Power From The Wall

[tuXguy15]

So our team so far in the off season has built several trainers for future members of the team. Our next goal is to run these trainers off of a wall outlet instead of a battery. I was wondering if anyone has built a device like this and tried it out? I know its just a power converter since its AC to DC but I'm not sure what else is involved if you guys have any tips for me.

[Alan Anderson]

Can you describe what these "trainers" are? How much current do they require, at what voltage?

[Mk.32]

Can you describe what these "trainers" are? How much current do they require, at what voltage?

The problem with running a robots/similar things off a AC-DC power supply is the spike in amperage (from motors) most psus cannot handle. I have some 1000w server power supplies I modded for use in bench testing electronics and running battery chargers.

[evanperryg]

There are ways to modify computer power supplies for custom use. If you're doing a lower power setup (i.e. control system and a couple small motors at most) you could be able to use a normal power supply out of an old OEM desktop. Regardless, for a just-good-enough system, look for a 350W+ supply, preferably made in the last 6-7 years. I have done this in the past with a couple of different OEM supplies if you are interested in how to do it.

[cadandcookies]

Can you describe what these "trainers" are? How much current do they require, at what voltage?

Seconded. I'm sorry, but "trainer" is not a standard term to my knowledge in FIRST. Even if it's a commonly used term where you are, I would sincerely appreciate it if you would explain precisely what you are referring you when you talk about a "trainer." I am relatively confident that you do not mean a shoe.

[Jared]

If you're just powering electronics, then a 12 volt supply good for a few amps should be just fine. I don't know exactly what type of power supply the PDB has for the cRIO, but I know some supplies designed for batteries don't like to be powered from another supply, so power your cRIO-II directly from the 12 volt supply.

If you're powering a drive base, then no, not unless you have a really expensive supply. At stall, you can draw a few hundred amps at 12 volts. Even if you did have a supply, it is likely that a robot's sudden current spikes could trip the typical 15 or 20 amp breaker that powers the wall outlets.

[DonRotolo]

I'm going to make a huge assumption here and guess you built a few basic robots that will be used to train drivers.

Can you run them off a wall outlet? Yes. As you almost said, what you want is a "12 volt power supply". Google it.

There are some considerations though:

1. A power supply with sufficient capacity to supply a typical FRC robot would need about 120 Amperes of current output. Such a power supply is likely to be VERY expensive.

2. You CAN put a battery into the trainer robot and also connect a power supply with a current capacity of maybe 10 to 20 Amperes. This will allow you to run a bit longer on that battery. But not forever. And not a LOT longer, maybe double or triple the battery-only time.

3. You can't leave the power supply connected to the battery if the robot won't be used for a while (more than 20-30 minutes), because it acts like a charger but isn't really set up like one - it can overcharge the battery, causing damage to many things.

4. Wires from the power supply to the robot will interfere with driving. Just like an Ethernet cord would do.

5. Running longer heats up motors more than when you give it a rest to change the battery, Be sure to manage the thermal aspects of your robot.

Another alternative is to get a larger deep-discharge type battery (sometimes called a Marine battery). It is not FRC legal, but it is a 12 volt battery and will have a far larger capacity than an FRC battery, allowing you to run quite a bit longer.

In the future, a clearer, more detailed question would help us help you better.

[tuXguy15]

There is an electrical trainer and a pneumatic trainer. The electrical trainer's highest load comes from the mini-cims without any load on them so the current draw is minimal. The pneumatic trainer's largest load is going to come from the compressor. I dont think current will be an issue with the electrical board, but I do think it might be an issue with the compressor since that does occasionally draw a lot of current.

[sparkytwd]

An ATX power supply should give you what you want. You can find several ATX to banana post breakouts online:

http://www.seeedstudio.com/depot/ATX-breakout-board-bench-power-supply-p-1222.html

If 15 amps is your power budget, you're pretty much set with any $30 supply you can find online.

If your budget is higher than that, you'll need to check the specs for the supplies, which should be easy to find as the beefier video cards require certain amounts of power delivery.

If you need even more than a single PSU rail can deliver, you can find multi-rail power supplies. In that case you need to remember you can't connect the 2 rails in parallel, they must be hooked up separately. Switch mode power supplies without explicit support for synchronous power conversion don't play well with each other.

