Open Source Multimeter

[poorchava]

Audience believe that the new Aura-Teflon is simply the most resolute and transparent capacitor ever engineered, regardless of cost.

 

[mrflibble]

Yeah, I love stuff like that!  Has me wonder though, what does that even mean? Their teflon has interesting optical properties?

Besides, audiences probably believe whatever they fancy after drinking Kool-Aid laced with lsd. Regardless of cost!

[poorchava]

I know that's still off-topic, but I couldn;t hold myself back:
http://www.lessloss.com/blackbody-p-200.html

[mrflibble]

LESS>LOSS to sanity.
MORE<LOSS to wallet.

I advocate heavy use of LSD to compensate for any "coloration of sound" by those pesky stray EM near fields. That's probably more cost effective, and better for humanity alround given the target group.

Anyways, I'll try to come up with something integrator related soon-ish so we can keep things on topic.

[branadic]

It's funny, isn't it? The link to the mentioned cap was only to show how expensive such teflon caps are and you won't find lot's of sellers for teflon caps this days, except for such audio freaks. I had expected some more professionalism in here, obviously I was wrong.

[mrflibble]

Mmmh? As said, teflon is pretty expensive and as such lets pick the not so expensive polypropylene option.

So you propose some caps, they happen to be on an audiophool website. Audiophoolery gets made fun of a little. The main point being that this particular incarnation of teflon is WAY overpriced. And teflon in general is quite expensive. And so what, teflon is not the only reasonable choice available. Polypropylene it is, problem solved, onwards to the next problem. No need to get all "I had expected some more professionalism in here, obviously I was wrong"-ey about things.

No slight intended towards you. But you can't blame people for having a chuckle at transparent and resolute capacitors.

[Stoney]

What about NP0?

[mrflibble]

As far as I understand dual slope, NP0 is not all that great for this application.

We don't care too much about the absolute cap value, and as such initial accuracy is not all that important. Even drift isn't too important. As long as the cap value is stable within an X amount of precision for the duration of the ADC conversion cycle (i.e short amount of time) we should be fine. The figure of merit for the integrator cap would be dielectric absorption. And as said, that's not all that great for NP0.

See for example: http://portal.national.com/rap/Application/0,1570,28,00.html and http://www.analog.com/library/analogDialogue/Anniversary/21.html

Essentially we're looking for a cap that's good for sampling & holding applications. I mention S/H  because the caps we're looking for are likely to be listed under "application: S/H, ..." at the cap manufacturer.

I currently have Vishay MKP1837 or MKP1840 as candidates. There are bound to be better ones, but I didn't want to spend endless hours on selecting just one component. I think those will be adequate for now. But if someone knows part numbers for affordable caps with lower dielectric absorption than those, then please do tell. Caps should be in the range of 0.01 uF to 1 uF I think. And as with all the other components, in stock with the major distribustors would be preferable ... but that almost goes without saying.

[robrenz]

TI 24 bit delta sigma evaluation board. I know this may not fit into the direction this thread is currently going. But these features caught my eye as usefull somehow.

Easy to use online and offline WEBENCH Medical AFE Designer tool
Clickable block diagram in tool
Pre-programmed 3-lead and 5-lead ECG application
Easy-to-use evaluation software for Microsoft™ Windows XP or Windows 7
Built-in analysis tools including oscilloscope and histogram displays, estimated device and measured system performances, and beats per minute calculation
Real time data streaming from the demo board to the PC
Ability to export data in simple test files for post processing

[mrflibble]

Indeed, that's quite a nifty board for a nice price. I have the ADS1258EVM which was about $50 at the time. Or was it EUR 50, I forgot.

For this particular project I intend to use that board to measure the integrator residue after an integration cycle, and see if there's a benifit to be had there. But at least for me the choice of delta-sigma vs dual-slope/multi-slope has been made. I started out thinking to maybe use delta-sigma, but after more reading I jumped ship to the dual-slope camp.

[Stoney]

Hey guys,

just wanted to let you know that mrflibble and I are working on the project and have made some progress. But we shifted our communication mostly to email/PM because we're chatting a lot about specific questions/typologies and calculations, and email is just easier in that matter.

