SMPS controller advice

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I'm thinking of building and constructing a SMPS circuit, and I have a general question. What controller chip to use?

When I did a search on the board, I found that the SG2525 and SG3525 is still very popularm but I have heard that they ain't the best chip for the job.

Any comments?

/Johan Ch
 
the LM3524/SG3524 were among the first dedicated switcher chips, so they are showing their age -- they are cheap, less than $0.50 here in the States. I use them for motor control -- it's easy to press into service to control an H-bridge -- like the LMD18200/1

it's not a "quiet" device by any means.

before you dig to far into it, consider reading a book like Pressman's "Switching Power Supply Design" , also take a look at the application notes at Texas Instruments, On-Semi, Linear Tech, Nat Semi websites -- if you spend a little more money on the chip you just might well have a more robust, quieter design.
 
I ended up with a simple way of controlling pulse width which uses a 74C14 and an optocoupler. I find it provides stable performance. The differentor method it uses is resistant to shutting off the pulses to the mosfets completely which keeps the transformer resonating, thus boosting efficiency. The other methods which use a timing ramps need special care to avoid that problem.
 
Hi jackinnj,

That observation about my link is perceptive. I need some format that is easy, free, reliable, and solid. A regular Web-site is a problem for me because I am not talented at programming--Basic computer language was hard enough for me. My focus of interest may be narrow and almost fanatical since I do not really have much talent with computers at all. My work on the project is totally non-profit. If something happens to me, I hope my discoveries can live on through the Yahoo Group. I am actually hoping to find some good folks who can find the same passion about the concept and design to become moderators also. I would hope that could allow the Group to live on even if I were no longer able to tend to it.
 
Re which controller to use

You don't mention what voltages you are planning to extract from your SMPS controller and your input voltage. These parameters are kind of important :), as is desired power level and any preference on topology that you might have.

Now, I would recommend that you start looking at the Linear Technology data sheets first, simply because the "black art" is reduced to simple explanations illustrated with formulae and design examples. Also, the very latest data sheets from TI (previously Unitrode) are very easy to evaluate for the same reasons.

Unitrode devices (now TI) are typically considered the reference standard for whatever reason - probably performance (it certainly is not the original data sheets). If you are really keen, you can download a bunch of application notes from www.ti.com (search for SEM600 and keep incrementing/decrementing the number by 100 until you have exhausted all the SEMX00 series). There is A LOT of information there.

Linear Technologies www.linear.com (I have been warned about stability of these devices from a member of this site) end up being my choice in the design process pretty much every time - typically based on the collection of features which just happen to fit well with my needs. They certainly have some highly specified units, some of which just happen to have very appropriate features for me. I guess you can call that well researched product planning which I am grateful for.

You should also head on over to www.infineon.com and look at their application notes / reference designs - they are very helpful, particularly their 200W demo board. They do have some controller IC's which seem OK but they seem hard to get, and to be frank, I was only swayed in their direction based on the incredibly helpful application note on the aforementioned demo board. If your power needs are low, I would strongly recommend that you consider their series of integrated devices (controller + switch) but you may be topologically or power level constrained.

Based on my limited experience, other manufacturers don't seem to have great controller devices (subjectively based on my relatively specialized needs). Your selection will probably end up being based on the features you desire. In my personal case, my desires just about force the selection of manufacturer. If your needs are more general, you should be able to get by with any manufacturer's device.

Petter
 
How much power do you need? What is the application?

Well, the first application where I want to try a SMPS PSU is in my current tube preamp. The SMPS should be used to supply the B+ voltage, around 300VDC and very little power.

I have been following the discussion that started a while ago in the Pass forum about a SMPS for power amplifier. Enlightning reading. But, is there acually much to gain over a conventienell design when talking about this much power?

/Johan Ch
 
Johan_Ch said:


Well, the first application where I want to try a SMPS PSU is in my current tube preamp. The SMPS should be used to supply the B+ voltage, around 300VDC and very little power.

/Johan Ch

I appreciate your interest in the switchmode idea. I think that even though the circuit versions I made up for the Web-site are high power renditions, low power versions are simpler and quicker to construct. For example, the IRF7343 buffer can be omitted and IR2113 can drive IRFIB7N50 mosfets directly through 10 ohm resistors. Using the p2p technique and 1/8w full-leaded resistors trimmed to needed length, a very compact circuit is possible. The primary power supply could use a full wave bridge charging a 22uF or 47uF capacitor. There is no need for +/- primary voltages. Use a power line LC filter to stop switching noise from getting to the AC mains. It shouldn't be hard to totally shield it and isolate any switching noise. Also, I've been thinking that the driver/conrol circuit can be completely potted in silicone or epoxy if desired.

SMPS circuits are very good for producing high voltages. A step-up secondary for a small power supply could use 28awg wire and take up little space on a 1 1/4 inch ferrite toroid core. Just place the winding on a layer of electrical tape. Each winding pass would go on a new wrapping of tape. You can wind the primary on one end of the toroid core, with the secondary on the opposite end, with at least a 1/4 inch gap between the primary and secondaries for isolation. Then use a full wave bridge, output choke, and filter capacitor. Remember to fuse the mains power input, but you can easily current limit the output by placing something like a .005uF capacitor in series with the secondary winding.
__________________
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.