5-Way DC-DC Flyback Converter

[blueroomelectronics]

Have you ever built an inverter before?

 

[iss407]

Have you ever built an inverter before?

Nope. I haven't done much power related stuff, mostly signal level and mostly digital. The last big project I worked on was an ultrasonic range finder. I adapted some existing circuits to use PIC microcontrollers. I want to use digital controls everywhere I can both to reduce components and because it is what I know. I also like to be able to add functionallity with only a software update.

 

[tcmtech]

This Post is admittedly mostly speculation here so keep that in mind.
More reasonable concept!

being this is a small EV design right 4- 10 hp drive system. (3 -7.5 kw)
I would be tempted to check out a golf cart speed controller. I dont know much about them hands on experiance yet, but I do know there are some fully programable and most likely comuter interfaceable units all ready available on the market. A fast look on eBay put them in the $150 to $700 range for 36 to 48 volt high amp with built in regen functions. And some sort of progamming power curve reference was made on a few as well.

Not sure about the nessesity for the ultra caps though. I dont have much experiance with them either for motor regen or power. Could be very effective or it could be a waist of time and money too. I just dont know.
If the caps are directly in series with the batteries I doubt they will be of much real world use. when you regen they may charge up abit but the batteries will suck them back down in mere seconds. Same with the putting back part of the cycle they will drop abit but so do the batteries. Again no real experiance to work from. I could be way off on the concept myself.
I have seen regen braking systems with batteries before and none used big capacitors for power storage.
Also on dumping back to them try this. convert your theoretical wattage you could put back on a single regen cycle. Now at what voltage your system is working at, what amount of amp hours did you get? Just guessing but I bet its not a lot. Compair it to the useable AH in the battery system you are using. Bet its down in low percents range! not enough to cost justify the number of ultra caps and extra gear to make it work. Possible money better spent on a bigger or better battery system?
Real life cost outlay VS efficiency gained.
as far as the inverter. Again assuming you went 24 - 48 volt system. There are many comercial units out there already. Sine wave output too! Some searching and you may get what you want used for pennies on the dollar.
Think back up power in stores and business places. I have scrapped a few over the years and my junk buddies have done many more. Some of the higher end ones were full solid state like the little cheapies but did a true sine wave and at 4 to 15 KW from 48 to 60 volts battery sources.
Actual used price? We have gotten a few at "take it out and clean up the mess you can have it price" most were scrap value but you may end up with having to buy a pickup load (or two or three!)of bad lead acid batteries just to get the big suitecase sized inverter.
Used and probibly less efficient but at how much real world money and time saved?

 

[iss407]

The system I want to build is a master/slave configuration with one master system controller that has no power electronics but manages and synchronizes the slaves. There will be multiple slaves in a moderately sized EV. I would probably want 3 or 4 (12-16 Hp cont, 40-50 peak) in the Saturn Coupe I'm converting now (it will use a "normal" DC system available today though). The key to the master/slave system is to have the master match frequency and phase of all of the slaves. The 120V 60Hz PSW inverter will be able to match power from the utility. The motor controller will match each slave so they can combine power to the motor. It will also be possible to have different slaves run different motors in the case of driving individual wheel motors, which are coming down in price. They can be run together, but then when turning the inside motors draw more power and the outside motors draw less which is the opposite of what you would want for good car control. I figure the slaves can reduce the amount of power they put to the motor by delaying the phases just a little bit. That way the master can match load and state of charge on all of the batteries.

As for the ultracaps, they would be run in parallel to the batteries. In my latest design a slave could have all batteries, all ultracaps, or a combination. The 4 DC storage connections can be configured independently. Ultracaps would be purely and option. They only make sense for cars that do a lot of stop and go driving, in a city environment. They will mostly store the energy of a stop and return it when the car goes again, but that will also reduce the max load on the batteries, which will extend their life. The batteries can store the energy from stopping as well, but not as efficiently as the ultracaps, and the extra high current loads on the batteries will reduce life, even as the extend the range on a single charge.

