Buck power modules to replace 7805.

[westfw]

Is there something around  that is "in between" in quality compared to the common "mini360" $0.40 Aliexpress buck converter modules ( https://www.aliexpress.com/item/32614063352.html ) and the multitude of $3-$6 converters (recom, CUI, adafruit, Pololu, etc)?  Something in the $1-$2 range, perhaps?
(> 16V input, 5V and/or 3V output, >1A output.  Not significantly bigger than a a TO220 layout (similar pinout "desirable"), but not ultra-tiny SMT (like an MPM3610) either.)  The mini360 modules would probably be fine if I trusted them more (although fixed output voltage is somewhat desirable.)

(an IDEAL solution would provide both 5V and 3.3V in less space than 2x TO220, with I_3.3 >300mA (and (I_3.3 + I₅₎ > 1A))

Also, has anybody tested the Mini360 modules to see how they perform?  What tests would you do (with a "reasonable" amateur lab.)

[coromonadalix]

cant recall where i saw them 
but there is some modules who has almost the same footprint than a to-220 regulator but at higher current

[T3sl4co1l]

Is there a reason you can't integrate the regulator on board, yourself?

I think... there's kind of three use cases for those modules:
1. Maintaining/improving old designs.  Don't have the board files, or it would be a PITA at least to update them; might as well slap in an improved version instead.  The similar outline facilitates this.
2. Inexperienced, or lazy -- in effect, you must pay a tax.  Either of the pricier brand names, or the cheap who-knows dice rolls of the unknown brands.
3. You truly don't have the room for an onboard regulator, and the vertical module saves horizontal space.  You could just as well build a riser board yourself, but honestly you might as well go with the already-proven and optimized design.  Attractive alternatives are those SMT and even IC modules/regulators, which are getting quite small and powerful these days.  In this case, again, you pay the tax, but you're likely able to accommodate that given the push for high density.

So, I would think there's not really a middle option, because there's a substitute: placing it yourself.

As for dual outputs, don't know of any offhand.  There are multi-output regulators of either type (linear or switching), but nothing as familiar as a 7805; and as usual, you'll be taking up the space using them.

If an onboard design is an option [and you're more just feeling out your other options here?], and you have some power dissipation to spare, I wouldn't flinch at a 5V switching reg + 3.3V LDO.  At up to 1A, it's under 1.7W, not a huge burden for SMT to dissipate, or you can use a traditional TO-220 still.

Also, hacks, like with multi-winding inductors, probably won't work well here (the voltages aren't a convenient ratio apart, and you likely need good enough regulation on both rails?), you'd likely use independent or cascaded regulators.

Tim

[HackedFridgeMagnet]

If you could place them in a new designTI webbench can throw up some interesting options.
Heres a couple.
The layouts don't often make sense though.
 

[bobbydazzler]

A SOT 223 ldo can easily do 500ma@3.3v when supplied with 5v so I 2nd T3's advice to use a buck for the 5v then an ldo(ams1117 or variant are super cheap and probably good enough if a buck converter is good enough).

[westfw]

Is there a reason you can't integrate the regulator on board, yourself?
2. Inexperienced, or lazy

Well, yes.  I am inexperienced with SMPS design, and I hear that it can be "tricky."  I am not particularly interested in being an SMPS designer.(although I guess many of the modules don't seem to be as carefully designed as data sheets imply they should be, either.)

Also, I don't think I believe that I can buy the components for the sort of price target I'd like to reach.  That takes mass production (or at least more "mass" than the ~10 boards at a time that I am likely to make.)

[SiliconWizard]

Is there a reason you can't integrate the regulator on board, yourself?
2. Inexperienced, or lazy

Well, yes.  I am inexperienced with SMPS design, and I hear that it can be "tricky".

Nah. If you're using an off-the-shelf buck converter IC, there's nothing much to know. Just follow the datasheet and pick a suitable inductor, mainly.

Designing a SPMS completely from scratch, especially if you're after high efficiency and high power, can be tricky indeed. Using a commercial buck IC for low power stuff is not.

[james_s]

I doubt there's much of a market for something in the middle price range. I've only ever used these 3 pin buck converter modules to retrofit older equipment in which case the $6 (or even $15) price is not an issue. If I'm designing something new I'll put the regulator components on the PCB with everything else.

[Siwastaja]

As SiliconWizard says, pick a modern fully integrated switching regulator, with fixed output voltage and internal compensation, that requires no other external components than the inductor and capacitors, and runs at high frequency enabling full MLCC solution. Design will be a matter of following the recommended layout section of the datasheet. Easy even for a beginner.

[chris_leyson]

Diodes Inc AP62150 synchronous buck converter will manage 1A at 5V. Max Vin 18V and typically 90% efficiency.

