PDA

View Full Version : looking for a power supply



John Stevenson
12-12-2009, 07:47 PM
I need 1.8 volts DC at 6 to 7 amps.

Any ideas on getting this, I only need one so continuity of supply isn't a problem.

I was wondering about an ATX computer power supply but have no idea what they put out.

.

MTNGUN
12-12-2009, 08:19 PM
Not sure if this one is DC or AC output.

http://cgi.ebay.com/PT5404N-POWER-SUPPLY-1-8V-6A-T-H-ROHS_W0QQitemZ110466743650QQcmdZViewItemQQptZLH_De faultDomain_0?hash=item19b854bd62

nheng
12-12-2009, 08:21 PM
John,

If you can handle a small surface mounted package, Linear Tech has their micro-module series. A chip/module like the LTM4616 puts out dual 8A supplies at 0.6 to 5 volts. Only problem is that the input is 2.7 to 5.5v.

You can purchase the eval kit online. #DC1245A is $65 and is basically a mounted device (proper thermal design, etc.) with all I/O and control lines brought out to human sized connectors. www.linear.com

1.8v is a common core voltage on various processors but I haven't seen any power supply modules at that voltage. It is usually generated locally as needed.

Spoke too soon ... here you go but not cheap: http://www.acopian.com/store/9-%287%29.aspx?min=1.5&max=5

MTNGUN put up a nice TI PT5404 module that is thru hole, similar to the Linear Tech but definitely easier to handle for a one up use. Same issue in that the input is low voltage dc.

Den

Weston Bye
12-12-2009, 08:23 PM
I need 1.8 volts DC at 6 to 7 amps.

Any ideas on getting this, I only need one so continuity of supply isn't a problem.

I was wondering about an ATX computer power supply but have no idea what they put out.

.

What for? .

2ManyHobbies
12-12-2009, 09:13 PM
ATX won't cut it, closest you could get is 1.7V and that is using 3.3V as a reference and 5V as a supply.

+3.3 Volts
+5 Volts
-5 Volts
+5 Volts
+12 Volts
-12 Volts

http://www.helpwithpcs.com/courses/power-supply-basics-inc-pinouts.htm

1.8V can be found on motherboards for memory controllers, etc, but that is probably a DC-DC bridge that runs on 3.3V or 5V. 1.8V@7A is going to see quite a drop over any distance. What kind of efficiency do you need? It would be easy to kludge something together that would feed from an ATX supply if you don't mind some heat loss and know more about what you are powering.

Forrest Addy
12-12-2009, 10:18 PM
My first concern is regulation? How close to 1.8 volts? Does the load vary? Is the load reactive or have characteritics that interfer with a regulated or a digital power supply? Will an analog supply do?

How about a variable dropping resistor from a computer 5 V supply? 3.2 volts drop at 8 Amps is 2.5 ohms @ 25 watts and at 6 amps the volts will drift up to 2.4 volts. Will that hurt?

Plating supplies are variable, low volts, and frequently reference from the tank busses. They might be a good first choice.

dp
12-12-2009, 10:42 PM
The forward voltage drop of diodes is reliably 0.7vdc. The collector/emitter voltage of a transistor in saturation is 0.3vdc, depending on the transistor. So if you wire two power diodes and an NPN power transistor in series with the base of the transistor connected to the collector you will have a voltage of 1.8vdc that is highly regulated. You then put a current limiting resistor between the collector and vcc, put your load across the active componnents, and turn it on.

The limiting resistor has to pass the full current of the load plus some waste current through the diode/transistor devices. This is a simplified circuit - a current limiting resistor is also needed in the base-collector connection.

http://thevirtualbarandgrill.com/machinery/LV-Supply.jpg

Astronowanabe
12-12-2009, 10:55 PM
What for? .


I suspect he is in a pickle ...

