Forums

Wall wart spurs

Started by Phil Hobbs November 15, 2016
So this quantum limited photoreceiver again.

After spending yesterday finding several fractional-decibel bits of
buried treasure, it's shot-noise limited above 30 nA in a full 1 MHz
bandwidth, which is pretty good.  CPH3910s rock, and the optimal drain
current is a tradeoff: as you increase it, e_N goes down slowly but
C_iss goes up rapidly.

It has an LM2594 switcher right in the box, which produces spurs too
small to see easily on the spectrum analyzer, even at narrow bandwidths.
 Soft edges, a toroid, a cap multiplier, an electrostatic shield, and
2-1/2 inches of distance get rid of the junk pretty well.

However, some wall warts make the box produce larger spurs--for
instance, a Thinkpad supply produces a spur about 10 dB above the noise
floor in 5 kHz bandwidth.  I can see that this is going to be a generic
problem.

I have some things to try, e.g. a CM choke and so on, but I thought I'd
ask what you folks do about it.

Thanks

Phil Hobbs

-- 
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203
Briarcliff Manor NY 10510

hobbs at electrooptical dot net
http://electrooptical.net
On Tue, 15 Nov 2016 10:44:02 -0500, Phil Hobbs wrote:

> So this quantum limited photoreceiver again. > > After spending yesterday finding several fractional-decibel bits of > buried treasure, it's shot-noise limited above 30 nA in a full 1 MHz > bandwidth, which is pretty good. CPH3910s rock, and the optimal drain > current is a tradeoff: as you increase it, e_N goes down slowly but > C_iss goes up rapidly. > > It has an LM2594 switcher right in the box, which produces spurs too > small to see easily on the spectrum analyzer, even at narrow bandwidths. > Soft edges, a toroid, a cap multiplier, an electrostatic shield, and > 2-1/2 inches of distance get rid of the junk pretty well. > > However, some wall warts make the box produce larger spurs--for > instance, a Thinkpad supply produces a spur about 10 dB above the noise > floor in 5 kHz bandwidth. I can see that this is going to be a generic > problem. > > I have some things to try, e.g. a CM choke and so on, but I thought I'd > ask what you folks do about it. > > Thanks > > Phil Hobbs
Buy your wall-warts from the ASPCA? -- Tim Wescott Wescott Design Services http://www.wescottdesign.com I'm looking for work -- see my website!
On Tuesday, November 15, 2016 at 7:44:10 AM UTC-8, Phil Hobbs wrote:
> So this quantum limited photoreceiver again.
> It has an LM2594 switcher right in the box, which produces spurs too > small to see easily on the spectrum analyzer, even at narrow bandwidths. > Soft edges, a toroid, a cap multiplier, an electrostatic shield, and > 2-1/2 inches of distance get rid of the junk pretty well.
Depending on your switcher frequency, making the 'electrostatic' shield thicker, by many skin depths, gives you a magnetic shield benefit. Seamless (cast or deep-drawn) is best, for obvious reasons. I've seen very low magnetic leakage from cup-and-lid inductors, it may be possible to reduce size (and therefore dipole field) and save money by replacing the toroid.
> > I have some things to try, e.g. a CM choke and so on, but I thought I'd > ask what you folks do about it. > > Thanks > > Phil Hobbs
Are these wall warts Powering your receiver or just nearby and powering something else? m
On 11/15/2016 02:22 PM, whit3rd wrote:
> On Tuesday, November 15, 2016 at 7:44:10 AM UTC-8, Phil Hobbs wrote: >> So this quantum limited photoreceiver again. > > >> It has an LM2594 switcher right in the box, which produces spurs too >> small to see easily on the spectrum analyzer, even at narrow bandwidths. >> Soft edges, a toroid, a cap multiplier, an electrostatic shield, and >> 2-1/2 inches of distance get rid of the junk pretty well. > > Depending on your switcher frequency, making the 'electrostatic' > shield thicker, by many skin depths, gives you a magnetic shield benefit. > Seamless (cast or deep-drawn) is best, for obvious reasons. > > I've seen very low magnetic leakage from cup-and-lid inductors, it may > be possible to reduce size (and therefore dipole field) and save money > by replacing the toroid. >
