EEVblog Electronics Community Forum
Electronics => Metrology => Topic started by: martinr33 on August 24, 2016, 07:46:59 am
-
How about this:
http://www.linear.com/product/LTC2508-32 (http://www.linear.com/product/LTC2508-32)
Datasheet
http://cds.linear.com/docs/en/datasheet/250832f.pdf (http://cds.linear.com/docs/en/datasheet/250832f.pdf)
32-bit ADC. That's - theoretically - good enough for 9 1/2 digits. Plus, it is only $12.
The question is - how far can anybody push this thing? It is one and a bit digits better than anything we have today, it needs 100 picovolt quality designs.
Looks like a fantastic part to let us look into the finest tools we have today,
-
One thing to look at is the linearity, which is the one I love the best for 3458A(0.05ppm), and LTC2508 is ten times worse(0.5ppm typ). Another thing to look at is the noise, which is also much inferior than modern 8.5 digits DMM.
However in the past, there are some people making DMM by using this kind of ADC to achieve results rival commercial 6.5 digits DMM or even 7.5 digits.
-
The linearity of a delta-sigma converter is limited by its quantiziser which is why I am always skeptical when people suggest dithering an ADC or DAC while expecting greater accuracy. The first INL Error versus Input Voltage graph shows this nicely with 16 peaks from a major bit transition.
Usually INL error can be calibrated out but that will be much more difficult in this case compared to a converter which uses a single bit quantiziser like the LTC2400 which has a simple INL error curve. Maybe it could be done with a precision ramp or low distortion sine wave and the 22 bit composite code.
-
The linearity of a delta-sigma converter is limited by its quantiziser which is why I am always skeptical when people suggest dithering an ADC or DAC while expecting greater accuracy.
I think you see a lot of people talk about dithering for greater accuracy, because most people easily confuse accuracy and resolution.
-
Btw, seems LT also has unreleased (yet?) LTC2512-24 ADC. It's probably similar to LTC2508 arch, but faster filter and lower resolution, only 24 bits.
That would be nice for precision ~100kHz AC measurements! 32bit one is too slow for that frequency range.
LTC2512-24 mentioned in DC2222A demoboard manual, but LT site has no product page for LTC2512-24 ^-^
-
Btw, seems LT also has unreleased (yet?) LTC2512-24 ADC. It's probably similar to LTC2508 arch, but faster filter and lower resolution, only 24 bits.
That would be nice for precision ~100kHz AC measurements! 32bit one is too slow for that frequency range.
LTC2512-24 mentioned in DC2222A demoboard manual, but LT site has no product page for LTC2512-24 ^-^
The 0.5ppm INL will always be the limiting factor. Linear Tech also came out with a 20-bit SAR ADC (the LTC2378-20 (http://www.linear.com/product/LTC2378-20)) that also has this 0.5ppm INL limitation, but it will take 1 million readings per second. So, you can over-sample and average to remove noise, but you will never be able to get beyond this 0.5ppm limitation.
If the INL is stable, it can be calibrated out. Linear Technology mentions this in a couple of places and an example of blind first and second order correction of an LTC2400 is shown in application note 86 (http://www.linear.com/docs/4177) with the mentioned caveat that "The strata is becoming rarified when “error contribution” is delineated in fractional parts-per-million and the yearly drift rate noted". I wonder though how heroic an effort it would take to do this with a SAR ADC.
AC measurements should be made with a thermal voltage converter [TVC]. There is just nothing that I know of that is commercially available that can outperform a TVC for ability to handle super high crest factors, pretty much any wave shape, and frequencies from DC to 10's of MHz, all while getting 10ppm of accuracy. Multi-junction thermal converters are available now commercially, and they can get you down to maybe <=1ppm-- so, AC accuracy to 6.5-digits!
Who makes these?
The Datron 1281 [aka: Fluke 8508A] does resistance measurement the "Right Way", and if you are going to copy a resistance measurement circuit, then copy this one. Better still, would be to build in a DCC for lower resistances [<=10K\$\Omega\$] and an electrometer arrangement (similar to your high-resistance bridge) for everything >=10K\$\Omega\$. The DCC could be dynamically re-purposed to provide DCI and ACI readings to 100A (again to 1ppm, which is phenomenal).
DCC?
-
The 0.5 ppm INL is quite good, but not that much better than lower cost slow ADCs like the LTC2400. So it might be an option for something like a battery powered 5-6 digit DMM.
With adjustment or extra calibration points one might be able to reduce the INL, at least the slow varying part of it that is stable.
Using one of the faster LT converters to do RMS conversion in software (just like the 3458, keysight meters call this true volts) is now a really realistic way for high resolution AC measurements. The bandwidth may be a little lower than the analog RMS chips, but it's more accurate and stable. One could even have the option to do filtering in software and thus measure RMS values with intentionally lower bandwidth, which could be interesting for low voltages. The crest factor for thermal converters is also limited by the input amplifier - unless you work without it, with lots or care and a low impedance source.
I would expect the high end digital lock-in amplifiers to use such convert(s).
-
Btw, seems LT also has unreleased (yet?) LTC2512-24 ADC. It's probably similar to LTC2508 arch, but faster filter and lower resolution, only 24 bits.
That would be nice for precision ~100kHz AC measurements! 32bit one is too slow for that frequency range.
LTC2512-24 mentioned in DC2222A demoboard manual, but LT site has no product page for LTC2512-24 ^-^
The 0.5ppm INL will always be the limiting factor. Linear Tech also came out with a 20-bit SAR ADC (the LTC2378-20 (http://www.linear.com/product/LTC2378-20)) that also has this 0.5ppm INL limitation, but it will take 1 million readings per second. So, you can over-sample and average to remove noise, but you will never be able to get beyond this 0.5ppm limitation.
If the INL is stable, it can be calibrated out. Linear Technology mentions this in a couple of places and an example of blind first and second order correction of an LTC2400 is shown in application note 86 (http://www.linear.com/docs/4177) with the mentioned caveat that "The strata is becoming rarified when “error contribution” is delineated in fractional parts-per-million and the yearly drift rate noted". I wonder though how heroic an effort it would take to do this with a SAR ADC.
You will drive this from a 5V or 2.5V reference voltage. A degree change in temperature will have several dozen LSB change in the output code. So your calibration table will be changing constantly. With these SAR ADCs you need to be aware that they need a stable, low impedance, high bandwidth (several MHz, really) reference source, and the ADC will have different load to the reference based on the input voltage. Same goes for driving the input. So you cannot really just filter it with RC, you need a fast, high current, not drifting opamp, which cannot really be zero drift, because the chopping frequency is just too low.
