In my search for a 8 to 10V, 500A source, needed for 2 seconds, I've been looking at LiFePO4, and LTO (Lithium Titanate) cells, in 38120 and 66160 sizes. All types of suppliers, plus completed battery systems, with charge balancing electronics. Sheesh, it's enough to make one's head spin. BTW, a simple "12V" battery has a bit too much voltage, an extra burden on power MOSFET(s), so cell-by-cell connectivity would be nice. Probably gonna be at least $500, no matter. Gotta ask Rob Legg what he thinks.
Good idea, but it'd be a 5kW switcher, and with enough bulk capacitance to last through the 8ms troughs. Each of six phases running at almost 100A. Sounds like more work than putting together a battery.
18650 Li-Ion cells, while they look impressive, are actually really wimpy. 38120 LiFePO4 cells look even more impressive, but still have rather disappointingly-high internal resistance. Stack four in series, esr is 4x higher yet, so you need lots of parallel stacks. 66160 cells, of either LiFePO4 or TLO chemistry have attractive low esr. A two or three-volt drop at 500A isn't too bad. Whoa, they're huge. I like TLO's safe operation.
I'd need four of them to get to 8 to 10V. And I have a pile of surplus half-rack 5V 200A supplies, could wire up four of them. Also a few years back I made a 6V 400V rack-mount supply, that could probably do 500A for a few seconds. Some parts I'd like to characterize could get by with the lower D.U.T. voltage.
Now you've changed it to 8 to 10V, before it was just 10V. LFP cell voltage is 3.2V, so you only need 3, given cells with low enough internal resistan ce to keep the voltage from dropping below 8V. What you won't get is an int egrated BMS, so you will have to take appropriate precautions. I suggested an external resistor to take part of the load off of the MOSFET.
The simpler the requirement, the simpler the solution.
If all you needed was to duplicate the 8mS half-sine forward surge current rating ( a non-repetitive test ) for rectifiers, a pulse-forming circuit could do it - but this is not really a linearly controlled method, it's just a resonant hammer that gets adjusted to suit pulse losses in the DUT.
The period is set by the non-adjustible PFN. The standard specifies a peak value of the half-sine. At 1Hz, you might lose 40W in the circuit and DUT.
A PWM buck would reduce the current stress on the source - so you'd not really be looking for a 500A source, you'd be looking for something that can supply the energy delivered to the DUT, and losses.
How this could be integrated into a linear test circuit like the RS-796A, is a puzzler. Perhaps PWMing the switch driver, outside of the linear control loop, with a choke and rated freewheeling diode might do it; but you'd have to sense the current in the DUT, not the switch current. This would defeats the RS-796A test modality, which assumes Idut = Iswitch.
Sorry if slow on the draw here. Figured it was a good time to catch up on paperwork, taxes etc.
Since you need 2 seconds and you have three phase available on site, why not use a three phase transformer and a six pulse rectifier ?
If you can't find a three phase transformer (or three single phase transformers) with exact voltage, get a transformer with slightly higher secondary voltage than needed and replace three or all six rectifiers with SCRs and adjust the firing angles for voltage fine adjustment.
A 5 to 6 Vrms phase to neutral secondary voltage should be quite suitable. About 200 A/phase secondary current would be required, With a 120 V / 6 V turns ratio that makes 10 A/phase on primary side.
Using wye/delta configurations on primary and/or secondary side might increase the number of suitable ready made transformers available.
Why would you want or need only 0.1V drop? 1V or even 2V drop would surely do. 3kF capacitors are fairly inexpensive from digi-key. The ones I have are rated for about 1.9kA and if I recall correctly about 0.3 milliohms series resistance. You'd need at least 3 in series, meaning at most 1kF total capacitance from a single string but that should be fine.
Yes, we already agreed that's a viable solution. But looking around for suitable transformer(s), I wasn't able to find much. Yes, looking in the wrong place, no doubt.
I can handle 2V drop, and while the peak current will be 500A, the RMS current will only be 177A. About $112 worth of capacitors would do for a 33ms 2-cycle budget - simple and very useful. But a full 100-cycle capability looks like it would be much more expensive than batteries.
The main power transformer from one piece of broadcast NTSC video equipm ent I maintained in the late '80s had a 1000A, regulated 5VDC power supply that was over a half ton. It ran on 208, three phase. It was in a Vital Ind ustries 'SqueezeZoom. It was the first electronic special effects system of its kind, and it sold for US $250,000. One was used at the studio that pro duced the Sonny and Cher show, where they came back on stage in all the cos tumes they wore throughout the show, with no loss of video quality. Sadly m ost or all were scraped when TV went digital. Some industrial surplus store s might turn up similar transformers. Mendelson's, in Dayton Ohio used to h ave tons of large, low voltage transformers from early mainframes.
Wow, impressive. But I actually may not need a high VA rating, because I'm only drawing current for 2 seconds. I do need low winding resistance. 350A for 8ms, 177A for 33ms, down to 20A for 2 sec. Hopefully no more than 3V drop at 350A = 8mR max. Probably not a low number compared to your story.
So if you had 4x Maxwell BCAP3000P270, total price $256, that would give you 750F at 10.8V initial voltage. After 2 seconds of 500A DC, it would discharge by 1.33V to about 9.46V. The internal resistance would drop another 580mV, so it would be down to 8.88V at the capacitor terminals whilst the current is still on. You would not be running 500A DC for the whole 2 seconds if you are doing half-sine waveforms, so it would not be as bad as that. Also if you got keen, you could upgrade it to do 1.9kA for shorter periods with the same capacitors.
Those capacitors are very light too. I was thinking of making a cordless spotwelder with a series parallel array of those capacitors, but for say
10kA welding current, the MOSFETs to regulate the current (or even just turn it off at the end of a weld) look like they will be expensive or big, and I haven't had time to work on it.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.