Black_Moons
12-12-2009, 11:21 PM
Id recommend this 100lb supply I just got at auction. Its rated for 0~150v and 0~12A. that fits your needs doesnt it? j/k.
For some more useful information...
www.allelectronics.com has verious supplys you could start off with an add a simple 1 chip SMPS.
Linear regulators keep amps the same, ie you'd need to start off with a 5v+ supply at 8amps.
SMPS incress amps at the expense of volts, ie they maintain wattage, more or less (70~90% is rather typical)
So, a 1.8V 8A SMPS would only need 5v at 4A~, and also get a lot less hot and need less heatsinks.

dp: Shunt regulators are not a great solution at high power. Try considering a series PNP whose base is controlled by the 1.2v.

dp
12-12-2009, 11:54 PM
Id recommend this 100lb supply I just got at auction. Its rated for 0~150v and 0~12A. that fits your needs doesnt it? j/k.
For some more useful information...
www.allelectronics.com has verious supplys you could start off with an add a simple 1 chip SMPS.
Linear regulators keep amps the same, ie you'd need to start off with a 5v+ supply at 8amps.
SMPS incress amps at the expense of volts, ie they maintain wattage, more or less (70~90% is rather typical)
So, a 1.8V 8A SMPS would only need 5v at 4A~, and also get a lot less hot and need less heatsinks.

dp: Shunt regulators are not a great solution at high power. Try considering a series PNP whose base is controlled by the 1.2v.

He's British - he's even worked around Lucas electrics - he's got to be used to this :). He's probably swapped out a few zener diodes on Triumph charging circuits, in fact.

And if the limiting resistor is chosen well there will be little shunt current needed. If 5v is the supply then the voltage across the resistor is 3.2v. 3.2v @ 8 amps is 0.4 Ohms which is just a few turns of wire wrapped around a beer can for a heat sink. Leave room for .25A through the semiconductors and you have 0.45W waste heat and the current limiting resistance is now 0.39 Ohms.

I presume the load is always attached so there is no need for the semiconductors to suck up the full 8 amps. Total power is now 40 watts or about as much as a refrigerator light. About 14 of those watts is in the attached device leaving 26 watts to heat up the beer can.

EVguru
12-13-2009, 04:35 AM
Look for a Vicor module. Quite expensive new, but they turn up surplus on ebay (usually in the US). Very reliable and with a wide output trim range, so a 3.3v or 5v module could be trimmed back.

Evan
12-13-2009, 05:07 AM
What for?

One cell of a lead acid battery is about 1.8 volts under load.

EVguru
12-13-2009, 05:13 AM
What for?

One cell of a lead acid battery is about 1.8 volts under load.

Not any of the lead acid batteries in my workshop!

They'll all hold 2+ volt well into the discharge.

Of course, they are all AGM lead acids (Optimas or Hawkers).

Evan
12-13-2009, 06:05 AM
I guess that depends on the load... :D

Ever seen Gates lead/acid D format gel cells? I wish I could find some of those at a reasonable price. Nothing beats lead acid for monster amps under short term load.

I currently (:D) have 2 18 amp hr VRLA batteries and 3 55ah VRLAs in my collection for EV experimentation. The 18ah batteries are rated for 248 amps max!!

Weston Bye
12-13-2009, 06:12 AM
How much regulation is needed? A variac feeding a 6 or 12 volt center-tapped transformer with a current rating greater than 7 amps, then feeding a full-wave rectifier and large capacitor would do it. Adjust the output voltage by adjusting the variac.

A non-center tapped transformer could also be used with a bridge rectifier and capacitor for less ripple.

A high current, low voltage transformer could be had from a battery charger.

Your Old Dog
12-13-2009, 06:46 AM
You re-plateing the bar stools or somethin? :D

EVguru
12-13-2009, 07:04 AM
Ever seen Gates lead/acid D format gel cells?

The Optima batteries are based on the Gates/Hawker licence.

I'll see your Gates 'D' cell and raise you a Boulder TMF 'C' cell; 1000 Amps for 10 seconds.

They didn't find commercial sucess in the end.

bob ward
12-13-2009, 07:23 AM
An LM138/LM338 voltage regulator is adjustable down to 1.2V DC output assuming a 12VDC input, and is rated at 5 amps continuous. A pair of circuits running in parallel will provide the amps you need. Apart from the regulators you need standard capacitors, resistors, pots, experimenters circuit board and a soldering iron. Its a cheap way around the problem.

