Regulated power supply

I have made and tested chip amp from 3886 by National Semi, and tested it with 20V-0-20V 100Va transformer.

Actually now i got a transformer rated @35V-0-35V rail, but after bridge i get 50V DC rails.

How can i use a regulator for getting 35v. I dont wanna go in for linear regulator design as it will have quite high losses & heat dissipation.

I need a schematic design of SMPS, using 3525/3524 and mosfets, to get regulated +-35V 4-5Amps
 
h_andree said:
I think the LM338 can get hot. I have them
regulating 38V down to 29V. I've attached
them to a 2" heatsink. You can touch the
heatsink but not too long.

If you can't keep your hand on the heatsink then you need more heatsink or a fan.

A hand's touch is usually a good judge when a part is being stressed by heat.

If you don't pay attention to that, the part may have quite a short life.


Carlos
 
mateo88 said:
An lm338 or lt1084 are surprisingly efficient and don't require a big heatsink, either. I would reccomend giving one of those a shot, since the circuit will be very simple and very cheap (and very good, too!).
This type of regulator is UNEFFECIENT by nature and can go down to almost 0% effeciency. How much heat you must burn away has to do with the load and how much voltage drop you have.
 
carlmart said:


If you can't keep your hand on the heatsink then you need more heatsink or a fan.

A hand's touch is usually a good judge when a part is being stressed by heat.

If you don't pay attention to that, the part may have quite a short life.


Carlos


Calos,

I you can touch it but not hold it, it is probably close to 50 degrees Celcius. Nothing to worry about!

Jan Didden
 
peranders said:

This type of regulator is UNEFFECIENT by nature and can go down to almost 0% effeciency. How much heat you must burn away has to do with the load and how much voltage drop you have.


Corretc. The dissipation in Watts depends on the product of voltage the regulator has to drop times the current it is passing. Different implementations have different dissipation requirements. One man's cool chip can become another man's burning hell ;)

Jan Didden
 
janneman said:

I you can touch it but not hold it, it is probably close to 50 degrees Celcius. Nothing to worry about!


You are certainly right. I am probably exaggerating in my heat concern on electronic parts.

Some say a bit of temperature is good for the sound.

Bottom line: it's important to see if the heatsink metal close to the part is about the same temperature as the part itself, to see if the heat transfer is being properly done.


Carlos
 
carlmart said:
[snip]Bottom line: it's important to see if the heatsink metal close to the part is about the same temperature as the part itself, to see if the heat transfer is being properly done.

Carlos


Yes, that's a very good tip! But beware: for a plastic part, like TO220, the exposed top may be much hotter than the metal tab connected to the heatsink, even if the mechanical connection is good.

Jan Didden
 
Linear regulators are very inefficient, and would heat more that the amplifer itself

SO...............
Here's my plan fo regulated high current & voltage regulator
using PWM Chip and using IR2110 + Nchannel Mosfets for Pos & neg voltage regulation.

I need to know exact operation and connections using the IR2110.

Would try to scetch it on a paper first.
 
crissty said:
Linear regulators are very inefficient, and would heat more that the amplifer itself

SO...............
Here's my plan fo regulated high current & voltage regulator
using PWM Chip and using IR2110 + Nchannel Mosfets for Pos & neg voltage regulation.

I need to know exact operation and connections using the IR2110.

Would try to scetch it on a paper first.


I have to join P-A to warn you. if you cannot even get the basic connections yourself, chances of succes are no more than -60dB. Not to discourage you, just to make sure you know what you are getting yourself into.
There is MUCH more in the successful development of a SMPS than is shown in the schematioc.

Jan Didden
 
Jan Didden,
I hope you also had a time when you learnt about what is MOSFET DRIVER.

Now is the time for me to learn this.

But one has to learn something sometime, might be it would take some time to get over it, especially if people like you all would support and help, new ideas would be fullfilled.

Do you think the idea i have thought is wrong, and is not possible.
 
If you check around you'll see no good DIY SMPS projects. Why? Much development is needed and also special parts. I'm afraid but this you'll have to forget unless you have the capacity yourself in designing.

peranders,
Its all about Improvements, so keep on improving

That is what we can do here, with each others ideas & experiences.

Some good circuit design for motor controller would serve as a regulator, with some additional circuit (like error correction etc...)

Correct me if I am wrong.
Any comments.......
 
crissty said:
Jan Didden,
I hope you also had a time when you learnt about what is MOSFET DRIVER.

Now is the time for me to learn this.

But one has to learn something sometime, might be it would take some time to get over it, especially if people like you all would support and help, new ideas would be fullfilled.

Do you think the idea i have thought is wrong, and is not possible.


No, your idea is a good one. And I really respect you for having the motivation and guts to attack such a project. If you bring it to a successful end, you will have learned immensely and you'll feel pretty good ;) . It's the only way to learn, but I just wanted you to realise what you are getting into.

Jan Didden
 
If you have 50 VDC and want to take it down to 35-40 volts, it may be done. I'll guess you have to poke around at Maxim or Linear Tech to find some suitable IC but when you are ready and have two good working switching regulators including necessary filters, I'll suspect that it would have been easier and cheaper to change the transformer to 2 x 24 V or 2 x 28 VAC.
 
As P-A said, research the switch mode controller offerings of the common vendors - National Semi and Linear Tech. Nat Semi, has on their site, a feature where you plug in your desired Vin, Vout, Iout and it recommends some devices for you.

The downside is that there's no negative switcher. You could do what the LT1085 fanatics do and use floating grounds with two positive switchers.

Since you have the willingness to experiment, please do so. It's fun to learn about output switching artifacts, output filter design, poor output impedance over freq, etc and how they affect audio quality.

Alternatively, you could try the easy way out. Use off-the-shelf 48Vin brick type switchers. They are cheap, compact, well documented and tweakable to get your desired Vout. Oh, they're fully protected and probably perform better than any home brew switcher.

Yet another option is to go hybrid between the two approaches. Use linear regulators but with a switching tracking pre-regulator. You get high efficiency (but not as great as a pure switcher) and low noise. Now that'll be fun.
 

sr2002

Member
2005-07-08 4:14 pm
Texas
but why do u want a regulated supply?

gainclones are usaully best suited with unregualted snubberised psu's.

and second thing if ur using lm3886 isnt 100va toooooooo low.

i mean your bass will be cut down.

and i wouldnt suggest using +/- 35v. use 25-30v so that the amp is flexible to different speaker loads.

if sometime u want to use a 6ohm speaker the amp will get all sweaty.


are u building a mono amp? for wat aplication are u using it?