EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: janaf on April 18, 2015, 10:57:25 am
-
New thread for the discussion on low frequency, very low level, DC biased, noise measurements, like uV, nV noise from voltage references.
A paper on a JFET front amp:
Ultra low-noise preamplifier for low-frequency noise measurements in electron devices
http://www.researchgate.net/profile/Bruno_Neri5/publication/3087527_Ultra_low-noise_preamplifier_for_low-frequency_noise_measurementsin_electron_devices/links/00b7d52dd4825a9b72000000.pdf (http://www.researchgate.net/profile/Bruno_Neri5/publication/3087527_Ultra_low-noise_preamplifier_for_low-frequency_noise_measurementsin_electron_devices/links/00b7d52dd4825a9b72000000.pdf)
-
This is a very interesting circuit in the article. The good thing is that it works with a moderate size capacitor (thus foil types are available) at the input and also allows relatively high source impedance.
However there are two minor drawbacks:
1) the used JFETs are old types not available any more. Choosing the right Fets sets the 1/f noise corner. Finding good modern replacements is needed. The 2SK369 may be a candidate.
2) Performance likely very much depends on the thermal layout and setup. So it's more than just putting the parts together on a board. Especially at low frequencies thermal effects can add extra noise.
-
This is a very interesting circuit in the article. The good thing is that it works with a moderate size capacitor (thus foil types are available) at the input and also allows relatively high source impedance.
However there are two minor drawbacks:
1) the used JFETs are old types not available any more. Choosing the right Fets sets the 1/f noise corner. Finding good modern replacements is needed. The 2SK369 may be a candidate.
2) Performance likely very much depends on the thermal layout and setup. So it's more than just putting the parts together on a board. Especially at low frequencies thermal effects can add extra noise.
Even the 2SK369 is also obsolete? There is the Linear systems LSK369 which should be equivalent, also available in dual in one can.
The 2SK369 is available on ebay. Seller polida2008 who many have good experiences of, has singles in TO-92. Should be matched and thermally coupled....
-
Even the 2SK369 is also obsolete? There is the Linear systems LSK369 which should be equivalent, also available in dual in one can.
The 2SK369 is available on ebay. Seller polida2008 who many have good experiences of, has singles in TO-92. Should be matched and thermally coupled....
Also on Reichelt:
http://www.reichelt.de/2SK-369/3/index.html?&ACTION=3&LA=446&ARTICLE=2198&artnr=2SK+369&SEARCH=2sk369 (http://www.reichelt.de/2SK-369/3/index.html?&ACTION=3&LA=446&ARTICLE=2198&artnr=2SK+369&SEARCH=2sk369)
With best regards
Andreas
Edit: I would use FETs only if I really need a high input impedance.
-
JFETs have to be matched for this amplifier, precise Vgs matching is not critical since it's not a differential amplifier but if they're too far apart they won't be biased right. That said, the switch they use to charge the input capacitor in reasonable time would load a LTZ1000 too much as Andreas pointed out (the low value input resistor in Jim Williams application note probably does too).
If you just want to throw expensive parts at it the LT1128 is so good though I'd say just use it as a buffer. After that you can use a HPF with a low value resistor to the next stage without issue and use a switch to quickly charge the capacitor as well. Then use a LT1028 as a 10000x amplifier and a standard 0.1-10 Hz measurement setup (ala this (http://itist.de/2013/10/op-amp-0-1-10hz-noise-measurement/)). It's not elegant, but should work (LTZ1000 has low dynamic resistance right?).
-
(LTZ1000 has low dynamic resistance right?).
Hello.
Not really compared to a LM399 (< 1 Ohm) or LT1027 (< 0.05 Ohm) reference.
The Datasheet indicates 5mV change for 1mA change at the output = 5 Ohms.
The zener alone is even worse: around 20 Ohms.
If you also want to measure ordinary (precision) zeners (1N829A) you will have around 10-20 Ohms.
So for a noise amplifier I would not go below around 1000 Ohms as input impedance.
With best regards
Andreas
-
Still not enough to push the LT1128 current noise into relevance.
-
The trouble with the current noise is the impedance of the coupling capacitor. This will be likely in the 1 K - 10 k range unless really huge caps are used. So using a high current noise amplifier will need a capacitor of considerable size. Allready the OP27 would still need something like 5 mF. So like a LT1028 would like something like 50mF or more. Finding a good cap may not be easy, as noise and leakage at 10 V are not the usual parameters for electrolytic caps.
I see several option that may work.
1) the JFET amp with good discrete JFETs and good thermal setup, just like first link showed. This may use a foil cap, like 10 µF. Just finding good Fets (low 1/f noise) can be a problem. The old 2SK147 seemed to be exceptional good, especially the 1/f crossover.
2) Using a large mF range cap and an BJT based OP, lke LT1037. The choice of Amplifier scales with the cap. This was proven to work, if good caps are found. Protection of the source is a little difficult, but possible.
3) use a compound amplifier, like Jim Williams in AN124. This at least good a the higher frequency part, e.g. 10 Hz - 1 kHz. Though I don't see much advantage in the 1/f range of both amplifiers, like below about 3 Hz.
4) Using DC amplification to a high voltage (e.g. 100 V) and AC coupling after this. The following stage is less critical and could use a foil cap and a less critical JFET amplifier. This may be tricky with battery supply.
5) Possibly using a Chopper amplifier like LT2057 or similar, with AC coupling in the 10 µF range. Here things are not so clear, as there can be more tricky effects like noise foldback and higher EMI susceptibility. However the data look promising so far, though data on current noise are not as complete as one would like it.
-
(LTZ1000 has low dynamic resistance right?).
The LTZ1000 itself, as mentioned some 20ohm but for the whole circuit, it's very low, milliohms at most as far as I can see. I was surprised to hear that the output took so long to recover after a mA surge. Can't really understand it (not doubting that you are right!)
I put some 10mF+ elyth caps on charge overnight. It looks like the leakage is surprisingly high in the order of 50Mhom at 10V below rated voltage (63 and 80V caps). Typically around 1uA leakage at 50V. It seems the resistance does not increase at lower voltage but it's really hard to measure.
I also had a smaller 1000uF/50V at rated voltage, now running it with a 9V battery. Leakage should be proportional to capacitance, some nA. I'm measuring on it now. The good news is that the noise seems very low. While there is a leakage, the current seems quite stable, drifting very slowly, much below 0.01Hz as far as I can see. So it may cause some offset on amp input, but offset drift should be slow. Right now, after a few hours on battery, measuring -88nA (current flowing from cap to battery), with drift in the order of 10pA/s, on the last digit of my DMM resolution. The current is still drifting towards zero. I have no idea where it will stop. It could be the battery leakage current that is seen. By the looks of it now, it may take days to stabilize. It may of course also be temperature drift. But then the mass of the caps is high so drift should be slow. The voltage is now dropping by less than 1uV/s (10Gohm input DMM).
-
The trouble with the current noise is the impedance of the coupling capacitor. This will be likely in the 1 K - 10 k range unless really huge caps are used.
Agree. Noise will be very dependent on input impedance / input resistor cap and that in turn dependent on amp input current leakage.
Right now I'm trying to see if E-caps are usable (as indicated from the German forum). It looks like it would be very possible. Again, there are tradeoffs as leakage is proportional to capacitance, inversely proportional to voltage rating vs voltage used. So a big cap but not too big.
-
Could someone who has his/her math's fresh in mind do humanity a favor; come up with a simple function for ball park calculating RMS noise from 0.1Hz to 10Hz. I would really like to be able to calculate it, based on (integrating) white noise level and 1/f corner frequency. 8)Possibly also it's derivate, unless that's constant of some physical reason.
-
The current in the cap I'm testing, could be dielectric absorption too?
Which makes me wonder how that influences filtering.
It's now at -46.9nA, still with a very steady drift towards zero current (Has a 9V battery connected)
-
Calculating the RMS noise shloud be realtively easy:
One can integrate the power (e.g. voltage square) over frequency. We can also treat the white and 1/f parts separately, as it is not correlated.
So the white part just gives bandwith (BW) times square of noise-voltage.
The 1/f part takes one more step: The Integral of 1/f² is -1/f. Integrated from f_low to high frequencies this gives 1/f_low.
At the corner frequency f_c, the white and 1/f part are of same size.
So with just an white a 1/f part and integrating to well abov f_c one would get:
U_noise² = U²_n(white) * BW + U²_n(white)*f²_c/ f_low = U²_n(white) * ( BW + f²_c/f_low)
So due to the 1/f part the noise power is inceased by a factor of ( 1 + f_c² / (f_low *BW)).
For the noise density (volts per sqrt(Hz)) take the square root of this factor.
For the LF amplifier there may well be a 1/f² part, e.g. from current noise time capacitor impedance. So we have a third anlog contribution
and get a faktor of
Sqrt( ( 1 + f_c² / (f_low *BW) + f³_c2 / (2*f_low² *BW))) ) to muliply the white noise density.
A "leakage" current in the 50 nA range sound good. The current itself is not so much of a problem, its just the noise that is possibly associated with it that may cause trouble. For filtering, the input AC coupling should have a time constant considerable lower than the measurement limit (e.g. 0.1 Hz). Its better to have a later stage or software filtering will set the lower limit. So 50 nA of leakage at a 10 K resistor to GND gives something like 0.5 mV of offset and limits the amplification of the first stage to something like 1000 times, possibly 100 times, which is OK.
-
Hello,
when I selected my caps (according to the cirquit from german forum)
I had around 2-3 out of 10 capacitors (new old stock) with suitable low leakage current
after two days forming.
According to "branadic" you should use normal standard 85 degree types.
So no low impedance switch mode types which are optimized for low ESR but no low leakage.
Also I cannot understand that someone recommends the "OS-Con" type capacitors:
they are also not optimized for low leakage.
A higher nominal voltage than needed (e.g. 25V) will also help to reduce the leakage current around 7V.
But be aware that any heating (soldering) will increase leakage current.
So cool the pins of the capacitor with a pincer during soldering.
With best regards
Andreas
-
https://www.mikrocontroller.net/topic/207061?page=2#3410803 (https://www.mikrocontroller.net/topic/207061?page=2#3410803)
100 mHz - 100 kHz
.5 nV^2/Hz white noise share
.7 nV^2/Hz over full bandwidth assuming white spectrum, i.e. it also has a very low 1/f corner
-
Not a simple or one device-design fits most topic. To achieve the lowest noise performance requires a LOT more than just circuit design, low noise devices, construction techniques, low noise power supply and ....
The design criteria must be very well defined for the device or item being measured and what the expected signal needs to be for the rest of the measurement system.
"Cook book" or off the shelf IC solutions often never achieve the lowest noise, highest signal fidelity and best overall performance. The best low noise designs are optimized for a specific application.
Bipolar transistors can be low noise depending on what the signal input might be. The do have a problem of noise current dependent on collector current. Data books once published noise contour maps to help designers optimize a particular device to a specific circuit and system.
JFETS don't really have this problem except the larger the JFET's gate area which results in lower device resistance and lower noise caused high input capacitance. One example of a large gate area JFET would be the Interfet process NJ3600 spec'ed at 0.35nV/root-Hz at 30Hz. The 30Hz spec is significant as low noise is difficult to achieve at low frequencies. Design trade off cost for this device, input capacitance of ~600pF at VDS =10V.
Other low noise JFETS are ones made by Moxtek
http://moxtek.com/wp-content/uploads/pdfs/n-channel-ultra-low-noise-jfets/X-RayJFETs.pdf (http://moxtek.com/wp-content/uploads/pdfs/n-channel-ultra-low-noise-jfets/X-RayJFETs.pdf)
These offer low noise and low input capacitance when operated at -100 degrees C.
The Physics folks tend to use germanium JFETS cooled to -xyz degrees C to achieve low noise for target arrays.
http://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/20000031730.pdf (http://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/20000031730.pdf)
Microwave folks often use GaAs FETs and numerous other FETs to achieve low noise. LNA's also appear as parametric amplifiers, to Masers.
Single ended input devices are 3db lower noise than a diff pair front end, except the single ended input is more difficult to manage than using a diff pair.
Going beyond low noise devices, there is an entire world of grounding, shielding, field management and powering to allow optimum low noise performance. Suggested reading: Grounding and shielding techniques by Ralph Morrison. First editions of this book were focused more on noise, grounding and related, later editions omitted some of what was covered in the first editions but added sections on digital related problems.
Trying to get an extremely low noise front end to live with digital back end is always difficult. More often than not, the digital section is located else where from the input section and put to sleep as the digital switching noise tends to get back into the inout section causing added noise and other problems.
