What are good quality step-up converters out there?

[HomeExperiement]

Hi!

I was wondering if you can suggest any decent DC-DC step-up converter that would actually withstand 100W output? My intention is to convert 12V to 23-24V and feed it to 100W consumer. I got myself 150W step-up module that according to specs should withstand 150W or 6 Amp at output (when you google it, the one with red board and 35V capacitors should come up - not wanting to link e-bay here). However, at around 90W it just melted down. So I returned it to local store and they replaced mosfet and when I got it back, capacitor exploded when I powered it up with no load (anyone can guess what it could be?). I returned it to shop, they gave me money back and suggested 200W step-up. But when I googled it, reviews quickly sow that 60W load can kill it in few minutes.

I was wondering if any of you guy have used step-up converters in your DIY projects and if you know by experience what are the good ones? Or is the only way to get something good to build it myself?

 

[DaveE]

1) In general, you will get the quality that you pay for. The world of DC-DC converters is swamped with low-cost low-quality products, especially if bought from places like eBay or hobby suppliers. Try looking online at places like digikey.com, newark.com, etc. (where industrial R&D people buy stuff). They will have good search capabilities to help find what you want. I would only look for manufacturers (not suppliers, manufacturers) that have a real website that will provide a good data sheet.

2) Read the data sheets associated with the product. For reliable operation you must meet the requirements that the product was designed to operate with. At this power level, cooling is key; not necessarily difficult, but it can't be ignored. You will probably have to supply a heat-sink or airflow.

3) If your power source isn't clean (well regulated, surge/spike free) then you may need to think about adding or buying something with input protection. Look for CE certification, it means they (should have) done some testing for input disturbances.

4) Derating is a good thing for reliability. In my 30+ years in the dealing with power supplies, I don't think I ever specified a PS that just barely provided enough power (or current, op. temperature, etc.) to meet my requirements.

 

[DaveE]

So I returned it to local store and they replaced mosfet and when I got it back, capacitor exploded when I powered it up with no load (anyone can guess what it could be?).

Undoubtedly it was unqualified people doing a mediocre repair. It isn't unusual that a repair like this wouldn't work out. In the real world, we normally don't repair this sort of thing we just replace it. There are many many ways that these things can break and it often isn't worth the analysis effort to understand what parts need to be replaced. That effort is better put into figuring out if the next one will also break because it isn't an appropriate part for the application.

 

[Rive]

My intention is to convert 12V to 23-24V and feed it to 100W consumer.

Was that 12V some kind of battery?

Did you try any active cooling? Additional input/output buffer cap?

 

[HomeExperiement]

About cooling: Makes sense. But in my opinion, specs should still say that it is needed? Otherwise it's kinda false advertising. Or is it standard that customers who buy these are expected to know when they need cooling? With this 150W unit the local shop only stated it was 150W lol. When I googled it, sellers on ebay/aliexpress asked to use cooking if output power is more than 100W. But I still destroyed it with about 90W. But to honest I didnt expect that need myself either - power supplies that plug in mains and give you 12V DC out already have fan built in. For example if you take PC PSU or some more powerful LED power supply. Since it was my first step-up I actually expected that if it comes fanless, then it can be used fanless. But you were right. I once even had incident where PC PSU was running fanless with also about 90W and it was stable around 75C. Brom the other hand with that particular step-up that I had, I found youtube video where someone used it within allowed output current/watts and had fan on it and it still died :P

I was also curious to know about cooling: If I want passive cooling without fan does the number of mosfets matter? I mean in theory having model with 4 mosfets instead of 2 should distribute heat more evenly over the heat sink? Or is the number of mosfets defined by something?

Later, I also tried to look up some components it had. For example it had IIRF3205 mosfet that got replaced by the shop. But I am trying to understand it: http://html.alldatasheet.com/html-pdf/1008000/ISC/IIRF3205/63/1/IIRF3205.html

Here this sheet has 110A and 390A but what to make of it? Is this amount of current that can pass through it? I am confused because in my case 4A made it so hot that it caused plastic to melt. Just looked it up becaust it was the part that got destroyed and I wanted to learn more about it.

