TheBackShed.com

Forum Index : Electronics : Ozinverter 135v input

Author

Message

stevebequik
Newbie

Joined: 20/07/2016
Location: Australia
Posts: 12

Posted: 01:54am 24 Jul 2016

Copy link to clipboard

Print this post

Hi shed team. Newbie here.

I have ordered clockman book and started ordering parts from china for my inverter build. However, due to my battery bank, my build needs to be a bit different.

Without boring you guys too much, my bank is a 2014 chevy volt battery pack I imported from the states. this is a 3s96p 16kWh lithium pack. Very popular with DIY EV community but I am using it for my off grid setup.
My initial idea was to keep the pack whole and build a transformer-less inverter. Since decided this is out of my league and then there is still the problem of isolation and handling high voltages (336 to 384v)

The pack can only be broken down into smaller packs of 3s12p
So this gives me options of 45v nominal (42-48v range)
90v nominal 84-96v range
135v nominal 126-144v range
I am direct charging from my PV using a simple PWM controller. My panels have a Vmp of 26 to 29 volts. I live in NQLD so get a bit of derating due temperature. I like the idea of 5 panels connected to a 135v pack for maximum efficiency.

So that how I desire a ozinverter with a 135vdc input.

This should give about 90Vac then I can rewire my toroidal to suit.

So I am really after help with redesigning the Fet board to suit.
Power fets and caps would need to be rated at say 250v.
Thoughts on say using a IRFB4332 fet. This is a 250v 60a mosfet that looks similar, but Rds, Vgs, gate charge, input capacitance are all a bit different. Possible problems with the drive chip and driver circuit?

10000uf 250v caps are huge and expensive, Miminium size caps required?

I already have a regulated 12 and 5V supply.

Any other possible problems or thoughts on the idea in general?

Any help greatly appreciated.

Madness

Guru

Joined: 08/10/2011
Location: Australia
Posts: 2498

Posted: 02:19am 24 Jul 2016

Copy link to clipboard

Print this post

Hi Steve,

Will be interesting to see how you go, if you are using Aero Sharp inverters for donor parts they have a heap of 450 V Caps, will have to check capacitance. You will have trouble fitting caps that you require in between the heat sinks if you use the standard configuration there.

No doubt you are aware of the precautions you need to take around the voltage levels you are planning to use. I have had a few kicks off my battery bank when it is up around 60 volts, at >100VDC it would get a whole lot more serious.

What does a battery pack like that cost landed here?

Gary

There are only 10 types of people in the world: those who understand binary, and those who don't.

oztules

Guru

Joined: 26/07/2007
Location: Australia
Posts: 1686

Posted: 01:06pm 24 Jul 2016

Copy link to clipboard

Print this post

I expect your drive will be fine, up to 600v, and the fets will work ok.
Capacitance can come down significantly, and I would run the aerosharp ones for a start.

The chinese boards use only 10000uf for the total of their fet boards, so with your current requirements so much less I expect if you can get up in the 10000uf and more total, then your good to go, but it seems forgiving with ripple, so thats why I expect the caps from the aerosharp will run it fine.

Take note... I have not had time to solve the short circuit protection yet, but if you don't short very definitely ( solid short circuit), your unlikely to find it failing.... Thats why I didn't wake up to it... I couldn't break it in normal operation, and generally I don't short out 240v... so I missed it.... but it still needs to be looked at.

...........oztules

Village idiot...or... just another hack out of his depth

stevebequik
Newbie

Joined: 20/07/2016
Location: Australia
Posts: 12

Posted: 01:55pm 24 Jul 2016

Copy link to clipboard

Print this post

Thanks Oztules and Madness.

I am a retired electrician, so fairly use to the precautions that are needed with these voltages.
Interestingly, the bank is currently ungrounded and sits at about 70 volts plus and minus to earth. Still enough to get a tingle if I am working in bare feet. Only did that once.

The pack cost me $1500US from a wreckers. Used a company called usatoaus.com to freight it to Brisbane. Shipping was dirt cheap but took 6 months in total. They wait until they have a full container of cars and parts. I think I was a little unlucky with this. Landed in Townsville it owes me just over $3000.
Was starting to worry that the pack could self discharge to dangerously low levels during this time. But pack arrived with about 80% SOC and all cells were perfectly balanced. In the states Chevy has a 8 year warranty on these packs.
These packs also contain 2 HV Panasonic contactors, 3 LV relays and a LEM curent
sensors that are reusable.
Eventually I wan to get two of them. Once re-configured this will give me 30kWh total and about 22kWh usable.

