Author Topic: Amplifier for PSU ripple and noise -- design considerations?  (Read 900 times)

0 Members and 1 Guest are viewing this topic.

Offline guymo

  • Regular Contributor
  • *
  • Posts: 56
  • Country: gb
Amplifier for PSU ripple and noise -- design considerations?
« on: April 18, 2021, 12:03:04 pm »
I'd like to make some ripple and noise measurements of a few PSU designs that I'm comparing and I'm wondering how best to do it.

My scope (Rigol DS1054) seems to have about 1mV pp noise whatever I do. For example with nothing but a 50 Ohm termination attached to the input, that's what i see.

Measuring my cleanest PSU with the best probing I can manage shows about 1.4mV pp noise. I've used the "tip and barrel" probing method, the ground spring method, and a coax with 50Ohm series at the PSU end, followed by a DC-blocking cap and into 50Ohm termination at the scope end. In all cases I see about the same noise. There is clearly something more than zero there, but I can't get a good reading on it.

So I'm considering constructing an amplifier to gain up the ripple/noise of the PSU so that it's more readily measured. At the levels I'm looking at (perhaps 500uV ripple and noise?), I don't think I would need an extremely low noise design -- or would I? I guess bandwidth would be the main concern because "official" ripple and noise readings are measured up to 20MHz so I'd probably better not roll the measurement off lower than that.

Is this a sensible plan? If so, what kind of approach would be best? Linear technologies AN-83 just uses two op amp stages for the gain and one idea I have is just to rig up something simple like that, with overall gain of perhaps 500, on a little battery powered board of some kind. But I've also been wondering about instrumentation amp type topologies, and wondering if I should try for something very low noise that could be useful if I decide I want more performance later... I am in design paralysis now because there are so many possible ways to go with this.

Any guidance from the collective wisdom of the forum would be much appreciated!

 

Online jonpaul

  • Frequent Contributor
  • **
  • Posts: 503
  • Country: fr
Re: Amplifier for PSU ripple and noise -- design considerations?
« Reply #1 on: April 18, 2021, 12:07:07 pm »
Hello the Chinese clone scopes are notorious for marginal BW and noise.

Short the input BNC and see what noise is present at full BW and lowest V/cm.

Repeat on BW limit.

PARD measurements are very easy, with most setups.

We use a coax RG 198/u to BNC short cable wired direct at PSU output terminals.

Never use a probe (1X or 10X) and never used 50 Ohm.

Kind Regards,


Jon
Jon Paul
 

Online jonpaul

  • Frequent Contributor
  • **
  • Posts: 503
  • Country: fr
Re: Amplifier for PSU ripple and noise -- design considerations?
« Reply #2 on: April 18, 2021, 02:21:33 pm »
Hello

"You have a problem with Chinese gears...."

Fine copies

(;-:)

j
Jon Paul
 

Offline strawberry

  • Frequent Contributor
  • **
  • Posts: 346
  • Country: lv
Re: Amplifier for PSU ripple and noise -- design considerations?
« Reply #3 on: April 18, 2021, 03:02:34 pm »
"You have a problem with Chinese gears...."
bottom end scopes have weird software glitches
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 12900
  • Country: us
  • DavidH
Re: Amplifier for PSU ripple and noise -- design considerations?
« Reply #4 on: April 18, 2021, 07:51:41 pm »
Achieving low noise is the easiest part; somewhere between 5 and 15 nV/SqrtHz yields 22 to 67 microvolts RMS or 110 to 335 microvolts peak-to-peak over 20 MHz so exotic parts are not required.  Oscilloscopes have more difficulty here because the devices needed for higher bandwidth are also inherently noisier and drift is an issue.

The real difficulty is probing and low frequency performance.  Single ended measurements tend to be corrupted by common mode noise and low frequency measurements have to contend with flicker noise, drift, and the difficulties of AC coupling.  The ATX power supply specifications recommend making differential measurements even without a more sensitive oscilloscope for this reason.

If I did not go for a discrete design, the simplest thing I would try is an AD8129 difference amplifier with AC coupled medium impedance inputs and a gain of 20 to support a double terminated 50 ohm load.  I think this will do what you want with the advantage of easier probing, although not with oscilloscope probes.  However it will not work down to DC and low frequencies; the low frequency cutoff would not be much lower than 300 Hz depending on the details.

The next step up would be the same AD8129, but with a pair of FET buffers to raise the input impedance allowing for a much lower cutoff frequency and the use of x1 oscilloscope probes which I think would be worthwhile.
 
