Tantalum capacitors in parallel?

[ChildOfVision]

Hi!
I'm interested, what is difference if I use, say, one tantalum capacitor of 10uF or if I use two 4,7uF in parallel? I need two tantalum of 10uF but I don't have them right now. I have, 4 capacitors of 4,7uF, so I can connect them in pairs (capacity is not critical). For ESR I know: it will be lower (twice) or I'm wrong? But what with other good characteristics of tantalum? What about noise, "stray" inductance and so?
I allready soldered 4,7uF (through-hole component) so I can easily add another one on the copper side (even SMD variant), but I don't know will that "2x4,7uF" combination be worse than a single 10uF?

Also, maybe stupid question, what if I connect (generally speaking) small NP0 ceramic capacitor in parallel with (any) tantalum? Can I get some improvements (say, 10uF tantalum + 10nF NP0)?

TIA!

 

[KeepItSimpleStupid]

One question becomes, what for or how is it used? Bypassing is an art. Yes, we know that capacitance's add in parallel. ESR might be lower than one of the paralleled size. You may find 3 or 4 different capacitor types in parallel such as electrolytic, tantalum and ceramic because no one capacitor is ideal.

 

[ChildOfVision]

Yes, I forgot to mention intended use. I'm not an expert so I can't tell you very precisely, but that capacitors are part of LCesR-meter - they are in the path of the signal toward PIC (C5 - the first one). Second tantalum (C4) is connected from PIC toward GND and I don't know what is its purpose. Here is schematic of that meter, maybe you can figure something out!?
View attachment 63670
Generally, they should be low noise, low inductance, low ESR, low noise, as much as possible.

You may find 3 or 4 different capacitor types in parallel such as electrolytic, tantalum and ceramic because no one capacitor is ideal.

Yes, that's why I asked about NP0 combination!

 

[KeepItSimpleStupid]

It looks as if C4 is basically a filter. High ripple current capability might be a good reason to use a tantalum. Low ESR secondary.

C5 looks like ESR, and ripple current are the prime reasons. Leakage, I'm not so sure. Two tantalum in parallel might be better for C5 to lower the ESR.

Most LCR meters of any sort will have the ability to null out external effects with the short and open tests. The short test compensates for lead resistance and probably inductance and the open test will compensate for stray capacitance. I'm not sure where to put the inductance compensation.

The short and zero tests are done at all frequencies that the meter uses.

 

[ChildOfVision]

Thank you for explanation! And what do you think: can I put DIFFERENT values for each of them, e.g. to leave 4,7uF for C4 (for filtering, as you think), and put two 4,7uF in place of C5 - to get about 9uF and lower ESR? Of course, nothing will explode (I hope so) if I try that, but I just worry to not ruin something (PIC!).
According to documentation, that tantalum caps have NO influence to calibration process. Allegedly, lowering from 10 to 4,7 (or even 2,2)uF (slightly) improves the upper limit for C-measurements (I'm not sure about that).

 

[crutschow]

......................................
Also, maybe stupid question, what if I connect (generally speaking) small NP0 ceramic capacitor in parallel with (any) tantalum? Can I get some improvements (say, 10uF tantalum + 10nF NP0)?

You can put an NPO ceramic in parallel with a tantalum but the NPO (low temperature coefficient) will not be of any use since it is overwhelmed by the tantalum temperature coefficient, which is large. But it is common to connect a standard ceramic cap in parallel with a tantalum cap to reduce the high frequency inductive reactance and improve high frequency noise filtering.

 

[ChildOfVision]

@crutschow: Thank you very much.

 

[dr pepper]

You often see a ceramic accross a larger cap, whether electro or tant, a 10u electro is not all that good at bypassing rf so a ceramic across it makes an improvement, and it can be cheaper than using a single high quality cap.

I'm interested in your esr and q meter, can I have a link.

I've built a couple of lc meters using a pic16f628, with a fairly similar circuit, one used tants and the other used old electro's salvaged from a tv.
I couldnt really tell any diffo in accuracy between either of them.

 

[KeepItSimpleStupid]

I've used the big guns like one of these: https://www.testequipmentdepot.com/...kard/impedanceanalyzers/hp4274aandhp4275a.htm and a relatively recent (about 5 years old) capacitance meters. We had to do measurements with DC bias on an active device. Both had issues.

I just purchased an Agilent U1733C which had just about everything I wanted EXCEPT a low oscillator level. I was hoping for about 0.3 V RMS. The most expensive of the series does DC resistance. They all have an auto identifying mode.

I read somewhere, where someone build a very fancy LCR meter with a sine wave oscillator and quadrature detector.

 

[ChildOfVision]

@dr pepper:

As the first, thanks for your opinion!

I must "warn" you that this "Hungarian LCM3" meter is NOT Q-meter if you mean to Q-factor. But it is very simple design, it demands to choose right parts, though, and that can be "tricky", so you must be patient. It is able to measure: L, C, small resistances (resolution up to 1mΩ), ESR, leakage of electrolytes (and bigger value non-electrolyte capacitors), "memory-effect" (or dielectric absorption), parasitic capacitance of larger inductance etc. Originally it is projected as low-accuracy instrument, but people from around the world, especially from Russia, improved that instrument so you can with some patience get 1% or even better accuracy!

I will give you link to original site - it is in Hungarian, so I recommend you to use google-translate to ENGLISH. Hungarian is VERY tough language and even with google translate it is very difficult to get some infos (and, to make things even worse, there are many typos)!

BUT, I will give you another link to Russian forum, dedicated to this project! Google-translate to English works surprisingly well! So far there are about 22 pages, but it is very easy to follow it!