The switcher fundamental is down below 400 fA RTI, which makes it invisible over the Johnson noise of the feedback resistor in a 100 Hz bandwidth. I'm seeing something at the fourth harmonic, near 570 kHz, but it's only 10 dB above the Johnson noise in the same bandwidth. It would be convenient if I could dither the switcher a bit, but unfortunately the 2594 has an internal oscillator that I can't get at. No other harmonics are visible. The inductor is a nonprogressive toroid, and the only place the shield seems to help is wrapped around that. It's made of 4-mil copper tape, which is couple of skin depths at 600 kHz. It's a bit inconvenient to chase, because those numbers only apply when it's in the box with at least one mounting screw installed. I bodged in a small loop of wire-wrap wire in series with the inductor, so that I could investigate magnetic effects, but there wasn't much difference regardless of how I moved it. The switcher takes +16V and makes -5 for the op amps and -23 for the photodiode bias. The switching waveform is beautiful--nice slow edges, zero ringing or overshoot. The proto has been through a lot, including soldering and desoldering about 30 JFETs to do noise comparisons. It's getting flaky, so I'll ask BLH to stuff another one and we'll try again. Cheers Phil Hobbs -- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC Optics, Electro-optics, Photonics, Analog Electronics 160 North State Road #203 Briarcliff Manor NY 10510 hobbs at electrooptical dot net http://electrooptical.net
On 11/15/2016 02:23 PM, makolber@yahoo.com wrote:
> >> >> I have some things to try, e.g. a CM choke and so on, but I thought I'd >> ask what you folks do about it. >> >> Thanks >> >> Phil Hobbs > > Are these wall warts Powering your receiver or just nearby and powering something else? > > m >
Powering the photoreceiver. It turns out it's internal to my box--what fooled me is that I'm only seeing the fourth harmonic (600 kHz), so I thought it was the computer wart. There's nothing with a switching supply turned on nearby, and it does the same thing when powered from an old school 60-Hz transformer + rectifier supply. The board lives in an uber-fancy Hammond RF shielded box (1457L1201EBK), which drops the switcher stuff down almost but not quite into the noise. There are four mounting screws in the edges of the box, so the grounds are pretty good for ~1 MHz use. Cheers Phil Hobbs -- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC Optics, Electro-optics, Photonics, Analog Electronics 160 North State Road #203 Briarcliff Manor NY 10510 hobbs at electrooptical dot net http://electrooptical.net
On 2016-11-15 07:44, Phil Hobbs wrote:
> So this quantum limited photoreceiver again. > > After spending yesterday finding several fractional-decibel bits of > buried treasure, it's shot-noise limited above 30 nA in a full 1 MHz > bandwidth, which is pretty good. CPH3910s rock, and the optimal drain > current is a tradeoff: as you increase it, e_N goes down slowly but > C_iss goes up rapidly. > > It has an LM2594 switcher right in the box, which produces spurs too > small to see easily on the spectrum analyzer, even at narrow bandwidths. > Soft edges, a toroid, a cap multiplier, an electrostatic shield, and > 2-1/2 inches of distance get rid of the junk pretty well. > > However, some wall warts make the box produce larger spurs--for > instance, a Thinkpad supply produces a spur about 10 dB above the noise > floor in 5 kHz bandwidth. I can see that this is going to be a generic > problem. > > I have some things to try, e.g. a CM choke and so on, but I thought I'd > ask what you folks do about it. >
If your equipment has to run in the presence of laptops that are plugged in there is only one way to make that work: Complete isolation of power into your box. The barrier (usually a toroid of pot core) sometimes even has to be driven sinusoidally. Also, the supply wiring on your side should be 100% symmetrical so it has the least chance of coupling with the outside world which is how such stuff gets around CM chokes. Most wall warts and laptop power supplies have issues with grounding and all their designers really care about is passing EMC. That is radiated and conducted, but conducted only towards the power line, not towards the load. -- Regards, Joerg http://www.analogconsultants.com/