The following users thanked this post: prasimix, guymo

Offline guymo

  • Regular Contributor
  • *
  • Posts: 56
  • Country: gb
Re: Amplifier for PSU ripple and noise -- design considerations?
« Reply #5 on: April 19, 2021, 11:14:24 am »
If I did not go for a discrete design, the simplest thing I would try is an AD8129 difference amplifier with AC coupled medium impedance inputs and a gain of 20 to support a double terminated 50 ohm load.  I think this will do what you want with the advantage of easier probing, although not with oscilloscope probes.  However it will not work down to DC and low frequencies; the low frequency cutoff would not be much lower than 300 Hz depending on the details.

This is interesting -- thank you. Can you quantify "medium impedance" for me please?

I'm having trouble understanding the reason for the 300Hz lower limit. AD 8129 seems to have input impedance of 1MOhm / 3pF. If I use for example a 68uF cap on the input, why would performance suffer so badly at lower frequencies?

Does "double terminated" mean 50Ohm to ground at both ends of the output cable? The datasheet specifies 150Ohm loads and greater for AD8129 -- is it ok driving something like that?
 

Online jonpaul

  • Frequent Contributor
  • **
  • Posts: 503
  • Country: fr
Re: Amplifier for PSU ripple and noise -- design considerations?
« Reply #6 on: April 19, 2021, 11:31:33 am »
We use TEK 7904 mainframe with Diff amp 7A22 has a 10 uV /div max and BW limiters for LF and HF cutoff. Very handy.

Jon
Jon Paul
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 12900
  • Country: us
  • DavidH
Re: Amplifier for PSU ripple and noise -- design considerations?
« Reply #7 on: April 19, 2021, 11:02:50 pm »
If I did not go for a discrete design, the simplest thing I would try is an AD8129 difference amplifier with AC coupled medium impedance inputs and a gain of 20 to support a double terminated 50 ohm load.  I think this will do what you want with the advantage of easier probing, although not with oscilloscope probes.  However it will not work down to DC and low frequencies; the low frequency cutoff would not be much lower than 300 Hz depending on the details.

This is interesting -- thank you. Can you quantify "medium impedance" for me please?

It would take hundreds of ohms to allow a reasonably sized AC coupling capacitor but then 1x oscilloscopes probes could not be used because they have too much series resistance.  Directly connected coaxial cables might be acceptable if they are short.

Quote
I'm having trouble understanding the reason for the 300Hz lower limit. AD 8129 seems to have input impedance of 1MOhm / 3pF. If I use for example a 68uF cap on the input, why would performance suffer so badly at lower frequencies?

There are two limits.  The input current noise creates a noise voltage through a high source resistance spoiling the low noise so the part must be used with a low impedance source for low noise.  The input bias current creates an offset voltage though the source resistance which adds to the input offset voltage.  So the input resistance should be kept below roughly 4 kilohms.

Quote
Does "double terminated" mean 50Ohm to ground at both ends of the output cable? The datasheet specifies 150Ohm loads and greater for AD8129 -- is it ok driving something like that?

No, double terminated means 50 ohms in series at the source and 50 ohms in parallel at the load, so the source sees a 100 ohm load, but signal is attenuated by 2x which has to be made up for with extra gain.  The AD8129 is intended for video loads which are 150 ohms when double terminated but 100 ohms will work.  If it is a problem, then a separate operational amplifier or other circuit can buffer the output which is what I would do anyway, but it should not be strictly necessary.

The input impedance of the feedback network also increases the output load so this can be increased to help a little bit.

We use TEK 7904 mainframe with Diff amp 7A22 has a 10 uV /div max and BW limiters for LF and HF cutoff. Very handy.

The 7A22 is fabulous for these types of measurements but limited to 1 MHz maximum bandwidth.  The 7A13 has the bandwidth but is pretty noisy at about 120 microvolts rms over 100 MHz because of its complex bootstrapped differential input which allows it to have a +/-10 volt input range.  Its high noise level is why the 7A13's bandwidth limit is 5 MHz instead of 20 MHz; 20 MHz is not low enough to take advantage of its 1 mV/div sensitivity.  The 2225 series of oscilloscopes with 500 uV/div sensitivity are another example where Tektronix lowered the bandwidth limiter to 10 MHz because 20 MHz was too high.
« Last Edit: April 20, 2021, 04:59:42 pm by David Hess »
 
The following users thanked this post: guymo

Offline alm

  • Super Contributor
  • ***
  • Posts: 1380
  • Country: 00
Re: Amplifier for PSU ripple and noise -- design considerations?
« Reply #8 on: April 20, 2021, 07:21:07 am »
There's an old project description with schematics and PCB artwork of an amplifier for this exact purpose here: https://tangentsoft.net/elec/lnmp/

This is focused on low frequency noise (up to 10Hz), but it could also be of interest: https://www.eevblog.com/forum/metrology/diy-low-frenquency-noise-meter/

Neither come close to 20MHz, though.

