It is currently Sat Jun 23, 2018 1:15 am

Resistor tolerance and measured values

 
Posts: 4092
Joined: Wed May 02, 2012 3:35 pm
Location: Worksop

Resistor tolerance and measured values

Post by Refugee » Thu May 14, 2015 10:45 pm

I have had reason to search for high tolerance carbon film resistors in cases where I have none in my spares.
2% was what I was looking for and all I had were several tapes of 5% parts.
I set about finding a couple that were 2% or better but was very frustrated in the end.
Each tape was either 2.5% high or 2.5% low and all the resistors on the tape were within 1% or better of each other but not the marked value.
So modern resistors must all be made close tolerance and the better ones must them be creamed off and marked with the better tolerance while the remainder are sold as 5%.
I had a kit of NOS Russian resistors bought in the 1970s and what was left of them had the spread of values I was expecting and a couple of the parts I needed were found there.
I have no chance of finding 6.2M 2% until I place another parts order though. I will order 1% as replacements for those.
Has anyone else tried this?

 
Posts: 917
Joined: Sat May 26, 2012 5:41 pm

Re: Resistor tolerance and measured values

Post by Valvebloke » Thu May 14, 2015 10:56 pm

Refugee wrote:... Has anyone else tried this?


Not with close-tolerance resistors, but retailers who sell matched pairs and quads (and even sextets !) of valves pick the ones from the middle of the distribution of course. The consequence is that people who buy a few odd ones, on the basis that they'll save the matching fee and that most of the random picks will probably be somewhere near the middle, tend to find that they get only hot and cold ones.

VB

 
Posts: 27
Joined: Sun Feb 10, 2013 11:34 am
Location: Forest of Dean

Re: Resistor tolerance and measured values

Post by Stockden » Sat May 16, 2015 8:13 am

Refugee wrote:So modern resistors must all be made close tolerance and the better ones must them be creamed off and marked with the better tolerance while the remainder are sold as 5%.


I think that it may always have been that way. Certainly, when I was at college 40 years or so ago the electronics professor told us that this is the way it was done. I remember having my doubts at the time but your experience would seem to support his assertion.

Hugh

 
Posts: 138
Joined: Sun Apr 13, 2014 2:47 pm
Location: Middlewich, On the canals, & Philippines

Re: Resistor tolerance and measured values

Post by Boater Sam » Sat May 16, 2015 10:24 am

I gave up trying to get close tolerance resistors long ago. Instead I buy in bulk and use series/parallel combinations to get the exact value. Not as neat but possibly better in the long term as value drift may balance out?
Certainly cheaper, I object to paying 10 or 20 times the price just because a robot somewhere has selected the close value ones.

 
Posts: 358
Joined: Sun Jul 29, 2012 8:41 am
Location: Duffort, France

Re: Resistor tolerance and measured values

Post by BluePilot » Sat May 16, 2015 4:07 pm

Refugee wrote:So modern resistors must all be made close tolerance

I didn't think they made resistors to any tolerance. I thought they just made them and measured them afterwards to see what values they had. If you look at say the 10% range then the values are such that 10% above one value is (about) 10% below the next value. Whatever value a resistor has, it will be within 10% of some standard value. There are more standard values in the 2% range so that any given resistor will be within 2% of some value or other. It doesn't seem logical to sort out all the resistors withing 2% of a standard value in the 10% range because the ones which are between 2.1% and 10% off will still be within 2% of some 2% standard value.

 
Posts: 4092
Joined: Wed May 02, 2012 3:35 pm
Location: Worksop

Re: Resistor tolerance and measured values

Post by Refugee » Sat May 16, 2015 5:28 pm

Carbon film resistors are made by coating a ceramic rod with carbon and then cutting a spiral grove in the carbon to make the track longer. The number of turns that are cut into the carbon dictates the value.
I doubt that the entire range of values could be selected from one batch of resistors.
Composition resistors were made up "bucket and spade style" and would be a different matter all together.

 
Posts: 3712
Joined: Sun Feb 12, 2012 7:43 pm
Location: North Hykeham, Lincolnshire and Ilford, Essex (but not for much longer ...!)