There are a host of power supply related, system related problems that hinder low noise and signal fidelity performance with any added digital processing or control system seriously compounding this problem.
This is a rather complex question with a lot of different and complex answers.
Bernice
-
These are mostly forgotten today, Analog Devices 301 Parametric Op-Amp.
http://www.analog.com/library/analogDialogue/cd/vol1n2.pdf (http://www.analog.com/library/analogDialogue/cd/vol1n2.pdf)
http://www.analog.com/library/analogDialogue/cd/vol1n3.pdf (http://www.analog.com/library/analogDialogue/cd/vol1n3.pdf)
Bernice
-
The Art of Electronics, third edition, has a very long chapter on low noise techniques. It contains many practical and theoretical results, some of which are well-known, some of which came as a surprise to me.
For example, I would never have guessed that (for some applications) the lowest noise transistor would be a medium power BJT!
-
Input cap leakage, an epiphany, surely not known anywhere on earth before :blah:
1) Voltage change due to discharge / stray resistance of caps is proportional to stray parallel resistance, RP and capacitance, C. The RP/C is typically constant for a capacitor type, regardless of capacitance.
This means that the voltage in a cap under self-discharge is a characteristic, regardless of size. IE all caps in a series should show the voltage declining at the same rate, regardless of size!
This characteristic rate can be simply measured for one size of cap, and should then be valid for all sizes in the same series (with the usual fine-prints about voltage ratings, reality, different manufacturing lines, boundary effect of can size.......)
Better: it also means that you can compare any two capacitors, different series in a very simple way. Pich for example a 10.000 uF e-cap and a 10uF polyester, charge two to the same voltage, keep one side connected, for example the negative, and measure the voltage difference or current flow between the two. The direction tells which cap leaks most, which is best, regardless of capacitance size!
So you can simply have a shootouts between series, see which one is best. Then if you like, select another cap size from the same series and it should have essentially the same RP/C.
Finally you can of course have a shootout between caps of the same size&series.
This is only valid assuming your measuring device does not add/subtract any current. A good o'l analogue panel type meter would seem the safe choice.
-
Here is another link to an LF JFET based Ampl.:
users.cosylab.com/~msekoranja/tmp/04447683.pdf
Its a little newer (2008) than the old one using obsolete 2SK147. There best choice of JFET was a rather expensiv IF9030 - at least it's still available. The Sk369 (still available , maybe as LSK369) is there second choice - still only 15 nV/sqrt(Hz) at 0.1 Hz.
If this really turns out to be true for new devices, a JFET amp with the 2SK369 should be a rather good solution, likely limited by thermal effects.
-
About getting rid of the expensive capacitors, there's an interesting article on Electronicdesign.com (http://electronicdesign.com/analog/build-your-own-capacitor-free-high-pass-filter).
They replace an integrator with a DAC in a feedback loop ... they used a differential amplifier, but the concept should work for a normal non inverting amplifier as well (you use a non inverting integrator to create the reference voltage for the feedback network).
PS. dropping the stage amplification to say 100x and and replacing Rg with some resistors and a wirewound pot across a 9V battery could work too (amplification would shift a bit with trimming, but nothing too bad).
-
For the expensive capacitor, it seem to be possible to use some types of normal AL-Caps, after selection for low leakage as well. So I don't see a real need for an expansive Ta cap. So it can come down to testing a few caps.
Such a large cap, still has the small problem, that there will be some "drift" due to temperature and time dependent "leakage". Since the low frequency limit should not come from the AC coupling anyway, this is not a big deal.
The large cap will likely need some protection circuit for the source and possible for the amplifier as well (in case of a short). In addition we may want a build in battery to keep it under voltage, even if not used. Still the size of the cap matters, especially if we want to go lower than 0.1 Hz as well: 1 mF at 0.1 Hz has an impedance of around 1.5 kOhms. This is significant for the current noise of a BJT amplifier. A FET based amplifier may get away with a more moderate foil type cap of something like 10µF.
Having a second low noise source to compensate the DC voltage may be an option. However we get the extra noise if that source, though this may be rather low. For this path I think a stack of 1.5 batteries may be enough: the fist DC coupled stage would then only see something like 1.5 V maximum and could use a 10 fold DC amplification at 20 V supply. After this amplification things a relatively simple.
If we don't mind some extra time for measurements and digital calculations afterwards, one could also use the classical way of using 2 separate amplifiers and (now digital) cross correlation. So the "amplifier" box would likely have two ADCs inside and give out digital values (e.g. 16 bit at 60-500 Hz). This also gives RMS noise as a function of frequency, but not direct peak to peak however. The good thing is, that it works even for values much lower than the amplifier voltage-noise - the limit is set by averaging time.
To a limited extend one can also calculate back the amplifiers noise in a single canal setup. This is especially true for an amplifier with low current noise, like a JFET one. However this may introduce systematic errors if noise levels change over time.
-
Did anyone ever try something like a Dicke switch?
The technique sounds very promising, but I really don't care near enough for 1/f ultra low noise to develop something in that direction ; an extended/modified version of the mikroncontroller.net link above serves me just fine for my needs.
(Although I wouldn't mind a similar spec'd differential amplifier with good n×50 Hz CMRR, but that's really only a minor nuisance for me).
Edit: switch selection might be an issue at <1 Hz (thermal offset drift due to coil heating), but should be no issue with low thermal EMF reed relays.
-
For relays, use the latching type with a short pulse to change over, like the lymex design in one of these precision threads somewhere.
-
another low noise fet is the BF862 suitably low capacitance for the gm it provides, and low noise too, just put 4-8 in parallel.
-
I've seen a report that the noise corner for new BF862 is pretty bad (China vs. Taiwan). They have two SKU's for it, one of which is impossible to get ... murphy's law suggests it's the good one.
J113 doesn't seem half bad, very cheap and tightly enough specced that you can parallel it with source resistors without matching.
-
The die is doped and scribed in europe, it is then sent to one of several packaging plant, I think it is just a poor batch or two. On another forum I visit, one of the members their explained that they are only assembled in asia.
-
Another option may be using some batteries in series to bing down the voltage from the ref., than have a DC coupled amplification (e.g. 10 fold with an OP27) and have AC Coupling only after this.
I like the battery idea :-) But wouldn't batteries generate noise, even if current is virtually zero?
Yes, but not much depending on which sources you believe. From:
[url=http://tf.nist.gov/timefreq/general/pdf/1133.pdf]http://tf.nist.gov/timefreq/general/pdf/1133.pdf (http://tf.nist.gov/timefreq/general/pdf/1133.pdf)[/url]
Using this low noise measurement system we have characterized the voltage noise of 5 different battery types and compared them to a popular voltage regulator and a high performance power supply. We found that the voltage noise of the chemical batteries measured was many decades better than that of traditional power supplies. The lowest noise battery tested was a AA Ni-Cd , with Vn = -205 dBV/Hz at 1kHz. Different battery types exhibit vastly different noise voltage. In the batteries measured the broad band noise voltage appears to be approximately equal to the Johnson noise of their internal resistance.
The internal resistance of Ni-Cd cells is extremely low (a few milliohms typically, depending on capacity) – their results show that Ni-Cd cells have extremely low noise levels - .2nVrms/sqrt(Hz) at 1Hz.
Another source which has almost identical results for an AA Alkaline is page 218 of:
https://books.google.co.uk/books?id=DctAAQAAQBAJ&pg=PA218&lpg=PA218&dq=alkaline+battery+low+frequency+noise&source=bl&ots=zOKnuaYH1x&sig=1yZkjU90_N9JHZcEYGJ2rPQXnGY&hl=en&sa=X&ei=qZMxVfFsy99q_P6BwAM&ved=0CEsQ6AEwCTgK#v=onepage&q=alkaline%20battery%20low%20frequency%20noise&f=true]https://books.google.co.uk/books?id=DctAAQAAQBAJ&pg=PA218&lpg=PA218&dq=alkaline+battery+low+frequency+noise&source=bl&ots=zOKnuaYH1x&sig=1yZkjU90_N9JHZcEYGJ2rPQXnGY&hl=en&sa=X&ei=qZMxVfFsy99q_P6BwAM&ved=0CEsQ6AEwCTgK#v=onepage&q=alkaline%20battery%20low%20frequency%20noise&f=true]https://books.google.co.uk/books?id=DctAAQAAQBAJ&pg=PA218&lpg=PA218&dq=alkaline+battery+low+frequency+noise&source=bl&ots=zOKnuaYH1x&sig=1yZkjU90_N9JHZcEYGJ2rPQXnGY&hl=en&sa=X&ei=qZMxVfFsy99q_P6BwAM&ved=0CEsQ6AEwCTgK#v=onepage&q=alkaline%20battery%20low%20frequency%20noise&f=true (https://books.google.co.uk/books?id=DctAAQAAQBAJ&pg=PA218&lpg=PA218&dq=alkaline+battery+low+frequency+noise&source=bl&ots=zOKnuaYH1x&sig=1yZkjU90_N9JHZcEYGJ2rPQXnGY&hl=en&sa=X&ei=qZMxVfFsy99q_P6BwAM&ved=0CEsQ6AEwCTgK#v=onepage&q=alkaline%20battery%20low%20frequency%20noise&f=true)
However this source has completely different, much noisier, results:
www.hoffmann-hochfrequenz.de/downloads/NoiseMeasurementsOnChemicalBatteries.pdf]http://www.hoffmann-hochfrequenz.de/downloads/NoiseMeasurementsOnChemicalBatteries.pdf]www.hoffmann-hochfrequenz.de/downloads/NoiseMeasurementsOnChemicalBatteries.pdf (http://www.hoffmann-hochfrequenz.de/downloads/NoiseMeasurementsOnChemicalBatteries.pdf)
Since all three sets of results plot the noise on different scales, and the last has results for 4 cells in series, I have normalised them in the attached spreadsheet. I have highlighted in blue the AA Alkaline results from the 3 sources.
Hoffman’s results are pretty much identical to the others at 100Hz but increase very much more rapidly below that. Hoffman used his own amplifier design using 20 paralleled ADA4898-2 op-amps to reduce the voltage noise, described here:
http://www.hoffmann-hochfrequenz.de/downloads/lono.pdf]http://www.hoffmann-hochfrequenz.de/downloads/lono.pdf]http://www.hoffmann-hochfrequenz.de/downloads/lono.pdf (http://www.hoffmann-hochfrequenz.de/downloads/lono.pdf)
He measured his amplifiers's voltage noise density to be 220 pV rms/sqrt(Hz) at 1kHz, rising to 1nV rms/sqrt(Hz) at .1Hz. The current noise will be rather high though with sqrt(20) x the noise of one op-amp (1/f characteristic – 2.4pA/sqrt(Hz) at 1kHz rising to 10pA/sqrt(Hz) at 1Hz) so only suitable for low impedance sources as he notes in his article. However I believe his amplifier to be seriously flawed – I calculated that the noise voltage generated by the current noise flowing through the 10k//160uF input filter will be around 45nV rms/sqrt(Hz) at 1Hz rising to 800nV at .1Hz! This goes a long way to explain the huge discrepancy between his results and the others for the Ni-Cd cells. However Hoffman’s results for AA Alkaline batteries are so much worse than the other sources, (397nV versus Walls’s 4.2nV at 1 Hz) that there must be something else going on as well.
Note that his article has a chart showing the very respectable noise floor with the amplifier input grounded; however the short is applied after his 160uF, 10kohm DC blocking filter so only the voltage noise is shown. The significant current noise can be seen in the same chart showing the noise floor with the input grounded via a 50 ohm resistor – again after the input filter. At .1Hz this is 18nV/sqrt(Hz) compared to 10nV for4 the voltage noise alone; this implies the noise due to the current noise in the 50 ohm resistor is sqrt(18nV^2 – 10nV^2 – 1nV^2), or 14.7nV/sqrt(Hz). The DC filter has a resistance of 7kohms at .1Hz so the current noise totally dominates.
He’s not alone though – looking at the data sheet for the ADA4898-2, on page 14 Analog Devices seem to have made exactly the same mistake! Their test circuit shows an AD743 used to amplify the noise x1000 using an input filter of 1uF/1612kohms!! The current noise of the AD743 will swamp the ADA4898-1’s noise - as shown in their 500nV pk-pk result shown in Figure 45.
The real value .1 to 10Hz AD4898-1 noise I calculate to be around 82nV pk-pk. Since AD generally know what they are doing, this may be a simple mistake showing the wrong values in the schematic etc. but I’m not convinced given that their 500nV pk-pk noise is almost exactly what I calculate it would be from the AD743 current noise spec into the 1uF//1612k input resistance. Don’t believe everything you see in the datasheets.
