Battery Powered PSU for Portable Rigol DS1000E/Z

[Mechatrommer]

when making a post in Battery pack for Rigol DS1054Z i started working on this, later made the pcb order and halted due to covid attack for more than a year, on and off and now i managed to get it working for the first time on my old DS1052E...

currently doing rev2, there are few improvements and corrections to be made. the resulting DSO's noise floor got higher 40mVpp, PSU got heated at some places, forgot to add cooling fan(s) etc... the measured efficiency is shy on 62%, 1.3A consumed from 12.6V 3S LiPo battery (28W in, 18W out) but i guess any attempt to squeeze more efficiency will result in more components needed and hence wont fit in $5 jlcpcb. anyway, any suggestions are welcomed, boost and buck converters all are using MC34063, i guess its the general purpose SMPS controller that has current limit feature etc. and its popular, cheap and easy enough to get in online market.

the psu features low batt protection circuit, AC Line trigger is taken cared off to take ±30V approx (i've not tested how much BW the optocoupler circuit be able to take), battery charging capability, but its just simple fixed boost converter from 5-8V wallwart charger/psu (but i figured its a little bit weak, got hot very quickly if forced on quick charging so i have to add limiting resistor to avoid self destruction). mains power connection/extension is provided in case we want to add dedicated offline lipo battery charger. currently i just stick out the battery connector so i can do full dedicated balance charge from outside. full/charge led indicator proved to be buggy (but its not the end without it). the circuit/pcb was designed to power both DS1000E series and the newer DS1000Z series (with different components population and values), but currently i only made for my old DS1052E because i need it portable to probe other projects outside the lab quickly, now i have new purpose for it instead of laying around collecting dust, my DS1054Z still stuck to the mains on the bench, how it meant to be...

attached are the resulting DSO noise and 1KHz cal signal that i managed to test, some pictures of how it looks and installed, and schematics of old, currently built pcb rev1 (i'm updating the schematics to rev2 but work in progress) sorry if the schematics drawing is not up to your standard, i've made personal notes here and there and i was not born with ee cert on my lap, this is just hobby grade work. any comments feel free to post, ymmv and cheers...

[Zero999]

I found my oscilloscope would work down to 36V.

Why not skip this complicated power supply and connect many, smaller lithium ion cells in series?

[Mechatrommer]

managing (charging) many lion batteries is a bit tedious on my side... 36V will need at least 9S cells, and that for a single line (the strongest i've found for 18650 batteries is 2600-3300mAh quality brand at $10 for a cell cheapest, ie $90 for single line, if want to get stronger battery add another line of 3.3Ah is another $90, i dont think 18x18650 cells will fit in the enclosure. if i do that anyway i will expect some complaints about how to properly balance them... using a single 3S lipo battery is easier for me and cheaper. the one i used in the above pictures is 7000mAh (china spec i guess) 3S 30C lipo about $45...

Why not skip this complicated power supply and connect many, smaller lithium ion cells in series?

not complicated at all. i've read a guy glued few buck/boost converters to a board for this purpose. this is just a few buck/boost/pump converters, but since i've designed myself i need to tune some components value myself, cheers.

[viperidae]

Surely switching to a more modern SMPS IC would give you much better efficiency. 80% shouldnt be hard to hit, without any more components. The inductors and capacitors will probably be smaller too, running at a higher frequency.

[Mechatrommer]

80% shouldnt be hard to hit, without any more components. The inductors and capacitors will probably be smaller too, running at a higher frequency.

the power sucker i guess mostly are the boost and inverter charge pump converter, not really specific to the smps controller. if anyone can point out efficient boost/inverting converter app-note, posting here is appreciated. the spec for standard DS1000E PSU from datasheet is consuming 50W of power, my PSU in real life is only 16W (12.6V @ 1.3A), the 28W previously mentioned was using slightly low value resistors across power rails to test/simulate slightly worse than what a DS1052E actually consumed... ymmv.

btw now another problem emerged. while tinkering last night, my DS1052E got crashed and now CH1 and CH2 traces are missing, i did self calibration but didnt fix it. if anyone know how to fix it or any thread that i missed, i will be glad if someone can point out, thanks. last night i reinstalled the PSU after fixing few aspect, the last i tested, noise floor improved significantly...

[m3vuv]

 last night i reinstalled the PSU after fixing few aspect, the last i tested, noise floor improved significantly.
what does that mean in english?,as that doesent make sense!,what are aspects,are they part of the psu?.