[SiliconWizard]

As to the modules shown in the OP's link, yes those are sold everywhere (Aliexpress, eBay, ...) and are all pretty much the same except for soldermask color. (Talking about this exact module of course.)

Note that they are erroneously stated as being based on the LM2596 IC. Which they are not. As can be seen on the pictures, and on real boards (I actually have a handful of them - that can always be handy for a quick prototype), they are using the MP2307, which is this: https://www.monolithicpower.com/en/mp2307.html
It's actually not bad at all, and overall I've found MPS products to be quite good.

Note that those boards are very poor when it comes to heat management of course, so don't try and get too much power out of them. 500 mA to 1 A @5V shouldn't be a problem at all though.
Also note they come with a (poor) pot for adjusting the output voltage. IIRC they have solder jumpers to use fixed resistors instead, but in any case, that's what I do. I wouldn't trust setting the output voltage with those pots for long-term operation, even if you block them with some resin.

[TomS_]

I've used the Murata OKI-78SR series before and they seem to do a good job. Basically a drop in replacement for 78xx series regulators.

Otherwise go to places like RS, Mouser, DigiKey and the likes and search for non-isolated DC-DC converters and you can find a whole lot more.

[PaulAm]

Also note they come with a (poor) pot for adjusting the output voltage. IIRC they have solder jumpers to use fixed resistors instead, but in any case, that's what I do. I wouldn't trust setting the output voltage with those pots for long-term operation, even if you block them with some resin.

I've use those modules; you can get the whole module cheaper than I can buy the chip but those pots are absolute crap.  I replace them with a Bourns multiturn pot (which probably costs almost as much as the module)

[rcbuck]

I've used the cheap modules. I do like SiliconWizard does. I use a multiturn  pot to set the voltage. I then remove the pot, make measurements, and replace the pot with nearest value 1% resistors. Output voltage is always very close to what the pot produced.

[SiliconWizard]

I've used the cheap modules. I do like SiliconWizard does. I use a multiturn  pot to set the voltage. I then remove the pot, make measurements, and replace the pot with nearest value 1% resistors. Output voltage is always very close to what the pot produced.

You don't have to go through a pot adjustment though. You can calculate the resistors value directly using the datasheet.

[John B]

These days I would go for a MOSFET based integrated converters. I use a lot of XLSEMI components, which are reasonably priced. Again the ebay units are totally fake, there's a reason you can buy the whole module for less than a legit 2596 or 2576 IC. The fake ones have a different switching frequency from legit parts, and I have them fail quite easily. Even if you get a legit one, there's still the darlington output switch which still gives you significant power dissipation, hence why a MOSFET based one is the better option.

[MIS42N]

Also, has anybody tested the Mini360 modules to see how they perform?  What tests would you do (with a "reasonable" amateur lab.)

I have done some tests.

As pointed out, it is hard to set the output voltage using the tiny pot on the board. I set one up to deliver 5V @ 1A from a 12V supply and let it run for a few days. It didn't get warm. I wired up the same board to deliver around 200mA to a circuit and it wasn't stable. Possibly I had not provided sufficient capacitance on the load side. A 10uH inductor is not enough to ensure continuous flow at 200mA. I removed the 10uH and replaced it with 100uH and it was fine. I accidentally shorted the output, producing a nice little arc for a fraction of a second. I thought that would kill it but it seemed unaffected.

I checked the circuit against the reference circuit for the chip. The reference recommended 20uF chip capacitors at input and output, the ones on the board are tiny and I doubt anywhere big enough for the claimed 3A but maybe enough for lower currents. Some of the other components are different to the reference, but not greatly so.

I replaced the pot with a fixed resistor for a different purpose, delivering 8.5V.  Works fine.

I haven't done it yet, but I figured for a low voltage output the strategy would be to put the pot in parallel to a fixed resistor. Its maximum value seems to be about 180kohm. There are pads to mount a fixed resistor.

If you have access to Silicon Chip magazine, they ran an article in August 2020 about substituting a buck converter for a 78xx.

[bobbydazzler]

I forgot to mention I have used the mini360 modules before, all I can recall is they did 700ma pretty easy - only got a little warm, I think they'd be good for 1amp at least.  No idea of the ripple quality etc though I don't have a scope.

[tszaboo]

TPSM84205
https://www.ti.com/product/TPSM84205

2.5 USD in quantity, so in-between the price.

[SiliconWizard]

Also, has anybody tested the Mini360 modules to see how they perform?  What tests would you do (with a "reasonable" amateur lab.)

I have done some tests.

