Why do some circuits use bizarre resistor values?

[Chris Wilson]

I'll cite an example with the schematic attached. I quite often see circuit schematics for what appear to be fairly simple circuits that cite none preferred or even downright odd resistor values (and cap values, sometimes). Are these from some sort of modelling and it's up to the reader to change them to preferred values based on his / her experience of what's critical, and what's not?

My current project is building a pre amp based on just before OpAmp U7 to the RF Out socket. It will be amplifying a 136kHz 0dBm signal to about 17dBm. I have started to build it (and what a fiddly thing that TLC072A is when you need to mount it on strip board...), and the schematic shows  5.1, 49.9 an 8K87 resistors. Oddly enough I have some 5.1 Ohm jobbies, but 49.9 Ohms? 8K87 Ohms? Do I need to special order them or can I use something based on preferred values here? Thanks!

[Skimask]

http://www.brannonelectronics.com/images/STANDARD%20VALUE.pdf

All the resistor values in that circuit fit the 'standard' 1% chart just fine.

[Matje]

I'll cite an example with the schematic attached. I quite often see circuit schematics for what appear to be fairly simple circuits that cite none preferred or even downright odd resistor values (and cap values, sometimes). Are these from some sort of modelling and it's up to the reader to change them to preferred values based on his / her experience of what's critical, and what's not?

My current project is building a pre amp based on just before OpAmp U7 to the RF Out socket. It will be amplifying a 136kHz 0dBm signal to about 17dBm. I have started to build it (and what a fiddly thing that TLC072A is when you need to mount it on strip board...), and the schematic shows  5.1, 49.9 an 8K87 resistors. Oddly enough I have some 5.1 Ohm jobbies, but 49.9 Ohms? 8K87 Ohms? Do I need to special order them or can I use something based on preferred values here? Thanks!

The values are the way they are for a reason (no, really ;-).

49.9 is used where actually 50 is wanted - but that value doesn't exist in the common E series. You can use a slightly more common value - at the cost of somewhat worse impedance matching, which at higher RF might become a problem.

The 8.87k resistor seems to be used to set a bias voltage, probably to a value maximising linear dynamic range. You can change it if you understand what the consequences will be.

So, as almost always, the answer is: it depends...

[Chris Wilson]

OK, OK So they aren't "bizarre", just not in my Ebay pack of 5% tolerance preferred value 1/4W jobbies   I just put an order into RS for stuff yesterday, and was scratching to reach the £20 free delivery minimum. So now I have to pay carriage on 150 resistors. It hurts, means I won't sleep well tonight I think I shouldn't fiddle with the biasing resistor (8.87K) and as I don't have any I might as well get the 49.9 and 680 ohm ones, too. Thanks, I told `er indoors this was a cheap, relaxing hobby <LOL>.

[Howardlong]

You're gonna hate me, but I keep those values in stock, but only in SMD 0603 and 0402, and for just the kind of thing you have there, i.e., precision filters. However I'm not sure how precise they'll be as the all pass phase shifting filter network is using 1nF and 10nF caps in the RC network with those E96 parts, so I'd have thought that care in the choice of those parts needs to be paid attention to as well including temperature coeffiicent as well as tolerance.

If 0603 or 0402 are OK for your project, I'll mail you some FoC.

[Zero999]

Couldn't you have just made the correct values out of the usual E24 values in series/parallel?

[dannyf]

Do I need to special order them or can I use something based on preferred values here?

It depends. Most of the times, they are not critical, at least critical to a point where it will make a big difference.

In filter applications, it can be critical sometimes. Understanding what the resistors do and then you have a lot more confidence as to if you can deviate from the specified values.

[Chris Wilson]

Thanks for more replies. and for Howard's very kind offer. This is all through hole, so I'll pass on the SMC parts, but that gesture was very good of you.


Just to clarify I am only building and needing the part of the circuit below. Not any filter bits. Cheers

[ludzinc]

If it's just the 49.9 and the 8k87 resistors I'd suggest:

x2 100R in parallel for a 50 ohm output resistor in place of the 49.9 and
A 10k trimpot set to 8K87 (every junk box should have a few 10 turn trimmers in it!).

[T3sl4co1l]

A better question:

Why do some circuits use bizarre resistor values?

Meaning,

I've seen commercial prototypes where, whoever designed it, probably several engineers in piecewise fashion, picked an entire spectrum of different resistors, and I mean just for useless crap like pullups.  3.52k, 5.1k, 10.2k, 14.7k, etc.  Often used only once at a time.

It's such a pain trying to maintain and work on designs like that, because you have to check every part that appears in the BOM, that the MFG/PN are present, correct, and consistent with the schematic value.  And the pain propagates all the way through purchasing and assembly, even just for prototypes, because the BOM is polluted by all these useless single-quantity parts that could all be combined as a single line item of 20 x 10k resistors or whatever.

The worst part is when you try asking one of the "responsible" engineers, and they don't give you any justification at all.  Because they know they have none.  Or worse yet, it's some dumbassed excuse that's not even wrong.    You ask your project manager to change them, but then you'd have to ask the customer, because you're screwing with a bunch of stuff beyond the one part of the circuit you were asked to change.  And obviously, all those little changes will cost more hours on the project, and time is money, right?