One big benefit of the flexible DC storage idea is that different types of batteries can be added. NiMH batteries have good energy density and improve range, but they have very poor high current capabilities. Lead Acid batteries can push lots of current, but the energy density is lower. Put a mix of batteries in and the LA can be used for acceleration and the NiMH can be used for cruising. The best of both worlds again. Since the connections are isolated normally they can be tied to a battery in parallel to increase the max power drain from the battery.

This whole idea is for something that could be manufactured in quantity for a reasonable price. The more universal I make it, the more places it can be used in and the more that can be sold, bringing the price down more. That's why I added the inverter function for example. Making it useful for solar/wind storage broadens the potential market even more.

Paul

 

[tcmtech]

Much better explanation. I see where you are going with this now. Thats where the multi splitting controls and multi directional stuff come in on this.

Bit far off from my area of expertise. I just make broken things go again. And sometimes break things that already went!

Power handling and stuff not a problem. finding the parts to make it, not a problem either, All that integration and handling stuff is outside of my area!
Someone else has to field those questions!

 

[tcmtech]

I do need someone elses input on this.
As I understand it one farad is one amp for one second. So, one amp for 3600 seconds is one amp hour right?
what do big ultra caps have for real world values right now?
what volts and what farads capacity?
And what dischsrge rate do they work up too? Amps per second or something like it.
And what actual cost?
They seem like interesting devices but how big do they need to be to match a small battery in terms of usefull energy.
Say physical size per theoretical amp hour of an ultra cap bank Vs different battery types.

Any Ideas out there?

 

[iss407]

Power handling and stuff not a problem. finding the parts to make it, not a problem either, All that integration and handling stuff is outside of my area!
Someone else has to field those questions!

The control system is my strong suit. I'll use serial digital communication between microcontrollers for most of the control and sensor readings. Each slave will be semi-autonomous. A slave should shut down if sensors indicate a problem or communication with the master is lost for more than a certain amount of time. (250ms sounds about right) The crucial timing information will be sent on dedicated timing connections from master to each slave. I'll probably use differential twisted pair wires for noise immunity. Besides, twisted pair ethernet cable is cheap. The master will have the feedback sensors from the motor(s) and do all of the PWM timing so the slaves all sync up.

The master will also log lots of data for serious system monitoring. An optional touchscreen interface would be a good addition. A connection to a laptop for system setup and maintenance seems a certainty. I'm a firm believer in open source software, and if I have my way I'd make the hardware open source as well. If I want to get a manufacturer interested however I might have to keep the rights, at least at first. Look at MP3Car.Com for an idea of how much open source stuff is out there. Tying the master controller into systems like they build seems natural.

For the Inverter I was trying to figure out how to set it up for 240V split-phase with a common. I couldn't see an easy way to do it. Then I realized that if 240V is required the electrically isolated inverter outputs from two different slaves can be tied together out of phase at the master. Just need an even number of slaves if 240V is needed. Heck, with three slaves tied together I could even program it for 3-phase power! That's not something you can find cheap, I'm sure!

So what other uses could this be used for? DC motors are not a problem with the right programming. I'd need extra sensors on the master for multiple wheel motors, but then traction control and anti-lock brakes are possible. I guess with some extra wiring it would be possible to hook up a trailer with extra batteries and slave controllers as a range extender. Maybe the trailer could have its own master controller that would be superseded by the master in the car when pulled as a trailer. Then when not needed for the car the trailer could sit in the garage as a whole-house UPS and/or power conditioner.

The possibilities are truly endless with some creative thinking. What kind of voltage/current is needed for Arc Welding?

Paul

 

[iss407]

I do need someone elses input on this.
As I understand it one farad is one amp for one second. So, one amp for 3600 seconds is one amp hour right?
what do big ultra caps have for real world values right now?
what volts and what farads capacity?
And what dischsrge rate do they work up too? Amps per second or something like it.
And what actual cost?
They seem like interesting devices but how big do they need to be to match a small battery in terms of usefull energy.
Say physical size per theoretical amp hour of an ultra cap bank Vs different battery types.

Any Ideas out there?