Now, taling about 2nd time when capacitor exploded: Does the power source matter? Is there difference between battery and something else? I used PC PSU as I power source. I had it with dummy load on it so I doubt it could have caused some high peek voltage on plugging in. I even tested it with multimeter and saw it climbing up to 12V within about 1 second after plugging in. I didn't see any occasion where startup voltage were above 12V - I guess it's well designed since it is built to power sensitive electronics. But one question I still have - is it normal that every time when I connected this 150W step-up to power source (aoso used another 12V adapter, same result with that) I saw spark at the moment of connecting it and heard the click that was almost as loud as you get when you short circuit 12V? The shop I asked explained it away saying that capacitors draw a huge current at startup. But is it true? In my PC PSU, there are way more powerful capacitors and I don't hear click when I plug it in, nor do I see sparks at the moment of connecting it to mains.As for repairs, in my country shops have right to attempt to repair it up to 3 times before they are foreced to give you money back :S So I had to convince them that I can stay within specified limitations and still blow it up easily if they wanted to repair it 3 times so badly and then they gave up 2nd repair attempt and returned me my money :)

Also thanks for suggested websites. Do you know any good brands that you would recommend? I wa curious: Here is something that seems to be boost converter: https://www.digikey.com/product-detail/en/abb-embedded-power/ABXS005A4X341-SRZ/555-1443-1-ND/7784416 but how to you exactly increase the output voltage ? At least on photo I don't notice the screw.

 

[Tom.G]

Datasheet at:
http://apps.geindustrial.com/publib...=Data Sheets|ABXS005|PDF&filename=ABXS005.pdf
That one from Digikey meets your stated requirements. The first page of the datasheet (paragraph at bottom of page) states that output voltage is set by an external resistor, about 10.55k to 11k for 24V to 23V output. It also shows 24V and 5.4A load capability without cooling at ambient temp. below 70°C.

You may learn about a bunch of stuff to use it. For instance the datasheet shows it really wants 18 surface mount capacitors and two electrolytics to work correctly. Then there is the 150mV output ripple, that may or may not be a problem for whatever you are using it for.

Cheers,
Tom

 

[Rive]

About cooling: Makes sense. But in my opinion, specs should still say that it is needed? Otherwise it's kinda false advertising

Well, most likely it has some of that in it... Those cheap modules from the internet are great to have available, but you better take the datasheet values with a bit truckload of salt.

You used it just a bit below the maximal power limit for ambient air, and it has two heatsinks hugging the capacitors (which are heat sensitive - their expected lifetime will be affected seriously). Allowing it to get hot is asking for trouble anyway. Next time watch out for the cooling.

Does the power source matter? Is there difference between battery and something else?

Yes. Battery has a voltage drop under load. A car battery would just shrug for 10A load, but any small size sealed-lead acid battery would give out less voltage: down to 8-9V, worst case. And step-up converters has one serious issue, that the switching element deals with the power at the input voltage. So when you want 100W at 12V it'll be ~10A, regardless the output voltage - but if the input voltage falls to 9V, it'll rise to ~ 13A (with regards to the losses and my laziness to deal with decimals). It'll increase the heat production and might just cook the device.
The same goes with the cables. Your PC PSU might has adequate output capability at 12V, but just one single cable won't be enough to deliver it: you will get a voltage drop on the cable -> higher current -> more drop -> more heat.
So, at 10A you have to watch for your connections and cables too.

Also, these typical modules are often underdesigned, so (next time) you better patch up those in- and output buffers too if you want to use them around the datasheet capacity.

 

[Asymptotic]

About cooling: Makes sense. But in my opinion, specs should still say that it is needed? Otherwise it's kinda false advertising. Or is it standard that customers who buy these are expected to know when they need cooling?

In general, I've observed an inverse relationship between the thoroughness of product specifications to vendor quality. Companies that stand behind their products tend to provide a wealth of application information. Those less so inclined either make obtaining detailed information difficult, or are mute on such subjects.

