, but over complicated. You'd have to run a huge diode on the output or isolate the alternator feed. I went with the 2.7 volt caps because they could handle the higher voltages and give a little margin. 15Volts is too near operating voltage for my taste.

A solar charger is another good option, but I highly recommend having a small aux battery. You don't realize how often you sit in the car with the engine off and the radio on, or the AC still blowing, or the headlights on, until you have a ultra capacitor instead of a battery.

Edit: Considering the 58F capacitor has only about 5500Wattseconds of power I'm amazed it can last overnight even if the S2 standby draw is much lower. That's the equivalent of a .1AH battery.

badinfluence

I had someone who work with supercaps give me the "real"numbers from Maxwell.
The 3kFs top out at about 2.9 and blowup at 3.0. the 350s are more flexible to voltage but less to current. So at 2.7(2.5) you better back off the current draw or pop.


The S2 with AKE if the transmitter is >10 feet away is 60ma at max according to Mazda. I got this from one of the files I got from NormalException, and it was something Mazda did to reduce the battery and starting issues of the S1. The Starter is also more tolerant to lower voltages, but I am not certain of the solenoid.

I am not sure how you got 5500watt/sec. I don't think that can be considered accurate since the caps don't behave like batteries. Part of the problem with the numbers I am throwing down is, the have a linear pattern to them unlike a battery. IE
58 Farads is only 58 Farads at 16.2 volts. at 8 volts it is about (8/16.2)*58, but not really. So you would have to take that, converter to Joules or Colcumbs, then devide by voltage to get amps, then *60*60 to get amps/sec. Or the VoltAmpures way. Not trying to be a dick, but I don't think 5500watts/sec can be used since watts are Amps*volts and volts change rapidly with the caps discharging so it would be a mess to try and figure out the math on paper without a linear curve/graph and then some of the formulas on the web are inaccurate. Add all the passive and active balancing **** onto it, and it is a stupid long annoying constant math problem just to figure out the discharge rate. I said **** that after the 2nd hour and just went and tried it. The general rule someone gave me was Actual Farads=Amps@Voltage. So 50F@14.5volts = 50 Amps per hour. According to every datalog I have since I added the caps, my alternator voltages (request and actual), have never hit >14.8 with or without the battery in the back. I am somewhat scared to just run 1 58F cap with heavy driving and city driving, which I why I bought two of those on ebay. According to the vendor of the BCAP kit (engineering firm in Cal), their sizer said I needed 2 to run properly.

There is also now 8 generations of Maxwell 3K/3.5K Farads which makes it a bitch to try and figure out. Knowing what I know now, I would order the Brand New 2.9 volt versions and go with 5 instead of 6 and modify the Alternator duty cycle tables for the PCM so they never exceed 14.5. It sounds scary, but those caps are supposed to be +/- 10%, so that is .3 volts average per cap. The 1st and 2nd caps are the ones would be be in trouble more than the 4th and 5th since they get the leftovers from the current rush. I have considered doubling or tripling up on the balancing circuit on the 1,2,3 caps and going with a single on the 4th,5th, and 6th caps. (one of the reasons I am going with 3 58f instead of just 1(18caps vs 6 caps)


Also with Supercaps, if they are not balanced, you will have insane parasitic draw and never actually settle at the voltage. What voltage you do get won't be accurate so that makes it harder. Active Balancing is the answer, and that is what my 5V charging is based on without all the Opamps, mosfets, and such. Lots of ANALogic to those circuits, which I need to get better at. Right now I am using either an ATTINY85 or an 84 depending on the number of pins I am using with an AUX battery. It sounds like you have the built in Maxwell bcap active balancing circuit. Someone steered me away from that because they said it is constantly settling and drains the batteries. I do NOT know that to be the case, so don't take that as gospel, but I do know that passive and active balancing are far less than perfect and most people ignore it completely now after some horror stories on the internet. I have a 16.2v 58F cap kit blow up on me because the switching diodes were to small and all I got from the designer of the circuit was, you did something wrong..... (he claimed to be the designer, but LaserHackers of LaserHacker fourms was the real designer, so I just got told a lie, stay away from eBay kits)

Ways to balance I am referring to are.