So, using Ni-Cd batteries to offset the LTZ1000’s 7.2V so that the first stage of the noise amplifier can use a DC-coupled low noise bipolar op-amp such as an LT1028, AD797 or LME49990 should be feasible. However as always there is a catch. The battery voltage is very temperature dependant. As an example I applied the shank of a screwdriver (shank approx 3mm dia.) warmed in my hand, across the side of an AA alkaline battery approx half way between the ends to minimise problems due to thermal EMFs at the connections. Within 3 seconds the voltage was changing by around 1uV/second. Not very scientific, but I believe it would be difficult to control the battery temperature sufficiently for making sub-microvolt measurements over periods of more than few seconds without the battery drift obscuring the results. Using 5 or 6 batteries in series increases the problem proportionally. A bigger battery with its larger thermal mass has slower drift, but also takes longer to stabilize when the test environment changes – if at all.
-
medium power BJT being used as a "low noise" device is rather conditional. It appears "low noise" due to sheer die size which lowers it's bulk resistance resulting in what appears to be low noise. Being a Bipolar, it will have current noise dependent on collector current (beta related) adding to this leakage currents can add to this effect making this a very source impedance dependent "low noise" device. Other problems are die size which has an effect on it's BW, collector base capacitance will be high and if used as a common emitter amplifier miller effect will stunt's if frequency response. Both collector-base capacitance and beta modulates with current and temperature changes. All these reasons an more are why medium sized bipolar transistors are not often used if at all for low noise input stages.
Devices like the LM394 would have been use if a circuit needed a device like this. Alternatively, super beta match diff pairs are also use as low noise input devices. Super beta bipolars goes a ways to reducing noise current, but does not eliminate it.
Want to use individual matched FETS or bipolars as a diff pair/ They will need to be thermally coupled to foster tracking. In the case of FETS, their IDSS or VGS will need to be matched over a range of temperature or there can be serious temperature drift problems. This problem is less with bipolars, but not to be ignored. Oh, any thermal gradient between the two diff pair devices aggravated this problem.
As for power supplies, if the circuit cannot achieve low noise (less than 1nV/root-Hz @ less than 100Hz) the circuit design will be difficult to implement into many systems. Designed properly low noise input sections have good power supply noise rejection. Batteries powered low noise input sections regardless of they batteries internal impedance and possible low noise is simply not the way to get there.
Using a coupling capacitor, specially a electrolytic or similar will cause a host of distortion and signal errors at the signal levels being discussed. EE caps have extremely high dielectric absorption (put a volt meter on a few hundred uF EE cap and measure it's voltage. Charge and discharge it and measure it again) and do not behave any where near like a ideal capacitor. If forced to use a coupling capacitor, it must be polystyrene or teflon film with few exceptions.
Noise from grounding is a serious problem that cannot be ignored. The problem is more difficult and more serious for input sections with a chassis grounded-single ended input. If a differential input section is needed, the impedance for both + and - inputs must have the same impedance, frequency/phase response and gain to idealize common mode rejection.
Knowing what the source impedance is makes all the difference to achieving low noise. Designing an input section for a low ohmic value thermocouple is very different than designing a low noise input section for a electrometer, charge amplifier or similar very high impedance device.
Any digital device like a DAC (i.e., DAC in the FB loop) will produce glitches as they make their switching transitions. These glitch transients will find their way back into the input section where they will be integrated by the devices involved resulting in more noise and distortion. This is one of the many reasons why mixing digits with low noise high fidelity analog can be so extremely difficult to do properly.
A "low noise" device is only the beginning to making a good, low distortion, signal accurate low noise input section. More often than not, really good low noise, low distortion, high performance input sections do not come from an IC, these are done the old fashion way using individual devices.
Bernice
The Art of Electronics, third edition, has a very long chapter on low noise techniques. It contains many practical and theoretical results, some of which are well-known, some of which came as a surprise to me.
For example, I would never have guessed that (for some applications) the lowest noise transistor would be a medium power BJT!
-
Hi,
Did some leakage tests on caps.
First: the failed path: Measuring leakage current by having the cap charged at a constant voltage, measuring the current, turned out to be quite un-practical except for on the large and leaky caps. For others, leakages are sub-nano-amp and beyond my capability.
It's much more practical to charge them all up to some voltage, leave them to rest a while, measure the voltage, rest, measure againg, see the voltage drop, calculate the rest.
Attached you'll find some results. I have measured on some 20+ caps, ranging from 2.2uF up to 33.000uF. Voltage ratings from 35V to 700V. So far I have measured at 100V and 30V. And for a few, for other voltages.
One figure of merit is the leakage expressed in seconds (Mohm*uF). Some manufacturers publish this value. It's typically between 5.000 and 100.000 at the rated voltage. Divide this value by your capacitance to get the leakage resistance for your cap size. I give this value for the measurements I have made, but at the measured voltage. I've made other measurements (not included here) that show that the leakage resistance can be 10x higher when the voltage is half of the rated voltage.
So some results:
Polystyrenes and Polyprops are the best performing as expected. In my measurements, a Russian mil spec polystyrene 10uF was the very best (per uF) but an identical one ended last of all, obviously has a defect, damage. The 10uF polyprops are huge. A large e-cap, 33.000uf RIFA (now Kemet) did very well too. I'ts a very heavy duty, long life type. I did not have any polyesters at hand and not many regular ecaps either. The ecaps I tested so far where mostly either large power supply caps, 10.000 and up, or smaller 1000-3000uF ultra-high ripple/low ESR types.
Right now I'm preparing to measure at around 10V, might be done tomorrow.
Well, results attached.
-
EE caps have extremely high dielectric absorption
There is no significant voltage across it in frequencies of interest. The apparent leakage current it causes can easily saturate the amplifier and/or load the source, but other than that it's irrelevant for AC coupling.
-
Magnitude of signals involved are in nano volts & pico amps.
If the E'lytic or cap in question is theoretically ideal for a capacitor, there might be zero volts across it. Keep in mind a coupling capacitor is in series with the input source and the measuring device's input. In the real world, that E'lytic capacitor behaves more like a battery in series between the input source and measurement device input, at these levels than a device no longer has zero voltage drop due to the effects of dielectric absorption, leakage and other non-linear behavior real world capacitors exhibit.
Don't believe this, take a E'lytic capacitor, put a DC Volt meter with no less than 10 Megohm input resistance across it's terminals and measure the voltage across that E'lytic capacitor. Then charge it up, discharge it, rinse and repeat a few time and see of the initial voltage measured is the same..
Try building a precision sample-hold or precision integrator using a E'lytic or other capacitor with significant amounts of DA, does it work, yes, how well.. that becomes a question of what is acceptable or not acceptable.
Will an E'lytic cap work as an AC coupling capacitor, absolutely. Question is, how much signal error, distortion and more is acceptable?
This analog stuff is not about does it work, it is more a matter of how well does it work, how much does the circuitry add or subtract to the original signal and all of those realities. It is very much shades of gray and the deal one makes with nature with a full understanding that nature will do what nature does regardless.
Bernice
EE caps have extremely high dielectric absorption
There is no significant voltage across it in frequencies of interest. The apparent leakage current it causes can easily saturate the amplifier and/or load the source, but other than that it's irrelevant for AC coupling.
-
Try doing some dielectric absorption test at the capacitor's rated voltage.
Leakage test tell one small part of actual capacitor behavior.
Bernice
Hi,
Did some leakage tests on caps.
-
There will be a DC across these caps, if measuring noise on a voltage reference.
All caps have Dielectric Absorption (memory effects). The question is how much is acceptable. E-caps have much more DA than film caps. Teflon and C0G ceramics probably best, but both are available up to something like 0.22uF
So far, from what I measured and read datasheet, depending on cap availability upper limits, ball park values:
- Up to 1uF, NP0 or teflon? (not measured)
- Up to 10uF: mil Polystyrenes
- Up to 100uF, possibly up to 1000uF:
- Polyprops, not all huge "DC-link" types
- Mil rated Tantal-Silver? (not measured)
- Higher than 1000uF: select e-caps
-
Batteries/ecaps:
Direct current measurements I did on a 1200uF 35V ecap using a 9V PP3 battery;
The e-cap pre-conditioned for a day at 9V, the battery connected for another day.
Then the current flow was still from cap to battery (4nA), ie it seems the battery internal leakage / discharge and / or DA of cap still dominated over anything else...
Unusable cap or battery ???
-
Another thing realized
Without leakage, C=Q/V is constant. But C is temperature dependent, goes up with decreasing temperature. So the voltage over a cap without any discharge will go up when temperature drops.
At the levels measured, the voltage change caused by temperature easily swamps the voltage caused by leakage.
In my case, a one degree C change of the room temperature caused an voltage drift that was larger than the leakage over two hours. |O
Got to go temperature controlled :(
-
Batteries/ecaps:
Direct current measurements I did on a 1200uF 35V ecap using a 9V PP3 battery;
The e-cap pre-conditioned for a day at 9V, the battery connected for another day.
Then the current flow was still from cap to battery (4nA), ie it seems the battery internal leakage / discharge and / or DA of cap still dominated over anything else...
Unusable cap or battery ???
The battery voltage may drift enough with temperature to screw your measurements. I would also measure/log the battery voltage during the testing.
Jim Williams mentioned in some appnote that selected aluminium electrolytic caps can be pretty good for this type of measurement leakage vise but was having problems with short random noise bursts. Apparently there is something weird going on inside the aluminium electrolytics. ( I also remember reading some battery noise measurements where some battery types exhibited random jumps in noise or voltage)
Wet slug tantalum would be your best bet according to Williams but these are crazy expensive. See AN124f from Linear www.linear.com/docs/28585 (http://www.linear.com/docs/28585)
-
I have read what I can find on the subject, including the AN124
I could monitor the voltages, but it would require the volt meter to be switched in for short durations now and then. Any voltage measurement discharges the cap / battery. Even if the DMM has 10Gohm inner resistance, it seems that this is much less than the caps and battery.
I suppose I could build a Terra-ohm buffer 8)
I could probably use a stable voltage ref as voltage source. It would need buffering output and isolating he capacitive load.
Wet slug tantalums, aka tantalum-silver caps: there are Russian mil surplus parts on ebay: I bought a 300uF/25V $7.50, will post results when I get it. One of the first posters in this? thread used something similar.
I also found some cheep Taiwanese motor starter cap, claimed to be Polyprop?, 800uF, brand "ABC" at around $15 eBay auction: #http://www.ebay.com/itm/231334817420?_trksid=p2060353.m1438.l2649&ssPageName=STRK%3AMEBIDX%3AIT Will post results of that one too.
One can also look for "DC-link" polyprops. These high voltage caps may be really good when voltage de-rated.
E-cap "blurp", if it's rms measurements, not long-time ptp, then those instabilities could be rejected. The temperature dependency should show as very low frequency drift, should be OK. Dielectric absorption; cant quite wrap my head around how this will work at very low noise, low frequencies. It will add some non-linearity. The RIFA ecap I got good results with are DC-link listed. But I should need have a 1000uF -size to bring leakage current down.
Need a climate / temperature chamber too. Drifting away from the voltage reference circuit :(
-
Simulating a bit, you'll need about 22 mF to get close to the noise limit with a single BJT amplifier at 100 mHz (ten times less would be good enough for JFET ... simply because even a really good one is easily 10 times worse to begin with).
If you want high input impedance you will need either some input circuitry (inductor+comparator+multiplexer) to quickly fill the cap when necessary using a temporary buffer or use a manual method to switch in/out such a buffer (with risk of human failure). If you used Jim Williams his buffer on a 7V reference you'd be pulling 7 mA initially.
PS. actually if you attached 7V to Jim William's AN124 amplifier with a discharged input capacitor you'd pull ~30 mA, Q1 would get forward biased.
-
At least some E-Caps seem to work (much lower noise that amplifier). The leakage current is just one thing that may be a little easier to measure than noise itself. One can expect that leakage current will also cause some noise. So this is just a first test. The final test for the caps would be to check for the noise of the whole system with a low noise voltage source (e.g. battery). The second step that may be needed is to get a rough measurement of the battery noise, if we don't take for granted that they are low noise - this could be done by measuring several batteries back to back.
The noise level of the zeners refs. is not so extremly low, so the simple system with a e-cap and more or less conventional OP-Amp (e.g. LT1037) can be sufficient.
The effect of leakage is to give a small DC offset - unless the amplification of the first stage is really high, this should not be a problem. DA and temperature drift will just give a slowly varying background, so not much in the range of 100 mHz and up.