Be sure to check out the repo from time to time and feel free to comment on any of the contents!

So long,
Stoney

[branadic]

And AFAIK for this application teflon is indeed best and most expensive. Then 2nd best is polypropylene and after that polystyrene.

It's not a good way to make an opinion just by the words of a single person, even if it's Bob Pease. I found this website that says, that polystyrene has smaller dissipation factor and tempco is stronger specified:

http://conradhoffman.com/cap_losses.htm

So you better test both types of different manufactors (not with a single example!!!) to find the better one as they will not specify all interesting parameters in their datasheets.
Best choice would be an air or vacuum cap but the needed value for the integrating cap is a bit to high. Teflon is expensive and rare this days and my link showed that it seems to be reserved to the audio freaks. You shouldn't take care of their advertising slogans, but to the electrical parameters.
But if there is a possibility to get one of this teflon guys in the bay you should take that chance.  You better don't put several caps in parallel as the resulting cap will have a higher leackage.
Concerning soakage you can for sure use one of the shown compensation circuits, but an advantage on compensating one parameter will often have a disadvantage in another, bear that in mind.

[mrflibble]

It's not a good way to make an opinion just by the words of a single person, even if it's Bob Pease.

Agreed.

So you better test both types of different manufactors (not with a single example!!!) to find the better one as they will not specify all interesting parameters in their datasheets.

That's the plan. Didn't I post that earlier, or was that just in PM? In any event, it's a good suggestion and it's on the TODO list.

You shouldn't take care of their advertising slogans, but to the electrical parameters.

Maybe I am blind, could you send me a link where I can find the electrical parameters for that teflon cap? Because I did try, but couldn't find it. Even with a google "site:teflonvendorforgotwhatitwas.com" I came up empty. Because I was indeed curious to know what kind of specs one buys for that large chunk of money.

Concerning soakage you can for sure use one of the shown compensation circuits, but an advantage on compensating one parameter will often have a disadvantage in another, bear that in mind.

Indeed. The prototype won't have any such compensation scheme since we'd like to keep things medium simple.

[chickenHeadKnob]

Best choice would be an air or vacuum cap but the needed value for the integrating cap is a bit to high. Teflon is expensive and rare this days and my link showed that it seems to be reserved to the audio freaks. You shouldn't take care of their advertising slogans, but to the electrical parameters.
But if there is a possibility to get one of this teflon guys in the bay you should take that chance.  You better don't put several caps in parallel as the resulting cap will have a higher leackage.
Concerning soakage you can for sure use one of the shown compensation circuits, but an advantage on compensating one parameter will often have a disadvantage in another, bear that in mind.

Can some-one tell me what is the range of values you folks are considering for an integration cap? My recollection from dave's HP3457a teardown was that HP was using 10 nano farad, or was it 10 micro?. I am thinking this might be possible to make 'home rolled' from teflon sheet and copper clad. The extra physical bulk should be manageable as it is planar and can be mounted in parallel with the main circuit board, no? Copper foil isn't particularly expensive but teflon board sold to the RF guys can be. I suspect there are cheaper (and thinner) sources of teflon film as it is used as a gasket material. Capacitors aren't really all that magic, even if some are resolute and transparent.
 Yes just checked Mcmaster Carr they have Ptfe film. I don't normally look there as they no longer ship to new Canadian customers but here is an example 0.002 inch thick 6 inch wide $1.55 per lineal foot. - jeeze those americans and their inches.

[mrflibble]

We're talking about something in the 220 pF - 10 nF range, depending on what speeds and slew rates and such.

Heh, and you just had to mention a DIY teflon cap. Earlier tonight I was going over several datasheets and was moderately frustrated by the lack of information for this type of app. The manufacturer is perfectly happy to tell you that it's an awesome cap and you should use it for S&H application. And it has low dielectric absorption blah blah blah. But actually giving numbers? Nooooooo, that would be too easy. Some do give info, and those typically are the ones that actually have a do-able dielectric absorption. Or rather, the DA is such that it is worth testing to see if it's a good enough cap for the application.

Anyways, the so-so availability of information had me almost considering a DIY cap solution. But then I remembered that no we are not going to do that because if we want to enable other people to build this as well, then we should keep all the parts as simple and available as possible. And now you tempt me with a DIY teflon cap.

