Inspired by Using a DC-DC Isolator to allow single power input to case? (and to stop hijacking that post) I'm checking the output noise of a Traco Power DC-DC converter.
The device I have is the TEN 6-4822WINTEN 6-4822WIN. It translates 48 V into +- 12 V.
In this post, I'm using its +12 V to power my dc load in two situations: with and without fan.
I'm using a very naive approach, directly using the switcher output to power an analog circuit and the fan. And virtually no output filtering.
It is interesting to see what happens in that situation though.
First capture is with the oscilloscope in DC. 2 V and 10 ms / div.
2nd is the same setting, with fan (consumes approc. 150 mA).
3rd is with the scope in AC, 200 mV and 100 µs / div.
The last one i again the same, with the fan spinning.
There are a few takeaways here:
- powering the fan from the analog Vcc is silly. You can't do a proper measurement with the noise induced by the fan motor.
- the DC drops when the fan is on. Not a terrible thing in the eLoad, because no measurements are directly in relation to Vcc. There's a Vref for that.
Vcc needs to be high enough to give the OpAmps headroom to swing and the output FET to regulate. But 12 V vs 11.x V is not an issue. - also without the fan, there is enough noise to require filtering. While the absolute voltage (the DC component) doesn't impact the working, noise on the power lines (AC component) does.
The Traco will be a valid choice to power the load, if I put a noise filter in place.
The fan should not be powered from the analog supply. It either has to be powered from a separate supply, or from the Traco before the output filter.
Another option is to use the Traco's -12 V for the fan ... It'll still dump noise on the ground line but that would be before my filter ...
Thoughts?
Top Comments