Thanks to all for the ideas. Our team has done some research and found the battery puts out around 120A/1440W of power! If you were going to try to drive CIMs off of that with the rest of the control system, it looks like the cost is over $1K. Suddenly, lugging the battery around looks a lot more reasonable.
I really like the servo motor idea mentioned in the thread. That gives the programming team what they need without adding the extra expense.
Thanks again to all!
We have a desktop training system mounted on a 2’ x 4’ piece of plywood.
It has the cRio, the full electrical distribution system, pneumatics, window motors connected to the Jaguars’ etc.
It is powered by an Astron RS-12A power supply. It works find but I’d suggest something like an RS-35 or so.
As Al pointed out, do NOT screw around with electricity if you are not sure what you are doing.
The Astron supply is a rugged supply, UL approved and provides a reliable supply for what you are doing.
Go here and select ‘DC Power’ on the product type pulldown selector on the left menu bar.
M,
The battery is capable of much more than that for short periods of time. Short circuit currents run above 600 amps for full charge.
Thanks to all for the great ideas. In the end I found converters capable of the output of the 12V battery to be VERY expensive. Suddenly, lugging around a 12V battery seems reasonable. I think we’ll wire the system up for a 12V DC cell and convert over to a converter source if the programmers can use it without the more powerful DC motors. The servo idea is particularly interesting. Thanks again, Michael
If simulating the power supply of the battery is a concern with a power supply please keep in mind that you’d have to simulate the battery’s internal resistance characteristics.
Even if you could get a power supply that could withstand the surge currents that battery is cable of (I have some really big power supplies and battery chargers with high current engine start functions) chances are high that the output of the power supply will have different impedance than the battery meaning that the motors under various conditions might behave differently.
On a slightly different note if the issue is just constantly swapping batteries and since you don’t mind the cord perhaps you could rig up a charger for the battery with a switch / relay / contactor / MOSFET to disconnect and reconnect the charger more quickly or automatically in between using the robot.
Then the battery is no different than during competition, but the charging of that battery will be more consistent. Unfortunately then you’re tethered to the wall socket.
I’m reluctant to share this actually…
…but a blade server power supply could potentially be used for this.
They show up on eBay for well under $100 quite regularly, and push out over 175A at 12V.
David Edwards, a mentor for 1310, and I have thrown around the idea of modifying one of these to power a functional FRC robot.
But truthfully, I have never pursued it, because I simply don’t trust myself enough to modify something that puts out that much power.
If you have an experienced power electronics engineer at your disposal, then maybe you could safely give this a shot.
If not, please do not try this at home!
Using a single blade server power supply won’t work for a stangard FRC robot since A CIM can pull 132 AMPs at stall (which is where motors start at) and there are typically 4 of those plus other motors.
Sistering (pairing) these power supplies may or may not be possible depending on their overcurrent protection. Most overcurrent protection detects a high current (low output resistance) and shuts down the output, then will periodically turn back on the output and re-check for overcurrent. This results in a periodic pulse (normally called hiccup mode) on the outputs until the overcurrent is removed.
If there is no overcurrent protection on these supplies then pairing is possible using “or-ing” diodes capable of more than the supplies output (typicallly 50% or more). These work by having the power supply with the highest output voltage supply all the current until its output voltage sags to the level of the next highest supply, at that point both supplies share the load until they both sag to the level of the next supply (et cetera).
If a CIM is pulling 132amps wouldn’t PD 40amp breakers throw? And there is a 120amp breaker, so the robot should never pull more then 120 amps, and if it tires to wouldn’t the breaker just kill power?
Those power supplies look like the deal of the decade for high power applications, I might get one to run my battery charging array…
The breakers are thermal-limited to 40a or 120a.
This does not mean that any current greater than the trip point will instantly trip the breaker.
Current transients can heat the breakers for a short time without tripping them. For the 120a main breaker, this number is several hundred amps for many seconds - I seem to remember 200a for a minute, but I don’t have the spec sheet on hand. The 40a breakers can also take several seconds to trip under CIM stall load (which IS ~130a).
All motors will pull their stall current when starting (assuming supply voltage = spec voltage). This is because the current draw is proportional to torque, which is inversely proportional to speed. Since the speed is 0 when starting, torque is stall torque, so current is stall current. As the motor applies a force to the output, the output beings to move and the motor’s output torque and input current go down. The motor will still draw ~130a for a short period of time.
Assuming 4 drive CIM’s, then a forward/reverse launch would instantaneously draw ~520a.
Server power supplies are not designed for high current transient loads like motors. They are primarily designed for steady-state computing and cooling loads, which do not produce transient current draws anywhere near that of CIM motors. Unless I had data or specs to say otherwise, I would assume the spec output current is the peak output current, and a single power supply would not be able to handle the load of multiple CIM motors starting.