So yes, it might definitely be interesting to see how such a cap would stack up against off-the-shelf parts...

For ease of discussion, lets say a 1 nF cap... How much would one need and where would you easily obtain it? Preferably with non-crazy shipping cost to Europe. Or samples would be great too.

Edit: And to prevent misunderstandings ... this diy teflon cap would be purely to compare it to say a polypropylene, mainly just for the fun of it. It's a hobby after all. Not as the primary candidate for this design.

[branadic]

As was aksing them for an electrical datasheet but all I received was this:

We do not have datasheet as such, here is a link to the specs on our
website.
http://www.soundlabsgroup.com.au/c/Capacitors-Teflon/Capacitors+-+Teflon.htm

Best wishes,...

So I went on and found this page in the link posted previously:

http://conradhoffman.com/cap_measurements_100606.html

Isn't it critical to do a DIY cap by just using metal (copper) and ptfe foil and roll them togehter? You can have air between the several layers and the resulting cap is somewhat a result off paralleled air and teflon caps. I would have expected that the teflon is evaporate with the electrodes material to get best performance and this is a process one can't simulate at home. But however, give it a go and tell us about the results

[chickenHeadKnob]

We're talking about something in the 220 pF - 10 nF range, depending on what speeds and slew rates and such.


So yes, it might definitely be interesting to see how such a cap would stack up against off-the-shelf parts...

For ease of discussion, lets say a 1 nF cap... How much would one need and where would you easily obtain it? Preferably with non-crazy shipping cost to Europe. Or samples would be great too.

Edit: And to prevent misunderstandings ... this diy teflon cap would be purely to compare it to say a polypropylene, mainly just for the fun of it. It's a hobby after all. Not as the primary candidate for this design.

Hello Mrflibble

OK it is late here and I'm a little punch drunk. Here is a first pass assessment.

First to calculate size, teflon dielectric constant 2.1, polypropylene 2.3 so teflon maybe going to be 10 % bigger plate area if gap is the same. I use
 C = 1 x 10e-9 F desired capacitance
 e0 = 8.854 x 10e-12  F/m electric constant
 er  = 2.1 permittivity  relative of PTFE
 g   = 0.001 inch or 2.54 x 10e-5 plate gap in meters
 A = plate area in m² to be calculated

so C = (e0 x er) x A/g
or  A = 0.000537825 m² or 5.37825 cm²    this is do able, the first time I ran through this calc. I ended up with half a square meter for 1nF and thought that can't be right. please check as I am tired and making mistakes now.

There are multiple sources for copper foil 0.001 inch thick and teflon same thickness, let me continue to search for the best. Some additional notes: teflon sheet at 0.001 inch thick is listed in various places as having dielectric strength of 500 to 2000 V per 'mil' which I think means 1 thou inch thickness. Holy shit-snacks we can tie these caps across the mains! Also it is produced by 'skiving' which I think is the process where a very big solid log of teflon is put in a lathe and a thin sheet is shaved off by a very sharp cutter. Does not seem like good way to me, but what do I know

Prices are looking reasonable, more later. These materials can be purchased by one individual in some quantity and then after cutting and subdividing with a paper cutter  sent to other experimenters using ordinary envelope and mail service for economy.

Suggest to me the best form factor/topology, alternating rectangular plates for a rectangular package or rolled up into a cylinder. I fear rectangular might act like a patch antenna but I know nothing!

Looking at conrad hoffman's page there doesn't appear to be a great advantage to teflon over polyprop in adsorption still it is fun to experiment. I am almost 'having a volt nut' now, just don't want it to happen in my pants

goodnight my friends, going to bed

[branadic]

We're talking about something in the 220 pF - 10 nF range, depending on what speeds and slew rates and such.

This value seems much to small to me. In this range it should be no problem to get stuff in the bay. I expected a few hundreds nF (Notice to americans: this is a SI conform unit called nano farad, somewhat in the middle of micro and pico farad) up to serveral µF.

I ended up with half a square meter for 1nF and thought that can't be right. please check as I am tired and making mistakes now.

