power supply

[nilcum]

Hey, I've been trying to figure out a good way to do power.

The problem we have is that if we have just 1 power source and 1 device active that device gets ALL the power. When two devices come on they each get 50%, with 4 each device gets 25%. If we think of this in terms of voltage, with 1 device and a 9v battery, that one device will have all 9 volts. This is quite a load for a little DC motor. If there are two vibes, each device will have 4.5 volts. With 4 vibes each will have 2.25v. By the time we get to 10 devices each one will have less than 1 volt available from a 9v battery.

This becomes a bigger issue when you get in to devices with more sensitive power requirements; like a shocking device.

So, the problem we need to solve is how to control the power to every device so it has a constant 3v. This is typically how many batteries a bullet vibe requires.

With 10 devices we need a total of 30v available if each is to have 3v. That means we will need to step up the 9v battery. In the process we will end up draining the battery faster so, a wall power source might not be a bad option.

Next, we need a way to ensure that when only one device is activated it does NOT have 30v available to it. The means we need a regulator of some kind. I am a bit inexperience with these, but I’ve been doing some reading. Turns out you can make one from a Zener diode and a resistor. We could simply add one of these for each device. Maybe we could find an IC that has 10 already there. Haven’t found one yet.

The thing I’m doing now is to have a separate power supply for each device. I have a bunch of 2AA battery packs that each device gets. It takes up a lot of space and is sort of a poor implementation.

Another option is to handle all of this programmatically. We could degrade the frequency of the PWM given the number of devices currently active. These calculations don’t sound like a lot of fun and there is no guarantee that a user actually has a device in to draw current.

There are also variable voltage regulators that can be configured for 3v with a resistor. We could do one of these for each device too. Should be more efficient than a zener diode setup. I looked for some that were configure for 3v and only found 1 setup already.

Then there is perhaps a more eloquent solution. We can use the variable regulator and use arduino itself to switch the power. Meaning that we have one step up from 9v to 30v and use some of the other I/O pins with a transistor and a resistor to add those resistors in parallel to the variable regulator. Then we just have to count the number of devices turned on and select the correct resistance. There is still no guarantee that a user has a device plugged in to everything that is actually turned on. Programming it is easier though. The biggest down side is that we would have to have two IO pins per device.

Anyway, let me know what you think. I’m leaning towards a step on 9v to 30v then individual regulators for each device.

[Elle]

Hi,
Certainly if you have a given amount of power you are limited to that power. I'm not sure what you are trying to do exactly, but I guess that you are trying to supply power to one or more devices through a computer interface.
I can't help with the computer part but I will say that it is likely that if you are using several devices then they will be hooked up in serial or parallel. Each way requires its own consideration.
I don't mean to be overly simplistic for those who are knowledgeable with this but the post suggests an unfamiliarity with the subject.
Its like the Christmas lights problem. If you have a 2 wire system of lights then its set up in series. If you have a 3 wire system then its set up parallel. It its in series then the current is the same for all devices. If its in parallel then the voltage is the same for all devices. The 2 wire system is the one were if one light goes out all the light go out.
If you want to control one or more devices with the same source (read that as power source) then typically it is best to hook them up in parallel. Then each one will have the source voltage available but draw its natural current. Its natural current is dependent on its resistance. You can control the voltage to each device with a regulator in series with it, but typically it is best to decrease the voltage rather than increase it.
There's a lot more to this particularly if you are using AC power but I don't want to bore anyone.
If this has been helpful let me know and we can discuss it more. Otherwise good luck with your project.

[Jaberwocky]

I go with Elle. I've tried to make a schematic picture. (those white empty words are my way to create blanks)

9 VOLT, with the devices in parallel: Amount of voltage remains stable over every device, and is nine volt. (the current is not the same in every device)
+ batery -
| empty |
| empty |
| device |
| empty |
| device |
| empty |
| device |


9 VOLT, with the devices in series: The voltage over each device is devided. Equally over each device if the resistance in each device is equal. If not, the device with the highest resistance, will have the voltage over the device. (the current is the same in every device)
+ ----------- batery -------- -
| emptyemptyemptyempt |
| emptyemptyemptyempt |
| device -- device -- device |

[les]

Nilcum
Do you know the voltage and current requirements for each device that you are trying to control?
Indivdual regulators are the way to go but may need to custom for an individual, or range of outputs.
With the above information I may be able to help in that direction.
Seezya Les

[nilcum]

So, here is how things are setup. The USB link from your computer powers a microcontroller. The amperage from USB is really too small to be used for a vibrator -- let alone 10.

