110 VAC inverter installation

I agree and in theory, we could be looking at a peak power of 18 kW and sustained power at any level in between.

Bob Wilson

 

So I re-read the manual for the PWRI150012S:

	Column 1
0	[th]CP[th]SP[th]Eff[th]THD[th]NL[th]cable[tr][td]1500W[td]3000W[td]>85%[td]<3%[td]1.0A[td]4AWG

Source: http://www.aimscorp.net/documents/PWRI150012S.pdf

Going over the limits:

• Low input voltage - 9.8-10.2V - after auto shut down the red FAULT light illuminates. My VOM has a minimum and maximum voltage hold function so this should be easy enough to measure.
• Low output voltage, 106Vac - after auto shutdown, the red FAULT light illuminate. I measured with my Kill A Watt P4400, a true RMS voltage reader, 110 VAC at the end of two extension cords with no load. I also measured 110 VAC on a single, contractor cord with no load when testing with the resistance, space heater. If we find 110Vac at the inverter output plugs then this unit may be set for too low of an output voltage. I bought it last year and hadn't installed it until this weekend.
• Red FAULT lights on any of these conditions: low or high input 12V voltage, overload, or excessive temperature.
• There is no "OFF" position battery load identified. Given the small size of the ON/OFF switch, it probably operates some internal circuitry and I don't know what sort of load this looks like. When I turned off the unit with the circuit breaker, the lights dimmed slowly suggesting there is a 'big nice person cap' on the 12V side.

After confirming the output AC voltage is low, I'll contact the vendor about getting it trimmed up.

They also recommend a 2" air gap around the unit so I'll mount it higher.

Bob Wilson

 

Ok, we have some metrics:

	Column 1
0	[th]Out W[th]In W[th]Bat V[th]Inv A[th]Gal/Hr[tr][td]0W[td]21W[td]14.28V[td]1.45A[td]0.068[tr][td]486W[td]491W[td]13.73V[td]35.73A[td]0.178[tr][td]773W[td]925W[td]13.15V[td]69.50A[td]0.192[tr][td]1221W[td]1493W[td]12.65V[td]118.0A[td]breaker tripped[td]

At a 1.2kW load, the voltage decreased as the 12V battery discharged over a 2-3 minute interval. As the 12V voltage decreased, the current increased and circuit breaker tripped. The 0W, 486W, and 773W power levels were sustained for a one hour interval.

I was also able to measure the circuit breaker and two Anderson connector resistance and losses:

	Column 1
0	[th]Out W[th]R_cb[th]R_a+[th]R_a-[th]W_cb[th]W_a+[th]W_a-[th][tr][td]486W[td]0.00168[td]0.00056[td]0.00128[td]2.14[td]0.71[td]1.64[tr][td]773W[td]0.00158[td]0.00067[td]0.00117[td]7.64[td]3.24[td]5.67[tr][td]1221W[td]0.00178[td]0.00475[td]0.00117[td]24.78[td]66.08[td]16.30

The voltages were measured relative to the car body ground: B+, circuit breaker+, inverter+, and inverter-. Then the inverter current was measured.

The data suggests my swagging the Anderson B+ terminals failed and they should be soldered. However, I'm really not happy with their relative resistance. I will look at a direct wire and eliminate the Anderson connector. However, the circuit breaker worked as expected given the size of the inverter.

If an inverter has a smaller maximum load rating, say 1.2kW, ~120A, the inverter itself limits the current draw by shutting down. This is what happens with our NHW11, 2003 Prius installation. But our ZVW30, 2010 Prius, has a more powerful inverter with a greater surge, 3kW. However, the circuit breaker and Anderson connectors are showing non-linear resistance which decreases the surge capacity.

Because the circuit breaker tripped under load, they often lose capacity. The internal arc oxidizes part of the contacts leading to greater internal resistance, heat and an earlier trip. I won't know until I retest if the 773W can still be sustained.

This data suggests the earlier surge was induced by the wiring, probably the Anderson connector. This suggests a surge design when you know the average load and source are close:

• circuit breaker protected, source - as we have now
• directly connected, surge battery - this only has to provide enough energy to handle startup load. An alternate approach might be a bank of ultracaps but the price/performance of each approach needs to be carefully weighed.

It is raining this morning so I'll have to wait until this evening to retest to see if the circuit breaker is still serviceable at 773W. I also need to model surge architectures including different battery chemistries. For example, an 11 cell, NiMH, surge battery, would nicely solve the problem:

• 14.63V - do not exceed charge, 1.33V/cell
• 13.20V - nominal charge voltage, 1.2V/cell
• 11.00V - discharged, 1V/cell

Bob Wilson

 

Some had asked about the voltage drop from the engine compartment inverter to the rear battery:

• Off
  • Rear: 12.41 V
  • Front (jumper post): 12.37 V
• Ready, inverter off
  • Rear: 14.43V -> 14.50V -> 14.52V
  • Front: 14.54V -> 14.55V -> 14.55V
• Ready, 750W inverter load
  • Rear: 13.44V -> 13.42V
  • Front: 14.51V -> 14.51V

The 1V drop for ~65A means 65W are lost from the front to the rear. For improved efficiency, a front-mounted, inverter will avoid losing 65W/872W ~=7.5%. However, there is an increased risk that if the battery tries to provide more than the 140A rated fusible link at the battery, it could blow. Looking at the DC-DC converter to the DC distribution, fuse box, there is a 125A fusible link. This is an ultimate limit for sustained draw if the inverter doesn't reduce voltage as it is loaded.