-
Using one of the faster LT converters to do RMS conversion in software (just like the 3458, keysight meters call this true volts) is now a really realistic way for high resolution AC measurements. The bandwidth may be a little lower than the analog RMS chips, but it's more accurate and stable. One could even have the option to do filtering in software and thus measure RMS values with intentionally lower bandwidth, which could be interesting for low voltages.
Tektronix apparently did exactly this with their handheld DMM series of multimeters which have RMS measurement capability better than their contemporary analog computing counterparts.
Fluke initially sued Tektronix for trademark infringement over the color of the protective rubber sleeve and then after Tektronix changed the color to blue, Fluke bought the entire handheld multimeter business from Tektronix to keep them from competing with their own technically inferior high end products. I do not know what became of them after that.
... So you cannot really just filter it with RC, you need a fast, high current, not drifting opamp, which cannot really be zero drift, because the chopping frequency is just too low.
Use the chopping amplifier to correct the drift of and remove the 1/f noise from a fast precision low noise amplifier.
-
Who makes these?
Best technologies, another one is made by a japanese company
[quote
DCC?
[/quote]
Direct Current Comparator.
Guildline, Measurements International, Isotech(resistsnce thermetry)
-
Multi-junction thermal converter data sheet attached.
Pretty nice! An idea of the price range?
-
DCC means "Direct Current Comparator", and this is used for current ratios in a resistance bridge, or for current ratios when measuring DC (and up to about 100KHz). See: Danfysik and LEM DCCT's...
I think these results might help someone. Was playing a bit with my LEM Ultrastab IT600-S and calibration equipment while ago.
Also have plan to redo similar test on my home gear, but that will come later on :)
Brief and non-destructive teardown of these zero-flux current sensors was done year ago and published here (https://xdevs.com/review/ultrastab_review/).
Ok, here are the results.
Current source 1: Fluke 5700A
Current source 2: Fluke 5700A + 5205A
Configuration : 7 turns thru LEM
Burden shunt: 2.5 ohm Vishay PG foil 1ppm/C 9W, CSNG
Meter: Keithley 2002 on 200mV range
All gear was freshly calibrated and kept within 2C of cal temp.
0A : 0 mVDC rel
214.1942 uA : 0.00250 mV
2.141942 mA : 0.02505 mA
21.41942 mA : 0.249960 mV
200mA : 2.33431 mV
214.1942mA : 2.49986 mV
1A : 11.6719 mV
2A : 23.3434 mV
2.141942A : 25.00005 mV
High current tests with Fluke 5205A (uncalibrated) + 5700A (calibrated)
1.000000 V(MFC) : 11.64732 mV
2.000000 V(MFC) : 23.31058 mV
2.142857 V(MFC) : 24.98688 mV
With 5205A:
3.000000 V : 34.99193 mV
4.000000 V : 46.66403 mV
5.000000 V : 58.33468 mV
6.000000 V : 70.00847 mV
7.000000 V : 81.68046 mV
8.000000 V : 93.35251 mV
9.000000 V : 105.024558 mV
10.000000 V : 116.69671 mV
11.000000 V : 128.36888 mV
12.000000 V : 140.04104 mV
13.000000 V : 151.71314 mV
14.000000 V : 163.37437 mV
15.000000 V : 175.05746 mV
16.000000 V : 186.72914 mV
16.300000 V : 190.23057 mV
5205A could not go higher, as it's not suitable for large inductive loads and was losing stability after 16A.
-
Very interresting TiN. Maybe we should start a topic on DCC/DCCT?
Some great introduction informations in this document by LEM : http://www.digikey.com/Web%20Export/Supplier%20Content/LEM_398/PDF/lem-engineering-applications-manual.pdf (http://www.digikey.com/Web%20Export/Supplier%20Content/LEM_398/PDF/lem-engineering-applications-manual.pdf)
-
The chain is as good as the weakest link.
32 bits sounds fun, but getting that accuracy in your resistor divider and reference voltage is gonna be a pain in the butt.
Especially if you involve temperature, noise and component tolerance.
-
I'm not complaining, 0.5ppm is very good-- in fact better than the specs for most 6.5-digit and 7.5-digit DMM's. Others have pointed out that the extra resolution can be useful for relative readings-- well, yes, it can, but maybe only one additional decimal digit.
What? What makes that one digit special? INL is a limitation of accuracy, but not precision and there is value in precision without accuracy (well, perhaps not in metrology, but outside of it, particularly if one is interested in trends).
Beyond that is "marketing hype", and that additional decimal digit cannot be relied upon to be absolutely accurate-- (kind of like "deceptive advertising" or an "attractive nuisance").
Unless they claim those 32bit to be accurate, I don't see how it is deceptive.
-
I'm not complaining, 0.5ppm is very good-- in fact better than the specs for most 6.5-digit and 7.5-digit DMM's. Others have pointed out that the extra resolution can be useful for relative readings-- well, yes, it can, but maybe only one additional decimal digit.
What? What makes that one digit special? INL is a limitation of accuracy, but not precision and there is value in precision without accuracy (well, perhaps not in metrology, but outside of it, particularly if one is interested in trends).
Beyond that is "marketing hype", and that additional decimal digit cannot be relied upon to be absolutely accurate-- (kind of like "deceptive advertising" or an "attractive nuisance").
Unless they claim those 32bit to be accurate, I don't see how it is deceptive.
I think you are confused. The INL is a measure of how precise the ADC is throughout the entire range of the ADC.
I might be confused, but it seems I'm not alone in that: Maxim calls INL and DNL parameters of accuracy: https://www.maximintegrated.com/en/app-notes/index.mvp/id/283 (https://www.maximintegrated.com/en/app-notes/index.mvp/id/283)
-
Hmmh, for precision we need monotonicity, but monotonicity is guaranteed only if INL <= 0.5 LSB [1]. Given that, I too, don't know what those extra bits are giving us. :-//
Update:
Well, strike that. It says right there that |INL| <= 0.5 LSB is a sufficient, but not necessary criteria for monotonicity. I.o.w. a larger INL doesn't preclude monotonicity (which strictly speaking isn't a requirement for precision either, but unless one can maintain a large translation table, it rather is a practical requirement). So I don't know.
