Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronic Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Register Log in

Strange Question : Why these values ?

Status
Not open for further replies.

superluminal

Member level 4
Joined
Jan 31, 2002
Messages
79
Helped
1
Reputation
2
Reaction score
0
Trophy points
1,286
Location
Inside a quantum well
Activity points
806
This is a simple question but I can not answer it.
Why usually resistors & caps are shipped in specific values like 47K,82K,91K & 47uF,2200uF ..etc , why these values & not , for example, 10 20 30 ..etc ?
 

buzkiller

Member level 4
Joined
Dec 31, 1999
Messages
75
Helped
1
Reputation
2
Reaction score
1
Trophy points
1,288
Activity points
602
These values are supposed to give the best possible choose when cascaded together to form other values.

regards,
Buzkiller.
 

Epegic

Member level 5
Joined
Mar 29, 2002
Messages
84
Helped
24
Reputation
46
Reaction score
19
Trophy points
1,288
Activity points
597
Values are defined to covered 10% torlerance range.
 

dkace

Full Member level 6
Joined
Jun 13, 2002
Messages
396
Helped
32
Reputation
64
Reaction score
7
Trophy points
1,298
Location
Greece
Activity points
3,557
Logical question for someone trying to design simple circuits for a begining. I think buzkiller gave an adequate answer on the point. If you check several schematics you'll find out that cirquit to ground connection, feedback in op amps etc, is accomplished with some of these values. If you also see the list of electronic elements in a lab, yo' ll see that also the fixed values of 10,20,etc are also there.
So a statistic from the use in real cirquits , will show you that this "odd" values are appearing more than often in the cirquits. What is more, cascade between a resistor of 40 Ω, 5Ω and 2Ω, to make a 47Ω one, can cause capicity or noise problems. So it is easier to make one 47Ω from the beginning! :idea:
 

Aoxomox

Member level 5
Joined
Jul 18, 2001
Messages
90
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,286
Activity points
1,294
hi dkace,

please correct me if I am wrong or maybe I only know half of the storry.

Well when being able to produce resistors with a tolerance of 10% then the absolute tolerance increases with increased absolute value.
Example:
R=10 Ohms deltaR = +- 1 Ohms
R=100 Ohms deltaR = +- 10 Ohms ( >1 Ohms)

so when producing a product portfolio how do you best cascade these values when you want to produce as few different components and still want to cover all values possible??

linear case with steppsize
Values : 10 20 30 40 50 60 70 80 90
Tolerances: 9-11 18-22 27-33 36-44 45-55 54-66 63-76 72-88 81-99

As you can easily see the tolerances in the beginning of the series don't overlapp and for increasing values (like 90) the tolerance edge meets almost the center value of the neighbour.

with a logarithmic stepsize you have:

Values : 10 22 33 47 68 82
Tolerances: 9-11 19.8-24.2 29.7-36.3 42.3-51.7 61.2-74.8 73.8-94.6

As you can see there are less single values needed and the tolerance edges almost meet for higher vaues within one decade.
Fewer different components are cheeper to produce, distribute, store.......

You are right in saying that fewer components in series have good affect to noise performance but for increasing accuracy in producing components I think it's not neccessary to combine them. 1% tolerated resistors are standard today and thoses series (E96) have much more single components than the E24 (10%) have.

hope it helps.
aOxOmOx
 

dkace

Full Member level 6
Joined
Jun 13, 2002
Messages
396
Helped
32
Reputation
64
Reaction score
7
Trophy points
1,298
Location
Greece
Activity points
3,557
Hi aoxomox,
you seem quite informed on the matter. I was speaking out of experiance-limited to say the truth-. I had the same question as superluminal, in my first steps in electronics. I think your answer is quite sophisticated, so I have to accept it!


