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

[MisterHeadache]

Here's my question: How can I measure the inductance (or inductive reactance) of through hole resistors of values from 1 ohm through 1 Megohm, fabricated from unknown construction, using instruments commonly available to the hobbyist?

Background: I work on a lot of old RF electronics that use carbon composition through-hole (axial leaded) resistors.  Carbon comps are good for those apps because they have low inductance, but they are hard to find nowadays (either expensive niche manufacturers or untrustworthy NOS).  Ceramic non-inductives are avaible but pricey (~$2 apiece).  Metal film I understand MIGHT be low enough inductance, but I have found only one (Vishay PR02-FS) which are spec'd as non-inductive, but they top out at 100 ohms and I need values upwards of 1 Megohm.

There's plenty of useless blather online about the 'relatively low' inductance of metal film and some carbon film through-hole resistors, but what I want is DATA.

So more specifically:

• Could I set up my nanoVNA to evaluate the inductance of my junk box of metal film/metal oxide/carbon film resistors?
• Could a relatively inexpensive LCR meter make this measurement? I understand most LCR meters cannot tolerate much resistance when making an inductive measurement.
• Could I make a resonant circuit with the resistor and a known capacitance and stimulate it with a step response and observe the response on a 'scope and determine the inductance?

Lastly, I don't really need super accurate measurements.  I'd settle with being able to do a side-by-side comparison to a carbon comp resitor of the same resistance value (which of course has lead length inductance anyway).

[vk6zgo]

Here's my question: How can I measure the inductance (or inductive reactance) of through hole resistors of values from 1 ohm through 1 Megohm, fabricated from unknown construction, using instruments commonly available to the hobbyist?

Background: I work on a lot of old RF electronics that use carbon composition through-hole (axial leaded) resistors.  Carbon comps are good for those apps because they have low inductance, but they are hard to find nowadays (either expensive niche manufacturers or untrustworthy NOS).  Ceramic non-inductives are avaible but pricey (~$2 apiece).  Metal film I understand MIGHT be low enough inductance, but I have found only one (Vishay PR02-FS) which are spec'd as non-inductive, but they top out at 100 ohms and I need values upwards of 1 Megohm.

There's plenty of useless blather online about the 'relatively low' inductance of metal film and some carbon film through-hole resistors, but what I want is DATA.

So more specifically:

• Could I set up my nanoVNA to evaluate the inductance of my junk box of metal film/metal oxide/carbon film resistors?
• Could a relatively inexpensive LCR meter make this measurement? I understand most LCR meters cannot tolerate much resistance when making an inductive measurement.
• Could I make a resonant circuit with the resistor and a known capacitance and stimulate it with a step response and observe the response on a 'scope and determine the inductance?

Lastly, I don't really need super accurate measurements.  I'd settle with being able to do a side-by-side comparison to a carbon comp resitor of the same resistance value (which of course has lead length inductance anyway).

Back in the early 1960s, our Lecturer at Perth Tech school demonstrated the inductive coefficient of carbon composite & carbon film through the hole resistors of various values using a "Q meter".

The carbon film resistors were trimmed to value at the factory by a spiral cut in the film, which added substantial inductance.
I think modern metal film resistors are a bit more sophisticated, but perhaps not.

[joeqsmith]

Maybe something here will help you out as the same basics apply:
https://www.eevblog.com/forum/rf-microwave/shunt-capacitance-of-1206-smd-resistors-jeroen-belleman-december-2010/

[mag_therm]

For a nanoVNA test jig:
Use a baseplate of about 75 by 50mm of un-etched 1 oz copper pcb
Solder two of SMA jacks AMPHENOL RF 132289 at required spacing.
That connector has easy access to solder the pins.
Solder in your network or single resistor suspended at required height over copper and do S11 and S21. (screw a shorting plug in to Sma2 for S11)

I also use the Vishay PR02 kit CCR-122 10R to 1 Meg.
I recall testing and using a series of 47000 and 50 Ohm  for 940:1 divider network which had low inductance at 14MHz etc but I did not record the values.

