3v isolated outputs from lipo, buck/boost isolating converters, or a better way?

[XOIIO]

Hi all, so annoyingly the post I had written up got nuked going through my history, I've been trying to find the chip that had looked good for the buck/boost side, but no joy so far.

Basically, I'm looking at a small project to rid myself of D cell batteries, and I want to replace two pairs of them with a lipo pack in a drop in case, so I don't have to modify the device in any way. Chances are it won't run as long but it will be massively more convenient.

So, I need two 3v outputs, fed from a lipo battery, that's fairly easy with a buck/boost converter chip, however, I need the two outputs to be isolated, as there's a tap off of the middle of the two sets of batteries, and connecting a common ground to the output of a converterwould be a dead short, no bueno.

So, I was looking for a magic chip that handles buck/boost and isolation duty to provide a 3v output from 2.7-4.2v in, but I haven't had any luck, so my next best guess is using a buck/boost converter to obtain the steady 3v (or possibly higher), and then run that through another chip which isolates the output.

I am hoping to avoid linear regulators to make this more efficient since I'll be going from the substantial rating of four d cells down to only 4000mah of lipo. I only need around 36ma on the output overall, so it's nothing particularly intensive.

It does seem that I may need to go for a 3.1 or 2.9v output since straight 3v isn't available for the things I'm currently looking at.

Unfortunately at the tail end of a long day so definitely struggling to find what I had before and probably missing something output. I had been hoping to figure this out solo but loosing my place didn't help. My 3d printing now layer shifting like crazy for no reason is really starting to piss me off too.

[Alti]

Aren't rechargeable D type cells available off-the-shelf?
If the tap current is small then you could split voltage in half/quarter with ICL7660.

[beanflying]

For a quick and dirty Isolated supply the Mornsun range is worth a look. I use a few of the 0505 range which is 5V 1W rated but there is some at 3.3V 1-2W here to look at https://www.mornsun-power.com/html/products/38/sip-dip-regulated-output--0.75-1w.html

If you need a stable 3V then add a LDO Reg after them will cost you a little in waste but still an easy fix.

EDIT: With the power of Beer another quick and sort of dirty option is a pair of these and two 18650's. Providing you remove any leads from the charge side they will be isolated then hook up to charge. They run DC-DC for the 5V boost and an LDO Reg for the 3V output.

Or one of them and add in one of the Mornsun modules for your second and power it from the Holder.

[David Hess]

As an alternative, many simple switching topologies can be configured with a second negative output using a dual inductor so +/-3 volts is possible.

A simple inverter could drive a pair of identical transformers to produce two floating unregulated 3 volt outputs, or a single transformer with a center tapped output to make unregulated +/- 3 volts.

[XOIIO]

Aren't rechargeable D type cells available off-the-shelf?
If the tap current is small then you could split voltage in half/quarter with ICL7660.

I'm sure they are, but then I need to have a dedicated charger for them and I'll no doubt loose it at some point. I have a usb type c lipo charger pcb with low voltage protection, and I can easily add a barrel plug as well, so that makes it dead easy to charge without having to dig around for something I haven't used in months.

18650's won't fit unfortunately, so I'm going to run four 1000mah pillows in parallel. I do already have a charging circuit solution but that one does look nice, cloning that circuit and adding the regulator in would probably be pretty nice. I suppose I could order a couple of those (so I had spare parts), then reverse engineer it to make a schematic and chop out/add in the bits I need, though honestly I'd be fine with paying more for new parts from digikey for the speed, aliexpress has been iffy for me lately with packages being returned to the sellers before they even leave china for some reason.


The batteries normally get discharged at slightly different rates, so this is why I'm running all four lipos in parallel, and wanting two isolated outputs by the way.

[beanflying]

I have a couple of those hanging around for random remote use. You might get some more ideas from Andreas's video on them and also for your eventual solution.

[ajb]

Do you know how the center tap on the device is used?  If the center tap is only used to supply a load between the center tap and the negative end of the battery string then you could use a conventional non-isolated 3V converter to drive the center tap and a conventional 6V converter to drive the whole string, with the grounds of the converters common with the negative end of the battery string.  If the center tap is loaded towards the positive end of the battery string the same idea could work with a negative converter, but if the center tap can be loaded in either direction you'd need something that can source and sink current to drive the center tap and a conventional buck or linear regulator won't work.  If the center tap current is low enough then a linear rail splitter may not be that bad. 

[tooki]

It's almost certain you don't need any isolated DC-DC converters.

Following on the great advice given so far: Is the device powered by the four D batteries ground-referenced at the negative of the four D cells?? If it is not, you could use a bipolar ±3V supply — the -3V would be your new "ground", the power supply ground would be your "+3V" and the real +3V would be your "+6V".

For example, in a recent project, I used a TI dual-rail DC-DC converter IC to generate ±5V. The adjustable version goes as low as +3.2/-2.0V: https://www.ti.com/product/TPS65130

Note that 3.2V isn't any problem whatsoever for a device intended for two 1.5V batteries: a new 1.5V alkaline cell is usually around 1.6V (and sometimes a bit higher!) and with a 36mA load, the voltage of a D cell is not going to sag under load, ergo the device must be able to handle at least 3.2V. (Also, if not otherwise stated, it's usually safe to assume a ±10% margin on power supply voltages.)

If it is ground referenced at the negative of the battery stack, you could just use two positive DC-DC converters. There are tons of them that'd fit the bill. You'd just have one for 3V (for the center tap) and one at 6V.

Try using TI's Webench Power Designer, where you can plunk in your requirements (e.g. 3–4.3V input, 3V 100mA output, no components smaller than 0805, etc.) and it'll spit out designs, complete with simulations, PCB layout, BOM (editable!) that you can then export to your PCB layout software. I used that tool to design 3 more different DC-DC converters in the aforementioned project. (Plus the ±5V, which isn't available in Webench.) It runs of a 1S LiIon battery and from the 2.7V–4.2V battery voltage and generates +3.3V/2A, ±5V/2x250mA, +7.5V/200mA, and +4.5V/3A–24V/2A (MCU-controlled dynamic output voltage).

[XOIIO]

Do you know how the center tap on the device is used?  If the center tap is only used to supply a load between the center tap and the negative end of the battery string then you could use a conventional non-isolated 3V converter to drive the center tap and a conventional 6V converter to drive the whole string, with the grounds of the converters common with the negative end of the battery string.  If the center tap is loaded towards the positive end of the battery string the same idea could work with a negative converter, but if the center tap can be loaded in either direction you'd need something that can source and sink current to drive the center tap and a conventional buck or linear regulator won't work.  If the center tap current is low enough then a linear rail splitter may not be that bad.

Ah, guess I should have posted this before, but here's the schematic, it's a CDV-700 geiger counter. There seem to be a few different versions but this schematic matches the one in mine, minus the weird random slightly gross mark on mine.

Ah, I wasn't aware of split rail regulators like this, that does sound like it might work, I need to hook the geiger counter up to my bench supplies and use my meter to check it a little more, though I'll wait till I haven't just woken up.

Negative voltages are one of those things I haven't really had to mess with as of yet, so perhaps this battery arrangement in the schematic is a common way of getting them and I just didn't recognize it.

[tooki]

Thanks for the schematic, that clarifies what you need. See how the negative end of the battery stack goes to ground when powered on? That means it’s conceptually not a -3V/+3V supply, but a +3V/+6V supply — as the test point voltages actually say. (The circle at the battery center tap says +3V, but the circle to the left, which connects to the positive end of the battery stack, says +6V.)

However, since it’s battery powered and never references any actual ground, you can supply it either way, it’s the same at the end of the day. (Voltage is inherently relative. “Ground” is, in circuits that aren’t actually connected to the earth, nothing but a convention for setting your 0V reference.) In the ±3V arrangement, consider that +3V is 3V more positive than 0V, and 0V is 3V more positive than -3V — and that consequently, +3V is 6V more positive than -3V. In the +3/+6V arrangement, +3V is 3V more positive than 0V, and +6V is 3V more positive than +3V, and that +6V is 6V more positive than 0V. See where I’m going with this?

[viperidae]

Looks like more than just a +3 and +6v supply.

T2 is powered solely by the top two cells.
T1 is powered by the bottom 3 cells, with an extra 0.1v supplied by a voltage divider off the top two cells to bias the feedback coil.

You will need +/-3V unless T1 always draws more current than T2, otherwise your +3V supply would need to sink current.

[tooki]

Good point, I overlooked that.

In that case, the ±3V (or ± 3.2V if using the IC I mentioned) is the way to go!