-Switching Diode + LED (Most common method, seems to be the most reliable)
-Passive Resistors (Risky because of current draw)
-Active Logic balancing (Maxwell Approved method)


Here is the full planned setup
18x 350BCAP@2.5v
Yoku 5,000MAH 16.8v Lipo battery with BMS (SMBUS)
Boost/Buck Converter (3A)
Boost/Buck converter (Syncronis with constant current)
Buck Converter (Constant Current)
DPDT Relays to switch from IGN to Charge Mode
Particle Photon for Warnings (overvoltage, undervoltage, isolation, issues, debugging, etc)
Passive balancing via small switching diode + LED+Optoisolator(to warn uC) (gated circuit)
Active overcharge protection via isolation mosfet, Burn off Mosfet, and Zener.
uC monitoring via 3x panels of 5v caps instead of 15v as a hole.
Undervoltage protection via Aux battery + Solar Charger.


So the plan is get the prototype built, replace the 16.2v cap and leave the battery in the trunk till I test it out a few times. I can't be ******* around if I get called into work with a dead battery. The amazing thing with this setup is though, if the main battery does go dead, I can turn on my 50A 14.4v Power supply, disconnect the battery in the trunk and that ****** cranks instantly. (48F + 50A)

Harlan

iTrader: (3)

It's not watts/sec or watts per second its a Watt Second, maybe I was unclear, and it is a very effective way of measuring the power in a capacitor in meaningful terms for us but you have to convert from joules to get there. I could have used Kwh instead.

As you said capacitors are not batteries, so the actual number of usable Watt Seconds is lower. A battery has a somewhat fixed voltage defined by chemistry so a 12volt lead acid battery will be most of the way discharged and still be at 11 volts. A capacitor on the other hand drops all the way to zero with useful power all the way down. So if your car stops being able to use the capacitor at 8volts then all the stored energy down there is lost.

This quickly cuts the effective size of a capacitor by around a third from the energy you can't use. So the 5500Watt Second charge is now less than 3600 Watt Seconds. And if you have a starter that draws 1.5kw or 1500 watts then how many seconds can it turn before the capacitor charge is too low to be used by the car...

Sure you can start it up, and if you don't let it sit too long it can restart as well, but it's waiting to bite you when you have the least amount of trouble or let it sit too long.

A HB4 headlight bulb draws 55 watts (at normal voltage), two of them are 110. If you have your cap bank it will run your headlights for less than a minute before you won't be able to crank the car. Try it yourself, and decide if running without a battery is a good idea.

badinfluence

Never seen Watt Second before, but I will look into that. What is the formula for that?

The whole idea of my Aux battery is get keep it at closer to 15. That is the entire idea behind the cap is get another 1.5 volts without the drawback of the load when the car is running. The alternator is 1440watts running at full tilt, so if the starter is 1500watts, then it would take very little time to recharge from 8 volts to 14.5v. 15 volts unbalanced might as well be 10 volts since the difference between each cap will kill you. This is why I targeted 3 smaller banks instead of 1 large bank, which each bank getting 5v .5a isolated from VDD/VSS. it will keep the car linear when it needs to be and when it doesn't need to it, wont do anything. That 24watts/hour of power would keep the cap constantly charge and keep the 60ma drain at bay, but also balance the car and properly charge the caps at the same time.

The narrower charge ceiling(15) greatly pays in my favor vs 16.2, because the density will be higher per volt, therefore give more power than a 16.2 only 3/4 charged.

For the first 100 cycles supercaps have to be broken in properly, otherwise you get the dead short problems with <1v for each cap. (6v) How you charge them is so important that IC manufactures are making PMIC units for them that charge, balance, discharge, and control the overall DC of the N channel. The Correct charging frequency DC, Constant Current, and discharging are critical parts of the cap's behavior and life based on Maxwell's labs and all the Electric Bus, Railway, Semi Trailer, and other electric parts they have done. It has been found that 48/54 350s is better than one 3000/3500.

IamFodi

iTrader: (1)

Isn't a watt defined as 1 joule per second?

So, a watt-second would just be a joule...

