Hi there,

I'm making a simple split /-9V supply for a project using an LT1054 (TI-model).

https://preview.redd.it/vunevgshzacd1.png?width=582&format=png&auto=webp&s=e0c18e97298b0065d04589badd61cb717109e20d

This is the circuit I initially built that worked ok. I used tantalum caps as recommended. A problem I noticed was that the output could occasionally become unstable, oscillating between 0 and -V. My suspicion was that C1 was perhaps too small, so I replaced C1 with a 10µF tantalum cap. I tried the circuit again and now it's not working. The OSC-pin still has its sawtooth wave at around 19kHz (biased at ~1.4V, 1.2V Pk-pk)å, the REF pin is 2.5V, FB is at 850mV and the CAP /- pins look like the pic below.

https://preview.redd.it/bzfqpu5vzacd1.png?width=1024&format=png&auto=webp&s=d8ab83b623c742414f8579eecc284dbdcd53edec

The symmetric waveforms are switching noise from my supply at around 65kHz. The negative waveforms I believe are from the IC at about 37kHz.

The output sits at a measly 150mV.

Switching for a fresh LT1054 exhibits the exact same behavior so I'm ruling out a faulty IC for now.

I feel like I'm missing some glaring issue or misunderstanding something?

Update: I *think* I found the issue? There may have been the slightest amount of solder jumping the pins CAP and FB/SHDN (I couldnt see it though, seemed clear to me). I just reheated the FB/SHDN pin and it seems to be stable now?

Update 2: When I switch it off and then on again it fails to start. By briefly touching the FB/SHDN pin it starts normally. If i leave enough time for the capacitors to discharge before switching on the power I need to bridge the FB/SHDN pin with the V pin to power it on again. My guess is that the shutdown feature of the IC is the cuplrit then. I'm not exactly sure from reading the spec sheet how to keep the IC in a constant on-state when powered on?