Re: Resistor tolerance and measured values

Post by Terrykc » Sat May 16, 2015 5:33 pm

I covered the subject of tolerances and value sequences in some detail in the 'In The Workshop' section of this forum - sorry, can't copy link on this phone - I stopped at 5% because that is the highest precision applicable to the majority of topics here.

 
Posts: 411
Joined: Thu Aug 04, 2011 8:50 pm
Location: Evesham

Re: Resistor tolerance and measured values

Post by Mark Hennessy » Sat May 16, 2015 6:08 pm

Some years ago I wanted closely matched resistors for a 4 channel differential amplifier stage. I was going to buy 0.1% resistors, but at nearly a pound each, that was a bit extravagant. Instead I opted for the 0.5% version (Dale RN55 metal film). I'd just treated myself to a brand new Fluke 189, so enthusiastically set about measuring the batch that I'd bought. The accuracy of the Fluke is 0.05%, with 1 Ohm resolution on the range that I was using. Happy to say that all resistors were comfortably on-spec, and they were scarily close to each other - which was the important thing for my application.

I repeated this elsewhere in the product with no problems at all. I was most impressed with those resistors!

Also in the same thing, I needed to match some 0.1 Ohm wire-wound resistors to within 0.5%. This was more tricky, and involved an external current source and using the DMM to measure voltage. But it worked well. Quite how you manufacture a 0.5 Ohm resistor with any degree of precision is beyond me, but the whole batch was well within the 5% spec.

Even the basic "Multicomp" 1% metal film types are surprisingly accurate when measured. Most of the time, the accuracy isn't needed - even though they are dead cheap, I still feel guilty using a 1% resistor to supply operating current to an LED. They've certainly got their manufacturing processes sorted out :aad

 
Posts: 4092
Joined: Wed May 02, 2012 3:35 pm
Location: Worksop

Re: Resistor tolerance and measured values

Post by Refugee » Sat May 16, 2015 6:52 pm

They sure are getting more accurate. I have not seen 10% resistors for sale new for some time now.

 
Posts: 3712
Joined: Sun Feb 12, 2012 7:43 pm
Location: North Hykeham, Lincolnshire and Ilford, Essex (but not for much longer ...!)

Re: Resistor tolerance and measured values

Post by Terrykc » Sun May 17, 2015 7:37 pm

Terrykc wrote:I covered the subject of tolerances and value sequences in some detail in the 'In The Workshop' section of this forum - sorry, can't copy link on this phone - I stopped at 5% because that is the highest precision applicable to the majority of topics here.

Here it is: viewtopic.php?f=15&t=2344

BluePilot wrote:... If you look at say the 10% range then the values are such that 10% above one value is (about) 10% below the next value. Whatever value a resistor has, it will be within 10% of some standard value ...

The transition from one value boundary to the next is much closer than 'about'!

As you can see from the diagram I drew for the above thread, the boundaries are eerily accurate - either exact or overlapping by a minute amount - considering that all the values in the 10 - 100Ω 10% range are whole numbers - see:

download/file.php?id=3902&mode=view

 
Posts: 67
Joined: Sat Sep 20, 2014 6:59 pm
Location: Bedford, UK.

Re: Resistor tolerance and measured values

Post by johntheboffin » Sun May 17, 2015 9:09 pm

There is a disadvantage to the reproducibility of modern manufacturing techniques. In the "10 % days", if you wanted, say, a 964 Ω resistor you could select from a batch of 1 kΩ 10 % resistors and the chances were that you would find one that was close enough. Not any more - they're all pretty much the same value!

John

 
Posts: 4092
Joined: Wed May 02, 2012 3:35 pm
Location: Worksop

Re: Resistor tolerance and measured values

Post by Refugee » Mon May 18, 2015 10:22 am

That is exactly what I tried to do. They were all so close that I just could not find one the exact value I wanted.
I have put 1% parts on my next parts order list.

 
Posts: 411
Joined: Thu Aug 04, 2011 8:50 pm
Location: Evesham

Re: Resistor tolerance and measured values

Post by Mark Hennessy » Mon May 18, 2015 10:52 am

LOL - the poor manufacturers just can't win :bba

Of course, if you selected a 964Ω resistor, how long would it stay at 964Ω? Modern resistors are a lot more stable than the old ones...


Return to General Work Shop Discussion



Who is online

Users browsing this forum: No registered users and 4 guests