EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: juanfermed on April 26, 2016, 07:07:54 pm
-
Hi Everyone! Hope y'all are doing great!
I would like to have your suggestions/thoughts on the following: where I work, we have some systems that have an COTS (commercial off the shelf) inverter, which has a 12VDC input and 120VAC output, rated at 500W and with a peak power delivery for 5ms of 1000W. Now, some market research stuff tells us that our "market" wants to use a fridge with our systems connecting it to the 500W inverter. As you might know, even a small fridge of 2ft.cu, may have 1000W peak power when the motor/compressor is turning ON (we have tried a range of fridges from 2ft,cu - 10 ft.cu and with all of them the inverter turned OFF when the fridge started). We are having the issue that our inverter is not capable of running almost any small fridge becuase it turs OFF automatically when it detects a peak power higher than its rating, which is good, just that we need a higher peak power.
So my question is, do you know any method that I can try to increase the peak power capacity of the current inverter?
The first thing that comes to mind is the use of output capacitors as one does with lower-power DC stuff, but being honest I dont know which type of capacitors can we use with 120AC or how to look for them. Any other methods and recommendations are more than welcome.
For the record, I am in need to solve this issue but I was not even working in this company when they first selected this inverter |O which was obviuosly not well selected, but now we have over 800 of this inverters practically sitting there with no use.
Thanks people!
-
Hi everyone,
Do you think this kind of "run capacitors" can be of any help?
http://www.temcoindustrialpower.com/search.html?category=AC_Electric_Motor_Run_Capacitors (http://www.temcoindustrialpower.com/search.html?category=AC_Electric_Motor_Run_Capacitors)
Thanks!
-
You're right, they should have bought inverters with integrated soft start current limiting (not that soft starting is necessarily great for the fridge).
The fridge will pull high power for multiple cycles. AFAICS you'd need to have a huge reactor (inductor plus capacitor) to smooth out the load. A passive solution is not an option I think. There are external soft starters (https://inverterdrive.com/group/Soft-Starts/Fairford-HFE1-230V-Single-Phase-Soft-Start-1100W/) which I assume simply cut off part of every cycle to limit current ... but I don't immediately see anything cheap (which is not to say it doesn't exist).
-
This is the down side to COTS. I'm not aware on any drop-in solution to this. you cant just add a capacitor across the terminal like in an DC device it would have active switching to make AC and at that point you built a inverter. single phase motor are also the worst load you could put on an inverter since they have such high starting currents. I see a couple options though.
Change the inverter to COTS unit capable of running a fridge
Build a custom fridge with a compressor the doesn't have such high starting current maybe a VFD with 3 phase compressor (ah smooth) or just run off the 12V
Hack the inverter you have by modding the protection circuity and then re-qualify it or build your own inverter.
-
The right solution is V/Hz ramp up soft start. That unfortunately seems to be restricted to expensive inverters (e.g. Samlex) despite it costing almost nothing to implement. In the case of a digital inverter, it's literally just a matter of a few lines of code. I suspect some company has a patent on it since Vector used to make cheap inverters with that feature but no longer do with no explanation.
-
The right solution is V/Hz ramp up soft start.
Shouldn't it just cut to 0V for the remainder of the half cycle when it hits a current limit? (Plus if it keeps tripping the current limit for more than a couple seconds turn off.)
It has no idea what the load will be after all, so it can't really ramp.
-
V/Hz control is a very simple way to control the speed of an induction motor. As the name implies, it holds a constant V/Hz ratio (except near the bottom to overcome resistance) so inductive loads don't saturate. The nice part is that it's open loop so it doesn't need to know about the load in order to work.
-
Does that still apply when the starter winding is switched in?
-
I actually have one of those Vector inverters with the V/Hz ramp up feature (albeit a 400W one that is too small for any sizable motor load) and the ramp up to 120V/60Hz is pretty quick - over about half a second. As it is a modified sine inverter, the voltage scaling isn't that good - the lower voltages look more like alternating spikes.
One trick to reduce power usage with motor loads is to use an inverter that outputs 100V/50Hz.
-
As it is a modified sine inverter, the voltage scaling isn't that good - the lower voltages look more like alternating spikes.
Even if it was a sine converter you would want this, using +/- 155V pulse trains is is the only way to keeps most other loads working during the ramp.
-
" now we have over 800 of this inverters practically sitting there with no use. "
Using two in parallel would likely work by connecting the two DC HV supplies together (Not AC outputs). I start a 100W compressor with a 12V 2000W inverter and it takes 118A peak to do it. You don't have enough inverter boost to start the motor. Protect circuits may also shut you down.
-
It has been done before.
https://www.youtube.com/watch?v=pwGBt2QB8Qg (https://www.youtube.com/watch?v=pwGBt2QB8Qg)
https://www.youtube.com/watch?v=GJhcsJdcQ4o (https://www.youtube.com/watch?v=GJhcsJdcQ4o)
For bonus points, tweak down the frequency and duty cycle on one inverter until it gets to about 100V/50Hz and label it "motor". Tweak up the duty cycle of the other inverter until the output is almost a square wave at 170V or so, add some series inductance to the output, and label it "electronics".
If you can find a way to vary the frequency and the duty cycle from a voltage signal, it would then be trivial to implement the V/Hz scaling, although I'm not sure about the patent situation.
-
If you're going to hack the inverter, start by simply tweaking it back to 100V/50Hz. Costs nothing in parts cost (or a trivial amount if you have to change resistors) and probably easy to do once you figure out the first unit.
-
A few years ago I bought a 500W inverter to power my laptop (big "desktop replacement" at the time), but even if that only had a 120W PSU it was simply impossible, the inverter's protection would always trip. No amount of reconnecting or repowering in any order would do, the PSU's caps would always have emptied by then.
Drove me crazy so I went into brute-force mode and just soldered a big fat wire on top of the current measurement shunt. No more protection, but at least it effing worked.
In your case I'd try something a but more refined and add a fat MOSFET instead with a control circuit that would turn it on and short the shunt for a few ms (time to determine experimentally) before restoring normal operation.
Or even better to avoid a big high current part otherwise futz the protection AFTER the shunt to delay response to a fast rise in shunt measurement, might be a simple RC low-pass filter between shunt and whatever measures it, or something more elaborate like an MCU that holds the voltage down for a while if needed.
If you've got 800 unused units you can dedicate a couple to experiment a bit.
-
That's a good way to blow up inverters. Overloaded power MOSFETs die faster than any other circuit component, and (assuming its a reputable brand) the over-current protection network will be designed to shut down the gate drive fast enough to stay within the worst case SOA power limit. If there was any margin there, the manufacturer would have already tuned the control loop to permit a greater surge rating.
Just about the only chance is a soft start kit fitted to the fridge compressor. A 1/3 HP compressor uses a nominal 246W when running, and that would be the upper limit of what would be practical. Obtaining fractional HP soft starters suitable for single phase fractional HP compressor duty may be problematic, and if they have to be custom ordered, may be more expensive than replacing the inverters.
-
In your case I'd try something a but more refined and add a fat MOSFET instead with a control circuit that would turn it on and short the shunt for a few ms (time to determine experimentally)
It's on the order of 0.5 seconds, far above the thermal time constant of the silicon in the output MOSFETs. Not above the time constant of their heat sinking though. You would have to mod the entire thing. Bigger MOSFETs (assuming the driver is strong enough to drive them), more DC bus capacitance ... and then change the shunt.