[1] http://www-inst.eecs.berkeley.edu/~ee247/fa08/files07/lectures/L12_f08.pdf (http://www-inst.eecs.berkeley.edu/~ee247/fa08/files07/lectures/L12_f08.pdf) (p. 20)
-
So, what I'm getting at, is that Linear Tech has a 24-bit version of this same part, and it is just as "precise" as this 32-bit part, because I think you will find that the lower 8 bits of this 32-bit part are nothing but noise and are useless-- even with very long integration times. I think that those extra 8 bits are just bullshit, for marketing purposes only. If someone can prove that the lower 8 bits of this part are repeatable and monotonic, then I will admit I'm wrong about this part.
I would also like to know how they tested the 32 bit part however the datasheet unambiguously says that all three outputs, 32, 14, and 8 bit, are monotonic. The 32 bit INL versus voltage graph shows the typical 1-bit delta sigma converter S shape overlayed with a 4 bit major transition of a higher resolution multibit modulator.
The DC histogram with the maximum decimation factor of 16k shows only 4 codes out of about 1300 samples. At 61 samples/second, that is 20 seconds which seems almost reasonable in some applications. With additional processing, it really does look like 32 bit resolution is available with an integration time below that.
I agree that both parts will be equally accurate as far as INL. I am surprised the INL is competitive with a delta sigma converter using a 1-bit modulator at all.
-
The limit to get a monotonic curve is DNL < 1 LSB. Even than if for few events there will steps that are not monotonic any more, a little bit more resolution of the ADC helps, as it could prevent additional rounding errors in the calculations the follow. The often state the resolution of the high resolution converters as the limit where you get no missing code, which would be more of less the DNL < LSB limit.
By some clever method, internal adjustment these SAR ADCs seem to adjust the steps very well. At least the DNL and INL curves look relatively smooth, with very little effect of MSB steps. For the 32 Bit Version the oversampling nature could be some help. The higher speed sigma delta converters also use muli bit DAC/ADC - so no more pure 1 Bit. So this one is more like the other extreme: 24 bit ADC and massive oversampling.
These 32 Bits ADC and the faster 24 Bit ones show some noise. So single conversions are not repeatable to that degree. But after enough averaging (could take more than 100 samples) the average should be OK, provided the reference, signal and temperature are that stable. If you want to compare noise, one should do it at the same rate: for the LTC2508 with 16000 averaging mode one is at about 60 samples per second. The DS gives an RMS noise of 0.02 ppm. The specs of the 3458 are about 0.08 ppm (RMS=SD) at 1 PLC (or 0.02 ppm at about 20 PLC) in the 10 V range. To be fair one has to add noise of the input amplifier. Still noise wise this ADC is really good.
So it may be a lower noise alternative to something like the LTC2410 in some applications.
-
Very interresting TiN. Maybe we should start a topic on DCC/DCCT?
Oh yes, please!
-
As in my previous post, I am still wondering about the practical implematations.
32 bits with a 2.5V supply voltage, means that eacht step is about 0.58nV.
Which is 0.000000023%.
Even if we forget things like temperature drift. Practiaclly it's almost impossible to get in the same order of magnitude with your supply voltage (and reference voltage) . On top of that the input stage and attenuation circuit.
In fact to do it right, you'll need these things an order of magnitude better.
-
How can you ask such questions?
Perhaps you accidently are in the wrong section.
> Metrology
> Everything metrology & Calibration related. This is where the Voltnuts hang out.
A real voltnut wants a 64 Bit ADC to play with.
With best regards
Andreas
-
As in my previous post, I am still wondering about the practical implematations.
32 bits with a 2.5V supply voltage, means that eacht step is about 0.58nV.
Which is 0.000000023%.
Even if we forget things like temperature drift. Practiaclly it's almost impossible to get in the same order of magnitude with your supply voltage (and reference voltage) . On top of that the input stage and attenuation circuit.
In fact to do it right, you'll need these things an order of magnitude better.
This is less of a problem with ratiometric operation which is likely for applications listed on the datasheet.
-
Does anyone know where to get a DC2222A-B Demo-Board which uses the LT2508-32? As far as i can see, the only shop which sells them is Digikey and they have none of them stocked. Also the sales representative couldnt tell me how long it would take Linear to ship the boards.
I just want to use the LT2508 as an ADC to accurately measure thermocouples and play around with the achievable low noise floor of the LT.
-
For a thermocouple you could use amplification before the ADC, as normally the voltage is small. So no need for the high resolution, just a low noise level. The LTC2508 is relatively low impedance at the input anyway and thus usually needs at least a buffer.
So for a TC it would be better to have a really low noise OP (e.g. ADA4522 or similar) for something like a 10-100 fold amplification. TC are not that accurate, so no need to worry about absolute accurate amplification. After that the ADC noise is less critical.
-
Does anyone know where to get a DC2222A-B Demo-Board which uses the LT2508-32? As far as i can see, the only shop which sells them is Digikey and they have none of them stocked. Also the sales representative couldnt tell me how long it would take Linear to ship the boards.
I just want to use the LT2508 as an ADC to accurately measure thermocouples and play around with the achievable low noise floor of the LT.
Buy directly from LT:
http://www.linear.com/solutions/7366 (http://www.linear.com/solutions/7366)
They ship in a day or two, normally.
-
There are cases where these 32bit ADC are useful. While yes the last few bits are pretty noisy but if you only want to look at a narrow frequency range and do a FFT on it you can see some incredibly tiny signals with it.
But yeah the analog front end design for these things is not a walk in the park. Just using a good spec single opamp can ruin everything with its own linearity. Also any grease on the PCB can affect it, slight changes in temperature will make the signal drift way off as at those levels, as resistors dance around in value with temperature, all sorts of connections become thermocouple junctions, microphonic effects might show up where you bump the table and the readings jump etc.
-
But yeah the analog front end design for these things is not a walk in the park. Just using a good spec single opamp can ruin everything with its own linearity. Also any grease on the PCB can affect it, slight changes in temperature will make the signal drift way off as at those values resistors dance around in value with temperature, all sorts of connection become thermocouple junctions, microphonic effects might show up where you bump the table and the readings jump etc.
+1
Modern high resolution ADCs' specs are amazing.
-
Thanks for the tip with the ADA4522 Kleinstein, i will take a look at it. But its just not about accurately measuring thermocouples, im also getting more and more interested in building a capable voltage-DMM based on the LTC2508. So i bought the DC2222A-B for testing purposes on digikey. Meanwhile i will be playing with the LTC1043 and maybe LM399/LT5400-configurations and see how far that gets me, while observing the results with the PREMA 5017 i have here at work.