On 2016-11-15 13:10, Phil Hobbs wrote:
> On 11/15/2016 02:22 PM, whit3rd wrote: >> On Tuesday, November 15, 2016 at 7:44:10 AM UTC-8, Phil Hobbs wrote: >>> So this quantum limited photoreceiver again. >> >> >>> It has an LM2594 switcher right in the box, which produces spurs too >>> small to see easily on the spectrum analyzer, even at narrow bandwidths. >>> Soft edges, a toroid, a cap multiplier, an electrostatic shield, and >>> 2-1/2 inches of distance get rid of the junk pretty well. >> >> Depending on your switcher frequency, making the 'electrostatic' >> shield thicker, by many skin depths, gives you a magnetic shield benefit. >> Seamless (cast or deep-drawn) is best, for obvious reasons. >> >> I've seen very low magnetic leakage from cup-and-lid inductors, it may >> be possible to reduce size (and therefore dipole field) and save money >> by replacing the toroid. >> > > The switcher fundamental is down below 400 fA RTI, which makes it > invisible over the Johnson noise of the feedback resistor in a 100 Hz > bandwidth. I'm seeing something at the fourth harmonic, near 570 kHz, > but it's only 10 dB above the Johnson noise in the same bandwidth. It > would be convenient if I could dither the switcher a bit, but > unfortunately the 2594 has an internal oscillator that I can't get at. > > No other harmonics are visible. > > The inductor is a nonprogressive toroid, and the only place the shield > seems to help is wrapped around that. It's made of 4-mil copper tape, > which is couple of skin depths at 600 kHz. >
Cover the whole switcher because the traces to the inductor and the leadframe in the IC are radiating as well. I normally use a suitably sized off-the-shelf shield can from Laird and make sure there is a full ground plane on the other side of any switching-noise carrying traces as well. IOW the whole switcher runs inside a Faraday cage. [...] -- Regards, Joerg http://www.analogconsultants.com/
On Tuesday, November 15, 2016 at 7:44:10 AM UTC-8, Phil Hobbs wrote:
> I have some things to try, e.g. a CM choke and so on, but I >thought I'd ask what you folks do about it.
If you run a switcher far below its rated output load, as would be the case if you're using a laptop supply to run a low-power gadget, it may either not regulate properly at all, or it may go into a cyclical operating mode with duty cycles and transition times that are fast enough to cause EMI problems. To catch this sort of problem reliably, the various regulatory agencies would have to require that the supply be tested with countless different loads. So the best strategy is to either specify a linear wall wart, or a switcher whose rated capacity is a good match for your device. Failing that you'll just have to test them individually until you find one or more candidates that you can live with. -- john, KE5FX
On Tue, 15 Nov 2016 10:44:02 -0500, Phil Hobbs
<pcdhSpamMeSenseless@electrooptical.net> wrote:

>So this quantum limited photoreceiver again. > >After spending yesterday finding several fractional-decibel bits of >buried treasure, it's shot-noise limited above 30 nA in a full 1 MHz >bandwidth, which is pretty good. CPH3910s rock, and the optimal drain >current is a tradeoff: as you increase it, e_N goes down slowly but >C_iss goes up rapidly. > >It has an LM2594 switcher right in the box, which produces spurs too >small to see easily on the spectrum analyzer, even at narrow bandwidths. > Soft edges, a toroid, a cap multiplier, an electrostatic shield, and >2-1/2 inches of distance get rid of the junk pretty well. > >However, some wall warts make the box produce larger spurs--for >instance, a Thinkpad supply produces a spur about 10 dB above the noise >floor in 5 kHz bandwidth. I can see that this is going to be a generic >problem. > >I have some things to try, e.g. a CM choke and so on, but I thought I'd >ask what you folks do about it. > >Thanks > >Phil Hobbs
I know the 2 wire AC input switchers are pretty noisy, I have been itching to try some 3 prong ones that at least have a earth shield and a proper common mode input filter. Output filtering and chokes are a good idea. Cheers