Thinking of such an arrangement I would expect nearly twice the cap value:

http://www.jogis-roehrenbude.de/Leserbriefe/Edelmann-Ko/Image182.jpg

[mrflibble]

We're talking about something in the 220 pF - 10 nF range, depending on what speeds and slew rates and such.

This value seems much to small to me. In this range it should be no problem to get stuff in the bay. I expected a few hundreds nF (Notice to americans: this is a SI conform unit called nano farad, somewhat in the middle of micro and pico farad) up to serveral µF.

You're welcome to join the club and help improve the design. If you go back in the thread you will notice that I mentioned an integrator cap in the 10 nF - 1 uF range. Those are the values I was looking at for a dual slope design. Ahem, pardon my terminology. An enhanced dual slope design (different runup/rundown resistors).

Right now we are looking at a relatively simple version of multi-slope with an ADC in the mix. The ADC is mainly there to drive up the BOM cost and create controversy should anyone check out the repo. It's not as if the idea is to do clever stuff with it and have fun while building this thing or anything. And the aim is to be able to do continuous integration.

For this type of multi-slope 10 nF actually is on the high side. And 220 pF is on the low side, so I would expect we end up somewhere along this range...

Oh and pssst chickenHeadKnob .. *cough* factor of 2.54 *cough* That, and what branadic said.

Regarding getting teflon caps on ebay, yes I did look at that. But it suffers from the same type of problem as DIY caps, namely that it doesn't help in making the design available to other people. If it's easier to obtain the required parts at a reasonable cost it's more likely someone will also try the design. So I'd rather not make the cap a hard to obtain part since it central.

Additional, on the subject of affordable parts ... the ADC currently in there is anything but affordable. But it's not strictly required, just for extra bells and whistles and to test some fun things. The design will (or at least that is the plan) work just fine with just comparators after the integrator.

[mrflibble]

Prices are looking reasonable, more later. These materials can be purchased by one individual in some quantity and then after cutting and subdividing with a paper cutter  sent to other experimenters using ordinary envelope and mail service for economy.

Suggest to me the best form factor/topology, alternating rectangular plates for a rectangular package or rolled up into a cylinder. I fear rectangular might act like a patch antenna but I know nothing!

Looking at conrad hoffman's page there doesn't appear to be a great advantage to teflon over polyprop in adsorption still it is fun to experiment. I am almost 'having a volt nut' now, just don't want it to happen in my pants

I must admit that it would be fun to try and make some teflon caps, and then see what the results are.

But after a good nights sleep ... while it would be fun to do, there is only so much time in the day. If I look at all the other work that needs to be done, it's best not to add to it unless strictly required. Case in point, take for example the caps... I would love to just be able to read the datasheets and then go "ah okay, this and this cap will do what we need, and it will behave such and such". All you can do is give it your best and make an educated guess based on the limited info that's in the datasheets. And then make a test jig and simply measure those suckers. But if I could I would rather not do the frigging test jig and just be able to trust the information in the datasheet.

"Simply measure everything" I hear some potential clever person say. A sound suggestion, but you don't make a test jig for every single component you put into your designs, do you? That would be a bit too much work. Since the performance of this cap is rather central I don't see any other easy way, but I would love to just plug in a known-to-work-for-this-app capacitor.

The best actually quoted in the datasheet numbers for DA have been 0.05% DA (typical). That number together with the loss tangent over frequency gives you some idea, but that's about it. And before someone says, well <fill_in_material> is better than <fill_in_other_material>, gimme a link to a datasheet that has an actual number for dielectric absorption that is better than 0.05%. And that is datasheets from the manufacturer, not measurements by a third party. Actually measurements by a 3rd party are also welcome, because for example that conradhoffman stuff also was useful. But DA numbers in a datasheet are nice. So far the most informative datasheets (for this app) have been by Vishay and WIMA.

I can only wonder how many types of caps (and other parts for that matter) a DMM manufacturer has to go through before they find the right ones. Then again, they probably have a slightly bigger purchasing and R&D department.

Does anyone have any references for test jig to measure DA and such? Right now I am looking at doing something like this and this.

I tried finding more on the measurement method on conradhoffman.com but couldn't find it, but I am sortof hoping I just missed it. Anyways, information on such would be appreciated.