A large cap will need some protection for the source. However this may already be necessary at something like 1-10 µF used with a pure JFET amplifier. Even a DC coupled amplifier may need some kind of input current limiting, in case the amplifier is not powered. Nearly all low noise OP Amps have something like diode across the inputs.
-
Unpowered input protection could be a relay or a back to back MOSFET combo (or two if you want to keep it symmetrical). It's the only real way to do it with low impedance.
Leakage current will cause noise, but that's shunted by the capacitance ... unless it's modulated by some other source (sound/temperature) in frequencies of interest I don't see how it can contribute much noise.
-
Marco, could you please give some numbers on what you where calculating, what noise limit you where looking at? Current and voltage noise versus the LTZ1000?
All caps, especially ecaps, need to be biased for long periods to get rid of DA. Or are you talking about measuring DA too?
Edit: Batteries for cap test: Alkalines are really wrong. Then best are probably Lithium's (not Li-ion) like Energizer Ultimate and other brands of the same type which have ultra-low self discharge.
Simulating a bit, you'll need about 22 mF to get close to the noise limit with a single BJT amplifier at 100 mHz (ten times less would be good enough for JFET ... simply because even a really good one is easily 10 times worse to begin with).
If you want high input impedance you will need either some input circuitry (inductor+comparator+multiplexer) to quickly fill the cap when necessary using a temporary buffer or use a manual method to switch in/out such a buffer (with risk of human failure). If you used Jim Williams his buffer on a 7V reference you'd be pulling 7 mA initially.
PS. actually if you attached 7V to Jim William's AN124 amplifier with a discharged input capacitor you'd pull ~30 mA, Q1 would get forward biased.
Simulating a bit, you'll need about 22 mF to get close to the noise limit with a single BJT amplifier at 100 mHz (ten times less would be good enough for JFET ... simply because even a really good one is easily 10 times worse to begin with).
If you want high input impedance you will need either some input circuitry (inductor+comparator+multiplexer) to quickly fill the cap when necessary using a temporary buffer or use a manual method to switch in/out such a buffer (with risk of human failure). If you used Jim Williams his buffer on a 7V reference you'd be pulling 7 mA initially.
PS. actually if you attached 7V to Jim William's AN124 amplifier with a discharged input capacitor you'd pull ~30 mA, Q1 would get forward biased.
-
It was kind of what I thought of initially. I'ts much cheaper to throw some expensive LT1128/1028 at it than spend days making a more elegant solution. I'll take a look at the calculations again.
If you just want to throw expensive parts at it the LT1128 is so good though I'd say just use it as a buffer. After that you can use a HPF with a low value resistor to the next stage without issue and use a switch to quickly charge the capacitor as well. Then use a LT1028 as a 10000x amplifier and a standard 0.1-10 Hz measurement setup (ala this (http://itist.de/2013/10/op-amp-0-1-10hz-noise-measurement/)). It's not elegant, but should work (LTZ1000 has low dynamic resistance right?).
-
Thanks. Looks interesting. But my German is poor. Anything one needs to know except the schematic? Like, which components / values are critical, which are not. Not used to this kind of amps...
https://www.mikrocontroller.net/topic/207061?page=2#3410803 (https://www.mikrocontroller.net/topic/207061?page=2#3410803)
100 mHz - 100 kHz
.5 nV^2/Hz white noise share
.7 nV^2/Hz over full bandwidth assuming white spectrum, i.e. it also has a very low 1/f corner
-
Thanks!
So let's see if I get this right. Simplified and including only 1/fy, for y=1, we get;
Urms(0.1-10) = UW2x(10+fc2x1000)
where
UW = Uwhite / sqrtHz
fc = corner frequency
Have mercy, 30 years since maths classes...
Calculating the RMS noise shloud be realtively easy:
One can integrate the power (e.g. voltage square) over frequency. We can also treat the white and 1/f parts separately, as it is not correlated.
So the white part just gives bandwith (BW) times square of noise-voltage.
The 1/f part takes one more step: The Integral of 1/f² is -1/f. Integrated from f_low to high frequencies this gives 1/f_low.
At the corner frequency f_c, the white and 1/f part are of same size.
So with just an white a 1/f part and integrating to well abov f_c one would get:
U_noise² = U²_n(white) * BW + U²_n(white)*f²_c/ f_low = U²_n(white) * ( BW + f²_c/f_low)
So due to the 1/f part the noise power is inceased by a factor of ( 1 + f_c² / (f_low *BW)).
For the noise density (volts per sqrt(Hz)) take the square root of this factor.
For the LF amplifier there may well be a 1/f² part, e.g. from current noise time capacitor impedance. So we have a third anlog contribution
and get a faktor of
Sqrt( ( 1 + f_c² / (f_low *BW) + f³_c2 / (2*f_low² *BW))) ) to muliply the white noise density.
A "leakage" current in the 50 nA range sound good. The current itself is not so much of a problem, its just the noise that is possibly associated with it that may cause trouble. For filtering, the input AC coupling should have a time constant considerable lower than the measurement limit (e.g. 0.1 Hz). Its better to have a later stage or software filtering will set the lower limit. So 50 nA of leakage at a 10 K resistor to GND gives something like 0.5 mV of offset and limits the amplification of the first stage to something like 1000 times, possibly 100 times, which is OK.
-
It was kind of what I thought of initially. I'ts much cheaper to throw some expensive LT1128/1028 at it than spend days making a more elegant solution. I'll take a look at the calculations again.
Hello Jan,
if you read DN6 or DN140 of LT you will find that LT1128/1028 is only useful up to
a maximum input impedance of 400 Ohms below 10 Hz.
http://www.linear.com/docs/4187 (http://www.linear.com/docs/4187)
http://www.linear.com/docs/4345 (http://www.linear.com/docs/4345)
Above 400 Ohms the LT1037 or LT1007 is the better choice.
With best regards
Andreas
-
Marco, could you please give some numbers on what you where calculating, what noise limit you where looking at?
No, I meant to get close to the noise limits of the transistor ... none of this is really relevant to the LTZ1000 measurement.
For the LTZ1000 measurement I found the OPA140 ... similar voltage noise to the LTC2057 but no current noise to worry about, so the impedance of an input HPF becomes irrelevant. I'd use 220 uF worth of film capacitors with 100 kOhm. Only thing which you'd have to check is what kind of current the OPA140 input pulls when there is a large differential voltage (ie. while the input capacitor is charging).
Unless you really want to build something out of discrete components :)
-
There is an update: App note 355 where they made separate plots per frequency range. Up to 1komh, 1028 still in the lead but any of those mentioned should probably do. Noise on the non-inverting input, coming in with the DC cap and input impedance, is not included in any of those graphs. Must be added. Will do some calculations tomorrow.
http://cds.linear.com/docs/en/design-note/dn355f.pdf (http://cds.linear.com/docs/en/design-note/dn355f.pdf)
-
Thanks for the OPA140 tip, new to me.
Right on the first page of the datasheet: 250nVPP, 0.1Hz to 10Hz . Should be very close to the total for that amp and is some 14dB down from the LTZ1000, should be enough! And supplies up to +/-18V :-+
Marco, could you please give some numbers on what you where calculating, what noise limit you where looking at?
No, I meant to get close to the noise limits of the transistor ... none of this is really relevant to the LTZ1000 measurement.
For the LTZ1000 measurement I found the OPA140 ... similar voltage noise to the LTC2057 but no current noise to worry about, so the impedance of an input HPF becomes irrelevant. I'd use 220 uF worth of film capacitors with 100 kOhm. Only thing which you'd have to check is what kind of current the OPA140 input pulls when there is a large differential voltage (ie. while the input capacitor is charging).
Unless you really want to build something out of discrete components :)
-
Don't get fooled, noise is actually 8nV/root Hz @ 10Hz (fig. 11, Input noise density -vs- freq). In the world of low noise amplifiers, this is not even close to being low noise.
http://www.ti.com/lit/ds/symlink/opa140.pdf (http://www.ti.com/lit/ds/symlink/opa140.pdf)
Compare that with a truly low noise device, Interfet process NJ3600. 0.4nV/root Hz @ 10 Hz. The op amp is not even close.
http://www.interfet.com/process/NJ3600L/ (http://www.interfet.com/process/NJ3600L/)
To achieve truly low noise, the amplifier will need to be designed and built with extreme care using discrete parts, no way around this fact.
The coupling capacitor can be easily deleted by applying an offset to the feedback summing junction of the amplifier if the input section has enough dynamic range (discrete devices using +/- 30V rails or higher, Forget about saving power, it is a matter of performance). In the case of measuring voltage reference noise, the signal of interest is small but sitting on a DC offset and the offset voltage should be quite stable, low drift over time and the impedance of the voltage reference should be low. Key aspect of this measurement is the input section. Once the input signal has been increased with good fidelity, then it can be processed as required (filtered, peak detected or etc).
Heat increases noise.
Voltages and currents involved can be viewed as water with random particles floating in the water representing noise. A simple numeric figure does not indicate or represent how much or type of these particles (noise) are in the water. This is why dealing with noise as a simple numeric figure is not accurate or representative of what could actually happening.
Get a Tektronix 7A22, 5A22N or AM502 differential input amplifier or Stanford Research Systems SR560 differential amplifier as these are quite useful for this type of work along with filters and peak detectors as required.
http://www.thinksrs.com/products/SR560.htm (http://www.thinksrs.com/products/SR560.htm)
Bernice
Thanks for the OPA140 tip, new to me.
Right on the first page of the datasheet: 250nVPP, 0.1Hz to 10Hz . Should be very close to the total for that amp and is some 14dB down from the LTZ1000, should be enough! And supplies up to +/-18V :-+
Unless you really want to build something out of discrete components :)
[/quote]
-
OPA827, not much lower noise than the OPA140 at 6nV/root-Hz @ 10Hz.
See figure 1, input noise density -vs- frequency.
http://www.ti.com/lit/ds/symlink/opa827.pdf (http://www.ti.com/lit/ds/symlink/opa827.pdf)
How many op-amp can be paralleled before the problems occur? After the first pair, the noise does not go down much with successive op-amps added.
Bernice
OPA140 is great for measurements with RC highpass filters on the input. As the current noise of this device is very low, you can parallel devices to achieve even lower noise.
OPA4140 is the quad version of OPA140, great for paralleling.
From my analysis a OPA827 (Lowest voltage noise) is also even greater for an AC-coupled / low frequency / low noise amplifier.
-
Whatever comes out with >= 250nVPP, 0.1Hz to 10Hz and gives a suitable input cap size, is good enough for me for now.
For the OPA140 they have specified 0.5pA input bias & offset. Some quick calculations show that with 1000 ohm Rin and D/C blocking, the voltage noise still outweigh current noise with a large margin; paralleling pays off in this case.
-
250nV at the output between 0.1Hz to 10Hz implies gain, question is how much gain?
Very generally speaking, FET input op-amps have low current noise (this is why FETs are used for low noise amplifiers more often than not as it allows lower noise to be achieved over a broader range of source impedances. See my initial reply).
LT1007 has lower noise. What matters is the 1/f corner freq. not just the volts/root-Hz or advertised output noise volte pk-pk. In all cases, spec sheets are designed to sell parts and play the game of spec-manship to move parts.
http://cds.linear.com/docs/en/datasheet/100737fbs.pdf (http://cds.linear.com/docs/en/datasheet/100737fbs.pdf)
or
AD797:
http://www.analog.com/media/en/technical-documentation/data-sheets/AD797.pdf (http://www.analog.com/media/en/technical-documentation/data-sheets/AD797.pdf)
Do consider NOT using an coupling capacitor and using a low noise current at the feedback summing junction to offset the DC input as required, Not using the coupling cap avoids a host of problems. With +/- 15 volt supplies, there should be enough dynamic range to accommodate a 10volt ref. given the op-amp has enough common mode range.
Do keep the feed back impedance low as this will go a ways to keep the overall noise low and keep in mind the relationship between feedback loop impedance -vs- source impedance.
Grounding, shielding, overall layout, powering and all those tin details will make or break how well it works regardless of device specs.
Devices in parallel gains a few db in lowering noise with rapidly diminishing returns as the number of devices increases.
Lower impedance generally lower noise, Lower temperature generally lower noise, Lower bandwidth lower noise. For higher source impedances, bipolar input devices are at a dis-advantage.
Beyond this gets into what type of noise aka spectra and a host of other specifics.