-
Thanks for the tip with the ADA4522 Kleinstein, i will take a look at it. But its just not about accurately measuring thermocouples, im also getting more and more interested in building a capable voltage-DMM based on the LTC2508. So i bought the DC2222A-B for testing purposes on digikey. Meanwhile i will be playing with the LTC1043 and maybe LM399/LT5400-configurations and see how far that gets me, while observing the results with the PREMA 5017 i have here at work.
For thermocouples, LT has much nicer chips. I've seen the LTC2983 in action, it is impressive. If you really want higher accuracy, just change the TC to an RTD, and to 3 or 4 wire, for even higher accuracy.
-
Indeed, really interesting. Would be interesting to compare the chip to the standards we have here at work. Would only get better if one could couple ADT7320/74 to the LTC2983, since those have a really good accuracy as cold junction sensors.
-
I bought a temperature measurement system for work with two sensor heads based on 2x PT100 inside each sensor head. It's realized with a fast digital approach based on a LTC2440, that means the ADC is supplied with 5V, reference voltage is created by a resistive divider out of this 5V supplying the Ref-pin of the ADC and the 4 dividers consisting of a PT100 element and a resistor of low TC (5ppm/K). This system is then calibrated over a temperature range of -40°C ... +150°C.
Work pretty good :-+
System accuracy after 6 point calibration: <0.05K over a range of -40 ... +140°C
-
The LTC2508-Dev-Board arrived. Sadly the purchased E34465A will not arrive till mid-november, so i cant really make precise measurements/comparisons at the moment.
In the picture we see the noisefloor of the Board, covered in insulating foil, powered by two Meanwell-Supplys RS-25-15 and regulated down to +-9V with two LM317 (low-noise bypass cap version). Powered with batteries the noisefloor is about the same (mainly 400nVpp with short peaks to 600nVpp). The decimation factor is 16k, sample interval about 300ms. Seems the LTC2508-QuickEval-Example is not capable of 60 samples per second.
-
Is this with slowest SPS speed? How about full-scale test, perhaps at 1V input?
-
That histogram does look pretty good. Pretty similar to Texases 32bit ADC that i used for calibration in a product.
-
I couldnt get a faster sampling-speed than 300ms with the QuikEval-software. Next week i will try to interface the board to a generic µC, to get the 61sps at 16k DF while logging it. I can also test it at 2,5V, since i have a LT6655-2.5 laying around.
-
I meant if possible to test on slow speed.
-
The ADC itself seems to be made for 60 sps as the lowest speed, unless you choose a slower system clock. For slower readings one can of cause always average or low pass filter in the µC.
If you use an external 2.5 V reference to measure, you will get a combination of noise: the ADC, the reference on the ADC board and the external reference. It is well possible to get most of the noise from the two references, especially in the lower frequency range where the 1/f noise of the references will dominate.
So this would be more like a test for the two references.
-
As far as i understand the data sheet, the conversion time is around 600ns for a single conversion and the ADC can sample max 1 megasample/s, which gives 61 samples/s at 16k decimation factor.
@ Kleinstein: Yes, of course the second measured reference will add noise. Im still looking into schematics which reduce the noise of references while maintaining the absolute stability of the reference (because i would like to use LM399s later on, but lower noise would be nice). But i guess the only realistic solution is the use of parallel connected references. I really need to learn more about op-amp-error contribution to decide for myself if a given schematic is capable of maintaining stability, like the following: http://electronicdesign.com/energy/filter-trims-ultra-precision-voltage-reference (http://electronicdesign.com/energy/filter-trims-ultra-precision-voltage-reference) ???
-
One should avoid trimming the reference - this only adds possible drift. There is plenty of resolution to do trimming in software.
The linked filter circuit looks promising. Depending on the OPs used and the voltage level, one might get away with a simpler supply for the OPs. E.g. an LTC2057 could directly work with a 8 or 15 V supply. One might be able to combine the filter circuit with other functionality like buffering after a divider.
The noise of the OPs should not be that bad, at least it should add without extra amplification. Low noise OPs could be important as the ADC's specified noise is about at a level of a good AZ OP. For A1 one could use a lower noise type even of this means more bias and offset, it is not that critical. For C1 a more normal capacitor (e.g. MKS) might be acceptable.
For the very low frequencies it takes a low noise and stable reference. The LM399 is an option, possibly even two of them in parallel as noise levels could be still better. The MAX6655 variants could also be an option, at least for low noise. Another option for low noise might be the Chinese 2DW23x, if you get a good one - at least the chances are there and they are cheap. It could get a real challenge to build a reference with a noise lower than this ADC, at least in the low frequency range where one can hardly filter.
-
As far as i understand the data sheet, the conversion time is around 600ns for a single conversion and the ADC can sample max 1 megasample/s, which gives 61 samples/s at 16k decimation factor.
Keep in mind that the 61 samples/second are not independent of each other. With a 16k decimation factor, page 24 of the datasheet says that the actual digital filter length is 144k or 9 x 16k with a total duration of about 150 milliseconds. So when comparing it to a no latency integrating converter, the sample rate is more like 6.78 samples/second or 8.85 60 Hz power line cycles. This explains why the frequency response shows nulls below 50 and 60 Hz.
Too bad there is no way to drive the sampling clock to 983kHz for a maximum null at 60 Hz or 819.2kHz for a maximum null at 50 Hz. The filtering is so good though that maybe it would not make any difference.
-
Maybe someone can help me with a problem regarding the DC2222A-B-Board:
Im searching for the best way to convert my +-5V-single ended-input signal to a suitable signal for the LTC2508, while the DC2222A-B itself takes only fully differential signals (as far as i understand).
The simplest solution seems to be the following: populating RP1 and U21 (matched resistor network and single ended to differential driver) and desoldering both LTC2057 U10/U18. Is that correct or is there a better solution to maintain the specifications of the ADC? Also: for RP1 i would use a LT5400-resistor network, but i dont know which single ended-to-differential-OP is should use for U21. Does anyone can recommend to me the most suitable one?
Edit: http://cds.linear.com/docs/en/demo-board-schematic/DC2222A_REV05_PCA_SCHEMATIC.PDF (http://cds.linear.com/docs/en/demo-board-schematic/DC2222A_REV05_PCA_SCHEMATIC.PDF)
Edit: Just ordered a LT5400 and a LTC6363 to populate on the board. Lets see how far that gets me.
