Isolated, Unregulated DC-DC Converters - The Pitfalls

Oi, want some galvanically isolated power rails? You’ll like them, they’re fun …
↓↓↓ Complete description, time index and links below ↓↓↓
Ever had an offer like that? Well, as responsible user you shouldn’t rely on the marketing hype. Instead you should check the datasheet, check the footnotes in the datasheet, and verify what’s written in the datasheet.
And that’s exactly what I did. I read the datasheet and measured the output voltage, the efficiency and the output ripple. The results were mixed. You really need to know what you’re getting into when using such a thing.
00:00 Intro - everything was already on the bench
01:08 Application - to use or not to use, that is the question
08:17 Minimum load - idling these converters is a bad idea
11:01 Measurements - some values for different loads
13:16 Efficiency - those converters won’t save the planet
15:55 Regulation - there is none, don’t trust the graphics
21:13 Output ripple - not too bad, but read the fine print
25:33 Back ripple - it’s a current ripple and it’s EMC related
28:54 Wrap-up - dirt cheap solution with serious drawbacks
Tutorials: • Tutorials
Mailbag: DC-DC Boost Converter Module (incl. Test with Partial Characterization): • Mailbag: DC-DC Boost C...
#robertssmorgasbord #tutorials #tutorial #how-to #isolated #dc-dc #converter #module

[@andymouse]

Very informative, It's great to really get into the numbers, cheers!

[@robertssmorgasbord]

You're welcome!

[@Benivator2]

Thank you. You saved my circuit!

[@robertssmorgasbord]

You're welcome!

[@prashanthb6521]

Very nicely explained sir. Thank you.

[@robertssmorgasbord]

You're welcome!

[@iKenFlyPPG]

3S rc batteries are nice to have around for quick and simple isolated power supplies with unlimited amperage, so use circuit protection.

[@robertssmorgasbord]

Well, not really unlimited amperage, but enough to melt most circuits 😅

[@iKenFlyPPG]

@@robertssmorgasbord Hehe ya just a little exaggerated ;) a shorted lipo is going to dump all she's got, interesting some battery cell manufacturers have built-in to the individual battery cell, a fuse for over current, more for a battery pack with high current capability beefcake busbars and all.

[@robertssmorgasbord]

@@iKenFlyPPG 🙂 I'll would never use a "naked" LiPo battery (singe or multiple cell) without at least some of those cheap Chinese protection circuits (just over current and under/over voltage).

[@cccmmm1234]

Many of these specify a minimum power draw. If you use less then they will be over voltage.

[@robertssmorgasbord]

You're absolutely right. That was one of the points I made in this video.

[@jugnu361]

I use 3.3v LDO at output

[@robertssmorgasbord]

That's a good approach when you require a really stable output voltage.

[@safaesendemir3494]

you're a perfect detail.

[@robertssmorgasbord]

Thank you very much :-)

[@gazsto9510]

Can you possible help with this equation pls. 12v (?a) using x2 of 5v 3.1a (x2 = 5v 6.2a/ ) sockets and one 5v .50a (5v 500ma) what size adapter would be needed to power these 3 independant units each has it's own built in step down from 12v to 5v (amps as shown) - I am getting 5v 3.1a (15w x 2 = 30w) and 5v 500ma (2.5w) combined total 32.5watts. So if this is correct what power 12v (amp?/watt?) supply via these step convertors could manage the load if all sockets were used and connected. Appreciate anyones technical knowledge. Thanks Gaz

[@robertssmorgasbord]

OK, so your three step down converters (from 12V) have an combined output (!) power of. 2*5V*3.1A + 1*5V*0.5A = 2*15.5W + 1*2.5W = 31W + 2.5W = 33.5W But - these step down converters don't work with 100% efficiency! In the following I'll assume 80% efficiency for the 15.5W converters and 60% efficiency for the 2.5W converter (you'll have to look up those numbers in the datasheets for your converters). So the required input (!) power is: 2*5V*3.1A/80% + 1*5V*0.5A/60% = 31W/80% + 2.5W/60% = 38.75W + 4.17W = 42.92W Add another safety margin of about 20% and you'll end up with a 12V/51.5W/4.29A supply. But I think that in this case e.g. a Mean Well 12V/50W/4.2A power supply would do the job. Hope that helps!

[@gazsto9510]

@@robertssmorgasbord That's brilliant and thank you so much. Really appreciate the help and advice.

[@robertssmorgasbord]

@@gazsto9510 You're welcome :-)

[@gazsto9510]

@@robertssmorgasbord Robert if I removed one of the 5v 3.1a would a 12v 3a supply be fine. I have been informed that the pcb board needs a min of 5v 1a (not 5v 500ma). So that's one 12v 3a power supply to feed, one 5v 3.1a socket and one pcb board 5v 1a, with 12v to 5v step downs. You have been extremely helpful and can't thank you enough.

[@robertssmorgasbord]

@@gazsto9510 Lets again assume 80% efficiency for the 15.5W (5V/3.1A) converter and 70% efficiency for the 5W (5V/1A) converter. Note: These are just assumptions. In general the efficiency of a converter goes up with its power. You should really try to look up those numbers. Anyway, the required input power is: 5V*3.1A/80% + 5V*1A/70% = 15.5W/80% + 5W/70% = 19.38W + 7.14W = 26.52W Again, add another safety margin of about 20% and you'll end up with a 31.82W/12V/2.65A supply. So you should be fine with a 12V/3A/36W supply :-)

[@Gocast2]

To solve the problem of it being unregulated, couldn't you just add another buck converter, or voltage regulator to the circuit? e.g. you have a NON-ISOLATED and regulated 240v mains to 5v DC power supply powering a 5v to 5v ISOLATED and unregulated power supply... The isolated power supply then powers a NON-ISOLATED and regulated 5v to 3.3v step-down buck converter. Then you can power a thermocouple module from the 3.3v pin with a regulated voltage. Would that work, as well as give all the benefits of being isolated? Once again, thanks for the all the useful info🙂

[@robertssmorgasbord]

I thank you for your feedback and input! And yes, you could do it that way. In most cases even a simple linear 3.3V regulator could do the job (your just dropping 1.7V, so for each 100mA you would burn 170mW). An alternative would be to use a regulated and isolated 5V to 5V converter in the first place, but those are expensive.

[@Gocast2]

@@robertssmorgasbord thanks! seems like a no brainer to add in a regulator. I have a couple of pricey thermocouple module's that I don't want to wreck.

[@robertssmorgasbord]

@@Gocast2 Yup, and a linear regulator also filter out some noise from the switch mode DC DC converter, so your measurements might be a little more precise (though I don't think that's measurable).