As pointed out, it is hard to set the output voltage using the tiny pot on the board. I set one up to deliver 5V @ 1A from a 12V supply and let it run for a few days. It didn't get warm. I wired up the same board to deliver around 200mA to a circuit and it wasn't stable. Possibly I had not provided sufficient capacitance on the load side. A 10uH inductor is not enough to ensure continuous flow at 200mA. I removed the 10uH and replaced it with 100uH and it was fine. I accidentally shorted the output, producing a nice little arc for a fraction of a second. I thought that would kill it but it seemed unaffected.

You're right about the inductors on those boards. Beware that there are dozens of similar boards on the market, and they do not all have the same inductor. But that's something to have an eye on.

As I said, OTOH, the buck IC on those (which is normally the one I mentioned) is good. As long as it's genuine. I doubt they would bother using fake MP2307's - but who knows. If you buy those ICs through common distributors, they are actually not that cheap, even in large quantities. So, I don't know how they would manage to sell the whole board for $0.40 only.

[james_s]

I could never figure out why they use a fake LM2517 or whatever on those 99c buck regulator boards. The fake part works fine but it's not even the same operating frequency as the part it claims to be. I'd be perfectly willing to buy it for applications where it can do the job, and if it had its own unique part number I wouldn't feel sketchy about it.

[SiliconWizard]

I could never figure out why they use a fake LM2517 or whatever on those 99c buck regulator boards. The fake part works fine but it's not even the same operating frequency as the part it claims to be. I'd be perfectly willing to buy it for applications where it can do the job, and if it had its own unique part number I wouldn't feel sketchy about it.

As I mentioned, the IC they use is perfectly fine. IIRC, it even has slighly better efficiency than the TI part. Of course, assuming the part they use is actually a genuine MPS product, but I have no reason to think it isn't (at least on the boards I have, it seems to be genuine and its measured characteristics match those of the datasheet.)

It's NOT a fake part whatsoever. I gave the link to the product. It's a fine buck converter that doesn't pretend to be something it isn't.

As with any other device sold on those platforms, many (but not all) sellers use deceptive practices. The mention of a TI part is a lie. It's not a like for what would be an equivalent part either. It's absolutely not equivalent. Obviously the LM2596 is a completely diffrent package, can dissipate a lot more power, and has a different switching frequency. Those boards again use a MP2307.

It's just done to catch buyers looking for this part number while searching. But I have also seen sellers mentioning the proper IC used on those boards. Sellers are to blame. Not the boards themselves or whoever manufactures them. For instance, same board with the right IC in the title: https://aliexpress.com/item/1005001617985227.html?spm=a2g0o.productlist.0.0.421256b43sZTne&algo_pvid=e3d3db09-630f-4874-a8ba-79cce6d9d050&algo_exp_id=e3d3db09-630f-4874-a8ba-79cce6d9d050-0

I've also found similar boards using the same IC, but with no awful pot on them - just a set of resistors and solder jumpers to select the output voltage. Much better. I also have a few of those, and would recommend looking for that instead.

[rcbuck]

I use a lot of XLSEMI components

John B, who is a reliable distributor for the XLSEMI components. Especially looking for a U.S. distributor. I do not consider WinSource as a reliable distributor.

[John B]

John B, who is a reliable distributor for the XLSEMI components. Especially looking for a U.S. distributor. I do not consider WinSource as a reliable distributor.

I get them from LCSC in China, I notice the local Aus distributors don't have them either.

I could never figure out why they use a fake LM2517 or whatever on those 99c buck regulator boards. The fake part works fine but it's not even the same operating frequency as the part it claims to be. I'd be perfectly willing to buy it for applications where it can do the job, and if it had its own unique part number I wouldn't feel sketchy about it.

I've had them fail multiple times at way less than their claimed 3A or whatever output current. This was an "LM2596" in a TO220 package, with the typical incorrect switching frequency. Others have had better luck with the surface mount variants. That's  the lucky dip I guess.

[tonyh88]

I've used the Murata OKI-78SR series before and they seem to do a good job. Basically a drop in replacement for 78xx series regulators.

Otherwise go to places like RS, Mouser, DigiKey and the likes and search for non-isolated DC-DC converters and you can find a whole lot more.

This. Drop in replacement for LDO or browse the wonders of non-isolated DC-DC from any distributor.

[rcbuck]

Thanks John B. My guess is shipping charges from LCSC would not make ordering from them practical. I guess I will stick to the TI and Microchip regulators that are available from Mouser and DigiKey. Some of them are currently out of stock but I have enough on hand for any projects I will be making over the next year or so.