-
I bought a temperature measurement system for work with two sensor heads based on 2x PT100 inside each sensor head. It's realized with a fast digital approach based on a LTC2440, that means the ADC is supplied with 5V, reference voltage is created by a resistive divider out of this 5V supplying the Ref-pin of the ADC and the 4 dividers consisting of a PT100 element and a resistor of low TC (5ppm/K). This system is then calibrated over a temperature range of -40°C ... +150°C.
Work pretty good :-+
System accuracy after 6 point calibration: <0.05K over a range of -40 ... +140°C
Since the PT100 of the system seem to have a slightly different behavior than the ones used since then the manufacturer decided to change temperature points for the 6 point calibration. The error is now reduced to 0.02K in the tested range of -40°C ... +140°C. A very nice pice of gear for 1.100€ with Modbus interface. :-+
-
I recently dug out my LTC2508-32-Eval-Board and got it to work. It now delivers 61 samples per second with a decimation factor of 16384 and a sample clock of 1MHz. The ADC is buffered by the two onboard LTC2057, each connected via 100kR (R35 and R36 in the schematic) to Ref/2 = 2.5V.
Schematic: http://www.analog.com/en/design-center/evaluation-hardware-and-software/evaluation-boards-kits/dc2222a-b.html (http://www.analog.com/en/design-center/evaluation-hardware-and-software/evaluation-boards-kits/dc2222a-b.html)
The Eval-Board sits in a tinned steel-box, communication via Linduino-Board and gets its supply via a linear supply. I take the 61SPS and average it over 244 samples, so i get a result every 4s.
Here are the results:
Noise-Tests:
Nothing connected (ADC only sees buffered Ref/2 via the 100kR-resistors at the LTC2057): 0.5µVpp measured over 4s (61SPS, 244 averaged samples)
Fluke 5440B Calibrator via shielded CAT6-network-cable, tested with different voltages within +-5V ADC-input-range: 3µVpp measured over 4s (61SPS, 244 averaged samples)
LTC6655-2.5 powered with battery, LTC6655 sits in the enclosure and is connected with short twisted wire to the ADC-buffer: 0.75µVpp measured over 4s (61SPS, 244 averaged samples)
Linearity-Tests:
Test were conducted with Fluke 5440B connected to the ADC-box and with a 3458A in parallel to check the voltages (wasnt necessary, accuracy and linearity of my 5440B is still wonderful ^-^ ). Both ACALed just before the test with the ADC-box, voltages tested from -5...+5V in 1V-steps.
After measuring the ADC-voltage-reference (a LTC6655-5), calculating offset and gain-correction-values i measured the maximum errors 6µV at +5V and -14µV at -5V compared to the 3458A in 100NPLC-mode. The values between +-5V were pretty much linear, with a few bumps which seems to be the INL-error.
I need to improve a few things to actually see and correct the INL and measure the stability:
-change LTC6655 against LTZ1000-LTC1043-divider to get 4.8Vref, this will improve the necessary long term drift
-put the ADC-box in my TEC-controlled-climate-chamber, unfortunately the TEC-controller died :palm: , but i hope to have a replacement in about two weeks
-improve my Python-program
-wait for the heatwave to pass by, so the temperatures for the ACALed devices are more stable
-
The LSB with a 5V reference is 1.2nV, your readings are 500/750/3000, thats like 24 bits, with heavy filtering
Whats the point of the extra 8 bits?
If you were to eliminate DC using a band pass or high pass filter, to restrict bandwidth, can you get a better signal ratio? Or does the noise from the v-ref carry over anyway?
I don't mean to offend the chip, but why would you need 8 more bits then the noise floor? How much can you possibly average the noise down and remain practical use? Is there any kind of signal you can look at without the stupid high averaging?
Lets say you built a baller high pass filter with some nice expensive capacitors. What can you get from it? I heard people say that its useful for low frequency AC application, but does anyone have any numbers for me? 61 SPS would only be useful for infrasonic AC frequencies.. What is the best number of bits you can get from it in what type of measurement circuit? Or is it meant for doing sampling on repetitive signals like a sampling head on a scope?
-
Its mostly marketing, 32Bit sounds better than ~26Bit, the resolution currently achievable with the best ADC-ICs. You get a tiny bit more resolution with a 32Bit-ADC (under the right circumstances and if no squirrel farts in a 1km radius during the sampling time) than with a 24Bit-ADC. The more relevant question is: is the ADC long-term-stable and is it possible to correct the INL.
Since i only ever focused on DC-measurement i cant really answer you questions sufficiently.
In the DC-measurement-case you use a low-pass-filter in front of the ADC and you need a low noise Vref. User NandBlog mentioned the possibility to mostly filter out Vref-1/f-noise in the 0.1-10Hz region if you multiplex the ADC-input fast enough between Vref/Ground AFAIR.
-
I was hoping the LT company was above 'mostly marketing'. Thats what I feared.
What are these best available ADCs?
-
Even if many of the 'extra' bits are just noise, 32 is a nice number for handling data these days. Extra effort and circuitry would be required to truncate those extra bits away, and any one decision as to which bits are extra and can be truncated will probably piss off someone with some obscure application down the road. So, I can see the point in letting the user decide what to do with the raw data, and provide honest specs as to the real performance of the converter, not just the output format.
-
Let me correct myself: its more like ~25Bit which are achievable at the moment with the ADCs.
The lowest noise ADC which i know of is the AD7177-2 which claims 24.8 noise-free bits at 5 SPS in Table 7 of the datasheet. One Could average more samples, but at this point one really needs to test the ADC in reality, to see what is really achievable under the best circumstances, since many IC-manufacturers try to hide/obscure the little errors of their components.
Sadly theres no Appnote like the AN86 in which they get silly and test contemporary ADCs with the best methods and references on a long term scale.
http://www.analog.com/media/en/technical-documentation/data-sheets/AD7177-2.pdf (http://www.analog.com/media/en/technical-documentation/data-sheets/AD7177-2.pdf)
-
The LSB with a 5V reference is 1.2nV, your readings are 500/750/3000, thats like 24 bits, with heavy filtering
Whats the point of the extra 8 bits?
If you were to eliminate DC using a band pass or high pass filter, to restrict bandwidth, can you get a better signal ratio? Or does the noise from the v-ref carry over anyway?
I don't mean to offend the chip, but why would you need 8 more bits then the noise floor? How much can you possibly average the noise down and remain practical use? Is there any kind of signal you can look at without the stupid high averaging?