Harlan

iTrader: (3)

Damn my HS physics teacher would be disappointed. Yeah I was a capacitor calc and just changed the output. Yes a joule by definition is a watt second.

Yes it does charge quickly, but that doesn't matter. You will have maybe enough energy to sit overnight and start the car, but very little extra margin if any. If it doesn't start first try you will be getting a jump. May as well put a hand crank powered charger in your trunk because you will need the power someday.

badinfluence

I have said I have an auxiliary battery with a balance charger like 3 times man. Also a simple 24watt solar panel setup (4 panels) that also runs a auxiliary fan inside the car to cool it down in the sun. (that I get in a parking garage....)

Realistically I need 6 of those 24 watts most of the time to keep up with 180 farads. The alternator charges the Aux battery back up with 16.9 volts from a boost converter.

badinfluence

Ok time for a recalibration and retrofit.

Here are the planned changes.

15v total system voltage
24 Capacitors total
Modules will be 2x4Caps (700Farad 5V packs) 4(parallel) + 4 (parallel) to make 8 in series at 700 Farads.

Each 5V module will have a safety bypass that activates at 4.75V and starts a fan cooling the capacitors and a resistor to burn off current till it goes below 4.8v. (yes I know this set point is higher)

So the crowbar circuit will be a bit different than a normal crowbar because I can't drain the caps down to 0v or it would start blowing fuses, so I will be using a combination Zener/Resistor divider instead of a SCR.

Each set will also have 4x passive LED+Diode circuit for balancing activated only if the voltage exceeds 13v, otherwise it will be 1 or none.

Each set will be charged with a LM338T set for 5.4v with a blocking diode to charge to 4.75v safely, then shutdown. This will be a fixed wattage of 5A@Vcap so it doesn't cook the TO220. This would be the equivalent of a jump start, and will run from 3 isolated power sources so they cannot interfere and nothing is between pack A-B-C (current would be to high for most parts to withstand without melting including buswire and tabbing)

I am testing the circuit with a 5V module that I made when I constructed a jig for mass producing them and to keep them standardized so they don't melt in the heat of the engine bay. They will eventually be wrapped in Teflon, then Kapton Tape, then finally Aluminium tape (the kind that doesn't melt and burn). There will be room for air to enter through the fresh air duct on the battery, so the caps don't suffer melting solder issues.

There will also be a remote indicator that will go to the dash area with 3 LEDs and resistors letting me know the charge level with the LEDs and a double 7seg display.


Sounds complicated, but really it is a pretty simple setup, just thought out a lot. Soldering the caps is a bitch, so I may end up getting or borrowing a welder to weld them or use my new "freakin laser" to weld them if I can figure that part out.

badinfluence

Ok, did a couple test cells on this one and found a serious issue unknown to me from before. Charge rate on these does not follow what I would consider making sense for normalize electric charging rate or current flow.

with 6 capacitors, number 2,4, and 6 overcharge ALOT to almost 2.8v/2.5v with 14.0 from a high current charger. Even with my Safety circuit of a switching diode and an LED, they came on very quickly. After reaching 12.7 I got scared and pulled the current source and let the cap bank sit for 4 days. The LEDs are still on for 4 and 6, which is concerning to me. I am thinking for the 350s I need to burn test them by charging and discharging each cap a few times with some kind of Automated break in process to normalize them to each other.

What this means for anyone else is, I would stick with at least 150F total capacitance, because 58 is just unpredictable without active capacitor balancing. If space isn't an issue the typical 3000F/3500F at 2.5/2.7 would be for you. The new 2.85v would be even better! (you would only need 5 of those and the balance to go in between.

So I am building a little test cell to charge all these up at a fixed rate, discharge them slowly, and recharge them a few hundred times with a simple arduino setup.


For the main system I am going to limit the current coming from the alternator, but not limit the current coming from the caps to the starter. I may possibly run a line from the caps in the back to the starter itself and leave the existing wire from the battery post to the fusebox using it as a junction point to a small battery and cap setup. I think the PCM need some kind of power supply itself in case the power drops and the NVram is lost.

9krpmrx8

iTrader: (46)

I really don't think anyone else would attempt something as ridiculous as this. I think it's safe to say that you and harlan are alone with this type of thing.