[chickenHeadKnob]

Right now we are looking at a relatively simple version of multi-slope with an ADC in the mix. The ADC is mainly there to drive up the BOM cost and create controversy should anyone check out the repo. It's not as if the idea is to do clever stuff with it and have fun while building this thing or anything. And the aim is to be able to do continuous integration.

For this type of multi-slope 10 nF actually is on the high side. And 220 pF is on the low side, so I would expect we end up somewhere along this range...

Oh and pssst chickenHeadKnob .. *cough* factor of 2.54 *cough* That, and what branadic said.

Hello everybody

Some suggestions in no particular order:

suggestion1 - to disambiguate  give the inner ADC an official designation, I would chose 'residue ADC' otherwise this nested recursive ADC situation could be somewhat confusing. It looks like the idea from that american patent set in the 1990's. I like it, don't worry about cost optimization so much early on, there is more elastic expense and controversy in the FPGA.


suggestion 2 - choose the best fitting easily sourced integration cap you want, just make it through hole. Once this is documented the elve(s) at Hobby Posse Capacitor Company can fabricate a near fac simile from copper/PTFE and mail to you and stoney if you want it. I am pretty sure I can duplicate most electrical parameters but will require greater physical bulk than an off the shelf polypropylene cap. The last thing I want to do is waste your time. I think you and stoney are smart guys and have similar design taste and esthetic to myself (but me not so smart  ) and I very much want you to succeed.
I am continuing the materials search but have to interleave it with personal time.

 @branadic - I looked at that diagram, and yes that is the obvious way to make a rolled up foil cap, however I see a potential improvement; instead of soldering on leads at the end of the sheets/planes wouldn't it be better to solder closer to the center? The leads would have to be offset from one another if radial leads are desired, but axial is also easy. I wanted to roughly calculate the amount of material needed so I only considered single plate area. What you said in another post about included air in the dielectric gap is entirely true and further muddies  the capacitance calculation.

@Mrflibble - * cough 2.54 cough *  I thought I took this into account. One of the annoyances is that the natural areal unit for capacitors is m² but buying sheet goods in north america is still bound to imperial units so I am always converting back and forth.

[branadic]

You're welcome to join the club and help improve the design.

Thanks, but I think you've chosen your way, the design is already fixed and there is less place for other/new ideas, that are not common. I'm not familar with FPGA and think this solution is an overkill for me. But I'll give some input from time to time.

however I see a potential improvement; instead of soldering on leads at the end of the sheets/planes wouldn't it be better to solder closer to the center?

My gut instinct tells me that the cap will have less performance compared to second-hand market teflon caps, primary concerning long-term stability characteristics as you are not able to incapsulate them well against humidity and with regards to the fact that you will use a copper foil instead of directly plating the teflon foil with the electrodes. Only a few critical points, but I'm sure interested in your test results. Just show me the opposite.

[glatocha]

Hi,
I went through almost whole topic and I can see a lot of discussion about the case and usb and screen, etc.

I am mostly interested in the analog front-end.
I am thinking myself of making a small Voltage and Amp meter. But let's say my budget is really tight. This will be just my project not for commercial.

I am totally zero on analog circuits so this should be my learning project.
Let's focus on voltage. I need the +-10V with two ranges 9.99V and 99.9 mV. ADC will be build in the uC. So I need to do the range selection, divide the voltage and add the offset to have 0-2V on the ADC input.

PGA's are expensive, analog multiplexers also

I am thinking to do the Voltage dividing on resistors and then summarizing op-amp for reference offset. Then will go to my ADC and do the selection inside uC. Do you think this is the easiest solution?

P.S. I could split that to another topic, but think it shares the interested people

[Rerouter]

why not have the reference into a voltage divider, for your main range you leave your micro pin high impedance as an input, then to switch to the low range pull it down to ground by outputting low, yes there will be variations in how closely to ground a micro can pull, but that can be worked out in calibration, as one cheap and very low parts method,... you should still have clamping diodes, but you can make the range switching software based

[glatocha]

Hi,

how do you plan to make the channel separation? I am thinking about the solution to measure Voltage on two channels (with range +-10V) but the potential difference on both can be high (100V).
How to deal with such problems?