Lets say you built a baller high pass filter with some nice expensive capacitors. What can you get from it? I heard people say that its useful for low frequency AC application, but does anyone have any numbers for me? 61 SPS would only be useful for infrasonic AC frequencies.. What is the best number of bits you can get from it in what type of measurement circuit? Or is it meant for doing sampling on repetitive signals like a sampling head on a scope?
This was exactly my point at the previous page.
Plus you'll need to have a very good PCB design as well.
-
@Coppercone: See also comment here: https://www.eevblog.com/forum/metrology/32-bit-adc-playground-for-precision-measurement-tasks/msg1237921/#msg1237921 (https://www.eevblog.com/forum/metrology/32-bit-adc-playground-for-precision-measurement-tasks/msg1237921/#msg1237921)
As far as i know these ADCs are also useful for seismic measurements were you need a very high dynamic range, see also this interesting pdf:
https://www.researchgate.net/profile/Nikolai_Beev/publication/325285614_Analog-to-digital_conversion_beyond_20_bits_Applications_architecutres_state_of_the_art_limitations_and_future_prospects/links/5b03dd96a6fdccf9e4f7c361/Analog-to-digital-conversion-beyond-20-bits-Applications-architecutres-state-of-the-art-limitations-and-future-prospects.pdf?origin=publication_detail (https://www.researchgate.net/profile/Nikolai_Beev/publication/325285614_Analog-to-digital_conversion_beyond_20_bits_Applications_architecutres_state_of_the_art_limitations_and_future_prospects/links/5b03dd96a6fdccf9e4f7c361/Analog-to-digital-conversion-beyond-20-bits-Applications-architecutres-state-of-the-art-limitations-and-future-prospects.pdf?origin=publication_detail)
and the reference [5]:
SH. Gao, B. Xue, J. Li, Z. Lin, Y. Chen and XY. Zhu, "High-resolution data
acquisition technique in broadband seismic observation systems," ->
"The dynamic range of the currently most widely used 24-bit seismic data acquisition devices is 10–20 dB lower than that of broadband seismometers, and this can affect the completeness of seismic waveform recordings under certain conditions. However, this problem is not easy to solve because of the lack of analog to digital converter (ADC) chips with more than 24 bits in the market. "
-
As far as i know these ADCs are also useful for seismic measurements were you need a very high dynamic range, see also this interesting pdf:
I think you might find a slight mismatch between the sampling speed of these ADCs, and the needs of seismic measurement.
-
I assumed the sampling speed was sufficient for measured waves between 1-100Hz. If thats wrong i stand corrected. Please explain what sampling speed is needed.
-
Why can't they use just audio ADCs for seismic stuff? ???
SNR of 120-130dB is even doable as well a THD+N -106 or better.
-
I assumed the sampling speed was sufficient for measured waves between 1-100Hz. If thats wrong i stand corrected. Please explain what sampling speed is needed.
The spec for the LTC2508 says 145dB Dynamic Range (Typ) at 61sps and 131dB Dynamic Range (Typ) at 4ksps.
Seismic waves have a low fundamental frequency, but are very rich in harmonics. If you've ever seen something like a seismic plot on the news when an earthquake happens you must have noticed the very sharp transitions which occur. The ADC needs to capture those. I don't know how high a sampling rate they use, but I assume its at least has high as good audio.
-
TI's ADS126x claims 26.9bits ENOB (25 noise free bits) at 2.5SPS (datasheet page 28).
Afaik the earthquake freqs of interest are mostly below 3Hz.
-
I assumed the sampling speed was sufficient for measured waves between 1-100Hz. If thats wrong i stand corrected. Please explain what sampling speed is needed.
Table 1, page 21 of the datasheet shows that due to the digital filtering and downsampling, the 3dB bandwidth at an *output data rate* of 61 SPS, is only 7.5Hz. At 796 SPS bandwidth increases to 120Hz but the dynamic range drops by 9dB to 136dB.
The LTC2500-32 does about 3dB better and you also get to choose from a range of filters to trade off BW, settling time and noise - see table 2 page 31.
-
I assumed the sampling speed was sufficient for measured waves between 1-100Hz. If thats wrong i stand corrected. Please explain what sampling speed is needed.
The spec for the LTC2508 says 145dB Dynamic Range (Typ) at 61sps and 131dB Dynamic Range (Typ) at 4ksps.
Seismic waves have a low fundamental frequency, but are very rich in harmonics. If you've ever seen something like a seismic plot on the news when an earthquake happens you must have noticed the very sharp transitions which occur. The ADC needs to capture those. I don't know how high a sampling rate they use, but I assume its at least has high as good audio.
Some audio ADCs go up to 768k that should be more than sufficient.
Hell the default 48k is fine as well at these low frequencies.
-
@Coppercone: See also comment here: https://www.eevblog.com/forum/metrology/32-bit-adc-playground-for-precision-measurement-tasks/msg1237921/#msg1237921 (https://www.eevblog.com/forum/metrology/32-bit-adc-playground-for-precision-measurement-tasks/msg1237921/#msg1237921)
As far as i know these ADCs are also useful for seismic measurements were you need a very high dynamic range, see also this interesting pdf:
https://www.researchgate.net/profile/Nikolai_Beev/publication/325285614_Analog-to-digital_conversion_beyond_20_bits_Applications_architecutres_state_of_the_art_limitations_and_future_prospects/links/5b03dd96a6fdccf9e4f7c361/Analog-to-digital-conversion-beyond-20-bits-Applications-architecutres-state-of-the-art-limitations-and-future-prospects.pdf?origin=publication_detail (https://www.researchgate.net/profile/Nikolai_Beev/publication/325285614_Analog-to-digital_conversion_beyond_20_bits_Applications_architecutres_state_of_the_art_limitations_and_future_prospects/links/5b03dd96a6fdccf9e4f7c361/Analog-to-digital-conversion-beyond-20-bits-Applications-architecutres-state-of-the-art-limitations-and-future-prospects.pdf?origin=publication_detail)
and the reference [5]:
SH. Gao, B. Xue, J. Li, Z. Lin, Y. Chen and XY. Zhu, "High-resolution data
acquisition technique in broadband seismic observation systems," ->
"The dynamic range of the currently most widely used 24-bit seismic data acquisition devices is 10–20 dB lower than that of broadband seismometers, and this can affect the completeness of seismic waveform recordings under certain conditions. However, this problem is not easy to solve because of the lack of analog to digital converter (ADC) chips with more than 24 bits in the market. "
Thanks. Additional search on the web gives another interesting links.
https://converterpassion.wordpress.com/category/data-converter/adc/adc-survey/page/2/ (https://converterpassion.wordpress.com/category/data-converter/adc/adc-survey/page/2/)
https://converterpassion.wordpress.com/2012/08/21/adc-performance-evolution-walden-figure-of-merit-fom/ (https://converterpassion.wordpress.com/2012/08/21/adc-performance-evolution-walden-figure-of-merit-fom/)
-branadic-
-
Hmm, checked my working LTZ1000-reference and my 5440B-calibrator against my 3458A. Both deliver reasonable ~1.5µVpp, as expected. I build a LTC1043-divider to get about 5V from a LTZ1000, to use it as a reference for my ADCs. But the relative large pp-noise bothers me.