Side curiosity, LVA zeners. They are low noise by virtue of their low dynamic impedance, sharp knee at low zener current.
http://ams.aeroflex.com/metelics/micro-metelics-prods-zener-LVA.cfm (http://ams.aeroflex.com/metelics/micro-metelics-prods-zener-LVA.cfm)
Bernice
Whatever comes out with >= 250nVPP, 0.1Hz to 10Hz and gives a suitable input cap size, is good enough for me for now.
For the OPA140 they have specified 0.5pA input bias & offset. Some quick calculations show that with 1000 ohm Rin and D/C blocking, the voltage noise still outweigh current noise with a large margin; paralleling pays off in this case.
-
Whatever comes out with >= 250nVPP, 0.1Hz to 10Hz and gives a suitable input cap size, is good enough for me for now.
I don't think you can go to <250 nVpp(!) with op amps. Even the very good discrete design linked earlier in this thread has about 300 nVrms, about 3000 nVpp input referred noise.
https://www.mikrocontroller.net/topic/207061?page=2#3410803 (https://www.mikrocontroller.net/topic/207061?page=2#3410803)
100 mHz - 100 kHz
.5 nV^2/Hz white noise share
.7 nV^2/Hz over full bandwidth assuming white spectrum, i.e. it also has a very low 1/f corner
-
I'd also guess that it could be quite challenging to devise a suitably low noise, low drift comparison / subtraction voltage. It probably makes more sense to have a I-controller with very low bandwidth to do that (eliminates drift). As integrators with low bandwidths become unhandy quickly a digital controller would be an option, too.
-
I don't think you can go to <250 nVpp(!) with op amps. Even the very good discrete design linked earlier in this thread has about 300 nVrms, about 3000 nVpp input referred noise.
I'm sure it's not trivial to build it to that quality, but there are scope shots in the datasheet.
These Burr Brown opamps seem to be just plain magic ... so were the Linear BJTs at the time (still are really, by a factor 2x over the competition) and so is the LTC2057. I wonder if Jim would even have bothered going discrete if Linear made an integrated JFET/CMOS amplifier that good :) (JRC does a CMOS amplifier with similar specs, the NJU77806, but it's lower voltage range and like most of their parts unobtainium.)
PS. if you don't go overboard with the amplification I don't think you need to worry about DC and ultra-low frequency effects too much, with the PP film caps you can use in this case leakage won't be an issue (as I said 220uF is plenty, that gives you a 0.01 Hz with 100kOhm and thus less than 10 kOhm equivalent noise over the entire relevant spectrum). Any remaining DC offset can be cleaned up in a subsequent high pass filter.
-
250nVptp 0.1Hz to 10Hz IS input referred level and comes from the datasheet of the OPA140, compared to 1200nVptp for the LTZ1000. Paralleling a quad brings it down to half, 125nVptp.
Influence by gain is Rf only, as it's noise gain is unity, I.e. typically totally marginal.
Check the datasheet for this amp, voltage noise is unusually low for a FET amp. With "normal" low noise FET op-amp data, this noise level at low frequency would be impossible.
With DC blocking, the input impedance becomes a serial noise source (like a serial resistor on the inverting input) With FET amps, it can be kept low while paralleling improves voltage noise.
Ultra-low noise amps like LT1028 have <50nVrms (0.1-10Hz) but the current leakage ruins it with DC blocking.
The problem with biasing voltage is that the bias voltage has to have lower noise than the one measured or the bias voltage noise will ruin the measurements. Unless one comes up with some noise cancelling magic.
Also, I'd want high gain in the very first stage, to get out of the noise sensitive domain.
250nV at the output between 0.1Hz to 10Hz implies gain, question is how much gain?
Very generally speaking, FET input op-amps have low current noise (this is why FETs are used for low noise amplifiers more often than not as it allows lower noise to be achieved over a broader range of source impedances. See my initial reply).
Whatever comes out with >= 250nVPP, 0.1Hz to 10Hz and gives a suitable input cap size, is good enough for me for now.
For the OPA140 they have specified 0.5pA input bias & offset. Some quick calculations show that with 1000 ohm Rin and D/C blocking, the voltage noise still outweigh current noise with a large margin; paralleling pays off in this case.
-
I think we are past the general good advise, now hard calculations and real tests count.... :)
-
So did I get this right?
Urms(0.1-10) = UW2x(10+fc2x1000)
Thanks!
So let's see if I get this right. Simplified and including only 1/fy, for y=1, we get;
Urms(0.1-10) = UW2x(10+fc2x1000)
where
UW = Uwhite / sqrtHz
fc = corner frequency
-
Don't get fooled, noise is actually 8nV/root Hz @ 10Hz (fig. 11, Input noise density -vs- freq). In the world of low noise amplifiers, this is not even close to being low noise.
http://www.ti.com/lit/ds/symlink/opa140.pdf (http://www.ti.com/lit/ds/symlink/opa140.pdf)
True.
Get a Tektronix 7A22, 5A22N or AM502 differential input amplifier or Stanford Research Systems SR560 differential amplifier as these are quite useful for this type of work along with filters and peak detectors as required.
http://www.thinksrs.com/products/SR560.htm (http://www.thinksrs.com/products/SR560.htm)
And yet the SR560 is a lot noisier than the OPA140 below 10Hz - 42nV/sqrt(Hz) @ 1Hz compared to 15, approx 590 nV pk-pk 0.1 to 10Hz v 250nV? (I extrapolated the SR560 noise down to .1Hz using Vn = 1/f^.5 * 43.7 - 1.7 which is a close fit to the 1 to 10Hz curve)
-
Don't get fooled, noise is actually 8nV/root Hz @ 10Hz (fig. 11, Input noise density -vs- freq). In the world of low noise amplifiers, this is not even close to being low noise.
No, but it's close to the limits of what the high pass filter can do without the capacitors getting ridiculously expensive ... so it's good enough. The only way to do much better without ridiculous capacitors is to make a differential amplifier and servo the output to zero ... for a measurement where an opamp with a relatively cheap passive high pass filter suffices that's just wasted effort.
-
These low noise JFET OPs are an option. However one may have to careful, as the 250 nV_pp noise number is a typical value. The 1/f noise of JFETs is known to vary between wafers. So not all OPs may meet that number. It likely would not help much getting 10 from one supplier - these are expected to be similar (all good or bad). Chances are better when getting one OPA4140 and two OPA2140.
Using the typical numbers an amplifier with 4 of the OPA140 would be at about 125 nV_pp, which is similar to what was achieved with a 3.3 mF coupling cap. and a LT1037 (German thread), and slightly better that what LTs AN124 got (about 150 nV_pp). All these numbers are well good enough for an LTZ1000 and likely most other references except the chemical cells - even a single typical OPA140 would work. To a limited extend one can calculate back the noise of the amplifier: So even if the noise of the source is as low as the noise of the amplifier one can still get some useful data.
One can not directly compare the numbers of OP-amps to a bare JFET. The ground referenced JEFT amplifier has much higher thermal effects - so thermal setup is much more demanding. Even the differential circuit will likely have a higher temperature dependence and already 3 dB more noise. The problem with typical numbers also applies to individual JFETs - they can be good ot bad unless expensive special FETs for LF use are used. The rather good IF9030 is something like $16 at mouser.
-
Do you have the resistor values used with these?
Using the typical numbers an amplifier with 4 of the OPA140 would be at about 125 nV_pp, which is similar to what was achieved with a 3.3 mF coupling cap. and a LT1037 (German thread), and slightly better that what LTs AN124 got (about 150 nV_pp).
-
The circuit with the LT1037 used 3.2 mF and 1 K ground. Feedback for 100 times amplification via 10 K and 100 Ohms. Source protection is through a resistor at the input, that is manually bridged with a switch.
If higher frequency stability is a problem and no need for amplification above 1 kHz, a LT1007 (compensated for G=1) may be the better option. The resistor to ground could well be larger, as the 0.1 Hz lower is set by the 2nd stage. A larger resistor would not increase noise in the 0.1 Hz and up range, but only the settling time at the start.
Link to simplified circuit:
http://www.mikrocontroller.net/topic/207061#2060389 (http://www.mikrocontroller.net/topic/207061#2060389)
-
Hello,
Interesting is also the noise floor (including leakage current when measuring 8*Eneloop)
around 100-120nVpp (0.2uV/Div)
(http://www.mikrocontroller.net/attachment/101086/ENELOOP.JPG)
And a LTZ1000 can be easily measured above the noise floor (1.02uVpp @ 0.2uV/Div)
(http://www.mikrocontroller.net/attachment/101087/LTZ1000_2.JPG)
A photo of the amplifier is shown in a follow up thread:
http://www.mikrocontroller.net/topic/250656#2679335 (http://www.mikrocontroller.net/topic/250656#2679335)
(http://www.mikrocontroller.net/attachment/144163/RED_0662.JPG)
with best regards
Andreas
-
Hi,
Thanks for the feedback!
What I'm looking for now is a Unity gain, very high impedance amplifier. Noise is not an issue here. An amp with 1x gain but 1000Gig+ inpunt impedance.
I'm measuring capacitor self-discharge and even the >10Gohm 10V range on my DMM is not high enough.
Is that something you could help me point at a design? Ideas?
I'll reply here because there don't seem to be any way to post photos in the PM area. This is part of the circuit I've been using for ultra high impedance radiation detector projects. IDK if this can be modified for your cap leakage measurements.
The single issue for me was the very high value (aka VERY expensive) feedback resistor and solved by taking advantage of leakage of other devices like JFETs and LEDs. The good news is in your case no feedback may be needed since you are interested in absolute voltage or current.
The bad news is if you need unity gain some reconfiguration probably required. Also there will be an offset. This should not be a problem since you are looking at rough values not precision. I think at room temp TC drift may not be a big issue there either.
Thanks again for the 2n4117. Those electrometer grade devices solved issues that were a problem with the 2n3819 parts tried previously. However for some strange reason they did not works as well as the 2n3819 for G0HZUs "Worlds Simplest Osc" in dannyf thread. Similar HF specs (transconductance, Vgs, etc) but something must be different.
-
A temperature controlled 2n7002 source follower if you really want no leakage (a really good JFET reverse biased junction diode holds off current nicely, but not as nicely as a layer of silicon oxide). A LMC6001 if you want an opamp with guarantees ... but I'd save some time and money and just get a LMC662.
-
I'd save some time and money and just get a LMC662.
Yes, a very good compromise between low noise and high input impedance to allow smalll inexpensive blocking caps. Exactly the same conclusion I came too when implementing the noise amp for a multiplexer/logger as seen in the top center of this photo from my own low cost voltage reference experiments:
(https://www.eevblog.com/forum/projects/low-cost-voltage-reference-experiment/?action=dlattach;attach=143884;image)
https://www.eevblog.com/forum/projects/low-cost-voltage-reference-experiment/msg638485/#msg638485 (https://www.eevblog.com/forum/projects/low-cost-voltage-reference-experiment/msg638485/#msg638485)
-
I'd save some time and money and just get a LMC662.
Yes, a very good compromise between low noise and high input impedance to allow smalll inexpensive blocking caps.
Really?
when I look at the datasheet I see 90nV/sqrt(Hz) at 10 Hz.
so effectively 300 nV RMS or 2uVpp for a 0.1 to 10 Hz bandwidth.
While we are looking for below 0.3uVpp.
Did you really measure the low frequency noise? pictures?
with best regards
Andreas
-
Yes, I did measure the noise level and yes, it is not on a par with some parts costing 10x or 100x more. Replacing a previous lower noise OP07 but I needed the higher impedance input to keep coupling cap cost down and dual part to make room on limited board size for a peak detector. I also needed r-r output yet better than 5v supply which is hard to come by.
However two points: First like in my previous photo, a much lower noise discrete is placed in front which is the real key to keeping circuit noise down. In my case biplolar transistors like S9014, NE85633, or even MPSA18 perform very well for a fraction the cost of special purpose devices.
Secondly I must admit to not being a volt-nut. In fact pretty much the opposite, more like a penny-nut. Instead of most expensive/complicated solution, simplest and lowest cost compromise. Achieving 90% or less of the benefits for 1% of the cost is what floats my boat. In the case of voltage references an order of magnitude less precision at thousands of times less cost. For example as seen in that thread 2 cent bandgaps or one cent zeners. No box full of 50$ LTZ1000 here.
Anyway like janaf noise measuring was nowhere near as demanding as the voltage readings. For me pretty much an afterthought.
ps. I do admit buying an LM399. Not so much for evaluation but rather to use as a maximum limit to test methods and circuits. Only because they were available at 4 dollars and change. For me an outrageous luxury.
-
In case it's not clear, I suggested the LMC662 as a high impedance buffer ...