[protoserge]

OP, no one can help you solve your problem until you clearly describe what you are trying to do. What is your definition of trainer? What is the quantity of motors in the trainer? How are they to be used?

If you are trying to run a robot as a trainer for new students, then what exactly precludes the use of the FRC battery pack? What about the battery limits you from using it as a reliable, representative solution? Part of training could include the limits of the FRC system, and battery life absolutely can be one of them.

[jimbo493]

Maybe it would be helpful if you posted a few pictures? Are these just platforms which you have but, say, a compressor, some storage tanks, and a few solenoids?
Or for the electrical a Full FRC layout? aka Main breaker, PDB, Motor controllers, some motors, maybe a sensor or encoder? Personally, I would see how having a power for the pneumatics "trainer" could be helpful, but I think its best to leave electrical as a battery, you probably won't use an entire battery in a day, and it helps students see the limits of the system, ect.

[adciv]

To do this "cheaply" What you would want is an unregulated 12V power supply. Basically, you'd parallel the output of the power supply with an FRC battery. The battery would handle the peak loads and the PSU would handle the base load & recharge the battery.

Theoretically, you can use a standard 12V automotive battery charger with this, although I'm not sure what current rating it would require. I'd recommend at least a 20A charger, it really will depend on your base load and your drive train.

[cgmv123]

To do this "cheaply" What you would want is an unregulated 12V power supply. Basically, you'd parallel the output of the power supply with an FRC battery. The battery would handle the peak loads and the PSU would handle the base load & recharge the battery.

Theoretically, you can use a standard 12V automotive battery charger with this, although I'm not sure what current rating it would require. I'd recommend at least a 20A charger, it really will depend on your base load and your drive train.

A 20A charger has the potential to do significant damage to both the battery and the power distribution board.

[Toa Circuit]

Older computer PSUs can output 20 amps of 12 volt power for cheap if you know how to hack them. Check your local junkyard.

You could possibly take an old UPS and use it. Just draw off of the battery terminals. The battery will act as a high capacity capacitor of sorts, so you can achieve high amperage output, but not have to worry about swapping in and out batteries.

But of course, all of this depends on your exact needs- how many amps you need, if that is continuous or not, and how much of ripple is acceptable.

[adciv]

A 20A charger has the potential to do significant damage to both the battery and the power distribution board.

I disagree that it could damage the battery, although I understand why people say that. But damaging the power distribution board? How?

[AdamHeard]

There is an electrical trainer and a pneumatic trainer. The electrical trainer's highest load comes from the mini-cims without any load on them so the current draw is minimal. The pneumatic trainer's largest load is going to come from the compressor. I dont think current will be an issue with the electrical board, but I do think it might be an issue with the compressor since that does occasionally draw a lot of current.

I would swap the motors for lower draw ones to make this easier.

From there, you can probably get away with surplus server power supplies.

Similar suggestions to PC ones above, but higher power, cheaper and easier to modify.

http://www.tjinguytech.com/my-projects/HP47A

is a good resource.

[cgmv123]

I disagree that it could damage the battery, although I understand why people say that. But damaging the power distribution board? How?

The manufacturer's recommended rated is ~5.5 amps. Charging the battery at 6 amps is pushing it, but allowed because virtually all chargers that are rated for 6 amps only put out 6 amps for a very limited time. Charging a battery at a constant 20 amps is asking for trouble. ([1] (http://www.chiefdelphi.com/forums/showpost.php?p=1366925&postcount=29))

As for the PDB, the short answer is that all chargers are not created equal. Some chargers will cause more damage to the PDB power supplies than others, since the supplies are more sensitive to voltage variances than batteries. ([1] (http://www.chiefdelphi.com/forums/showpost.php?p=1273320&postcount=14),[2] (http://www.chiefdelphi.com/forums/showpost.php?p=1272965&postcount=12))

[adciv]

Here's what you're not taking into account. With a 20A charger in the system, you're not charging the battery at 20A. You're using the 20A charger to supply the base load of the robot. When the power demand is for greater than 20A, the battery picks up the load. This will discharge it slightly, and once the load has been reduced, it will recharge at a low rate. The only way you exceed the 5.5A rate is if you have started to drain the battery. This could also be solved by using a motorcycle battery in place of the standard FRC battery.

As to the voltage spikes, that's an engineering problem which can be solved a few different ways. One way would be making sure you don't use that kind of charger, another would be to increase the capacitance in the system to dampen the spikes. I'd like to know the load on those chargers when the spikes were recorded.

[Ether]

There are ways to modify computer power supplies for custom use... I have done this in the past with a couple of different OEM supplies if you are interested in how to do it.

I'm interested. Have you got something written up that you could post or link to?

(Not sure what to do with all the wires; see attached pic)

[protoserge]

I'm interested. Have you got something written up that you could post or link to?

Here's a couple:


http://www.instructables.com/id/Convert-A-Computer-Power-supply-to-a-Bench-Top-Lab/
http://makezine.com/projects/computer-power-supply-to-bench-power-supply-adapter/



I've found one (http://tuxgraphics.org/electronics/201005/bench-power-supply-v3.shtml) that is variable and uses a PWM circuit vs. the typical LM317 adjustable regulator. Here's another (http://hackaday.com/2012/07/17/atx-benchtop-conversion-retains-safety-features-delivers-plenty-of-current/) one.

Now that I've been reminded, I have an old 480W PSU to repair (solder in some new 2200 uF caps) and convert to a bench power supply. It's a little more wallet-friendly than this (http://www.amazon.com/RIGOL-DP832-Outputs-Programmable-Supply/dp/B00ENX02GC/ref=sr_1_1?ie=UTF8&qid=1406117922&sr=8-1&keywords=rigol+power+supply)

[evanperryg]

I'm interested. Have you got something written up that you could post or link to?

There are a number of instructional videos on how to convert ATX supplies to benchtop supplies, but this is the one I used: https://www.youtube.com/watch?v=z2oSFpKh_Uw

I added LEDS to the "power good" and "5v standby lines" as status indicators. Not necessary, but it's helpful at times.

SIDE NOTE: I suggest you avoid Dell OEM supplies. They sometimes have different color codes.

[AdamHeard]

I'm interested. Have you got something written up that you could post or link to?

(Not sure what to do with all the wires; see attached pic)




I would say that unless you need the 5V rail, the use of server power supplies I linked are much more cost effective.

[evanperryg]

I've found one (http://tuxgraphics.org/electronics/201005/bench-power-supply-v3.shtml) that is variable and uses a PWM circuit vs. the typical LM317 adjustable regulator.

I doubt the LM317 would be capable of putting out the amount of current required for OP's test bed. The cRio alone is rated for 800mA, which is half the LM317's nominal current output. Add a sidedar and some motors, and you're well over the 317's 1.5A rating.

Here's some pics of my conversion. I've done a couple others, and they all work very well.
https://drive.google.com/file/d/0B4POpKC2IAD9LTlSeXFrNUZnVXFOeHRuazhJdUpNTkprUlpz/edit?usp=sharing
https://drive.google.com/file/d/0B4POpKC2IAD9Ylp0bHpvV2piNE40dXNnSHlHa2VFVjNDWlJz/edit?usp=sharing
https://drive.google.com/file/d/0B4POpKC2IAD9ai01UHZnWURSTkE3a3JXWDhSRWRKamFtYUhr/edit?usp=sharing

[protoserge]

I doubt the LM317 would be capable of putting out the amount of current required for OP's test bed. The cRio alone is rated for 800mA, which is half the LM317's nominal current output. Add a sidedar and some motors, and you're well over the 317's 1.5A rating.

Here's some pics of my conversion. I've done a couple others, and they all work very well.
https://drive.google.com/file/d/0B4POpKC2IAD9LTlSeXFrNUZnVXFOeHRuazhJdUpNTkprUlpz/edit?usp=sharing
https://drive.google.com/file/d/0B4POpKC2IAD9Ylp0bHpvV2piNE40dXNnSHlHa2VFVjNDWlJz/edit?usp=sharing
https://drive.google.com/file/d/0B4POpKC2IAD9ai01UHZnWURSTkE3a3JXWDhSRWRKamFtYUhr/edit?usp=sharing
Ether didn't ask for one capable of powering a robot ;)

That's a good point in case someone wanted to use it though.