Configuration as follows (Andreas buffer-design, battery-powered, in a cookie-box): LTZ 7.2V -> LTC2057-buffer -> LTC1043 / 3 -> LTC1043 * 2 -> LTC2057-buffer delivers 4.8V ~8µVpp.
Any advice how the noise might be improved, maybe a buffer-stage between the /3 and *2 -block?
-
Hello,
how do you measure the 1,5uVpp/8uVpp?
(Scope with 1/f LNA or with the DMM at some NPLC?)
I never built my cirquit idea (except in LTSpice).
Did you synchronize all LTC1043 clocks?
(That is the first thing I would try, I had already some strange temperature dependant noise behaviour with a LTC1050 after a LTC1043 together with a LTC2400. Most probably resulting from interferences of the chopper frequencies and the ADC sampling window. With a LTC2057 in the same configuration I have much less temperature dependant noise. )
The next step would be a buffer between the LTC1043s.
But you have to take care that the buffer can handle capacitive loads for the step up configuration.
with best regards
Andreas
Edit: lowest noise (without buffer) should be when the connections of the second stage are built in a way where all input output capacitors are connected in parallel during the same clock phase. (Eventually you have to swap input and output connections of the switch on the 2nd stage)
-
The clock synchronisation was a very good idea Andreas. :-+
With 600Hz at the clk-input of both LTC1043 i now get ~1µVpp at 4.8V. No buffer between the two LTC1043 was used in this case and since the result is already very good i wont build one in.
I measure the noise with my 3458A at 100NPLC and look for stable regions without thermal drift which are reasonable long, lets say 5min for example. Of course thats not as good as a LNA, but i dont own one.
-
Using the 100 PLC mode makes it look at very low frequency noise, more like an 0.25 Hz upper limit. To get a value more comparable to the usual 0.1-10 Hz data, it would be looking at windows of about 100 readings at something like 1 or 2 PLC.
-
I measure the noise with my 3458A at 100NPLC
Its always good to know how something is measured to estimate what is really going on.
By the way: is it the attached cirquit?
In this case you would need only one LTC1043 (so it is automatically synchronized).
(but the first buffer needs to be stable against capacitive loads).
with best regards
Andreas
Edit: explanation see here:
https://www.eevblog.com/forum/projects/ppmgeek!-5-5-digit-dvm-volt-ref-cal-(for-arduino-or-any-uc-w-spi)/msg296127/#msg296127 (https://www.eevblog.com/forum/projects/ppmgeek!-5-5-digit-dvm-volt-ref-cal-(for-arduino-or-any-uc-w-spi)/msg296127/#msg296127)
So with a real LTC1043 the 20nF are in reality only 10 nF.
-
No, the circuit uses two separate LTC1043, since i didnt found your circuit suggestion during my searches before building. Might be interesting to also test it and see if it delivers the same results and behaves as well as the circuit with the two LTC1043.
Im now using this 4.8V-LTZ-based-reference as the reference for my AD7177-Eval-Board to see how many bits i get out of it, with my low noise-Fluke 5440B as the signal source for the ADC. But maybe that belongs in another coming thread.
-
I just stumbled across this thread. Interesting.
I have made a small stamp-sized board for an LT2500-32 ( home-etched :-) )
that has the LT2500, 2 * LT3042 for analog & digital VCC, LT6655 reference
the differential driver and its neg. regulator, all in 1" * 1".
A xilinx Coolrunner on a second stamp generates the sampling clock / other timing.
But currently I'm stuck with the Beaglebone Black / Debian Stretch firmware.
Any access to the SPI interface results in a bus error. |O
I want to use the PRU of the BBB to collect the data and forward it to the Linux
ARM CPU via toggling buffers in 1Ksample blocks.
There is already a working LAN link that can read/write GPIB-like commands
by opening port 5000-sth. on 192.168.178.111 .
I have done some VERY low noise preamplifiers including choppers, but the
1/f noise of my Agilent 89441A that I use as an FFT analyzer is limiting me.
I do not want yet another boat anchor.
Integrating the ADC would make a very nice low noise digitizer with isolated
network access.
PS
Why can't this !§$%&!! board software reduce a cell phone photo itself
to the size it likes?
-
Is there a chance to see the schematics of the 2500-32 pcb? :-+
-
Please :)
That's not intended for publication, just a fast screen dump.
There are some comments what should be changed, and the
large row of pins is the BBB 16 bit multiplexed bus that is
obviously not on the stamp. More like a lab note with everything
important on one screen.
cheers, Gerhard
-
I just stumbled across this thread. Interesting.
I have made a small stamp-sized board for an LT2500-32 ( home-etched :-) )
that has the LT2500, 2 * LT3042 for analog & digital VCC, LT6655 reference
the differential driver and its neg. regulator, all in 1" * 1".
A xilinx Coolrunner on a second stamp generates the sampling clock / other timing.
But currently I'm stuck with the Beaglebone Black / Debian Stretch firmware.
Any access to the SPI interface results in a bus error. |O
I want to use the PRU of the BBB to collect the data and forward it to the Linux
ARM CPU via toggling buffers in 1Ksample blocks.
There is already a working LAN link that can read/write GPIB-like commands
by opening port 5000-sth. on 192.168.178.111 .
I have done some VERY low noise preamplifiers including choppers, but the
1/f noise of my Agilent 89441A that I use as an FFT analyzer is limiting me.
I do not want yet another boat anchor.
Integrating the ADC would make a very nice low noise digitizer with isolated
network access.
PS
Why can't this !§$%&!! board software reduce a cell phone photo itself
to the size it likes?
why not just use a PIC?
This is exactly why I hate digital design.