I think user blackdog may have also published a design here a number of years ago.
« Last Edit: April 20, 2021, 07:46:40 am by alm »
 
The following users thanked this post: guymo

Offline guymo

  • Regular Contributor
  • *
  • Posts: 56
  • Country: gb
Re: Amplifier for PSU ripple and noise -- design considerations?
« Reply #9 on: April 20, 2021, 04:36:43 pm »
There are two limits.  The input current noise creates a noise voltage through a high source resistance spoiling the low noise so the part must be used with a low impedance source for low noise.  The input bias current creates an offset voltage though the source resistance which adds to the input offset voltage.  So the input resistance should be kept below roughly 4 kilohms.

Input current noise is not something I've dealt with before so thank you for reminding me about that.
If the input stage were an RC highpass with 68uF / 1kOhm, the cutoff would be about 2.4Hz. Is there something wrong with that kind of setup?

I didn't imagine input bias current would be much of a problem here: won't it just lead to a DC offset in the mV range which I can AC-couple out at the output stage? An issue could arise if it's too high for the circuit to deal with after the gain is applied but I think we should get away with this -- 1uA input bias current becomes 1mV of offset, so even if I went nuts with gain it would still be manageable.

Quote
double terminated means 50 ohms in series at the source and 50 ohms in parallel at the load
Ah, perfect, just as I did with my original test setup.

You've mentioned scope probes or coax a couple of times, I think in respect of the input to this amplifier stage. I hadn't thought beyond using short wires to connect directly, or a twisted pair. If using coax, what would be the connections? The device under test has two terminals (it's output and 0V) and this amplifier has I suppose three (two inputs and its circuit ground). It's making me feel stupid that I can't figure out what you have in mind here.

Thanks again -- your advice is much appreciated.
 

Offline guymo

  • Regular Contributor
  • *
  • Posts: 56
  • Country: gb
Re: Amplifier for PSU ripple and noise -- design considerations?
« Reply #10 on: April 20, 2021, 04:42:35 pm »
There's an old project description with schematics and PCB artwork of an amplifier for this exact purpose here: https://tangentsoft.net/elec/lnmp/

Nice, thank you. That's a very simple design along the lines of one I had in mind -- very similar to one of the Jim Williams ones in fact. I suppose that more bandwidth can be obtained just by using op amps with higher bandwidth, right? Any pitfalls here?
 

Offline TimFox

  • Super Contributor
  • ***
  • Posts: 2433
  • Country: us
  • Retired, now restoring antique test equipment
Re: Amplifier for PSU ripple and noise -- design considerations?
« Reply #11 on: April 20, 2021, 05:27:04 pm »
Suggestion to cover the lower end of your frequency range:
Princeton Applied Research (PAR, later bought by EG&G) had several good free-standing differential preamplifiers, e.g. the 113  http://research.physics.illinois.edu/bezryadin/labprotocol/par113manual.pdf  that show up on eBay.  Note that the input capacitors in AC coupling are only rated to 200 V, and you have to be careful about the transient when connecting the AC-coupled input to a high DC voltage.  You could build a shielded coupling box with higher-voltage polypropylene capacitors (preferably matched) into suitable resistors to circuit common (ground), with a DPST shorting switch to absorb the initial capacitor-charging transient, to avoid the "damage" input level of +/- 7.5 V. 
For example, 100 nF 600 V capacitors driving 1 megohm resistors form a high-pass filter -3dB at 1.6 Hz.
These units have useful HP and LP filter settings and can drive an oscilloscope easily.  The maximum upper-frequency cutoff for this unit is 300 kHz.  Others are good to 1 MHz.  The minimum gain is x10, and the maximum input signal for linear operation is +/- 500 mV, at gain below 100.
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 12900
  • Country: us
  • DavidH
Re: Amplifier for PSU ripple and noise -- design considerations?
« Reply #12 on: April 20, 2021, 05:29:42 pm »
Input current noise is not something I've dealt with before so thank you for reminding me about that.
If the input stage were an RC highpass with 68uF / 1kOhm, the cutoff would be about 2.4Hz. Is there something wrong with that kind of setup?