The microncontroller has a series of digital io pins which are used to trigger transistors that close a circuit with a vibrator attached. The only two components in this circuit are the vibrator and power source. 0-10 of these circuits may be switched on at any given time. We are using PWM from the microcontroller to vary the voltage going to each device.

The problem we need to solve is how to have BOTH constant maximum voltage and amperage to each individual circuit with a single power source. I'm sure this has been done before -- I'm just not sure of the best way to go for this project.

The other problem that makes things a little more difficult is that vibrators are basically electric shorts.

Thanks for the posts so far!

[Executive Angel]

He just said vibrators are What? Electrical shorts....OMG this answers many questions, but also brings up more questions than there are answers !

[Elle]

Nilcum,
You need to determine the maximum voltage and current for each of your devices. Since it is likely that each device needs the same voltage (3Vdc) half of your problem is solved. Otherwise your power supply needs to be able to supply the maximum voltage. You can then trim the voltage to the other devices with a potentiometer.
The current problem may be a bit tougher especially if vibrators are really electrical shorts. They can't be true shorts of course, otherwise they would load down your supply so much that voltage would drop to nothing. I suggest that you use current limiting resistors in series with each device. This will keep your supply from being loaded down but you will have to compensate for it with a higher voltage. It will also make it easier to add up the currents of each branch for your total current. The down side is that you may not be able to operate the devices at their max.
You may already know this but your supply should have operating margins of about 30% (+/- 10%). Meaning if each device draws an amp and you have 10 of them then your supply should have a max current of about 13 amps. If you run it too close to its limits for too lone it may overheat and fail.

Hope this helps.

[nilcum]

Well, wiring in parallel seemed to help. There was still a noticeable drop in power delivered to the vibrator when new ones turned on but it was MUCH better than when wired in serial.

Anyway, with all this juice running to these vibes I've discovered a new problem. Specifically, my IC darling transistor array was overheating an performance was degrading. Looks like it could only handle 500mA. I did a bit more research and found lots of good reviews of a TIP120. They have a bigger form factor with a heat sink but can handle much more amperage.

[t.carey83]

Hi:

I've been lurking here for a while, but just joined up.

For the power supply I would recommend using a cheap PC ATX power supply. Even a 250/300 watt power supply can provide 25+ amps at 3.3V (which should power a lot of vibrators, etc.) and also provides high amperage +5 and +12. You can't build a better power supply for the price. You will need to connect the green wire on the motherboard connector to any black wire (i.e. ground) to get the power supply to turn on without being connected to a motherboard.

With regards to your power problem with using darlington pairs to drive the motors/vibrators, rather than adding a TIP120 or other high power bipolar transistor, I would suggest using logic-level N-channel MOSFET's. You can get small three-leg devices or use smaller surface mount parts and easily handle the current you need without any heat sinks and they can be driven directly from the micro-controller. The problem with the bipolar transistors like your darlington pairs and the TIP120, is that current output is proportional to current input. So you waste a lot of power and also the bipolar transistors aren't that efficient and you will generate a lot of heat. The FET's on the other hand have a very low resistance when fully on and don't get hot at all.

I would not recommend using current limiting resistors inline with the vibrators, since you will just end up with some huge ass high wattage resistors that will generate a bunch of heat and do nothing. The motors in the vibrators will limit the current just fine. The motors/vibrators while low in resistance are not shorts as they may seem.

[nilcum]

Hey t.carey83, glad you decided to post!!! Welcome.

Using a computer power supply is a great idea. I even have a few of them laying around -- perfect. I had purchased a variable DC converter. Sort of worked, but when connected to something like an electromagnet the resistance was so low that it went over the max amperage and would turn off... then back on.. then off.

I'll take a look at logic-level N-channel MOSFET's. Do you suggest any particular model? Thanks for the suggestion.

[Jabber]

I'll take a look at logic-level N-channel MOSFET's. Do you suggest any particular model?

I just did some experiments with this. At first I built the motor control circuit at Opendildonics, and it worked fine. I was able to drive my 3.3v, 100mA micro-bullet vibrator at full speed, using my Arduino as a PWM source.

But looking at the circuit I decided to make a few changes to keep the cost down: First I removed the protective Schottky diodes D1 and D2, as they are not really needed. Just remember to connect ground from the PWM source to the MOSFET source before you connect the PWM signal to the high-impedance MOSFET gate. And there was no need to use a relatively expensive Schottky diode for D3, so I replaced it with the cheaper 1N4004. Finally I replaced the power MOSFET IRL-520npbf with the cheaper BS170 in a TO-92 case.

And it all works fine with these changes. The BS170 MOSFET doesn't even get warm when running the vibrator at full speed for a long time. And I am able to power it all (Arduino, driver circuit and vibrator) from the USB port in my PC.