Inverted Power Rail

Attempt 1: LT1931 #

i initially tried to tackle the -8V power rail first. after some intense searching on the internet, i stumbled across the LT1931. i thought that this would be a good choice to do some testing with (it wasn’t) as it seemed like a very small package capable of delivering enough current to meet my requirements - the LT1931 advertises a 1A switching current, even at 50% duty cycle this should be enough to deliver my desired 400mA.

there was one circuit configuration in particular that stood out to me - it was almost exactly what i needed! LT1931 Circuit

LT1931 Calculations #

i calculated the circuit to provide a -8V output and searched for some components to suit the newly calculated values. unfortunately i don’t have the exact components and calculations on hand but perhaps i’ll consider redoing this at some other point.

LT1931 Assembly #

i ordered some of the LT1931 ICs and the components to suit, and some SOT-23 break-out PCBs suitable for use on a solderless breadboard (i think this was my first mistake). i didn’t design a PCB for this at this point in the project as i had never done PCB design before, nor was i ready to try and take a leap into the deep end. for context, this was the first piece of electronics design i had delved into in a good 2-3 years.

once everything had arrived i constructed the circuit on my breadboard and unfortunately observed no useful output. in fact, i tried with 3 different LT1931 ICs, 4 different solderless breadboards, and 3 sets of components all to no avail. i was quite stumped as to why this might not have been working, but for now i decided to take a break on this design. i did consider coming back to this IC at a later date but they are quite costly and not abundantly available - so that is the end of this attempt.

Attempt 2: LM46002 #

the next IC that caught my eye was the texas instruments (TI) LM46002. while this didn’t advertise an inverting configuration, i found some resources ( a report and a video from TI). from these resources, it seemed the LM46002 demonstrates some promising properties of the IC, and appeared to be aligned with my design requirements.

Inverting circuit and calculations #

the LM46002 sets the output voltage using a voltage divider set with a feedback voltage. the configuration can be seen in the image below.

LM46002 Voltage Feedback

where;

$R_{FBT}$ is the top feedback resistor
$R_{FBB}$ is the bottom feedback resistor
$FB$ is the feedback voltage pin

the datasheet defines the relationship between the output voltage $V_{out}$ and the feedback resistors as follows. $$ \begin{align} R_{FBB} = \frac{V_{FB}}{V_{out} - V_{FB}} R_{FBT} \end{align} $$ where $V_{FB}$ is the feedback voltage (typically 1.011V).

for revision A, i chose the values of $R_{FBT} = 68k\Omega$ and $R_{FBB} = 10k\Omega$. if we rearrange the equation to solve for $V_{out}$, we get the following;

$$ \begin{align*} V_{out} &= \frac{R_{FBT}}{R_{FBB}}V_{FB} + V_{FB} \\ &= \frac{68k\Omega}{10k\Omega}\times1.011V + 1.011V \\ &= (6.8 \times 1.011) + 1.011 \\ &= 7.89V \end{align*} $$

for revision B, the values i chose were as follows.

PCB Design; Revision A #

my first approach was to design the same circuit used in the resources previously mentioned onto a PCB - as an attempt to get something happening. the below images are my first revision of the LM46002 PCB. LM46002 Revision A PCB LM46002 Revision A PCB assembled these boards did appear to function, but not at the -8V value as intended (despite my calculations). i even tried to solder a potentiometer in-line as an attempt to adjust the output value (something i wish i did in this design anyway). for some reason, these boards would provide approx. -20V at the output.

there appears to be a common theme here; either i’m bad at maths, or the datasheets are incorrect (i’m probably bad at maths). since starting revision B of the LM46002 PCB, i went back to this revision to attempt to revise and review the inverted rail design but they all stopped working (as of 28/May/2024). more to come on that at some point. for now, i’d rather work with a solution that is already existing and reliable enough, which is where the third attempt comes in (LM2596).

Attempt 3: LM2596 #

i found this super super interesting video which may appear to solve my problems for this (and allow me to progress with the alternating voltage signal). i have ordered some of the boards to act as interim inverted rail power supplies during testing while i finalise an approach to the inverted rail. read over on the H-Bridge pages to see what i’m working towards there.

floppa net