-
Me too. Originally there was no gain after the discrete front end but only need for high impedance buffer to allow a cheaper film cap. Eventually put some gain in as I realized the dynamic range was excessive and I could afford it to get a little more resolution.
-
FET input follower (buffer) with"helper feed back to increase inout impedance. This is basically a charge amplifier.
Here is the data sheet for the 2N4117 (InterFet process NJ01), it is a tiny geometry FET with very small gate area and designed to have low gate leakage. The IDSS max is 600uA.
http://www.interfet.com/process/NJ01/ (http://www.interfet.com/process/NJ01/)
The 2N3819 is a very different FET with a IDSS of 2-20mA, about four times the input capacitance and no specification for gate leakage. This FET has much larger gate area and overall die size making it very different and not interchangeable with the 2N4117 in any way.
http://dpnc.unige.ch/tp/elect/spec-sheets/2N3819.pdf (http://dpnc.unige.ch/tp/elect/spec-sheets/2N3819.pdf)
-Or why interchanging these in the oscillator circuit is a no go.
Good high value resistors are available from Caddock. Decent parts are NOT cheap and should be use when needed.
http://www.caddock.com (http://www.caddock.com)
Bernice
Hi,
Thanks for the feedback!
What I'm looking for now is a Unity gain, very high impedance amplifier. Noise is not an issue here. An amp with 1x gain but 1000Gig+ inpunt impedance.
I'm measuring capacitor self-discharge and even the >10Gohm 10V range on my DMM is not high enough.
Is that something you could help me point at a design? Ideas?
I'll reply here because there don't seem to be any way to post photos in the PM area. This is part of the circuit I've been using for ultra high impedance radiation detector projects. IDK if this can be modified for your cap leakage measurements.
The single issue for me was the very high value (aka VERY expensive) feedback resistor and solved by taking advantage of leakage of other devices like JFETs and LEDs. The good news is in your case no feedback may be needed since you are interested in absolute voltage or current.
The bad news is if you need unity gain some reconfiguration probably required. Also there will be an offset. This should not be a problem since you are looking at rough values not precision. I think at room temp TC drift may not be a big issue there either.
Thanks again for the 2n4117. Those electrometer grade devices solved issues that were a problem with the 2n3819 parts tried previously. However for some strange reason they did not works as well as the 2n3819 for G0HZUs "Worlds Simplest Osc" in dannyf thread. Similar HF specs (transconductance, Vgs, etc) but something must be different.
-
At what cost to signal fidelity?
How much DC drift and other effects does adding any impedance buffer have on the overall system performance? Have a good look at what is inside that device, it is not a simple black box that is benign in every way.
Often taken for granted today, ICs specially analog ICs are considered nothing more than building blocks that are use to make up what ever is needed with little to no consideration for the finer details of what they could be doing to the analog signal's involved or how they might interact with the overall system. If the distortion or affect cannot be "measured" it is not a problem.. until it becomes a problem then there is a mad scramble to try and figure out what has gone wrong.
What is curious, the aversion to using and designing a higher performance input section using discrete parts that can offer lower noise performance over the IC solution.
After all this discussion, it turns out the goal was to measure the noise of a voltage reference over a bandwidth of about 10Hz and the noise levels are not that low after all.
The SR560, Tek 7A22, 5A22N, AM502 becomes useful once a low noise amplifier section is place in from of them increasing effective gain and allowing some LP-HP filter options with ease. DC drift over time will still need to be addressed in some way with any measurement device choice.
Bernice
In case it's not clear, I suggested the LMC662 as a high impedance buffer ...
-
At what cost to signal fidelity?
I don't think you followed the conversation, he wanted to measure the self discharge and dielectric absorption of the capacitor ... noise is irrelevant. Voltage offset and input current are relevant. JFET's can do a lot of things, but beat a good CMOS at input current is not one of them.
LMC662 is the go to part for high impedance buffering with some input protection.
PS. don't top post :)
-
Ah, now I get it! :)
-
Yes, I agree putting it after is very confusing and makes things hard to follow. :)
There are situations where MOSFET does a great job and 2n7000 is one of my favorites. But in this case a JFET has significant advantages due to gate biasing difference. For example in a single supply follower the MOSFET fails miserably near ground but JFET soldiers on. That's assuming you account for the offset but at least it works.
So MOSFET generally higher impedance and JFET allows negative gate bias. For my ultra low leakage measurements (ion chamber) the 2n4117 was best of both worlds. BTW thanks Rupunzell for the info vs 2n3819, it explains a lot.
JFET's can do a lot of things, but beat a good CMOS at input current is not one of them.
PS. don't top post :)
-
Thank you sir! :-+
The circuit with the LT1037 used 3.2 mF and 1 K ground. Feedback for 100 times amplification via 10 K and 100 Ohms. Source protection is through a resistor at the input, that is manually bridged with a switch.
If higher frequency stability is a problem and no need for amplification above 1 kHz, a LT1007 (compensated for G=1) may be the better option. The resistor to ground could well be larger, as the 0.1 Hz lower is set by the 2nd stage. A larger resistor would not increase noise in the 0.1 Hz and up range, but only the settling time at the start.
Link to simplified circuit:
http://www.mikrocontroller.net/topic/207061#2060389 (http://www.mikrocontroller.net/topic/207061#2060389)
-
Recap:
- Wanted to measure noise from the LTZ1000 after doing various changes. Noise around 250nVrms 0.1-10hz
- Needed a low noise amp for that, with noise below 250nVptp or 100nVrms 0.1-10hz.
- Need high gain, >= 100x to get decent signal output levels. So DC blocking needed.
- Ruled out the option to bias amp as that introduces new noise.
- Low input impedance needed on input to keep noise down and therefore very big input caps
- So I started to measure leakage on big caps.
- Needed an ultra low current amp (pA) to do that
- And a temperature controlled box for everything
All these subjects have been covered, are interesting in any case so the discussions kept rolling.
- I'm convinced I can reach the noise goals with op-amps, both bi-polars and jfet could be used
- I'm building some hardware now, results coming after weekend.
- Discreet front end could reach even better results, but would be too much of a learning curve for me. For now, not really needed.
The low noise and leakage current amps are still demanding. So suggestions need numbers, actual component values used, input impedance, type of op-amp/transistors, capacitor sizes, and results also for the 1/f domain. White noise on it's own is not really useful at 0.1-10hz.
There has been a lot of useful info in the thread! But no offence meant guys, for myself, my case for the LTZ1000 measurements, I feel like I'm beyond general good advise and opinions now, unless backed by fairly detailed descriptions, numbers, both for measurements and calculations.
Keep it rolling :scared:
-
As far as I know, mosfet (opamps) have high 1/f low frequency voltage noise/corner frequency. Are there new good devises out there?
2N4117, excellent current leakage specs but the high voltage noise kept me away. Also the IDSS was confusing. The uA value is there but in graphs, results go several orders of magnitude higher? I have a big bag full of these since a surplus shop closedown...
-
Yes, this was my impression. Some where along the way ultra low input bias op-amps got mixed up into the whole works and completely drove this "discussion" is beyond tangent ways.
Yes, CMOS input op-amps have low input bias current. They are not the answer to it all as the inputs can be zapped if abused. While the inputs are protected by a set of diode clamps, their leakage currents do not balance out at all power supply voltages resulting in an increase if input bias current and the capacitance at their inputs can result in frequency response and related dynamic problems. Their common mode rejection is not that great. Once the source impedance drops, their advantages disappear.
As for MOSFET type devices in general, the first large volume production of MOSFET based sensors were alpha particle based smoke detectors (circa 1980's) where the sensing was done using a P channel MOSFET like the 3N163/3N164. These devices are low inout current due the effectively insulated gate structure which resulted in input bias currents in the femto-amps over a broad range of temperature and conditions.
http://www.linearsystems.com/assets/media/file/datasheets/3N163-4.pdf (http://www.linearsystems.com/assets/media/file/datasheets/3N163-4.pdf)
There is a LOT more to what is required to making the noise measurement device than just the input device or op-amp.
**Voltage references are not high impedance devices which negates the advantages offered by extremely low input bias current devices, discrete or integrated.***
Back to the initial question of measuring noise from a voltage reference that has relatively low impedance and lowish noise over a bandwidth of 10 Hz, Correct?
Two examples have been presented here as starting solutions. Linear Technologies AN124 and the other was using a LT1012 and LT1037 which appears to have good and reasonable test results. The way to proceed would be to build either the test fixtures as outlined in AN124 or the LT1012 ad LT1037 and see if they produce the results required. If the test results are not good enough, then proceed to improve upon what has been done as needed.
It's that simple.
If there are improvements that needs to be made to improve specifics such as noise floor, DC stability or etc, that is when specific questions can be asked with possible solutions suggested.
Set the intellectual/Mental masterbation aside, check the engineering ego at the door as none of that is helpful towards solving the initial question.
Bernice
Recap:
- Wanted to measure noise from the LTZ1000 after doing various changes. Noise around 250nVrms 0.1-10hz
- Needed a low noise amp for that, with noise below 250nVptp or 100nVrms 0.1-10hz.
- Need high gain, >= 100x to get decent signal output levels. So DC blocking needed.
- Ruled out the option to bias amp as that introduces new noise.
- Low input impedance needed on input to keep noise down and therefore very big input caps
- So I started to measure leakage on big caps.
- Needed an ultra low current amp (pA) to do that
- And a temperature controlled box for everything
All these subjects have been covered, are interesting in any case so the discussions kept rolling.
- I'm convinced I can reach the noise goals with op-amps, both bi-polars and jfet could be used
- I'm building some hardware now, results coming after weekend.
- Discreet front end could reach even better results, but would be too much of a learning curve for me. For now, not really needed.
The low noise and leakage current amps are still demanding. So suggestions need numbers, actual component values used, input impedance, type of op-amp/transistors, capacitor sizes, and results also for the 1/f domain. White noise on it's own is not really useful at 0.1-10hz.
There has been a lot of useful info in the thread! But no offence meant guys, for myself, my case for the LTZ1000 measurements, I feel like I'm beyond general good advise and opinions now, unless backed by fairly detailed descriptions, numbers, both for measurements and calculations.
Keep it rolling :scared:
-
Hmm, but Janaf IS actually building stuff ... it's the rest of us mentally masturbating.
On that note, as I said before ... dielectric absorption is entirely irrelevant (http://pessina.mib.infn.it/Biblio/Biblio_Articoli/IEEE%20TNS%20V%2047%20p%201851%201856%202000.pdf) as long as it doesn't saturate the amplifier. Won't find capacitors with worse DA than EDLC, yet it's what they used to get pretty much to the JFET noise floor in actual measurements (I'm pretty sure they misdiagnosed the source of the extra noise with the normal high pass filter though, as far as I can see it's simply the 10K resistor Johnson noise no longer being attenuated by the capacitor).
-
If you're measuring the noise of things like voltage regulators/references which have low output impedances (particularly at low frequencies), (in)sensitivity to current input noise is such that bipolar input op-amps are suitable here. bipolar op-amps generally have much lower 1/f corners than do jfet op-amps. It's trivial to make a sub 1nV/rootHz amplifier suitable for measuring LF PSU and voltage reference noise with multiple bipolar-input op-amp stages in parallel. I've done 0.4nV with a ~10Hz 1/f corner with multiple AD8599. The current/voltage input noise crossover was for a source impedance of approximately 100 ohms IIRC.
-
A recap for the people who obviously didn't read the previous 5 pages of mental masturbation :) The problem is the high pass filter, the source might be low impedance but the high pass filter isn't and thus we end up with the OPA140.
You can put the filter in the feedback path of the opamp, but then you pretty much need EDLC's ... which means you need to keep amplification low'ish (100x range) so the leakage current doesn't saturate the amplifier.
-
I can offer some help in PCB, as interested in this topic too, for exactly same reasons.
-
Die impedance of AC coupling capacitor is a critical part. When using a BJT amplifier a large part of the low frequency noise can be from amplifier current noise times impedance of the capacitor. So a BJT based amplifier needs to use a large capacitor, likely an electrolytic one.
Leakage of the cap is not such a big problem. It may limit the amplification of the first stage, but not that much, as a large cap also allows for a lower resistor to ground. So even 1 µA of leakage at a 10 K resistor is only 10 mV at the input. Having a second stage may be needed for extra filtering anyway. At least down to 0.1 Hz the E-cap. and BJT OP is still feasible, it just needs a large cap with reasonable low leakage and noise. It gets increasingly difficult as the lower frequency limit is lowered. Also lower voltage noise OP-Amps will also need even larger caps, as they have more current noise.