:wink:
 

superluminal

Member level 4
Joined
Jan 31, 2002
Messages
79
Helped
1
Reputation
2
Reaction score
0
Trophy points
1,286
Location
Inside a quantum well
Activity points
806
When I was doing my first circuits long ago,I was asking this question & I was answered like what buzkiller & dkace said & I accepted this ( & still).Now it was my turn for a beginner to ask me this question & I answered him but he asked for manufacturing reasons for these values & thanks Aoxomox ; now I can answer him.
And thanks to answer me why metal film R's (1% tol.) are in many values than carbon R's (10 %).
 

jzaghal

Full Member level 2
Joined
Jul 31, 2001
Messages
130
Helped
1
Reputation
2
Reaction score
0
Trophy points
1,296
Activity points
955
Hi all,

Search for E12 and E24 with google and you will find the answer to your question on resistor tolerances.

:D :D :D

Best wishes.
 

j_o_n_i

Junior Member level 3
Joined
Mar 26, 2002
Messages
26
Helped
1
Reputation
2
Reaction score
1
Trophy points
1,283
Location
Between Apenines and Balcans
Activity points
210
Hey everybody, remember there are also the variable resistors to help. They're not only used to do adjustments ( eg. adjust the volume ), but also to help in achieving difficult resistor values. You could use one of them (usually in series with another resistor) instead of putting several resistors in series and/or parallel. It would help to resolve the problems dkace was talking about. :arrow:
 

Methanol

Junior Member level 1
Joined
Mar 23, 2002
Messages
15
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,281
Activity points
48
Hi everybody I agree with Aoxomox but let me add one more think. The usual resistors and capacitors values (10,12,15,18,....) come from the beginning of the radio,by that time resistors and capacitors were usually manufacturated with a 20% tolerance (fourth band gold=5% ,silver=10% ,without fourth band=20%).
So if you first value is 10 the next one should be 10+20% (if you want overlap) so the next should be 12. You can build and small table.
10
10x1.2=12
12x1.2=14.4~15
15x1.2=18
18x1.2=21.6~22
22x1.2=26.4~27
27x1.2=32.4~33
etc.
Regards.
Methanol
 

Kit-the-great

Full Member level 3
Joined
Mar 30, 2002
Messages
151
Helped
5
Reputation
10
Reaction score
3
Trophy points
1,298
Location
Russia
Activity points
1,318
'Strange' nominals of res/caps

Nominals of components choosing by the manufactures on this reason:

Minimum number of nominals to obtain all values (by serial/parallel combinations).
Just it easiest way to manufacturer.

Just a business, not personal !

With response,
Kit-the-great
 

RegUser_2

Full Member level 2
Joined
Dec 24, 2001
Messages
125
Helped
11
Reputation
22
Reaction score
9
Trophy points
1,298
Activity points
1,507
Oh well, the interesting question here is how precise the paralel/serial combination of a resistors is usefull. So far the only place such precision is needed is in the feedback loop of an active filter (to define the filter behaviour), yet if someone look on the capacitance tollerances (at best about 2%), the need of better matching for the resistors is pretty much questionable. In general the gain matching (as well as the frequency responce of a filter) are mapable via correct calibration procedures during the process of measurement, hence the need of high precision components (and in this sense the standard set of values) aren't much critical. Naturally in case of a digital design, the needeed set of resistors (and their matching) is much relaxed. In case of power supply applications, the desired tollerances won't need be much better than say 5%. So yes, why do we need even the 1% set values? :)

In my humble opinion the thermal properties (the TKC/TKR matching) are a lot more important than having extended set of values though ;)
 

Aoxomox

Member level 5
Joined
Jul 18, 2001
Messages
90
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,286
Activity points
1,294
Hi RegUser_2

you are right that for high resolution applications low tolerated components are not enough. there you need a calibration procedure where (in analogue case trimpots/trim caps, in digital case a frequency response and/or offset, gain etc.) correction values are defined to compensate the behaviour of the circuit.
This is usefull for low volume devices like high perfomance mesurement equipment with high prices.

But in all other cases where high volume products are assembled it will be much to expensive to adjust every single circuit if one can avoid this procedure just by using low tolerated components.
sometimes the expense for a very few capacitors with 2% tolerance will even be cheaper than manual or automatical adjustment of the circuit.

aOxOmOx
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Top