Please, feel free to let us know if this is a terrible idea. Basically, we currently have a battery shortage. Due to limited resources we have to come up with some *creative* solutions for testing purposes only, (idle work such as code-deploy or electrical testing).
We were hoping that we could power a robot, possibly with a computer ATX PSU or a bench supply, if we strap some powerpole connectors to the 12v rail. It is my impression that it would have to supply a maximum of 120A (As the main breaker is 120A), and the minimum being amount TBD. I attempted to check the DS logs to see our total current draw recorded by the PDP, but gave me wildly inaccurate numbers exceeding 120A, which doesn't seem to make sense.
I'm sure that due to the nature of FRC, current spikes would be frequent, so something that has a max amperage of 120 would be sufficient.
Any suggestions to make this work would be greatly appreciated. Let me know what you are thinking. Thank-You!
If you are just testing code with the motors freely spinning, you would only need a few amps per motor. But if you are actually driving around, you might draw upwards of 300A. These spikes are rather common, and have a good chance of tripping the protection circuitry in a nice power supply. A better idea may be to have a single battery in parallel with a charger (maybe with more circuitry), using the battery to suppress the spikes. Spending $40 on a battery to handle current spikes is a lot cheaper than $170+ on a decent power supply to handle the spikes.
By the way, seeing a sum of way over 120A through the PDP is not unusual. Because the main breaker is a thermal circuit breaker, it takes time for it to trip. It trips faster the more overload there is, but you can take 350A for 5s without tripping the breaker.
If all you’re using the bench testing setup for is deploying code, testing electronics, and spinning motors at free speed, 20A should be plenty. Most electronics draw currents only on the order of milliamps, and all FRC motors’ free currents are less than 10A.
If you plan on actually hooking up this setup to motors and mechanisms that will draw a robot’s worth of current, then 120A will likely not be enough. Those main breakers only trip after a few seconds, based on how much over the limit is being drawn. Based on the data sheets here, the break can sustain 162A indefinitely, and 720A for up to 3 seconds. You need to be careful here that you don’t burn up your power supply.
548 bought a 13.5V ~50A power supply to test motors/gearboxes. Even with a single motor, we had to connect the motors through cheap PWM speed controllers to not trip the short circuit detection in the power supply when the motors start turning.
These power supplies are likely to have self protection circuitry that acts fast enough to trip out on the starting surge of a large motor if the motor power is not ramped up gradually over say half a second.
We’ve had a really nice experience with the APS-100. It’s a 1.3kW 13.5v power supply, puts out about 100A, about $300 used (last time I checked). Not going to run your robot at full load, but if you do some current limiting, it’s pretty great for testing autonomous and subsystems.
We do the same - actually hacked up an old xbox power supply to get ~14A max. An ATX should do just fine. The key is just don’t have any motors hooked up. This should be sufficent for most software-checkout or electrical-training purposes.
Also, to follow up on the slow-blow nature of the main breaker mentioned above - Per the “Time vs. Rated Current” chart in the datasheet, even at 400% rated current (480A) the breaker still takes a second or two to trip.
We’ve recorded spikes in the ~350A range pretty regularly without ever tripping the main breaker.
Everyone,
Most of the motors have stall current ratings far in excess of 50 amps, the CIMs at 131 amps. Yes they draw that much when they start. I have tried testing CIMs and smaller motors with my big ham radio supply and tripped it most of the time. There is no substitute for the robot battery instantaneous current supply when fully charged.
Wow, I never thought someone else would think about doing this other than me. Basically, my idea was to build an industrial-grade full bridge-rectifier and then use it to power the robot. There are a few reasons this is a bad idea. First, connecting the robot to mains can introduce noise and other artifacts in the power, that would in theory get filtered by the capacitors of the rectifiers, but in practice things could get trickier. The second bad reason is pretty much as simple as letting Highschool students with not a lot of experience “play” with mains voltage and current. 1800 watts is a lot, and by my math you could pull about 150a from mains at 12v. if someone messes up and shorts something somewhere, something could go horribly wrong, with melted copper everywhere. Thats just my take, I’m no electrical engineer (yet) but feel free to offer alternatives to my idea all. Edit: the reason you are getting pulls in excess of 120a is because the main breaker is slow-blow and you can draw more for a certain period of time.
There shouldn’t be much noise to go back into the mains. If there is, it would already interfere with the radio and RoboRio.
You are correct that it would not be a good idea for high school students to build their own 1800 W, mains connected power supply as a side project.
It would be better to get a server power supply like this one. Poeple building e-bikes with large battery packs often use these to power their chargers.
Sorry, but I have to take a real stand on why not to do this. If this seems a little harsh it is because I really want to protect you.