Putting the filter in the feedback path does not really help, except possibly isolating the source from the charge stored in the large cap. However input protection tends to have something like diodes over the OPs inputs and thus input current can still be a problem - so not much gained with this. Also the noise problem does not really change, its just at the other input.
The high quality JFET OP (like OPA140) is an option in spite of the higher 1/f noise limit. The input cap. can be smaller (e.g. 10 µF) and thus a film type. Going lower in frequency only needs an linear increase in the cap size. Different form BJTs several OPs in parallel are a real option. It would need something like 3-4 OPA140 in parallel to get about the LF performance of an LT1037 with a 3 mF coupling cap.
So there are several (not just the 2 above) options that should work. Also having 2 ref. sources and a differential measurement is a way to get around the coupling capacitor at the input. For an extremely low noise source like a battery or chemical ref. cell there still is the option to go for two independent amplifiers and use cross correlation to compensate for the amplifiers noise.
-
The input cap. can be smaller (e.g. 10 µF) and thus a film type.
Hello,
this would imply to have a input impedance of around 200K for the resistor.
You forget one fact: a 250K resistor already has a noise of 0.2uV RMS (or 1.2uVpp) for a 10 Hz bandwidth.
http://www.sengpielaudio.com/calculator-noise.htm (http://www.sengpielaudio.com/calculator-noise.htm)
The noise that a volt nut tries to measure is the 1.2uVpp of a LTZ1000.
So the whole measurement system has to have at minimum a factor 3 less noise (or less noise than a 25K resistor alone).
So practical you cannot go above 10K with the input impedance.
With best regards
Andreas
-
Putting the filter in the feedback path does not really help, except possibly isolating the source from the charge stored in the large cap.
It decouples resistor noise and time constant from input impedance ... when the HPF is at the input all three are interconnected and you are limited in your choices. It doesn't really matter here since say a 100kOhm/220uF combo has a reasonable time constant, below OPA140 noise at 0.1 Hz and presents a reasonable input impedance (22uF would work too, but with 10's of nV/rtHz of Johnson Noise from the effectively 50kOhm input resistance near 0.1 Hz you start adding a bit of noise to the OPA140 there).
-
True, having the AC coupling in the feedback gives a higher input impedance. However input impedance already is relatively high - its essentially set by the resistor to ground. As noted by EmmanuelFaure above, a larger resistor actually reduces the noise. So Its good to have a lower frequency limit a the input, and use a later stage (or software) to set the lower frequency limit. A very large resistor makes settling slower, so I would not go much above 100s for RC (without extra provisions), but also not below about 10s. So even with a large 10 mF cap at the input, something like 10 K input resistance is reasonable.
Anyway a LF noise measurement is nothing to do fast. Thermal settling likely also take several minutes or more.
-
Set the intellectual/Mental masterbation aside
Yeah... but it feels sooooooo goooooood! ngggnnga...ngannnmnn...
-
This paper:http://pessina.mib.infn.it/Biblio/Biblio_Articoli/IEEE%20TNS%20V%2047%20p%201851%201856%202000.pdf
Bolometer = thermo energy to power converter (in this paper, the Bolometer is modeled as a current source with an impedance in parallel with it).
Tries to account for coupling capacitor problems as noted in equation (2) with partial success. Looking at their model, they do not account for many of the real world non-linear realties of capacitors. As the value of the capacitor increases, the problem increase.
Parametric amplifiers use the reactance of capacitors (pumped varactor diodes. capacitors are essentially noiseless when near ideal but can accumulate noise) to achieve very low noise. Electrolytic capacitors are a very different kettle of fish.
Much of this paper re-states what has already been discussed.
*Single ended input section delivers lower noise than a diff pair, about 3db.
*When the source impedance is low, bipolar input devices are low noise. They are ideal when their noise current and noise voltage meet at a narrow range of source impedance. Note in Fig. 5, bipolar transistors are lower noise at low frequencies if optimized for source impedance.
*Once the source impedance increases, JFETs gain an advantage due to much lower to lack of noise current.
*Capacitors can be a source of error-distortion.
*The Moxtek cooled JFET (about 100 degrees K) combines very low noise with low input capacitance. This was one of the devices initially mentioned with a link to the data sheet.
*MAT-03 (PNP diff pair), which originally were made by Precision Monolithics now owned by Analog Devices are "super match" bipolar pairs have low noise with low source impedances specially at low frequencies. This was one in diff pair offering in the MAT series from PMI. Much the same applies to low noise bipolar op-amps and bipolar input devices.
This is a typical academia style text book design. They could have used a servo to control the DC drift of cooled input device and used another device to add gain which would result in a input section with much better performance while retaining the low noise performance of the Moxtek JFET. This is a rather interesting device as it can be had in a "nail head" or bare die package that is designed to be used in a sensor device. The substrate is also intended to be biased to aid gate conduction.
They could have used a cooled Germanium JFET that would have resulted in lower noise performance than the Moxtek JFET.
No details on actual physical construction of this amplifier which can make the difference between a good low noise amplifier or one that does not work at all.
Once the initial gain has been achieved and the problems of signal integrity has been lowered, filtering and other signal processing can be done with fewer problems. Combining the initial signal gain and conditioning often results in a variety of difficulties and problems. Or applying the output of the input section to the Tektronix 7A22, 5A22N AM502 or Stanford Research Systems SR560.
Last, given the noise levels of the voltage reference to be measured, good low noise op-amps are good enough and if the dynamic impedance of the voltage reference is below say 300 ohms, a low noise bipolar input device or op-amp would do quite well. The more difficult problem is likely DC drift over time and temperature.
This measurement is just not that difficult to do.
Bernice
-
Another way to do this.
Use the voltage reference under test as part of the DC bias network for the input section of a discrete low noise amplifier complete with DC servo to compensate for DC drift over time. This way, there is no need for any coupling capacitor, can keep the feed back impedances low and can provide enough gain for further signal processing.
One more note-toid on parametric amplifiers from another time, they offer common mode range of +/- 200 volts with pico-amp input bias currents and very low noise. Mostly obsolete today, become a mostly forgotten technology yet quite interesting.
Bernice
-
Well then you have the big cap (t=RC) for the DC servo again, although it might not need to be as big with a good FET op amp. If you do the DC servo analog, of course. If it's digital not so much an issue.
-
In my endless search for low cost substitutes for expensive glass giga-ohm resistors I spent the morning upgrading my ultra high resistance measurement jig. It uses cap discharge time instead of actual voltage ratios and since this can takes many minutes or hours instead of display it logs the data for offline analysis.
Then it struck me as potential for cap leakage test and I gave it a try. Fixed R instead of cap. Sure enough perfect for the job. No problem calculating leakage for big or small caps to less than a percent. Basically just a FET feeding directly into the MCU analog pin. No need for the more complicated op amp circuit I posted earlier. Of course having an electrometer grade 2n4117 helped a lot compared to the old 2n3819 I was using before. Thanks again for the opportunity. Huge fun, this hi-z stuff.
-
Those Victoreen Giga ohm glass resistors originally intended for HV work?
Jennings vacuum capacitors also comes to mind with those HV parts.
http://www.jenningstech.com/ps/jen/prodsearch.cgi?action=capacitors (http://www.jenningstech.com/ps/jen/prodsearch.cgi?action=capacitors)
That charge amplifiers should do pretty good at measuring capacitor leakage as cap leakage is essentially measuring change in charge.
If you're curious and want to tinker, find some old electrometer tubes and give those a go as these were what was used to measure charge before the days of FETs. Some of the first electronic ph meters used these tubes. Continue on this path leads to devices such as Geiger–Müller tubes, photo multiplier tubes and eventually image intensifiers including micro channel plates. This is all really great and interesting stuff. Oh, condenser microphones with polarizing voltages and electret condenser microphones fit into this topic too.
As for using the voltage reference as part of the bias network. The VR would float between two current sources (source & sink) of near identical current. One end of the VR would connect to the FET gate, the return end of the VR would end up at the virtual or floating summing junction of the common source, active loaded FET. This summing junction would also be where FB is applied. A second single device (bipolar) is direct coupled to the FET's drain with the second stage's BJT's collector being actively loaded. This active load current source will have it's current adjustable to allow for output DC level. This is where the DC servo's control (Either a simple single pole integrator or sample-hold control type) can be applied. Add a current gain section to promote voltage gain and apply overall FB. Then use a low noise diff amp-op-amp to further increase gain and produce an output. The overall scheme is not that different from Jim William's AN124 with the exception of a different input section using discrete devices instead of a composite op-amp.
One last item, some where in the early Siliconix data book was a very small die FET with an internal cascode that was designed to be used as a charge amplifier. Now that my curiosity is up, I'll need to look it up later then week. Siliconix did not make these or very long. These could be a better front end FET than the 2N4117 as they are specifically designed for this kind of work as he 2N4117.
Bernice
In my endless search for low cost substitutes for expensive glass giga-ohm resistors I spent the morning upgrading my ultra high resistance measurement jig. It uses cap discharge time instead of actual voltage ratios and since this can takes many minutes or hours instead of display it logs the data for offline analysis.
Then it struck me as potential for cap leakage test and I gave it a try. Fixed R instead of cap. Sure enough perfect for the job. No problem calculating leakage for big or small caps to less than a percent. Basically just a FET feeding directly into the MCU analog pin. No need for the more complicated op amp circuit I posted earlier. Of course having an electrometer grade 2n4117 helped a lot compared to the old 2n3819 I was using before. Thanks again for the opportunity. Huge fun, this hi-z stuff.
-
Hi guys, been away for a couple of weeks.
As I wrote before, I did some measurements on capacitors / leakages, intended for DC biased measurements.
Now I made some more measurements, this time I charged the caps to 7.1V and let them stand for one week for self discharge.
The results are ballpark the same as before but "better" tanks to longer time and better procedure.
Still a polystyrene top the list followed by polypros. Still a RIFA (Kemet) PEH169 industrial elyth is giving really good results.
I attach results. The second last column is the cap type leakage characteristic, the last column scaled by capacitance.
Some previously tested caps have been excluded now. Before the tests, the caps had been charged and left to rest for two days. Still, several e-caps have had higher voltage now than a week ago, i.e. very strong memory effect / dielectric absorption. Those results are not included here. I will measure them again in a week but if the DA is that high, they will act as non-linear LP filters with very un-predictable characteristics.
I also added a few new caps to the test set but it seems too early to draw conclusions:
- Russian "wet slug" tantalum-silver, 25V, 300uF
- More RIFA PEH169, 2200uF 63V
- Motor starter polypro, 800uF
Preliminarily,
- one tantalum-silver cap was quite good, the other quite bad.
- the RIFA, one cap was quite good, the other quite bad.
- the motor starter polypro; really bad.
I will charge them up to high voltage / take them down and se if any of them recover.
A comment; one needs very high impedance meters for this. My DMM has >20Gohm input and this still discharges "small" caps, like 10uF in the order of 1mV / second which makes measurements a bit demanding, when the capacitor self discharge is only tens of millivolts in a week. Thinking of a terra-ohm input buffer and DMM synced FET switch.
-
6 pV/s is a spectacularly low droop rate in my opinion. Very nice results!
-
Hi again, not been very active here for a while. Now, I have a dozen caps that have been resting, charged, for a few weeks :)
For the cap leakage measurements, LMC6001 sure looks like a good candidate as well as OPA129 and LTC2054HV. Will build & test soon...
For noise, the new AD8428 looks really good. Instrument amp with 40nV ptp 0.1-10Hz. Excellent. Downside as with all low noise ams is input current, typ in the order of 200nA but I have an idea how to solve that.
It has a fixed G=2000 so I'd have to charge a big cap to within 1mV of an LTZ1000 and measure the LTZ1000 without DC blocking.
-
Having the extra cap is very similar to having the filter in the feedback loop. This moves the problem to the large cap and the current noise of the other input causes trouble. The 200 nA Bias current is also combined with significant current noise - so the capacitor would need to be very large (e.g. F range), a achieve a low impedance even at 0.1 Hz. The big question is than whether the voltage of the capacitor is that stable. Buffering the capacitor only moved the problem to the buffer.
The more realistic approach would be having a view low noise batteries (instead of the capacitor) to compensate most of the voltage. so hat less than 1 V DC is left.
Then a moderate (e.g. 10-20 fold) DC coupled amplification is possible, and AC coupling after that.
Batteries are known to have a rather good noise level. If in doubt, testing is possible by switching view batteries in series with different polarities. Large capacitors (likely super-caps) are not that well understood and can be more dependent on history.