One of my previous posts pointed to a Chinese supercapacitor I found online. It was a module rated at 2200 Farads and 2.7 Volts, for about $20 or so. The energy stored in a capacitor is E=0.5*C*V^2. So if a charged cap drops to half voltage, it has lost 3/4 of its energy. Fully charged to 2.7V the cap above will hold about 8,000 Joules, or Watt-Seconds. It has a max charge/discharge rate of over 1200A, so it can transfer energy FAST. I'd say it would make sense to run it between 1 and 2.5 Volts or so. Not much energy left below 1 Volt and getting too close to comfort above 2.5.

Caps in series work like resisters in parallel. So 5 of those caps in series act as a 440 Farad capacitor that can reach 13.5 Volts max. That would be a better voltage range to work with. The energy capacity adds, so 40 KJ or so. If max transfer is 1200A at 13.5V, that is over 16 KW. It falls off quickly though. If I use the above cap pack with one of my 7500 KW max controllers I could charge/discharge fast 40KJ in 5 or 6 seconds. If I use only one of the 4 DC connections on the slave I can only transfer 2KW or so max, perhaps more for short bursts. If I figure 30 KJ usable energy storage that gives around 10 or 15 seconds. That is just what I want for acceleration from stop to highway speeds.

Those numbers are SWAGs, of course, but I think they are in the right ballpark. Adding $100 worth of caps for each controller would allow it to work well for regen. A good sized Deep Cycle LA battery will be $100-$200. If I put 3 batteries and 1 cap bank on each, the cost of the batteries far outweigh the caps and would probably pay for themselves extending the life of the first set of batteries. NiMH batteries are many times that cost. They ultracaps make much more sense with them since NiMH batteries are so poor at max current capability.

Paul

 

[tcmtech]

hmm... theoretical super EV car. I put much thought into that one myself!
I stuck to my strong point and thought big high torque forklift traction motors at each wheel. I came up with a theoretical design once. With traction motors at each wheel I came up with 600 HP at the wheels, theoretical of course! I intended to use a 360 VDC system. Simple PWM controler at each motor 350 amp switch capacity per motor. completely doable for under $1000 per motor.
standard high cranking amp car batteries for the power pack. all show no distance! Just to put the EV slug myth to rest!
Show off at fairs and drag strips. good public turn out places.
Blistering acceleration and tire smokeability from all electric! but the $40,000 build tag is out of my reach!

I do see what you are after now. it sounds almost like it could be practical to do. but still expensive for making at low volume.
As a avid hauler of usless stuff I like the electric trailer concept!
As a hands on tweeker I also like the data logging possibility too!

Three phase power you say? trust me thats cheap to make!
I plan to do a post here some time soon about how to make three phase motors run on single phase. cheap and efficiently. I have many I use every day that are three phase running on single phase. My biggest one is my 15 hp commercial air compressor. its 240v and runs on a single 60 amp line!

I might get a bit of pokes at first but I can back it up!
Any one that wants to disprove me can try! I have made bets with profesional life long electricians that said its impossible. They never show up to take the bet though. Wont even take the bet when I offer to do it at there own shops with them suppying all the parts off there own stock shelves! I have built dozens and sell them from time to time too.

 

[tcmtech]

usefull info on the super caps. but still 13.5 volts is very low for driving any decent size car. My old EV was 96 volt and ran at 250 amps regulated. It had good power and could do normal road cruising at 65 mph most of the time. but the 16 6 volt 220 amp hour batteries still did not go that far. for real wolrd driving I would go double that voltage and amp hour and double the motor amp limit as well. 96 volts at 250 amps gave me 24 kw or about 30 hp at the wheels. With a bigger motor or multiple motors I could get 120 hp to work with.
given 3 electric hp is considered the same as 5 gas hp that would be like having a 200 hp engine. Or about like driving my mercury grand marquis.

But then to do that I get the same delema as you, battery type.

My brother plays with the lithium polymer batteries alot in hobby RC stuff. Given the average cell size he uses and 3.7 volt 5 ah capacity plus the 50 amp draw capability.
We sized up to that of a 12 100ah deep cycle battery and we came up with a theoretical cappacity of 14.8 volts and around 1000ah per battery block. With a discharging rate of well over 1500 amps! In theory of course!
Now High output EV with good real world driving range is doable.
( with a $30,000 battery pack!)