This has nothing to do with DC-DC converters per se, but a piece of gear I worked on used solid-state relays in a PCB-mount package rated for a maximum of 5A operated in free air at 40°C or lower ambient. The same semiconductor die was used in their 25A devices, with the difference being it was mounted in a panel-mount package, and must be used in conjunction with a heat sink of 2.5°C/W or lower thermal resistance.

Heat removal and semiconductor longevity go hand in hand.

 

[DaveE]

About cooling: Makes sense. But in my opinion, specs should still say that it is needed?

Cooling requirements and power specifications can't be fully defined by the manufacturer. The performance of the converter depends on how you use it also (i.e. how well you cool it). The key specifications for determining the power allowed are the maximum temperature specification and the efficiency. The efficiency determines the amount of waste heat that must be dissipated. Your cooling systems then determines what temperatures result from the waste heat. The Power specification is really just a guideline to help you choose products quickly; or, for less well specified products, it is an excuse for not providing the information that you really need.

Or is it standard that customers who buy these are expected to know when they need cooling?

Yes. Based on the product specifications.

 

[HomeExperiement]

Also, these typical modules are often underdesigned, so (next time) you better patch up those in- and output buffers too if you want to use them around the datasheet capacity.

How does this patching look like? Haven't heard this. Also, is it normal for step-up to draw a tons of current on first connect or is it also a sign of bad design? I already mentioned I saw sparks when I connected the one that exploded... With PC PSU, if I plugged it in when it was running, the short circuit protection just turned the PSU off on the moment I connected it The only way to get it work was to first connect it to PSU and then power it up - I suppose it was because at the startup the volts climbed up from 0V 12V and thus limiting the current. When the step-up was already powered up by something else and I then connected it to PSU, then it also worked fine. It did not trigger any protection mechanism. With an other 12V adapter, that ignored high current draw at step-up module's startup in addition to sparks it also left black burn mark on metal surface where I connected the wire.

 

[Rive]

Also, is it normal for step-up to draw a tons of current on first connect or is it also a sign of bad design?

No. The empty capacitors on the input generates high inrush current. It is normal. These boards are intended to be used in a system, if the inrush current is a problem then you have to handle it at system level.

How does this patching look like?

It is just additional capacitors on/instead of the original ones, in case they are inadequate at the planned current.

 

[HomeExperiement]

No. The empty capacitors on the input generates high inrush current. It is normal. These boards are intended to be used in a system, if the inrush current is a problem then you have to handle it at system level.

Does this inrush current also damage PC PSU or any other power supply I could possibly use? I heard this before too but I don't get on thing: in my computer PSU there are way more powerful capacitors but it doesn't act like that when I plug in in. Yow do I handle it in system level? Like starting it up with lower voltage to limit the current?

It is just additional capacitors on/instead of the original ones, in case they are inadequate at the planned current.

For more current I would take capacitors with higher capacitance and current? Is this to reduce heat that generates in capacitors and through that allow me to draw more current? From the last time I had it, I remember that the capacitors started to while/whistle when I applied 100W load. Was this bad sign?

 

[Rive]

Does this inrush current also damage PC PSU or any other power supply I could possibly use?

Hard to say: although PC PSUs are expected to be able to handle some hot swap, an unexpected inrush current of a decent sized capacitor might trigger the overcurrent defense - as it happened in your case. When it is just starting up with the device attached, then it is a 'soft start', when the voltage rises in slow slope. At system level it is the question of additional components, mentioned in the previously linked article.

For more current I would take capacitors with higher capacitance and current?

No: for an underdesigned device you can/should add more capacitors to safely get currents close to the nominal current limits.

From the last time I had it, I remember that the capacitors started to while/whistle when I applied 100W load. Was this bad sign?

Most likely that was the coil, but it is indeed a bad sign. Just keep it as a rule of thumb that for these unknown aliexpress and ebay devices just stay within 70% of the limits.