Measuring leakage of the caps is only a relatively easy indication in searching for low noise capacitors. Some leakage, comparable to the bias of the OP and even a little more is likely acceptable. I just would not spend much time in testing caps with leakage in the µA range, if you can get better ones.
Anyway the Noise from an LTZ1000 is not that low. It gets tricky only if it comes to measuring the noise of individual batteries or chemical reference cells.
-
As I said before, even a 22uF/100kOhm HPF is hardly going to budge the noise on a typical OPA140.
-
Hi ;)
Let me show you my design of a measuring set, it is designed for measuring on voltage references and lineair power supply's.
It is not finished yet, I'am building the preamp stage on a circuitboard for testing.
I have 6 different quad opamps ready for testing.
I placed the feedback SMD resistors mounted on the opamp circuitboards, the feedback resistor is 500 Ohms.
Because the signal level is low here, the load for the opamp is is therefore not high and not a problem.
The bandwidth of this stage is just above 100Khz.
The preamp...
(http://www.bramcam.nl/NA/NA-Noise-Measuring-Amp/NA-Noise-Measuring-Amp-03-sch.png)
Building the test circuit, left the Quad opamp in parallel mode and on the other side the 46dB amp (total 80dB gain)
Next step is de power supply for the quad opamp on the left site from the quad opamp.
This board wil be used to test different opamps for noise lever and bandwith.
(http://www.bramcam.nl/NA/NA-Noise-Measuring-Amp/NA-Noise-Measuring-Amp-Circuitboard.png)
20 pieces of WIMA goodness :-DD
(http://www.bramcam.nl/NA/NA-Noise-Measuring-Amp/NA-Noise-Measuring-20x-WIMA.png)
The filtersection
(http://www.bramcam.nl/NA/NA-Noise-Measuring-Amp/NA-Noise-Measuring-Amp-Filter-02.png)
I wil use a old Networdk switch box
(http://www.bramcam.nl/NA/NA-Noise-Measuring-Amp/NA-Noise-Measuring-Amp-33.png)
The wil be also a symetrical input, sorry no schematic yet...
Also there wil be a audio amplifier for lissening to the noise, yes, only in the 10Hz -100Khz mode :-)
Shoot @ it!
Kind regarts,
Blackdog
-
The precharging not really needs manual control, one might use a FET OP to sense the input voltage, so that no manual adjustment is needed. I think R8 should be larger, to have a lover frequency limit at the input. Input protection might need something like a small (e.g. 20 Ohms) series resistor at the OP inputs, as the voltage from the 1N4007s might be a little to high if power is off.
The OPA140 is not the best choice for the second amplifier stage and other low impedance parts. There are better (e.g. faster, lower power and cost) alternatives for that, and nose is not that critical any more.
There might be to much amplification before the filters - especially 50 Hz, but also wide band noise might already cause clipping at the filter input. At least some way of detecting clipping at that point is needed. So some of the amplification might be better after the filter, if it is needed at all.
Depending on the source, a 50 Hz notch might be a good idea.
-
Hi Kleinstein, :-)
Thanks for your remaks!
The are manny ways to do the precharge, also automatic, but I opted for manual way.
R8 can get a higher value, but I'm going to do the first tests with 5K.
About you remark for extra imput protection 20 Ohm, maybe, i will think about it.
The second stage i tested already, and works fine.
I will certainly test the NE5532a in this position, cost is not a problem, i want quality ;-)
I've thought about it, to change the position of filters, but for the fist test i keep it this way.
There is also a peak level detector, in order to show whether the signal remains within dynamic range.
And the dynamic rage is BIG, almost -+15V
This noise measuring system main input, is for low level use, noise/hum below 1mV RMS, so that i can use my HAMEG HMO3004 to view the noise of references and low noise power supply's.
There wil be a second input, symetrical, with more gain choices.
I also thought of a 50/100 Hz notch filter, but I have a lot of experience in the measurement of small signals and do not need it now.
There is something very wrong when I measure references and the 50Hz is dominant.
That means I can go back to school and then will have to re-learn how to make a measurement setup :-DD
I welcome your comments, they make me think about some points.
The project is not finished yet, this afternoon I'm going to build on the preamp and make some first measurements.
I fixt the gain of the test setup at about 80dB.
I have just received some extra samples quad opamps to test for the first stage.
The first stage wil be dominant for the noise, and afther the first measurements I will decide witch quad opamp i wil use.
OPA4140
ADA4004-4
MAX44252
MAX44243
OPA1664
OPA4188
OPA1654
ADA4077-4 (lower bandwith)
LT1885
LT1679
A lot of testing to do, low and high impedance source (10 Ohm and say 1K source impadance)
It wil keep me of the streets of Amsterdam ;)
Kind regarts,
Blackodg
-
"MAX44252"
These are just ridiculous value for money compared to the competition.
-
The MAX44252 is quite good when it comes to low frequency voltage noise, but it also has current noise and bias, in a range comparable to BJT based OPs. So I don't see an advantage over something like an LT1007 or OP27 in this application. Still an interesting chip.
-
All the really low noise choppers seem to be BJT based unfortunately (except discrete designs like Jim Williams's 40nVpp design idea).
So yeah, for this purpose the OPA140 is still the best bet unless you want to go discrete ... but in general the MAX4425x is ridiculously good value.
-
The MAX44252 is still MOS based. However due to charge injection and leakage in the rather low impedance choppers, there is a fair amount of current flowing in and out of the input. It's just coincidence to end at a current noise comparable to the LT1007. Most copper-stabilized don't have such detailed data in the DS, but some noise current is still there.
The discrete chopper stabilized circuits like Jim Williams's 40nVpp design idea will also have input currents and current noise. pC charge injection and kHz chopper frequencies end up in the nAs current range, even if much if this compensates this will be accompanied with noise. So a rather low chopping frequency might help.
A better alternative with lower noise current and good voltage noise would be an ADA4528. Still strange why there is no 1/f part in the noise current visible. So the max44252 is not that exceptionally low noise - it's a rather high speed AZ amplifier in first place.
With very careful trimming of switch controlling voltages and possibly with JFET choppers there may be a chance to get a little better, but there will be still a significant higher current noise than with a pure JFET amplifier.
-
Shoot @ it!
I am missing the battery compartment for the power supply.
And where is the cake box where you put all in (including DUT)?
The 1K resistor at the input seems to be too low when measuring a unbuffered LTZ1000.
And why do you use sockets (with thermal noise) and thick film resistors for the input stage?
With best regards
Andreas
-
However due to charge injection and leakage in the rather low impedance choppers, there is a fair amount of current flowing in and out of the input.
The rather ancient TLC2654 says it has fA input noise current (much higher bias current of course). Dunno if it's true though.
So the max44252 is not that exceptionally low noise
It's exceptionally cheap for a relatively high voltage chopper.
-
Hi Andreas, :-)
The power supply is in a another enclosure, and this includes a good double-insulated transformer.
Comming from a precision measuring instrument.
With this I hope to suppress the common mode as much as possible.
If this does not work well enough, then there is a battery supply consisting of AA batteries.
The amplifier wil be dubble shielded.
The 1K resistor wil be shorted if i start to measure a LTZ1000. (the d.u.t. will se then, 440uF in series en 5 or 10K to grond, if the 440uF is pre charged)
First i measure the LTZ1000, I precharge the 440uF capacitor to te LTZ1000 voltage.
Put the mute switch to te "ON" setting, the right side of the capacitor wil be at ground level,
the left site wil be via the 1K resistor connected to the d.u.t. (reference) to charge the last view tens of mV.
Maybe 1K is still to smal, not mutch experience with the LTZ1000, i have two parts on stock en i will test them :-)
Normal reference's (d.u.t.) wil be no problem, the 1K wil be a light load for the last part of charging the 440uF capacitor.
If the charching is ready i wil flip the mute switch and start measuring the noise.
There will be 2x4 LED with 8 comparators to sense the output of the amplifier to detect level of the output (Clipping enz)
I use sockets because this is a test setup circuit board, to test all the low noise quad opamps i have ready for this project.
I hope its more clear now.
Kind regarts,
Blackdog
-
Hi,
A little update of buildig the test board, not yet finished...
On the left site you can see the + and - 6V Low Noise Power Supply for the quad input opamp, the noise level of this power supply? about 1uV at 20Khz bandwith ;D
(http://www.bramcam.nl/NA/NA-Noise-Measuring-Amp/NA-Noise-Measuring-Amp-Circuitboard-2.png)
There is now already one part of the shielding on the board, ofcource there will be more shielding.
The red switches is voor testing the noise input impedance dependency, ik can switch R8 in the schematic from: 10K, 1K, 60 Ohm and "0"'
(http://www.bramcam.nl/NA/NA-Noise-Measuring-Amp/NA-Noise-Measuring-Amp-Circuitboard-3.png)
It takes a lot more time to think en build the board :-)
Tomorrow maybe more about this project.
Kind regarts,
Blackdog
-
There is still no real reason to do this with film capacitors by the way ... 0.002CV leakage electrolytics can do it just fine in the feedback path. A 22000uF capacitor in the feedback path with a 100k/1k feedback divider adds just adds just 4.4V on top with worst case leakage (after it stabilizes) which is no problem with the OPA140's 36V supply range.
-
There is still no real reason to do this with film capacitors by the way ... 0.002CV leakage electrolytics can do it just fine in the feedback path. A 22000uF capacitor in the feedback path with a 100k/1k feedback divider adds just adds just 4.4V on top with worst case leakage (after it stabilizes) which is no problem with the OPA140's 36V supply range.
But its not just the leakage current that is a problem - the open circuit voltage of an electrolytic is very temperature sensitive. Eg. I just connected a 10mm x 25mm 100V 220uF to a 34401A DVM set to 10G ohm input resistance, and left it to stabilize for 30 mins. It showed approximately 89mV, decreasing by approx 2uV over 100s. Then holding my fingertip approx 10mm above the body of the capacitor the voltage rose by 65us over 100s. Touching it with a fingertip caused the voltage to increase at around 3uV/s.
Not the most scientifically controlled test but it illustrates the point. Very careful thermal control, together with long stabilization times may be needed to avoid thermal effects swamping the noise measurements. Larger capacitors having larger thermal capacity will respond more slowly to temperature changes but require longer to gain equilibrium.
A 10uF 63V filmcap seemed to be much better but was harder to measure due to the increased sensitivity to mains pickup induced by approaching the capacitor/test lead. Using warm air from my laptop exhaust from a distance showed little change in the rate of voltage drift.
-
Using the OPA140 for the input, one may not need the large input capacitance, at least for the 0.1 Hz -10 Hz range. This may be different if 0.01 Hz lower limit is needed.
Getting something like 3 µV/s drift from touching the cap is not that bad. Thermoelectric effects on IC pins and drift of the chips are of similar size. So a stable thermal design is needed anyway. But it's true electrolytic caps can cause some trouble and at least need to be tested.
-
Getting something like 3 µV/s drift from touching the cap is not that bad. Thermoelectric effects on IC pins and drift of the chips are of similar size. So a stable thermal design is needed anyway. But it's true electrolytic caps can cause some trouble and at least need to be tested.
True, but thermal EMFs arise from temperature differences between connections which can be kept quite low by good layout and thermal design, whereas the electrolytic is sensitive to absolute temperature which can only be closely controlled by using an oven.
3uV/s has to be kept in context with the requirement to be able to measure noise levels down to 100nV or better over .1 to 10Hz, so drift needs to be less than 10nV/s - and preferrably rather better than that to allow for other noise and error sources in the instrument such that in total it is at least 4x better than the noise level precision required.
Touching the capacitor with a finger 15 degrees C warmer than ambient is a fairly extreme thermal event however, so as you say some testing would be needed.
-
The MAX44252 is still MOS based. However due to charge injection and leakage in the rather low impedance choppers, there is a fair amount of current flowing in and out of the input. It's just coincidence to end at a current noise comparable to the LT1007. Most copper-stabilized don't have such detailed data in the DS, but some noise current is still there.
The MAX4208/4209 (http://datasheets.maximintegrated.com/en/ds/MAX4208-MAX4209.pdf) have negligible input bias current period. Thinking about it, with some auto-zero architectures the opamp's inputs can simply be switched between the two transistors of a differential pair. This should create near zero average charge injection when both inputs are at the same voltage, due to the symmetry (assuming for a moment the transistors/switches are CMOS, charge can only be transported from one input to the other ... with symmetry why would it go one way and not the other, a couple nV combined with a couple of pF isn't going to transport enough coulombs to matter).
Of course there are a lot of different architectures, so MAX44252 might still be MOS.