[MisterHeadache]

Thanks for the replies!  I read Jeroen's article and I understand the VNA technique.  Guess I was overthinking it!  I'll build the jig and run some comparison tests and post my results.  Question on the proper way to do a basic SOLT calibration with the jig, would it be:

Short: solder the pin on port #1 temporarily to the ground plane on the PCB
Open: remove the short, have nothing between the pins on the ports
Load: solder an SMT 50 ohm resistor from the pin on port #1 to the ground plane on the PCB
Through: solder a length of bare wire from port #1 pin to port #2 pin

[T3sl4co1l]

The basic shape you're looking for is a single-pole response.

The basic setup will work in reflection (s11, resistor wired across RF jack from signal to GND), or transmission (s11, s12*) with a series (jack 1---resistor---jack 2) or shunt (jack 1---jack 2, resistor from trace to GND) arrangement.  Which one is best, depends on value: small values (under Zo) are better in shunt, large values in series.

*s22 = s11 and s21 = s12 by symmetry, but you can always check all four just in case, or, say, to reduce noise by combining curves.

Then, fit a curve such that the component under test, in the jig arrangement given, is modeled as an R+L, or even (R+L || C) (or (R || C) + L, but you may not have data to disambiguate the two possibilities, for example, full transmission line effects might take over by the time series/parallel LC comes into question, and then a much more developed model is required; this will be in the GHz, corresponding to lead length).

You'll find that carbon comp are more or less ideal, with some capacitance at high frequencies, and inductance given by lead+body length (major downside: excess noise, and value tends to drift up over time, and under voltage bias); carbon/metal film (normally spiral cut) aren't much worse (despite what you may think of the spiral), and, wirewound fall apart in the MHz.  Wirewound are available into the megs, where C will also dominate at such frequencies; these will most strongly call into question the L/C placement in the model, but more than that, C is distributed along the winding so you may end up with a more detailed model anyway, even without transmission line effects.

Tim

[jonpaul]

Carbon Comp, MF, dep F about same as for a wire of that legnth.

We use 1 nH/ft

Up to GHz its insignificant,

Measuer with HP4195A Impedance test fixture and Thru Hole binding post adapters.

Jon

[trobbins]

I use a Picoscope 4224A and FRA4PicoScope software to easily measure part impedance out to 1MHz.  For an R-X series network, using an appropriate reference metal film R, I'm able to assess small value inductors like ferrite beads with <1uH, where the gain/phase plot nicely shows the impedance change of the inductor, with no indication of reactance from the resistor.  What I haven't done is to measure the response for an R-R* network, where R* is say a WW or a vintage carbon comp to assess the response out to 1MHz.

Are you not wanting to use normal PRO-2 resistors, with nominal impedance curves out to 100MHz, or is their impedance change too much ?

[mawyatt]

Carbon Comp, MF, dep F about same as for a wire of that legnth.

We use 1 nH/ft

Up to GHz its insignificant,

Measuer with HP4195A Impedance test fixture and Thru Hole binding post adapters.

Jon

Agree a leaded component should have an inductance ~ of a straight wire of equal total length if the component doesn't have any geometry which allows field coupling or additional effective length.

However, a for a 1mm dia wire the inductance is ~387nH/ft 

Best,

[Smokey]

Would a basic LCR meter like a DE-5000 not be able to measure inductance in this case?

[vk6zgo]

The basic shape you're looking for is a single-pole response.

The basic setup will work in reflection (s11, resistor wired across RF jack from signal to GND), or transmission (s11, s12*) with a series (jack 1---resistor---jack 2) or shunt (jack 1---jack 2, resistor from trace to GND) arrangement.  Which one is best, depends on value: small values (under Zo) are better in shunt, large values in series.

*s22 = s11 and s21 = s12 by symmetry, but you can always check all four just in case, or, say, to reduce noise by combining curves.

Then, fit a curve such that the component under test, in the jig arrangement given, is modeled as an R+L, or even (R+L || C) (or (R || C) + L, but you may not have data to disambiguate the two possibilities, for example, full transmission line effects might take over by the time series/parallel LC comes into question, and then a much more developed model is required; this will be in the GHz, corresponding to lead length).

You'll find that carbon comp are more or less ideal, with some capacitance at high frequencies, and inductance given by lead+body length (major downside: excess noise, and value tends to drift up over time, and under voltage bias); carbon/metal film (normally spiral cut) aren't much worse (despite what you may think of the spiral), and, wirewound fall apart in the MHz.  Wirewound are available into the megs, where C will also dominate at such frequencies; these will most strongly call into question the L/C placement in the model, but more than that, C is distributed along the winding so you may end up with a more detailed model anyway, even without transmission line effects.

Tim

The old Philips spiral cut carbon film resistors were pretty terrible back in the 1960s.

[mag_therm]

I tested some Vishay PR02 leaded resistors on the nanoVNA jig mentioned in #3 above.
I did not cut the leads , and laid the resistor bodies on the copper baseplate between the SMA jacks to simulate how they might be used in old tube restoration.

Up to about 4700 Ohm, the resistors were inductive, jX was between 10% (R=47 Ohm)  and 80% (R=4700 Ohm) of R at 20 MHz.
Up to about 100 Ohm, the resistors were inductive, jX was between 10% (of nominal R=47 Ohm)  and 6% (of nominal R=100 Ohm) of R at 20 MHz.
The nominal 4700 Ohm at 20 MHz measures 2660-j2360 Ohm

Above 47000 Ohm, the resistor becomes a parallel network of R and -jX   : (R*[-jXc])/(R+[-jXc])
For example of 100000 Ohm resistor on the copper baseplate, at 20 MHz it becomes 307-j5590 Ohm.
When the resistor was lifted about 15mm above the copper baseplate: 406-j6350 Ohm

I don't have any old carbon resistors here, it is likely that the higher values would have had have similar capacitive shunting.

I also tested the only leaded RF Choke here, about size of a 3 Watt resistor, rated 294 uH at 0.2 Amp
It has a double pole at 4.97 MHz with a Q of about 16, and capacitive above resonance.
If replacing RFC in old tube units  best to check individually.

scrots of sweeps. Marker 2 (Green) is at 20 MHz
Sweep S11 of Vishay_PR02 4700 Ohm:
https://app.box.com/s/gj439ukiujaufjnhem655i8yq5qocrdz

Sweep S11 of Multicomp Size1206 4700 Ohm
https://app.box.com/s/mautip8as1kgts73mw6ggfwqacusq206

Sweep S11 of 50 Ohm Cal
https://app.box.com/s/1ejn2hc0zqn1baxrsgifoi9fhzfcvgj2

[MisterHeadache]

For a nanoVNA test jig:
Use a baseplate of about 75 by 50mm of un-etched 1 oz copper pcb
Solder two of SMA jacks AMPHENOL RF 132289 at required spacing.
That connector has easy access to solder the pins.
Solder in your network or single resistor suspended at required height over copper and do S11 and S21. (screw a shorting plug in to Sma2 for S11)

I built this fixture last night and did a SOLT calibration on it per my presumption above.  I soldered in a 51 ohm resistor that I know is carbon comp and measured an almost flat S21 response from 1MHz to 100MHz.  It was within a few tenths of dB to the theoretical value per Jeroen's formula (-3.6dB).  But I ran out of time, will experiment with other resistors tonight.

I tested some Vishay PR02 leaded resistors on the nanoVNA jig mentioned in #3 above.
I did not cut the leads , and laid the resistor bodies on the copper baseplate between the SMA jacks to simulate how they might be used in old tube restoration.

Up to about 4700 Ohm, the resistors were inductive, jX was between 10% (R=47 Ohm)  and 80% (R=4700 Ohm) of R at 20 MHz.

So if I did my math right, that 4.7k resistor has an effective inductance of 30 microhenries at 20MHz?  I'd think that would upset most 20MHz RF circuits!

[mag_therm]

You are correct.
At a frequency where the abs  jXL = R, the impedance of the resistor is 1.41 times nominal R.
Mostly not a problem.
Looking at a HF receiver here (1963 Lafayette HE30) there are no circuits where the RF or IF signals  go through a resistor except at the cathode of the 455 kHz Q multiplier which is a 10 k selectivity potentiometer.
The plates have tuned coupling transformers, the low value cathode bias resistors are all capacitor bypassed.
RF and IF gain and AGC are by variable grid bias.

[mawyatt]

With a 4.7K resistor the effective series inductance will be swapped by the R, however the parallel equivalent capacitance will have an effect tho, as just ~1.7pF has an equal capacitive reactance @ 20MHz.

For low resistance values the series inductance influences the overall impedance, for high resistances the parallel capacitance has more effect.

Generally one can estimate the inductance using straight wire equivalents for the overall lead wires and resistor body as the effective length¹, and the capacitance by the physical resistor body makeup.

(1) https://chemandy.com/calculators/round-wire-inductance-calculator.htm

BTW the physical resistor body distance of the PR02-FS (10mm) implies an equivalent inductance of ~5nH, add wire leads so the total length is ~40mm and the overall effective inductance becomes somewhere around 35nH.

Here's a simple lumped element model for a leaded resistor, you should be able to "fit" measured values over reasonable frequency ranges to such.

Best,

[MisterHeadache]

OK, thanks very much for the explanations.  I'm now getting my head around how these components behave over frequency.

Referring to mawyatt's lumped model, aren't L-sub-L (lead inductance) and C-Sub-R (body capacitance) largely the same regardless of ohm value within a particular manufacturer's series and body/wattage size?  For laser trimmed or etched resistors which have helical groove patterns, doesn't L-sub-R vary, possibly substantially, depending on how much of a helical pattern they have cut into them to get the specific ohm value?  That's the consistent caution that I've read about using metal film/metal oxide/carbon film resistors in RF circuits, and why I'm interested in quantifying their performance versus carbon comps of the same values.

I understand that one has to look at the schematic and understand what function that specific resistor is doing. Let's say that a 4.7k carbon comp resistor was used in a feedback circuit or a filter circuit that would see significant signal magnitude at 20MHz.  Wouldn't replacing it with that Vishay PRO2 with that much inductance likey upset the circuit?  A resistor with 4.7k ohms of resistance and 80% of that in inductive reactance (3.76k ohms) would also introduce significant phase delay, right?

[T3sl4co1l]

I would wonder firstly whether your measurement is accurate. Very little phase error is necessary to read inductance that far out of whack.  Repeat the measurement at other frequencies: does it follow consistently, or does it vary?  Is the instrument properly calibrated?  If there is an averaging option, might it prove useful?

Tim

[mawyatt]

Don't know how much a spiral cut resistor body inductance increases, but with a 4.7K Resistor Real part, the body Inductive Imaginary part should be significantly smaller at even 20MHz and usually be ignored. 37uH to equal the 4.7K impedance @ 20MHz is not realistic in any practical resistor, about 1000 times too large!!

However the body capacitive effects shouldn't be ignored, as a few 100s fF wouldn't be surprising!!

Best,

[mag_therm]

I am sorry I made a mistake interpreting my Vishay PR02 resistor sweep data in post #11
The crossover from inductive to capacitive when mounted near the copper baseplate occurs at nominal value below 470 Ohm (not 4700 Ohm)
I have struck out and corrected post 11 and added scrots of the nanoVNAsaver S11 sweeps for the Vishay PR02 4700 Ohm and for a Multicomp 1206 4700 Ohm and for a 50 Ohm Cal

[joeqsmith]

As pointed out already, a fair amount of the inductance is determined by lead length.  Based on your question, I have no idea if you are trying to remove the effect of the leads. 

While not really what you are asking,  this video attempts to demonstrate using a VNA to measure the impedance of a simple network using through hole parts and a breadboard.  At 35 minutes, I show a simple RLC network sweeping from 10k to 50M.   As I work through the problem, I eventually shorten the leads (37 min) and you can see what effect this has.   Then I change to surface mount (38 min).   

Depending on what you are doing, construction can be every bit as important as the component selection. 

[MisterHeadache]

As pointed out already, a fair amount of the inductance is determined by lead length.  Based on your question, I have no idea if you are trying to remove the effect of the leads. 

The goal of my effort here is to determine using data if I can or cannot substitute modern construction through-hole resistors for original carbon composition resistors in vintage gear that operates at RF frequencies.  Secondary constraints are defining the differences between constructions (metal film, carbon film, metal oxide), and upper limit of useable frequencies (my equipment would be primarly HF, so 30MHz and below, but up to 100MHz would be better.  Certainly don't care about GHz operation).

Since the applications must use leaded resistors, the natural inductance of the leads is expected, regardless of resistor construction.  Knowing just how much 'worse' a modern resistor is as compared to a carbon comp would suffice.

I am aware that ceramic resistors are available and have naturally low inductance, but paying $2 apiece gets painfully expensive in a hurry, and would be foolish to spend if there was a $0.10 alternative that would work fine.  As would something like - "the series XYZ are OK to substitute but only in values above 1k because below that they are too inductive."

Lastly, I emphasize chasing DATA over LORE - there's no shortage of 'OMG, don't ever replace your carbon comps with anything else!' LORE on the internet.  If carbon comps were still readily available, there'd ne no problem.  But AB stopped making them a long time ago, and Stackpole stopped a couple years ago, and only niche ($$$) manufacturers are left.

(come to think of it, wasn't chasing DATA over LORE a STTNG episode??

[mag_therm]

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.
Even if you don't use them for restoration, you could get a few of samples to compare on nanoVNA.
About 10 years ago in a 1959 FM receiver I replaced nearly all of the carbon with smaller 1/4 Watt metal film, keeping the leads about the same.
It re-aligned OK and is still running. L.O is 10.7 MHz above 88 ~108 MHz. At that time a 120V isolation transformer was added and the whole tube string was replaced with 6.3V equivalents.

[bson]

I would imagine for any resistor other than wire wound, the inductance is going to be dominated by the loop area it forms with the return path underneath.  If it has a circuitous return path or is part of a longer signal path the resistor itself probably makes little to no difference.

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

BTW, a wire wound resistor, like any coil, also has a capacitance in parallel with the inductance; this is the inter-winding capacitive coupling between adjacent turns.  This should be measured for a ww resistor as much as the inductance.

[joeqsmith]

I've bought from surplussales as well and have had no problems with them. 

The goal of my effort here is to determine using data if I can or cannot substitute modern construction through-hole resistors for original carbon composition resistors in vintage gear that operates at RF frequencies.  Secondary constraints are defining the differences between constructions (metal film, carbon film, metal oxide), and upper limit of useable frequencies (my equipment would be primarly HF, so 30MHz and below, but up to 100MHz would be better.  Certainly don't care about GHz operation).

Not many 1Meg parts will be stable from DC to 100MHz but I understand what you are after.   

Looking at Digikey, they appear to have some stock of CC parts.   You could stock up on the common parts while buying a range of parts for your experiment.   

If I had a stock pile of old CC and CF parts, I would attempt to run something.   I'm in the same boat and would need to order a matrix of parts to run it.   Looking forward to seeing what you come up with. 

[joeqsmith]

Using the LiteVNA64 to measure S21 for two sets of resistors.  One CC, the other CF.  Both sets are 1/2W.  From 1-100MHz, outside of the CC's DCR being off, their AC performance is similar.  Leads were left long and were measured at their ends.  No attempt was made to compensate for the difference.  CC measure roughly 87.3mm and CF 60.0mm, end to end.   Note the CF's spiral pattern.   The DCR for the 390 ohm CF measured 393 while the CC measured 400.  For the 750, the CF measured 745 and the CC 785.