Large coupling capacitors tend to create other problems like long settling times and temperature sensitivity, and they may contribute "excess noise", but you could try it.  In order of preference, wet slug tantalum (expensive), sealed dry tantalum, and high voltage low leakage aluminum electrolytic capacitors seem to be the best.  I have not tried any of the modern "polymer" aluminum or tantalum capacitors though.  I have done it myself for this exact application but not with a low noise preamplifier simply because I could always design the preamplifier with high impedance inputs so that large coupling capacitors are not required.

There is something to beware of when using a large AC coupling capacitor with a low impedance input; the current spike from connecting the probe without precharging the capacitor may damage the low impedance input.  This is why some oscilloscope inputs have a "precharge" function to be used before AC coupling is selected.

Quote
I didn't imagine input bias current would be much of a problem here: won't it just lead to a DC offset in the mV range which I can AC-couple out at the output stage? An issue could arise if it's too high for the circuit to deal with after the gain is applied but I think we should get away with this -- 1uA input bias current becomes 1mV of offset, so even if I went nuts with gain it would still be manageable.

You can do that, but ignoring the input current induced offset does not gain much since the input current noise still sets an upper limit on the input series resistance.

Something else to consider, which helps for once, is that with a low impedance source and a large coupling capacitor, higher frequency noise from the input shunt resistor is shorted out into the source so higher frequency noise above the cutoff is actually dominated by the source.  On oscilloscope high impedance inputs, this shows up as a rise in noise level above a relatively low frequency when there is no low impedance input connection but it is seldom noticeable.

Quote
Quote
double terminated means 50 ohms in series at the source and 50 ohms in parallel at the load

You've mentioned scope probes or coax a couple of times, I think in respect of the input to this amplifier stage. I hadn't thought beyond using short wires to connect directly, or a twisted pair. If using coax, what would be the connections? The device under test has two terminals (it's output and 0V) and this amplifier has I suppose three (two inputs and its circuit ground). It's making me feel stupid that I can't figure out what you have in mind here.

I have done it with twisted pair but always got better results using 1x probes or coaxial cable.  Sometimes I attach short coaxial cable "pig tails" which connect coaxially into the 1x probes.

With a differential input, "chassis" ground and "power" or "signal" ground do not have to be the same location.  Presumably in the DUT (device under test) there is a single point ground somewhere that the coaxial or probe shield can connect to.  If the DUT and oscilloscope share power ground, then connecting the coaxial shields to either other and nothing else may be best.
 

Offline guymo

  • Regular Contributor
  • *
  • Posts: 56
  • Country: gb
Re: Amplifier for PSU ripple and noise -- design considerations?
« Reply #13 on: April 25, 2021, 09:14:52 am »
I've been mulling this over a bit and still going round in circles about the basic design to choose.

A major decision point is whether to go for a differential or a single-ended input. Differential would presumably be intended to reduce the impact of common mode noise. Since I need to AC-couple the inputs, I'm concerned that the input RC filters would need very careful matching to achieve good CMRR -- mismatches in the response would result in different versions of the common mode noise being passed to the two inputs, and the differences would then be amplified. Is that a legitimate concern?
 

Offline TimFox

  • Super Contributor
  • ***
  • Posts: 2433
  • Country: us
  • Retired, now restoring antique test equipment
Re: Amplifier for PSU ripple and noise -- design considerations?
« Reply #14 on: April 25, 2021, 02:46:41 pm »
When I built a low-noise differential preamplifier, with AC coupling, I carefully matched the input capacitors (10 uF polypropylene for this design) by measuring them on a DER DE-5000.  The absolute accuracy doesn't matter, just the resolution of the meter (<0.1% for this unit).  Similarly, the input resistors were matched on a decent DMM.  Far above the corner frequency for the high-pass filter, the capacitance match is less critical.
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 12900
  • Country: us
  • DavidH
Re: Amplifier for PSU ripple and noise -- design considerations?
« Reply #15 on: April 26, 2021, 03:21:00 am »
A major decision point is whether to go for a differential or a single-ended input. Differential would presumably be intended to reduce the impact of common mode noise. Since I need to AC-couple the inputs, I'm concerned that the input RC filters would need very careful matching to achieve good CMRR -- mismatches in the response would result in different versions of the common mode noise being passed to the two inputs, and the differences would then be amplified. Is that a legitimate concern?

It is a real problem unless the cutoff frequency is much lower than any signals of interest.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf