How to measure the inductance of a through-hole resistor?

[joeqsmith]

Including S21 for the 24k and 220k CC & CF 1/2W parts.  The unloaded fixture was also included to show it's contribution to the capacitive parasitics.   

I do not have any other type of 1/2W parts I can use for a comparison.   

[joeqsmith]

Looking at various 1W & 2W CC.     That 1M part is not very stable.  Again remember the unloaded fixture is roughly 90dB. 

[joeqsmith]

The 1W 56 ohm resistor behaved unexpectedly.  It was highly inductive compared to the others.   Shown are the 87ohm 1W and 47ohm 2W along with the 56 ohm.  Inductance of the 47 & 87 appear similar but something is up with the 56 ohm.

[joeqsmith]

A rasp and sandpaper tells the story.   

The CF spiral has 3 turns or so compared with the WW.   This is why we don't see much of a difference between the CF and CC.  Now, if I snip the leads short the effects of the spiral pattern would be more pronounced.   

Good luck with your future repairs. 

[joeqsmith]

Not a great comparison but showing two MF parts compared with the 1W CC.   CC measure 77mm end to end where the 1/4W MF is 50mm.   

[joeqsmith]

Ohmite Red Devil 82 ohm.   I think these were CC.

[T3sl4co1l]

Wirewound in a phenolic case? Oooh, sneaky! I had no idea!

Is the core ceramic? (Steatite? Alumina?)  If you can check the Mohs hardness or something like that, that may be slightly identifying.

Tim

[joeqsmith]

Wirewound in a phenolic case? Oooh, sneaky! I had no idea!

Is the core ceramic? (Steatite? Alumina?)  If you can check the Mohs hardness or something like that, that may be slightly identifying.

Tim

It's a pretty old part and other materials may have been used.    To expose it we start with a diamond stone to get a decent surface and finish it up with some 3000.  Never grind anything that is unknown dry.    Note the different layers of brass and copper.   

[joeqsmith]

Dried and under the microscope, it has a woven pattern to it.   I suspect the substrate is made from asbestos.   

[joeqsmith]

Using a very sharp dental pick from my old dentist, I attempt to deform the substrate.  It was very soft and spongy.  You can see how I pushed a hole into it. 

Just a quick patent search for resistors and asbestos,  first hit:   

Since the helix was wound onto the support assembly after construction relatively large clearances were required, producing a loose ribbon-to-insulator fit. In addition the use of the asbestos in the resistor presented a possible health hazard to assembly workers.

Searching for lawsuits,  Allen-Bradley Co. v. Erie Resistor Corporation 1939

This invention relates to certain new and useful improvements in electrical resistors and refers more particularly to molded fixed resistors and the method of making the same.

Heretofore, fixed resistor units have been molded of two general types of compound, one comprising a mixture of phenol resin, asbestos fiber, and graphite, and the other a mixture of phenol resin, asbestos fiber, and carbon black

I would assume OP and anyone else working on old electronics and electrical devices is very much aware of the use of asbestos and follow some basic precautions. 

[joeqsmith]

Comparing a 47 ohm RCD 5W with the 47 ohm 2W CC and what we now know is a 56 ohm WW.   The 5W RCD part is roughly the same package size as the 2W CC. 

Datasheet may be found here:
https://www.rcdcomponents.com/wp-content/uploads/2022/01/100-Series-Datasheet_RevI.pdf

I doubt based on the comments about costs that this would interest OP but maybe someone else would find these useful:
https://www.rcdcomponents.com/rcd-announces-release-of-cm-series-carbon-matrix-resistors/

***
Front to rear
RCD 160 5W WW,  Ohmite 2W CC, Ohmite Red Devil 

[MisterHeadache]

Guys, thanks for the very detailed test results, especially you joeqsmith!  I've made a summary of the results so far:

poster   mfr/series   construction   value   measurement   reactance?
mag_therm   Vishay PR02   metal film   4.7k   S11   strongly inductive
mag_therm   Multicomp 1206   unknown   4.7k   S11   moderately inductive
joeqsmith   unknown   carbon comp   390/0.5W   S21   not significant
joeqsmith   unknown   carbon film   390/0.5W   S21   not significant
joeqsmith   unknown   carbon comp   750/0.5W   S21   not significant
joeqsmith   unknown   carbon film   750/0.5W   S21   not significant
joeqsmith   unknown   carbon comp   24k/0.5W   S21   weakly capacitive
joeqsmith   unknown   carbon film   24k/0.5W   S21   weakly capacitive
joeqsmith   unknown   carbon comp   220k/0.5W   S21   moderately capacitive
joeqsmith   unknown   carbon film   220k/0.5W   S21   moderately capacitive
joeqsmith   unknown   carbon comp   1k/1W   S21   not significant
joeqsmith   unknown   carbon comp   1.5k/1W   S21   not significant
joeqsmith   unknown   carbon comp   1Meg/1W   S21   strongly capacitive/fixture effects
joeqsmith   unknown   carbon comp   3.9k/2W   S21   moderately capacitive
joeqsmith   unknown   carbon comp   47/2W   S21   not significant
joeqsmith   unknown   wirewound in phenolic   56/1W   S21   strongly inductive
joeqsmith   unknown   carbon comp   87/1W   S21   not significant
joeqsmith   unknown   carbon comp   240/2W   S21   not significant
joeqsmith   unknown   carbon comp   270/1W   S21   not significant
joeqsmith   unknown   metal film   1k/0.25W   S21   not significant
joeqsmith   unknown   metal film   82/0.25W   S21   not significant
joeqsmith   Ohmite red devil   carbon comp(?)   82/0.25W   S21   not significant
joeqsmith   Ohmite red devil   carbon comp(?)   82/2W   S21   not significant
joeqsmith   RCD 160   wirewound   47/5W   S21   severely inductive

I see some key take-aways:

• Just because a resistor looks like a classic carbon composition resistor doesn't mean it's non-inductive!  Those results on that 56 ohm / 1 watt resistor are alarming.  I wonder how many folks over the years tried to use those instead of 51 ohm ones in a quick and dirty QRP dummy load and had frustrating results.
• The carbon film resistors so far look to have insigificant reactance.
• Any resistor over a couple hundred kilo ohms looks difficult to quantify (fixture effects, capacitance starts to dominate)
• And this - ultimately one might have no choice but to measure every candidate resistor before using it in an RF circuit to know with certainty whether its reactance will be minimal or not.

I've got my own measurements to do this week.  And just for grins, I sent an email to tech support at Vishay, asking if they have lab data on the PR02 series for reactance.  Won't hold my breath for a response...

There is a  range of carbon resistors at https://www.surplussales.com/ of Nebraska
I buy tubes etc and metal film there, but I wouldn't buy carbon.

Yep, I recently ordered some carbon comps from them and several had drifted more than 10%.  I was not surprised and expected that because they are NOS and who knows how old they are.  I will use some of them for my VNA comparisons.

If inductance really matters it's probably time to get acquainted with SMT...

My desperation plan is exactly that - make a tiny bit of PCB to hold a 1206 SMT resistor (or possibly two, one on each side in parallel for higher wattage), then solder wire leads to it to make a cobbled-up low-inductance resistor replacement.

[joeqsmith]

See 7.20
https://www.rcdcomponents.com/wp-content/uploads/2021/06/R-14-CF-MCF-Carbon-Film-Resistors.pdf
See 7.20
https://www.rcdcomponents.com/wp-content/uploads/2021/06/R-08-GP-MGP-Metal-Film-Resistors-.pdf

Any resistor over a couple hundred kilo ohms looks difficult to quantify (fixture effects, capacitance starts to dominate)

I had provided data for the fixture with no load attached just to answer that question.  While it is 15 or more below,  you would need to decide what your are  really trying to measure.  Parts with their stock leads unbent, bent, cut .....  In this case, the leads are straight and uncut.  If they were all cut short we would greatly reduce the parasitics.   As it is, they all have different lead lengths depending on the series. 

Keep in mind too that the voltage rating their ability to handle very high surge currents are a few reasons we may still use CC today.  It all depend on the application and it's really up to the person doing the repairs to understand what parameters are important for each component they are replacing. 

[bson]

I'm surprised at that Vishay PR02!  An inductive MF resistor...

It also seems to have a significant parallel capacitance that becomes dominant at higher resistance values... But we're only looking at like 10-100MHz here.  But then it is an axial THT part.

From the datasheet:

[MisterHeadache]

Guys, thanks for this additional info!  I have been looking at various data sheets and totally missed those impedance versus frequency plots on the Vishay PR01/02/03 data sheet.  That's pretty much telling me exactly what I have been chasing, at least for that particular metal film resistor series.  And a rep from Vishay did respond to me - and referred me to the data sheet. 

I did a quick search and found Allen Bradley's data sheet from the early 90's that also has impedance versus frequency curves for their carbon comp resistors.  I'll spend some time studying them versus the Vishay curves and the RCD curves for carbon film. That will help immensely with my understanding the expected differences between classic carbon comps and modern carbon film and metal film.

Also it is finally sinking in my aged brain what you guys have been trying to tell me: 

• The series inductance typically dominates on low ohm resistors, causing the Z/R ratio to rise above 1.0 with increasing frequency.
• There is a 'sweet spot' of resistances where Z/R doesn't deviate much from 1.0.  That is roughly between 100 ohms to several kiloohms, depending on a specific resistor series.
• Above the 'sweet spot', the parallel capacitance dominates, causing the Z/R ratio to decline below 1.0 with increasing frequency.

And mawyatt's model is a useful approximation to simulate that performance.

[MisterHeadache]

I took the 0.1-100MHz impedance versus frequency plots from the Vishay PR01/02/03 [metal film], RCD CF [carbon film], and Allen Bradley [carbon composition] data sheets and re-plotted the curves on the same scale.  I chose the Vishay and Allen Bradley 1W plots, but the RCD is 1/4 Watt, that's just the size they chose to show in the datasheet and they don't have anything else on their web site that I could find.

There are three groupings: <10 ohms, medium (~10k), and high (100k).  The values aren't exact for all three because all three didn't have curves for the exact same values.  And I had to visually pick data points from them, so this is just an approximation.  But it is more than adequate to show a few interesting observations:

RCD CF25 [carbon film] shows the least impedance change over frequency and over the largest range of resistances.  Maybe it's helped by being the 1/4W size instead of 1W size?
Vishay PR01/02/03 [metal film] has significant inductive impedance at low resistance (1 Ohm), but has the second best performance at 100k.
The beloved Allen Bradley [carbon comp] has virtually no change in impedance for values under about 100 ohms, but quickly becomes substantial with higher resistances.  It has the worst performance of the three as frequency increases, but the Vishay catches up to it at 100MHz/100k Ohms.

The biggest eye opener for me is just how poorly the Allen Bradley are shown to perform over frequency.  Maybe because their data is with leads of only 1/4" in length, so there's minimal lead inductance to mix with the body capacitance at higher frequencies?  Just a guess.

There is this mythos about carbon composition resistors being so precious for proper function in RF circuits versus modern resistors.  I can see that being true for low (a few hundred Ohms) applications, but that's it - the modern metal film seem to actually perform with less capacitive reactance above that.  I wonder now if some older RF circuits that used CC's in the signal chain were actually tuned to work with the impedance characteristics of a CC, and substituting a non-CC (which ironically might have less reactance) could de-tune or upset the circuit performance.

Here is a link to the AB datasheet that I found:
https://www.hificollective.co.uk/sites/default/files/allen_bradley_data_sheet.pdf