It seems my TIP120 gets too hot for the touch (can't touch it for more than 2-4 seconds). The current should be around 1A. The drop voltage is about 0.6v-0.8v for TIP120. So TIP120 dissipates "just" 0.6-0.8 watts, correct? Is it OK to have it ON 24/7 without a proper heat sink with these parameters?
(I'll get MOSFETs next time).
the datasheet says 2W power dissipation @ta=25°C so it will be ok.
but I allways prefer to cool the power components to keep a good life time.
The RoJA of this TO220 transistor is 62.5°C/W
so for your application with 1A and 0.8v drop you will have:
0.8W * 62.5 + Ta = 50°C + 25°C = 75°C
the transistor can work up to 150°C so it's ok but make sure it was not installed near electrolytic capacitors (to avoid short life time) and think about the still air temperature rise!
the 0.8V is measured or from datasheet? because perhaps the transistor is not in saturation mode. But it's a darlington one with Hfe = 2500 so it's quickly in saturation with low base current.
Hi, have you measured the voltage drop from emitter to collector?
I'd put a heatsink on it.
Tom.... 
I feed it around 5.8v, and it outputs around 5.1v under load to the device. So I guess the temperature is bearable and is within limits if I can touch it for a few seconds.
I have a weird problem now. I'm using this transistor to power an USB HDD 2.5 that requires 5v and 1A. It gets its power completely from this transistor (I cut the 5v cable from USB that goes to HDD). It works fine, but after some minutes (at random, could be 2, could be 10) of intensive hdd work (encryption with Truecrypt) I get a crc error where Truecrypt says something is wrong with the device. The hard disk is fine. So I assume the problem is with current. I know there's lots of things that could go wrong, but could the issue here be something like the lack of ~220uf capacitor right before HDD?
I'm using 5v, 2A adapter. Then this thing to boost the voltage to 5.8v:
5v goes to Arduino pro mini's 3.3v raw pin. Connected TIP120 to Arduino with 470 resistor.
You must use a PMOS for this application to avoid big voltage drop 
in your schematic the collector is connected to the 5V and emitter to the usb power out as I understand.
if the base is connected to a 3.3V out of the board the transistor can't be correctly driven.
and to answer your question it's possible that's the voltage drop increase when the HDD is working due to it's increased power consumption.
Genesis92:
You must use a PMOS for this application to avoid big voltage drop
in your schematic the collector is connected to the 5V and emitter to the usb power out as I understand.
if the base is connected to a 3.3V out of the board the transistor can't be correctly driven.
and to answer your question it's possible that's the voltage drop increase when the HDD is working due to it's increased power consumption.
You must use a PMOS for this application to avoid big voltage drop
Well, I'm using the stuff I have right now, which is, sadly, TIP120. What do you mean to avoid big voltage drop? If I get 5.0-5.1v when the hdd is under full load, shouldn't it be fine?
if the base is connected to a 3.3V out of the board the transistor can't be correctly driven.
Why not? I thought it's fine as long as the resistor is low enough (470 in my case).
it's possible that's the voltage drop increase when the HDD is working due to it's increased power consumption.
So 220uf capacitor soldered across + - on the hdd wouldn't help this?
More info:
This USB HDD 2.5 case has an USB port and a some small 5v DC jack. I've removed the DC jack and soldered 2 wires there. The negative wire from HDD goes right into the collector. Positive wire from HDD is connected to positive output of that DC converter chip I mentioned from ebay.
Hi,
in your schematic the collector is connected to the 5V and emitter to the usb power out as I understand.
What schematic?
Please do not remove things from earlier posts, it causes a discontinuity that make following the posts very difficult.
Can you post your diagram, complete with USB mods and HD feeds that you have mentioned in earlier posts.
Also a picture to see how you have layed out the components will be worth a thousand words.
Thanks.. just trying to simplify things and make the posts understandable.
Tom.....
I didn't remove anything, there wasn't a schematic. Well, here's not a schematic but something close to it:
Hi, have you actually measured the current to the HDD.
Is there a reason that you need to power up and down the HDD.
Some bypass caps may help.
Tom.......
TomGeorge:
Hi, have you actually measured the current to the HDD.
Is there a reason that you need to power up and down the HDD.
Some bypass caps may help.Tom.......
It works for (2-20) minutes under full load, so I think it gets enough current. I can't measure it right now without breaking wires (it's soldered on the perfboard).
Some bypass caps may help.
So soldering a 220uf capacitor on the HDD (USB-SATA chipset) might help after all? Is 220uf enough? Too much?
Is there a reason that you need to power up and down the HDD.
I'll use the HDD for automatic backups. The software on PC will turn it on and off as needed to make backups. It's more secure and safe if this backup drive is completely disconnected from PC. I could, of course, connect it manually when needed, but this is just a fun project I want to make.
Hi, 220uF is okay, but I'd suggest a 0.1uF cap as well.
Tom.....
"I think your schematic is designed like this" as I've understood if you prefer
Energia:
You must use a PMOS for this application to avoid big voltage drop
if the base is connected to a 3.3V out of the board the transistor can't be correctly driven.
draw a schematic and use arrows to show the Vbe voltage (it's a darlington one) and I think you will see something wrong but only a schematic can confirm that point.Why not? I thought it's fine as long as the resistor is low enough (470 in my case).
you are sure? with 5V rail and driven by a 3.3V signal?it's possible that's the voltage drop increase when the HDD is working due to it's increased power consumption.
yes, and the 5.8V can decrease too because your power adapter is not perfect so the end voltage for the HDD will be lowerSo 220uf capacitor soldered across + - on the hdd wouldn't help this?
No, even if it help for the first peaks the current peak goes thru the transistor so even if the HDD use more current the capacitor must be recharged too....by the same transistor!look here updated Mega 2560 schematic - Other Hardware Development - Arduino Forum
and see how T1 a PMOS is used to switch the power supply in the schematic
NPN and n-channel MOSFETs are used for low-side switching, not high-side switching.
Darlingtons waste lots of power, no-one uses them any more commercially, they are
legacy devices really as MOSFETs are superior in performance in all counts. They are
popular because most modern devices are surface-mount only, whereas chips like
the ULN2803 octal-darlington are still cheap and DIP packaged.
For power-switching a 5V 2A load a logic-level p-channel MOSFET with on-resistance
below about 0.1 ohms (ideally < 0.03 ohms) would be suitable.
Another thing a device that theoretically can operate at 150C should not be run at
150C unless you use a ceramic-packaged version mounted on a high-temperature
tolerant board, and the expected lifetime is vastly less operating at that temperature.
In a plastic/epoxy package thermal cycling from room temp to 150C will cause mechanical
failure of the packaging/bonding wires at an accelerated rate - its a no-no basically.
The lower the operating temperature the better, basically.
I agree 
It seems my problem was related to the USB hubs (I have 3 connected "in series"; some are of questionable quality). I connected the HDD to PC (no +5v wire on USB cable), and finally managed to finish that 2-hour drive encryption without errors.
I also added a 0.1uf ceramic capacitor, but I doubt it was the reason for the fix as I've tried 220uf before that, and it didn't help. Also added a few tiny heat sinks to TIP120.
Thanks for your input.
Hi, good mate, but I must say, a picture of your setup could have pointed that situation out to us.
And probably solved your problem much earlier.
That many hubs in series is going to be looking for problems, with switching delays and power for the hub circuits.
If you need so many USB sockets, may I suggest a HUB that is independently powered, you can get up to 8 sockets on those models.
Tom..........