[james_s]

I could never figure out why they use a fake LM2517 or whatever on those 99c buck regulator boards. The fake part works fine but it's not even the same operating frequency as the part it claims to be. I'd be perfectly willing to buy it for applications where it can do the job, and if it had its own unique part number I wouldn't feel sketchy about it.

As I mentioned, the IC they use is perfectly fine. IIRC, it even has slighly better efficiency than the TI part. Of course, assuming the part they use is actually a genuine MPS product, but I have no reason to think it isn't (at least on the boards I have, it seems to be genuine and its measured characteristics match those of the datasheet.)

It's NOT a fake part whatsoever. I gave the link to the product. It's a fine buck converter that doesn't pretend to be something it isn't.

As with any other device sold on those platforms, many (but not all) sellers use deceptive practices. The mention of a TI part is a lie. It's not a like for what would be an equivalent part either. It's absolutely not equivalent. Obviously the LM2596 is a completely diffrent package, can dissipate a lot more power, and has a different switching frequency. Those boards again use a MP2307.

It's just done to catch buyers looking for this part number while searching. But I have also seen sellers mentioning the proper IC used on those boards. Sellers are to blame. Not the boards themselves or whoever manufactures them. For instance, same board with the right IC in the title: https://aliexpress.com/item/1005001617985227.html?spm=a2g0o.productlist.0.0.421256b43sZTne&algo_pvid=e3d3db09-630f-4874-a8ba-79cce6d9d050&algo_exp_id=e3d3db09-630f-4874-a8ba-79cce6d9d050-0

I've also found similar boards using the same IC, but with no awful pot on them - just a set of resistors and solder jumpers to select the output voltage. Much better. I also have a few of those, and would recommend looking for that instead.

I had to check, the boards I'm talking about claim to be based on the LM2596 and they have a surface mounted TO220 part that is labeled LM2596 but it is NOT a real LM2596. It's a real buck converter yes, and it works fine, at least if you derate it to Chinese Amps but certainly isn't what the number printed on it suggests, it is not even the same operating frequency. Despite all this, it's a good value for the price if you just need a simple buck regulator and don't need to draw the maximum rated current.

[SiliconWizard]

I could never figure out why they use a fake LM2517 or whatever on those 99c buck regulator boards. The fake part works fine but it's not even the same operating frequency as the part it claims to be. I'd be perfectly willing to buy it for applications where it can do the job, and if it had its own unique part number I wouldn't feel sketchy about it.

As I mentioned, the IC they use is perfectly fine. IIRC, it even has slighly better efficiency than the TI part. Of course, assuming the part they use is actually a genuine MPS product, but I have no reason to think it isn't (at least on the boards I have, it seems to be genuine and its measured characteristics match those of the datasheet.)

It's NOT a fake part whatsoever. I gave the link to the product. It's a fine buck converter that doesn't pretend to be something it isn't.

As with any other device sold on those platforms, many (but not all) sellers use deceptive practices. The mention of a TI part is a lie. It's not a like for what would be an equivalent part either. It's absolutely not equivalent. Obviously the LM2596 is a completely diffrent package, can dissipate a lot more power, and has a different switching frequency. Those boards again use a MP2307.

It's just done to catch buyers looking for this part number while searching. But I have also seen sellers mentioning the proper IC used on those boards. Sellers are to blame. Not the boards themselves or whoever manufactures them. For instance, same board with the right IC in the title: https://aliexpress.com/item/1005001617985227.html?spm=a2g0o.productlist.0.0.421256b43sZTne&algo_pvid=e3d3db09-630f-4874-a8ba-79cce6d9d050&algo_exp_id=e3d3db09-630f-4874-a8ba-79cce6d9d050-0

I've also found similar boards using the same IC, but with no awful pot on them - just a set of resistors and solder jumpers to select the output voltage. Much better. I also have a few of those, and would recommend looking for that instead.

I had to check, the boards I'm talking about claim to be based on the LM2596 and they have a surface mounted TO220 part that is labeled LM2596 but it is NOT a real LM2596. It's a real buck converter yes, and it works fine, at least if you derate it to Chinese Amps but certainly isn't what the number printed on it suggests, it is not even the same operating frequency. Despite all this, it's a good value for the price if you just need a simple buck regulator and don't need to draw the maximum rated current.

Ah, then this is not the small board the OP linked to and that I was talking about here (which some sellers also wrongly claim as having a LM2596.)

[MIS42N]

Interesting. The boards I bought use MP1484EN chip, not MP2307. Looks identical to all the other mini 360 boards advertised. I went back to the seller's site, doesn't mention what chip. It's the equivalent of the 2307 with a lower maximum voltage (18V) and lower resistance in the switching MOSFETs  (85mΩ). I was looking to run off car battery voltages so fortunately the lower maximum voltage is not an issue.