Ping John Larkin: Lascar power supply noise

Jun 16, 2011 3 Replies
J

Say John... have you ever looked at the noise on your (switching) Lascar power supplies?



I'm using one here to power a PLL and noticed my close-in phase noise was about 20dB worse with the Lascar than with a linear power supply. I'm amazed it makes that much difference!



(It's power supply->maybe 3' of twisted-pair conductors->power connector on board->10uF bulk capacitor -- the current draw on the board is perhaps 25mA or so -- ... and then the various chips on the board have their own 100nF caps nearby. This is only a two-layer board, though.)


---Joel



Vote
Twitter Facebook LinkedIn
J

It's a switcher, so it will be noisy. Let me check...

OK, with a 10M scope probe touching one output terminal, I see about

150 v p-p of 60 Hz hum, and about +2 volt spikes riding on that, 20 KHz, and a smaller ringy thing at a couple hundred KHz, clearly discontinuous switching.

With the output coax-ed directly into the scope and a modest load,

5V/100 ohms, I'm seeing a confused noise mess, about 20 mv p-p, with 40 KHz spikes around 40 mv. Some of this stuff is probably ambient RF, from Sutro Tower. I'd expect it to have more low-frequency noise than a good linear supply, but I'd have to verify that, and a marketing meeting is in the way.

Not all that bad for a switcher.

I do wish it went down to 0 volts. But a pretty nice supply, overall.

John

Vote
Twitter Facebook LinkedIn
J

Thanks for measuring it, John... I would have myself except that my last day here is tomorrow, so I'm madly scrambling to finish things. :-)

(Next week is then trying to get the house ready to sell, and then it's back up to Portland for the next job...)

I bet it's the discontinuous switching that causes the increase in PLL phase noise -- I've seen that before on a DSP's internal DPLL-based clock source, where it would have a hard time maintaining anything approaching reasonable jitter without very careful attention to the DSP's power supply.

---Joel

Vote
Twitter Facebook LinkedIn
J

I think switchers have extra low-frequency noise, too. Nanoseconds or picoseconds of jitter in their PWM comparators translates to output voltage wiggle. Loop bandwidths are low, too, compared to what you can do with a linear regulator.

The 150 v p-p 60 Hz unloaded common-mode voltage is a little unexpected, considering that both the scope and the power supply have

3-prong line cords.

John

Vote
Twitter Facebook LinkedIn

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required