 

[iss407]

hmm... theoretical super EV car. I put much thought into that one myself!

I don't want a supercar, there are several of them already and they are way too expensive. I want to build something that is universal and inexpensive. I want flexibility most of all. This system seems to fit the bill. I love the idea of starting small, with two slaves and a few batteries so a car can drive around town. More power and/or range could be added at any time with a major redesign. Right now and EV will replace a whole string of batteries if just one is tired because replacing the one will unbalance the string. With this setup the batteries could be used until they were truly near the end. Also, the idea of adding different types of batteries is VERY appealing! I'd love to build all NiMH but there is no way I could afford it. With this type of system I could add NiMH incrementally, slowly replacing over time as the older LA batteries wore out.

AC motors have much more appeal than DC, but they are very expensive, as are the controllers. The smaller PMAC motors look great, but they can't power a car on their own. One for each wheel makes sense, but what about controllers? There are small PMAC controllers out there, and I have considered them. I want that max torque at start, since they won't be geared. The motors have to be oversized for the higher speed so they can perform well enough at the low speed. A 3-phase motor can do better at low end.

Well, I'm basically going way off course at this point, let me get back to the design.

What options are there for isolated gate control? I'd like something for the high side that didn't need its own power.

I've seen several options. I read about an optocoupler that could produce enough voltage at the output to drive a gate without additional power. There has to be a penalty though. Speed?

I've seen small transformers laid out with traces on the PCB to isolate a small gate signal. Seems tough to implement.

I've seen one module with an isolated internal DC-DC converter. Didn't get a price, but I figure it would be high.

What about having a small high-freq low-power transformer on board with a bunch of secondaries that could each power a high side gate driver? With a H bridge or 3-phase bridge is it possible to use one power source for all of the high-side gates or do they each need their own?

Anyone have any experience using a FET as an ideal diode? Can a simple circuit turn on the FET when the body diode becomes forward biased and turn it off when the current reaches zero? It would need power but it should be able to run autonomously, without central control. I'm using some high-current, low-voltage circuits for the batteries and the voltage drop across a read diode will really hurt efficiency and generate a lot of heat, both bad. The high-voltage DC stages won't suffer as much loss, but it can be done without a lot of hassle it makes sense to do it there as well.

That's enough rambling for now. Time to try and shut my brain off and go to bed.

Paul

 

[iss407]

usefull info on the super caps. but still 13.5 volts is very low for driving any decent size car. My old EV was 96 volt and ran at 250 amps regulated. It had good power and could do normal road cruising at 65 mph most of the time. but the 16 6 volt 220 amp hour batteries still did not go that far. for real wolrd driving I would go double that voltage and amp hour and double the motor amp limit as well. 96 volts at 250 amps gave me 24 kw or about 30 hp at the wheels. With a bigger motor or multiple motors I could get 120 hp to work with.
given 3 electric hp is considered the same as 5 gas hp that would be like having a 200 hp engine. Or about like driving my mercury grand marquis.

But then to do that I get the same delema as you, battery type.

My brother plays with the lithium polymer batteries alot in hobby RC stuff. Given the average cell size he uses and 3.7 volt 5 ah capacity plus the 50 amp draw capability.
We sized up to that of a 12 100ah deep cycle battery and we came up with a theoretical cappacity of 14.8 volts and around 1000ah per battery block. With a discharging rate of well over 1500 amps! In theory of course!
Now High output EV with good real world driving range is doable.
( with a $30,000 battery pack!)

Remember that supercap would be tied to only one slave controller. With each additional controller there would be another supercap bank. 1200A at 13.5V is good. With 6 slaves you get the total equivalent of 1200A at 80V. Now that could really move you!

 

[Ubergeek63]

Kit Information

 

[iss407]

Kit Information

The power kit still consists of seperate components. My goal is a combined, all-in-one system.

 

[Ubergeek63]

true, but it is a neat and relatively affordable electric vehicle that almost goes highway speed. I would love to have the money to get it and get a custom motor for it designed by these guys: **broken link removed**, preferable a hub motor, and LED lights.

Dan