Adjustable Current and Voltage Lab Supply. LM317

[TheLaw]

Smp stuff is great i use a modded atx to power radio gear, and its as good as gold. Transformers in the Uk are not to steep in price, chassis mount only go so high in VA ratings, teroidals go much higher, i can get a 300 VA unit for about a little over £ 30.00 not a terible price for a new unit, linear psu's are always going to be bigger bulkier in size, early smp's suffered a lot of noise, but current technology is far better today. Dave.

The thing is. I WANT to do SMPS, but I'm not 100% proficient in them yet. As in I'm still confused as to how to PROPERLY do adjustable current. I'm trying to find a design that it does it well so I can really study how to do it. Current sense? Ehhh I'm confused.

 

[davelectronic]

variable psu

I always build circuits properly, if not 100% sure i would not build it until i was, smp's are fine with today's components on offer and there efficiency. Not sure why you believe linear supply's are inferior, and or poorly regulated, a good transformer designed linear supply is as good as any smp in my opinion, both to me have there merits, smps are as stable and efficient as any linear, and the same the other way round, given a reasonable task. Dave.

 

[TheLaw]

I always build circuits properly, if not 100% sure i would not build it until i was, smp's are fine with today's components on offer and there efficiency. Not sure why you believe linear supply's are inferior, and or poorly regulated, a good transformer designed linear supply is as good as any smp in my opinion, both to me have there merits, smps are as stable and efficient as any linear, and the same the other way round, given a reasonable task. Dave.

Yeah if you only need 100mA, there's really no point to doing SMPS because the heat and power losses are pretty low. But once you get into the amp range, I think an SMPS supply starts to become advantageous. Though a lot of professional lab supplies are linear, in my opinion, I think an SMPS design is even more "professional". Not that I'm a pro...but especially with a lot of nice single chip solutions on the market, there's not much reason to go linear, imho again. Linear regs are like 50% efficient, and a simple 50Khz smp reg can be 80% efficient or better with very little tweaking or optimization. You can hit close to 90% if you know what you are doing. That means it can be smaller, lighter (not applicable here), make less heat, and be more "green". Good traits in my opinion.

Meanwhile, it is very hard to mess up a linear supply while in construction. And layout isn't really an issue though I advise you don't have 5 miles of wire between parts.

I dunno. I was hoping to have a +/- supply but I'm not sure how to do this either. I'd expect something with the feedback loop or something. Grr. Sorry to complain to you haha.

 

[davelectronic]

question

Question ? have you built a linear supply, before pulling out the big guns on smp's and there design, a LOT of thought needs to go in to an smp's spec to get your 90% + efficiency and you will still need to wind smaller transformers, or use stock products, your need a scope for smp design as to avoid excess ripple and keep a close eye on frequency switching efficiency, if you design your own spec unit, you really have your work cut out, previous post history tells me this.

You might start with a simple smp, there are a few really basic current designs using stock components, its a start, and you dont have to have a physicists qualifications. Dave.
PS I can remember a previous thread on psu's, not sure about a reclaimed transformers spec, with all the respect deserved, is not smp's design and build from scratch a bit much to chew, mmm. Dave.
PS, I have my head in the clouds sometimes, it nice to dream.

 

[(*steve*)]

Current sense is typically a resistor placed in series with the load and where the voltage across it is monitored.

There are other ways to do this, but this is pretty common and simple.

The current sense is typically placed prior to the voltage sense, so that the voltage drop across the current sense resistor does not affect the regulation.

In a simple case, the voltage across the current sense resistor is compared (using a comparator) with a voltage derived from a voltage reference, and some action takes place as the voltage across the sense resistor exceeds a set point (the voltage derived from the voltage reference).

The voltage dropped across current sense resistors is typically well under a volt. 0.1V and 0.6V are values commonly seen, the latter is pretty much the Vbe for a transistor, which suggests how it can be used. Lower voltages are typically inputs to comparators.

 

[TheLaw]

Yeah I was thinking of an LM2575 or LM2576 which are one chip, TO-220-5 package, only like 4 or 5 external components.

Steve,

Thanks for the input. It's just so hard to learn these things when you're trying to learn electronics without a proper class or course. Like I know what these things are but gahh, there's just so much to learn and so much to get sidetracked with, if you know what I mean. Between work and school, I just have so little energy to learn...My only real time is the weekends and I'm usually out. So thank you for the continued support. I really appreciate it. I'm just trying to sort through things. I've been reading your posts over and over.

 

[TheLaw]

Question ? have you built a linear supply, before pulling out the big guns on smp's and there design, a LOT of thought needs to go in to an smp's spec to get your 90% + efficiency and you will still need to wind smaller transformers, or use stock products, your need a scope for smp design as to avoid excess ripple and keep a close eye on frequency switching efficiency, if you design your own spec unit, you really have your work cut out, previous post history tells me this.

You might start with a simple smp, there are a few really basic current designs using stock components, its a start, and you dont have to have a physicists qualifications. Dave.
PS I can remember a previous thread on psu's, not sure about a reclaimed transformers spec, with all the respect deserved, is not smp's design and build from scratch a bit much to chew, mmm. Dave.
PS, I have my head in the clouds sometimes, it nice to dream.

Yes I've built many linear supplies. Perhaps I won't get 90%, and I kind of doubt it, but if I have some nice broad tracks and good quality components with proper ratings, I might be able to pull off some decent number. If I am to build an SMPS, I'm just copying the design straight off the National datasheet. Hell naw, I'm not designing my own. lol

 

[TheLaw]

Just a quick question. According to ST, the pot controlling the current limit on the L200 has to be a pretty low value. Like 50R...I can't find a decent 50R pot, so is it possible for me to use a 1K pot and a 50ohm fixed resistor in parallel to get roughly 50ohms? Is that a good idea?

 

[(*steve*)]

Well, you can, but a little thinking about resistors in parallel will show you that the result will be anything but linear.

The effect will be that the total resistance will vary little from 50 ohms until the very last part of the travel of the trimpot when it will vary rapidly.

If you can point me at the datasheet for the device you're using I'll take a look at it for you.