I just "fine-tune" LCM3, I'm not some expert but I have quite decent knowledge about it, so feel free to ask me whatever you want, here or to PM!

<HUNGARIAN SITE (original)>

<RUSSIAN FORUM>

@KeepItSimpleStupid (and maybe @dr pepper): There are AFAIK very, very good LCFX meter ("GO-RLC2") - it is quite complicated instrument, and demands a lot of parts (about 100€) and a proper instruments (oscilloscope, at least), but many people that builded it tell it is worthwhile! Maybe that is what you need, here is a link:
<GO RLC-2 (Russian site!)>

 

[dr pepper]

Thanks Mr Vision.

I think thats an excellent project, the developer is obviously keen and skilled.

I have an LC meter built from a web article, and as mentioned even though it has electro caps in the oscillator circuit it still gives around 1% accuracy. Its an excellent tool and has been well usefull on both repairs and development, its in a small hand held case so it goes in my travel pack and only costing 20 quid to build I dont have to worry about using it.

I can make out the translated hungarian page once its been translated just about, my english is very good I am english.
I wont be building this right away, there another couple of things on the list, however I'll report back on my findings.

 

[dr pepper]

Revived an old thread.

OK then, I want to build an esr meter, tired of fiddling about trying to find dead caps, and after some parts mining I have a biscuit tin full of low esr caps.

The LCM3 seems like an excellent lc meter that can also measure esr down to milliohms, very nice.

Has anyone either built one of these or have a link to a site in english, I'm finding it hard to find a good reference in english, I'm thinking of just throwing it together, when I find some styroflex caps, which in my 30 years in electronics I've never heard of.

 

[ChildOfVision]

dr.Pepper, I'm here to help you as much as I can! There is NO english documentation about this meter! It is all in Hungarian, which is one of the most toughest language in the world! I assembled it, with a lot of help of our colleagues from all of the world. It is very "tricky" instrument, every component is critical! This instrument also shows "memory" effect (or dielectric absorption) and leakage of electrolytes, as well as parasitic capacitance for windings above cca. 22 mH! This is hard to find even in very expensive professional instruments!

However, I will try to find some useful links, also I will send you my documentation (even on my language, but with google-translate, it should be much more legible than translation from Hungarian!) + some VERY IMPORTANT hints (tips&tricks) and pictures!

Just, please, be patient! If you wish, we can continue our conversation via emails - in that case contact me by PM (to exchange addresses)!

Cheers!
ChildOfVision

PS: for now, you can visit this Russian site (google-translate to English works pretty well), it is very useful:
https://vrtp.ru/index.php?showtopic=19662

 

[dr pepper]

Ta.
I think we should talk here for now, theres probably others that speak english who want to build this.

I allready built an lc meter and ported the code to another pic, it doesnt do esr which I want.

The circuit number of components wise doesnt seem complicated, I have the hex file so I'm thinking of dumping the code to a pic and assembling the circuit on breadboard and see what happens.

I can kinda make sense of the russian page thats been translated, its not so much the translation its sifting through pages and pages of threads.

Thanks for your offer of help.

 

[ChildOfVision]

dr.Pepper, just very shortly:

The circuit number of components wise doesnt seem complicated, I have the hex file so I'm thinking of dumping the code to a pic and assembling the circuit on breadboard and see what happens.

Did you mean on this Hungarian meter? Yes, the schematic is very simple, but that is big "illusion", so beware! Use only quality polystyrene capacitors for 1nF and 33nF!!! The inductor is much more critical and tricky! If you can understand enough from google-translate, it is no hard to follow the story, although there is 40+ pages! If you have some troubles, I will help you (my language is also in "Slavic language" group, and I can read (printed only) Cyrillic, so that helps me alot)!

About experimentig on the breadboard: one big NO! You MUST use PCB (FR2 is OK), with good layout, ground plane etc, but I do not reccomend you to use the original layout!!! Make your own layout, according to your parts. Switches also must be of very good quality (you can find it on futurlec.com). I found the best inductor is from some old CRT monitor - winded with "litz-wire" on drum-core (you don't have to use pot-core!) with appropriate inductivity (95-100 uH)!

That much, for now!
See you later!

 

[audioguru]

I used tantalum capacitors in the '70s. They all blew up after a while. I haven't used them since.

 

[ChildOfVision]

Yes, I hate them! But, sometimes there is no alternative! Maybe new Os-Cons...?

 

[dr pepper]

I have ordered up some tants, mkp's and resistors for this project.

I found a site that has a not particularly brilliant translation, however it makes more sense than russian or hungarian:

https://www.scribd.com/doc/96638253/Lcm3-Esr-Meter-Engl

Vision child, what is the latest software version, I'm guessing the one you can download is an older version.

 

[ChildOfVision]

Dear dr.pepper, I had some "internal" problems today, I will reply to your questions later (tomorrow)! Sorry! For now, yes, use any piece of documentation you can get! Unfortunatelly, I can't download this particular scribd-document because I'm not a "premium member", but as much as I could see - there is one cardinal mistake in design - the LCD panel should not be above inductor! If you wish you can specify what parts you ordered so I can comment these! You must use all metal-film resistors with (at least) 1% tolerance! Tantals must be with as low ESR as possible (200 mOhm is OK)! But more about that later!

See you soon!
Child Of Vision

 

[dr pepper]

OK then.
Caps are mkp, not sure about the tant beads, noted about the display, that will apply more to one with an inverter for a eld.
I'm not a member of scribd either, must be something else.