Still looking for my donor inverter. Don't seem to be a lot of these aero sharps in north Queensland. Great news that I should be able to reuse the caps.

Oztules, because the toroidal rewind will only have say a 2.6 to 1 ratio. Does this effect the core required? Or should two 1.5kw stacked or a 3kW still be sufficient?

Thanks

yahoo2

Guru

Joined: 05/04/2011
Location: Australia
Posts: 1166

Posted: 02:31pm 24 Jul 2016

Copy link to clipboard

Print this post

Hi Steve
from what I hear it is a combination of 3s12p and 3s6p however looking at some photos there are tie rods that hold the packs in bigger chunks. if you could split the pack electrically but not physically it would be a big help I think.

It would be handy to know what upper and lower voltages that Chev use the pack between.

the only reason I say that is you are in the ballpark of the 120v standard.

for example with LiFePo banks, a company like selectronics recommended 38 cells @ 3.2v nominal with their old firmware. that's 121.6 volts. there is a new revision in 2016 that uses 40 cells, I am not over that update yet.

So with 38 cells I set the charge and float settings at

charge 3.4875 per cell = 132.5 volts

float 3.3375 per cell = 127 volts

If I went to the max balance voltage I would see 3.7v per cell that is around 141 volts.

So you are in the ballpark with your 36 cell combination and specific battery chemistries higher voltage spec (I assume). It will help your Public Relations and avoid confusion if you can actually say you are installing a 120 volt system.

Cheers Yahoo
Keep us all in the loop, it looks interesting!

I'm confused, no wait... maybe I'm not...

oztules

Guru

Joined: 26/07/2007
Location: Australia
Posts: 1686

Posted: 11:14pm 24 Jul 2016

Copy link to clipboard

Print this post

"Oztules, because the toroidal rewind will only have say a 2.6 to 1 ratio. Does this effect the core required? Or should two 1.5kw stacked or a 3kW still be sufficient? "

As I understand it, the problem is an ampere turns problem, you will have less amps but a correspondingly higher number of turns, so magnetization should remain the same..... ie no change as far as the core is concerned... and ratio's make no difference to the core... it could care less.

In the case of the capacitors, there is a stark difference, as the energy stored is a square function of the voltage, not the capacitance... so double the voltage, we have four times the energy stored... your going to triple perhaps... so nine times the storage for the same uf..... as you can see we can lower the capacitance by a factor of nine and end up with the same storage.

The aerosharp has 600uf caps, x 6 =3600uf as against 40000uf for the standard 8kw PJ board... so similar sort of storage capacity.

They use the same for 24v systems for 8kw, and there we are way ahead even for 120v, as we have a 25:1 increase then... so not concerned... if it works for them, then it work fine for us.

energy = 1/2 E^2 x C from memory.... perhaps look it up to be certain.

.......oztulesEdited by oztules 2016-07-26

Village idiot...or... just another hack out of his depth

stevebequik
Newbie

Joined: 20/07/2016
Location: Australia
Posts: 12

Posted: 06:22pm 25 Jul 2016

Copy link to clipboard

Print this post

Thanks oztules for explaining. I have gone ahead and ordered the 4332 fets. Will be at least a few weeks before I can get started building. Plenty of time for you to find out a solution to the over current protection hehe.

Yahoo, these cells are a lithium manganese cobalt oxide chemistry. They are a nominal 3.75v. Chevy apparently plays it pretty safe and operate these cells between 3.5 and 4.05 volts.
The 135v volt input is a neat modules of 36 cells. The way the tabs on the pouch cells are welder together makes it virtually impossible to have any arrangement that isn't a factor or 6 or 12 cells.
This voltage works really well direct charging from 5 panels.

I will keep the forum updated. Having the option of a 48, 72, 96, 120volt input could have many advantages.

oztules

Guru

Joined: 26/07/2007
Location: Australia
Posts: 1686

Posted: 09:34pm 25 Jul 2016

Copy link to clipboard

Print this post

Actually, the over current works fine, it is the instant solid short that is iffy.

Ie. larger loads than the setting is not a problem, but an instant 20kw surge is not caught fast enough or something similar... It is tough to plant a short deliberately, goes against the grain big time...

...........oztules

Village idiot...or... just another hack out of his depth

Print this page

To reply to this topic, you need to log in.