[evanperryg]

Ether didn't ask for one capable of powering a robot ;) That's a good point in case someone wanted to use it though.

True :D If you're doing something smaller, a LM317 should work fine, just throw a heatsink on it (and, in a cinch, mayonnaise makes an effective thermal paste). The supply pictured is a 270W OEM supply from an early 2000s IBM desktop. I have also converted an off-the-shelf Antec 350W supply, and another OEM supply I had sitting around. All of them work, and I prepared all three in the exact same way, except for the dummy load resistors, which vary due to supply size/what I had sitting around.

[Chadfrom308]

The CNC Lathe at my work has a 20hp vector drive amplifier that powers a DC brushless spindle servo. The problem is that it takes 3 phase 240 volt ac power and cost $5,000.

In realistic terms, maybe try getting as big of a battery as possible? maybe a car battery or a really big lipo (very expensive)?

The thing is, batteries running out of juice isn't necessarily a bad thing. The motors get really hot (especially because they are hauling a 150lbs robot around), the compressor needs to catch up (it only has like a 20% ish duty cylce) and it can give you a way to take turns.

[Ether]

I would say that unless you need the 5V rail, the use of server power supplies I linked are much more cost effective.

It's just that I already have a half dozen PC desktop power supplies that I harvested and have been wanting to experiment with them, and possibly find a use for them (other than as a door stop).

[evanperryg]

In realistic terms, maybe try getting as big of a battery as possible? maybe a car battery or a really big lipo (very expensive)?

The thing is, batteries running out of juice isn't necessarily a bad thing. The motors get really hot (especially because they are hauling a 150lbs robot around), the compressor needs to catch up (it only has like a 20% ish duty cylce) and it can give you a way to take turns.
Converted ATX supplies are very realistic. They're designed to provide very consistent power for very sensitive components. The motor will only get severely hot if there is a significant difference between the forward voltage and back EMF. At free speed, the back EMF should match the forward voltage, meaning the motor won't heat up significantly. In other words, OP, don't put any force on the motor, let it spin freely.

[Ether]

SIDE NOTE: I suggest you avoid Dell OEM supplies. They sometimes have different color codes.

Bummer. I was going to use the Dell (pictured in earlier post (http://www.chiefdelphi.com/forums/attachment.php?attachmentid=17194&d=1406118108)).

Do you know of a resource anywhere for decoding the Dell color scheme?

The motherboard is long gone so I can't trace where the wires were going.

[Michael Hill]

Bummer. I was going to use the Dell (pictured in earlier post (http://www.chiefdelphi.com/forums/attachment.php?attachmentid=17194&d=1406118108)).

Do you know of a resource anywhere for decoding the Dell color scheme?

The motherboard is long gone so I can't trace where the wires were going.




I would imagine the pinouts are still the same since it will conform to the ATX standard. if you have the molex connector on there, check out http://en.wikipedia.org/wiki/ATX#Power_supply

Just beware that ATX supplies can be rather noisy (electrically speaking)

[evanperryg]

I would imagine the pinouts are still the same since it will conform to the ATX standard. if you have the molex connector on there, check out http://en.wikipedia.org/wiki/ATX#Power_supply

Just beware that ATX supplies can be rather noisy (electrically speaking)

Dell broke the color code with a number of models. Not all of them are like this, but many models are. I don't know the extent of the changes, but with some research you could use them.

[Michael Hill]

Dell broke the color code with a number of models. Not all of them are like this, but many models are. I don't know the extent of the changes, but with some research you could use them.

yeah, but I'm sure they still stuck with an ATX compliant motherboard meaning the pinouts should still be the same even if the colors are wrong. A DMM would be a true test.

[AdamHeard]

yeah, but I'm sure they still stuck with an ATX compliant motherboard meaning the pinouts should still be the same even if the colors are wrong. A DMM would be a true test.

Years ago I did a dell power supply in this fashion w/ the wrong color codes and your logic. Worked out just fine, as you suspect.