The mains voltage in the US is 120 V AC, in other countries is 220 V AC. These numbers are the RMS measurement of the actual peak to peak voltage of the sinusoidal waveform supplied to your house. The actual peak to peak voltage of 120 volts is 1.414 * RMS * 2 or 339 volts. A full wave rectifier across the mains will produce an output of ~160 volts DC. This is the standard input circuit for line operated switching supplies including the little box that supplies your laptop. The high voltage then is switched into a high frequency transformer to transform down to the desired low voltage.
There are two methods to get the 339 volts AC to 12 volts DC. A transformer will convert the voltage to something near 12 volts and then some form of rectifier and filter will make the output of your supply 12 V DC. The transformer will also isolate the line voltages from your test bed. The other method is to use a switching power supply which is a very complex device and also uses a transformer internally to make a regulated voltage and isolate the line voltage.
Under the right conditions, the supply you describe (full wave rectifier and filter of the mains) could end up elevating the whole test bed by 120 V RMS or either the peak line voltage (160 volts) or something higher. That means that everything you connect to the supply could be sitting at a very high voltage before you even connect anything. The wiring in most of the components on the robot cannot withstand that much voltage and the result is the high voltage will arc over to you and other components. This will result in very bad things happening. To quote one of my professors…“There will be a flash of light, a puff of smoke and a hole in the air where you used to stand.”
To build a transformer operated supply, you will need a transformer that is capable of at least 200 amps. If your supply were perfect and had 100% efficiency in conversion, 200 amps at 12 volts is 2400 watts. (yes that is the equivalent of 24, 100 watt light bulbs) At 120 volts, that would need 20 amps minimum. While some house circuits are 20 amp circuits most are 15 amps. The transformer would likely be several hundred dollars and weigh close to one hundred pounds. I found a 100 amp switching supply that cost $1500.
Thanks Al, these are a lot of biiig reasons why I didn’t go along with this project, and a lot of the theory you’re talking about is pretty interesting. Once I realized the size of the transformer I was going to need, I pretty much knew that was the final nail in the coffin. I have done crazier than this, (citing my experiments with charging myself up to a few thousand volts using a Van De Graaf generator and shooting lightning out of my fingers) but this, as I realized, would have been REAALLY dumb. so everyone else, please, heed our warning.
If you’ve got a disposable power supply kicking around, why not hook it up to a CIM and see what happens? If it runs, then try two CIMS… then three. Don’t bother with a speed controller… just connect them to the rails. No sense risking a speed controller.
I mean… you’ve got some excellent suggestions on here about what is likely to happen, but hey… give it a shot. At worst you lose a power supply that you were going to throw out anyway, at best maybe you discover that this will actually work for unloaded motors. Record it on video… the ElectroBoom channel on YouTube had to start somewhere, eh?
If the power supply ISN’T disposable, or if it becomes eminently disposable following this test routine then why not solve the root problem? If you’ve got a battery shortage, then it probably makes sense to solve two problems at once and just buy a battery. They aren’t THAT expensive and it sounds like you need one or two more anyway.
Jason
P.S. For code testing, there is no reason to run CIM motors… maybe less powerful motors designed to work at 12V will show you what you need to see.
We are currently using a 12 V, 3 A laptop charger (not the more common 19.2 V ones) to power a RoboRio for initial code tests. This worked well for a previous team I worked with. We are now looking to add the Talon SRX motor controllers. We tried the throttle motors we had laying around but found that they would draw up over 0.5 A. Several of these would put the laptop charger into current limit.
We are now looking at a small digital panel meter but are having a hard time finding one that will accept positive and negative input voltages. We have also considered using a bidirectional LED with a series resistor and a filter capacitor connected across the motor controller output to give an indication of the output polarity and average output voltage but we have not had a chance to try it yet.
I want to meet ElectroBoom on one of my next visits to Vancouver
He’s one of Vancouver’s YouTube celebrities. Celebrated online, but little known at home. Can’t say I’ve ever met the guy, but I’d bring my eye protection and wear an insulated, grounded suit if I did!
I have to admit… I’ve always wondered exactly how he manages to pull off his “mistakes”. They look too authentic to be fake. I mean, sure… he could just have an unusually high pain threshold, but I don’t wonder if he limits the current or something when he’s going to shock himself so that he gets the shock, but not the full blast of the current. He strikes me as one of those guys who has to be really, really smart in order to appear that stupid.
Hmm… wonder if he’d do a demo at the Canadian Pacific Regional?
I would be very afraid of him becoming a “bad influence” on the younger members of this community. Maybe he can do a demo for adults only.
Mehdi clearly knows his stuff well enough to play the clown. I don’t doubt that he does some testing when he can. Some of it is just pure guts like when he rips part of his moustache out using wax strips then sets the rest of it on fire with a mirror behind so you get two views…