Data on the voltage regulators and different circuits is here.
http://www.jaycar.com.au/images_uploaded/LM138.PDF

gld
12-13-2009, 09:59 AM
John
You did not say, you want to buy or build. If you want to build, you could modify the transformer winding in the home made spot welder (http://www.instructables.com/id/Homemade-Spot-Welder/) to get the voltage you need then add rectifiers and regulation to suite.

Just a thought.

psomero
12-13-2009, 05:19 PM
ATX won't cut it, closest you could get is 1.7V and that is using 3.3V as a reference and 5V as a supply.

+3.3 Volts
+5 Volts
-5 Volts
+5 Volts
+12 Volts
-12 Volts

http://www.helpwithpcs.com/courses/power-supply-basics-inc-pinouts.htm



ATX v1.2 and later don't have a -5V rail any more. only the ISA interface bus used it and since it was being phased out at the time, they stopped putting the -5V in PSUs.





To op: get a laboratory power supply. they can usually be found in 0-30V adjustable voltage output, but i'm not sure you're if going to find one with that much current output (for a good price).

I'd say get a regular fixed voltage 13.8V dc power supply, which are available in like 10-20A output for around $100,

then get a voltage regulator like a LM317 in a TO-220 package. you can get adjustable ones that will do like 0-15V based upon what resistor values you connect to them. they usually only can handle 1.0-1.5 A each, but they can be had for around $2 a pop and if you put them in parallel you can get enough juice for whatever it is you're doing.

by the way, what the heck are you doing?

Black_Moons
12-13-2009, 06:23 PM
The Optima batteries are based on the Gates/Hawker licence.

I'll see your Gates 'D' cell and raise you a Boulder TMF 'C' cell; 1000 Amps for 10 seconds.

They didn't find commercial sucess in the end.
I'll see your C cell and raise you a liquid sodium battery.
http://en.wikipedia.org/wiki/Molten_salt_battery
It might not have higher peak currents, but it is one of the coolest battery chemistrys I know of :)

J Tiers
12-13-2009, 08:24 PM
Yah, no good suggestions unles we know how good the 1.8V needs to be.

darryl
12-13-2009, 11:23 PM
Use a computer power supply to start with, since the 5v output will be capable of many amps, more than you need. The ps will have at least one negative output, so you can arrange a fairly simple circuit using a fixed regulator driving a power transistor on a heatsink to get your 1.8 volts. the ground lead on the regulator is put to a voltage divider coming from the negative supply, so the regulator still thinks it's putting out 5 volts.

More easily perhaps would be an adjustable regulator driving the power transistor. I think between the voltage drop from the power transistors base-emitter junction and the minimum over-voltage requirement of the regulator, you would arrive pretty close to the 1.8 volts, with no negative voltage needed to make the circuit operate.

There are several ways of downgrading the voltage from an existing power supply, most requiring that you waste the difference in voltage from in to out, times the current drawn by the load. In this case, it would be 3.2 volts difference (5 - 1.8) at the rated current, lets say 10 amps at most- that's a wasted power of 30 watts or so. That's not hard to handle with a single power transistor on a large enough heat sink.

If you used a single high current PNP transistor as the main element, you can maintain the full regulation capability of the adjustable regulator chip. If you used an NPN, you would lose some of the regulation, but the transistor would be more rugged.

Either way, if you mounted the (simple three terminal) regulator chip right opposite the transistor on the heatsink, you can take advantage of the over-temperature protection built into the regulator.

This is a very simple circuit. It would probably be as much work to mount some terminals for input and output as it would to solder the parts together to make it work.

You still have the option of using a power resistor to handle much of the current so your pass transistor wouldn't have to dissipate as much power. In this case though, you'd have to maintain a load on the output at all times, or the output voltage would rise when the current draw went below a certain point. At 6 or 7 amps output, the heat loss in the regulator is going to be no more than (7 amps x 3.2v ) = about 22 watts. Easy to deal with, easy circuit, no power resistor to fool with-