Engineering is often making the least bad of multiple choices. The front post would not work for me:

• 125A fusible link DC-DC converter - lose this link and you've got a 12V operated Prius
• 140A battery fusible link - lose this link from a reverse surge and once the car stops, it has to be jumped each time

So I'm thinking of a simpler and more robust approach:

• 120A DC circuit breaker replacing the 140A fusible link - the purpose is to stop any current draw approaching the 125A fusible link limit from the DC-to-DC converter
• battery-to-inverter cabling - direct connection to both the B+ and Gnd posts, shortest possible 4 gauge cable
• upgrading the 12V battery - there are sealed, lead-acid batteries with high cranking and deep discharge capability that can provide the motor surge capability

This approach protects the car from any surge demand, including the DC-to-DC converter. There will be a power loss from the inverter to the battery/inverter system but in the 7-8% range.

Bob Wilson

 

It looks like 4 awg will handle the steady state load. The surge should be less than a second so I don't see a need to upgrade, yet. Once I take out the Anderson connectors, we should be good to go. But I have one open question:

Anyone have the Toyota 12V battery specs?

In particular I'm looking for the equivalent of CCA and any other technical specifications.

I had replaced the 12V battery on our 2003 with an Odyssey 925, 28Ahr, even though I knew the Toyota 12V was listed as 36Ahr. I've had no problems, completely satisfied. But I'm building my wife's car out to handle startup loads. So for now, I'm assuming 1kW from the inverter and 500W from the battery. My options are:

	Column 1
0	[th]model[th]500W min[th]2.5kW min[tr][td]PC925[td]20 min[td]2 min[tr][td]ER30[td]30 min[td]2 min[tr][td]ER40[td]45 min[td]5 min

So I have an idea of how long I can sustain 1.5kW and the 400 ms 2.5kW is also within reason.

What I don't have are any metrics for the existing Toyota, 12V battery.

Now one approach, wire it up and test. If I meet my requirements, starting and sustained running of the air conditioner, no problem. We already know a 1.2kW load can only be sustained for less than a minute so I suspect the Toyota battery is inferior for my purposes to the PC925. But I'd like to check the specs to be sure.

Thanks,
Bob Wilson

 

I called EnerSys customer support and after looking at the options, I'm going to try for a PC1200:

• 540A - CCA
• 42Ahr - capacity
• 7.87 x 6.66 x 6.80 - dimensions
• 38.2 lbs - weight
• 4.5 mOhm - internal resistance

I've built a slightly oversized cardboard box sized for the PC1200 and will test fit it in the morning. I may need one of their mounting hardware kits but the results will be good:

• 458W, 45 minutes - with the Prius providing 1kW easily 1.5kW for over half an hour
• 1,992W, 5 minutes - with Prius 1kW, 3kW
• 3,580W, 2 minutes - without Prius 1kW, enough to handle the surge alone

I'll still have to mount the circuit breaker but when done, we should have the full 3kW surge as well as sustained 1.5kW long enough to trim loads down to the what the Prius can sustain. I've also figured out how I'll handle mounting the inverter.

I'm going to trace the 12V battery cover and make a plywood base exactly the same size. The inverter will be mounted above the platform and two/three holes drilled for alignment posts. That way I can move the inverter platform to the engine compartment _IF_ I decide to reduce the power loss from the inverter to the battery.

A 90/100A circuit breaker will protect the 125A fusible link to the DC-to-DC converter. I'll lose the battery energy buffer but gain 10%, constant power load, 80-100W.

Bob Wilson

 

When I put a 1.2kW load, I could see the input voltage decay as the Toyota 12V battery drained. It ran for about 30-45 seconds before the 120A circuit breaker tripped.

I have a clamp-on, DC amp meter so I'm good for measurements. <grins>

The PC1200 may not fit. But if it does, it'll provide near 2x the capacity of the existing 12V battery.

Bob Wilson

 

I have a modified sine wave inverter in the 2003 Prius and I found the 750W gas heater motor did run warmer but it ran. The real problem was the voltage drop over the length of the extension cords. The sharp dV/t transitions make the extension cord look a little like a capacitor and the voltage drop lead to higher currents.

That is on my list. <grins>

Bob Wilson

 

Thanks,

I have some ferrite cores that I could run twin conductors through to make a relatively, low resistance, inductor. But I've been concerned that it might induce higher, reflected dV/t back to the power electronics. Once I get my wife's inverter installed to my satisfaction, I'll be more interested in reworking the original installation. It worked four days and six hour when tornadoes gutted the TVA transmission lines.

Bob Wilson