Man I am thinking of migrating to the Z80. Is there a high speed easy to use processor that you can run off your own EEPROM with you own peripherals? I hate all that BS that I don't use on MCUs.
-
> why not just use a PIC?
Because I cannot ssh into a PIC, because I cannot edit & compile my sources there
because there is no gcc, no yacc/bison, gdb right on the target, because there is no Debian
Linux and file system on the target, because there is no way to run fftw on a substantial time
series with a NEON floating point accelerator, no el cheapo web server to show results?
>This is exactly why I hate digital design.
> Man I am thinking of migrating to the Z80.
I came from there 30 years ago. There has been some progress in the mean time.
A Z80 could not read out the ADC at 1 Msample/s @ 32 Bits, much less storing the data
in its 64 Kbytes.
> Is there a high speed easy to use processor that you can run off your own EEPROM
> with you own peripherals? I hate all that BS that I don't use on MCUs.
So what. The BBB has its own flash eprom, lots of it, a 1 GHz CPU and you don't even
need a programmer. There are 2 aux. risc processors @ 200 MHz to do I/O, and DMAs,
timers, PWM units, what do you want?
Plug in the network cable, log in and start working.
regards, Gerhard
-
Man I am thinking of migrating to the Z80. Is there a high speed easy to use processor that you can run off your own EEPROM with you own peripherals? I hate all that BS that I don't use on MCUs.
Ok, this is getting seriously off-topic and I wouldn't recommend one for professional use and I'm not in the retro-computing Z80 fan camp (gosh I'm glad those times are over -- remember CP/M? *shudder*), but if you're looking for an easy MCU w/o EEPROM, then there is Parallax' Propeller (not affiliated, just a satisfied customer). They target chiefly the educational sector. It's popular with some hobbyist and artists, but there have been some measurement devices made with it too. Its eight core ("cogs") architecture is a bit peculiar in that it doesn't support interrupts (or even stacks and subroutine calls), but that allows for accurate bit-banging. The manual is only 400 pages and there is a small, but friendly community. 32KiB internal RAM assures that software doesn't get unwieldy large. :-DD
-
The 2500-32 returns "24bits result from SAR core at full speed" and "7bits of Common Mode result" and "1 bit overrange indication".
How to combine all those outputs actually?
-
> why not just use a PIC?
Because I cannot ssh into a PIC, because I cannot edit & compile my sources there
because there is no gcc, no yacc/bison, gdb right on the target, because there is no Debian
Linux and file system on the target, because there is no way to run fftw on a substantial time
series with a NEON floating point accelerator, no el cheapo web server to show results?
>This is exactly why I hate digital design.
> Man I am thinking of migrating to the Z80.
I came from there 30 years ago. There has been some progress in the mean time.
A Z80 could not read out the ADC at 1 Msample/s @ 32 Bits, much less storing the data
in its 64 Kbytes.
> Is there a high speed easy to use processor that you can run off your own EEPROM
> with you own peripherals? I hate all that BS that I don't use on MCUs.
So what. The BBB has its own flash eprom, lots of it, a 1 GHz CPU and you don't even
need a programmer. There are 2 aux. risc processors @ 200 MHz to do I/O, and DMAs,
timers, PWM units, what do you want?
Plug in the network cable, log in and start working.
regards, Gerhard
Sounds like all those features the BB has are not working in your favor for some reason, since you cant seem to get any data.
I use the simplest PIC with the least peripherals and enough speed to spit out processed ADC data to serial cables logging to putty lol. Kinda wondering what will happen when I want to use faster ADCs.
My appeal behind the use of a MPU (never did) is that you can probe everything with a logic analyzer and do dumps to see whats going on without some kind of marginally useful JTAG device. It seems that it would cut down on the bullshit that comes with all the linux/other overhead.
I just had this gut feeling that using some kind of dedicated LINUX based PCB thing to program a MCU would result in hurt so I just stuck with programming PIC's with a pickit and using sockets. Even that had its fare share of misbehavior (particularly with the directives used by the compiler to configure the PIC. Would prefer separate parts rather then configurable ones, I don't care for most of the features I just like using them as ADC to PC or LCD converters).
I have less issues with MCU's when it comes to control systems though, then they are handy and their multi use is tolerable and they save time*. For home made equipment anyway it seems like using a PIC to do the control features like range switching, buttons, etc on its own LCD is a good idea to keep it away from the data processing system and having to deal with interrupts and everything... its all so cheap now\
I guess you need to slave it if you want to make some kind of self calibration but if you keep the control system separated from the measurement system it might make it alot eaiser. Just seems so much cleaner.
Seems to get alot uglier when you are trying to make some kind of source measure unit/digilog DSP feedback thing.. then you need shared memory and a feedback interrupt to make it follow some kind of known stable inaccurate looop that won't fuck up the process or completely halt while the memory is being updated
Yea the physical size of my projects would drive most people nuts LOL
*save time if you have the damn Microchip/ATMEL product family memorized otherwise you can get totally sidetracked trying to find the MCU of the correct form factor, peripherals location, etc.
-
The 2500-32 returns "24bits result from SAR core at full speed" and "7bits of Common Mode result" and "1 bit overrange indication".
How to combine all those outputs actually?
There is usually not need to combine the results. The main result it the differential input. This could be either raw data or internally filtered at a lower speed.
The common mode and over-range indication are more for debug or error checking. The common mode voltage indicates how well the signal is balanced to the inputs.
-
Ok, it means, for example, when they provide the phone number of a board member in an additional 32bit part of the "result" their marketing will call it 2500-64, a 64bit ADC :)
-
You can do something with the common mode thing like switch the input off if you can get data fast enough to protect it, or make a warning flag that some kind of transient EM event occurred
-
Ok, it means, for example, when they provide the phone number of a board member in an additional 32bit part of the "result" their marketing will call it 2500-64, a 64bit ADC :)
There are two types of possible output for the 2500-32:
One is fast (1 MHz) 24 bit data and the other is filtered data with a much lower rate (kHz range) but 32 bit resolution - however with still quite some noise. The noise also helps to avoid missing codes :popcorn:.
It is more like coincidence (could be the interface HW) that the 24 bit + 1 bit + 7 bit fast format is also 32 bit long.
-
Hopping in here.
There's another LTC2500 board available: "ADC 7 click" from MikroElectronica at around $100. It piggybacks on a mainboard, various versions available.
https://www.mikroe.com/adc-7-click (https://www.mikroe.com/adc-7-click)