If one looks back at the problems the tan jag had. One of them was that running them on a stiff bench power supply would kill them fast. While the BLDC24 is more rugged all motor controllers can be damaged by running them on a ac to dc switching power supply. The problem is the switching transients and back EMF can cause the switching regulator to output some high short term voltages that can go above the voltage rating of the controller components. Motor controllers and big motor currents and ac switching power supplies are a bad idea.
*FWIW, per the data sheets:
40amp breaker less than 1 second at CIM stall current of 133 amps.
120amp breaker less than 6 seconds at 4x133 amps.
Has anyone ever actually used a scope to measure the inrush current spike on a CIM (both unloaded on the bench and loaded on a drivetrain)?
I have checked the currents but did not record the data. As I remember, the current spiked as expected and loaded currents were not far from predicted. I used a four motor chassis with one motor on each side providing dynamic loads to the other motor. We connected various values of resistance across the driven motor including a dead short to simulate loading on the driving motors.
From observation, few teams will actually trip any breakers. Several teams appeared to trip the 40 amp breakers this season and a few also reported tripping the main breaker this season. Making some assumptions on wiring designs, the “must hold” data for 250% would likely be the most accurate for general use. From the linked spec sheet that would be 1.8 seconds max.
I would agree with Gary and add that motor start currents would likely trip this supply on over current even with just a single motor.
A few years ago when Circuit City was going out of business I snapped up a couple of these power supplies for real cheap:
http://www.audioauthority.com/product_details/978-100
They are 100A 12VDC power supplies and are what places like Circuit City use to run their big car stereo demo boards. Two of these units can be synced together with a simple patch cable to generate 200A of 12VDC.
We have a test station setup in our programmers cave with a CRIO, PD board, digital side car, speed controllers, spikes, motors, and a test pneumatics system, etc. We have one of these Audio Authority 100A power supplies setup to power the test board. It is way overkill but it is what I had handy and works well. We have never needed to connect 2 together for the 200A but have everything to do so if the need arises. It is much more convenient to just flip the power switch on and not need to go get a battery out of the rack (but we can use that option as well if needed).
Given that Circuit City went under, there are a bunch of these things floating around out there in the second hand market. I have seen them on Ebay and Craig’s List.
A battery usually is the best answer. If you’re willing to invest a little money you can buy a smaller battery that’s a little easier to move around. Take a look around on McMaster for lead-acid batteries- there are a variety of options that use the same sort of non-spillable design that the official battery uses, but in smaller capacities/sizes.
The official battery has a 18amp/hr rating. Around 6 amp/hr would probably run a tethered control system and a few small motors for quite some time while being considerably smaller/lighter. Expect to spend somewhere around 30-50 dollars.
You can charge the small batteries using the same chargers you use for the big batteries as long as you can limit the charge rate- 1amp would be the safe limit for most cases. Smaller batteries also have lower safe discharge rates, so trying to run a full size robot could overheat and damage it- so don’t do that!
Just a note of clarification so as not to confuse students:
Batteries are rated in ampere-hours, not amperes per hour.
Come back to the motor lab one afternoon and we can do that. The duration of the spike will naturally depend on the load and the load inertia – probably just a few milliseconds when the shaft is free.
Invitation accepted. I’ll contact you to work out the date/time details 
We had a similar situation last year. We borrowed a Tripp-Lite PR-60 power supply from my company. We connected it directly to the robot in place of the battery when we were trying to dial in the shooter. I just looked up the PR60 and it costs new for about $282 from Tripp-Lite (http://www.tripplite.com/en/products/model.cfm?txtSeriesID=841&txtModelID=244). If you do not need 60 amps, they have smaller power supplies.
Our team did something like this to tune our shooter motors. We used a Tripp-Lite PR-60 power supply I borrowed from work. We connected it directly to the robot in place of the regular battery. We were only running the shooter motors (had the robot up on blocks to prevent any moving) and it worked great.
A new from the manufacturer PR-60 is about $280.
PR-60 Power Supply
If you do not need a 60 amp supply they have a 10 amp for about $80.
PR-10 Power Supply
The 40 amp snap action breakers are a thermal device. When the current is high the device heats up, when it gets to s specific temperature it breaks contact. Since it’s an automatic resetting breaker, when the temperature cools down it will reengage the contact. The 40 Amp specification is a requirement that it will NOT break contact below that current. Above 40 Amps it WILL break contact, but only after a specified time (at 132 Amps it’s just under 1 second).
Do not use the above power supplies to charge batteries. They do not have a current control and do not have an automatic shutoff.