Alas, the 90-10 rule is present in all fields.  The average engineer knows little, and fakes along well enough to maintain employment.  The experts are few and far between, and often brought in to clean up messes like this (in "record" time and accuracy, no less).

On the plus side, I have no worries over my continued value...

Tim

[tszaboo]

The worst part is when you try asking one of the "responsible" engineers, and they don't give you any justification at all.  Because they know they have none.  Or worse yet, it's some dumbassed excuse that's not even wrong.    You ask your project manager to change them, but then you'd have to ask the customer, because you're screwing with a bunch of stuff beyond the one part of the circuit you were asked to change.  And obviously, all those little changes will cost more hours on the project, and time is money, right?

The average engineer dont give a fck about these things. It was copied, copied and copied over and over again, meaning the values werent revised for five iteration. It is very rare to see if someone calculates the right value, or the if the dissipation is good enough. I was able to reduce the component type count by 10-15% on an average project just by consolidating the BOM. Because there was 10nF NP0 cap for one thing, 12nF for the for something else, and 10nF X5R yet again.
I've seen 0603 CAN termination resistors. Pretty much something else should have been terminated.

[deephaven]

HP/Agilent/Keysight always seem to use a lot of obscure valued resistors, even in what appear to be non-critical places. I often wondered why.

[Chris Wilson]

Thanks guys, glad I asked, some interesting insights as to how various people design circuits, cheers.

[katzohki]

As others have mentioned, it looks like you have standard values here. In my experience there are two reasons that I have experienced in using non-standard values.

One is timing. In your phasing exciter schematic I can imagine a situation where a resistor needs to be a specific value in order to create a specific timing, generally an RC time constant. The other factor in this is that it is easier to get a specific resistor value than a specific capacitor value. I have a filter design application that I use and generally I set the capacitors to be very standard values and let the resistors be the more specific values.

The other thing I've experienced is people using potentiometers to dial in a circuit. Basically they threw in pots everywhere and twiddled the knobs until the damn thing worked. Personally I really feel that this is not good design engineering and is a sure sign someone doesn't know what they are doing. Don't get me wrong, twiddling pots is OK here and there where you need to get an idea of how something responds, but using them as a method of design is not OK.

[SeanB]

Often you will find strange resistor values because somebody bought a job lot of these for some reason, and then they were used all over for non critical values, then for a new design they continued with the same value, and when the odd value ran out then you are stuck 10 iterations down trying to figure out why this value was used.

It helps to specify non critical resistors as any part within a range, which makes consolidating future orders easier. You might be using a 0.1% high stability resistor as a pull up, just because it is used elsewhere and the cost of the resistor is lower than the extra feeder cost.

[German_EE]

1) Voltage. You may be building a voltage regulator for (say) 13,8V and there's a voltage divider on the output that steps it down to the 1,25V reference. Well, if you use a 100 ohm resistor on one side you will need a 1004 ohm resistor on the other http://diyaudioprojects.com/Technical/Voltage-Regulator/

2) Current. If you're making a meter shunt then you can get some really weird values and they are all on the low side. Right now I'm looking at a 0,018 ohm resistor that I made using some brass wire.

3) Timing. Capacitors come in standard values as well so if you want a strange frequency or period then either the resistor or capacitor has to be a non-standard value. An example, a 1750 tone generator in a ham radio transmitter will need a 555 timer, a 100nF capacitor and a 3,958K resistor. http://www.changpuak.ch/electronics/calc_05.php

4) Precision. Some time ago I repaired an electronic module that went into a sports car. There were no presets on the board because of vibration issues and about six of the resistors were select on test. Well, I needed a 1,52K resistor and the only way to make one was two higher value resistors in parallel.

[SeanB]

Avionics use it as well for analogue and timing, with some resistors that are soldered into turret pins riveted and soldered to the board. Parts list would have about 4 pages of acceptable values of resistors to use in selection on test for those places. I used the quicker option on a lot of just using a pot with 2 wires that you soldered in and adjusted to within limits, then unsolder and measure resistance and use the nearest value resistor.

On the test board I glued a small piece of veroboard, with the small preset pots soldered to it and then joined to the points with solid core wire wrap lead. You could not close the case any more, but as it was painted bright orange, and no pilot would fly with a case that colour in a rack, it was fine.

That unit also had presets that were loaded via the ADC section, using precision metal film resistors that had been selected decades ago in a voltage divider. Odd values, with the 5 digits hand written on the Vishay resistor boxes.

[tszaboo]

3) Timing. Capacitors come in standard values as well so if you want a strange frequency or period then either the resistor or capacitor has to be a non-standard value. An example, a 1750 tone generator in a ham radio transmitter will need a 555 timer, a 100nF capacitor and a 3,958K resistor. http://www.changpuak.ch/electronics/calc_05.php

That is the worst example you could make. If you are using RC for timing, and your Capacitor has +/-5% tolerance (that,i f the temperature is nice, and you dont connect DC on it) there is really no point to use 3.958K instead of a 3.9K (E12 series). Unless you put there a capacitor which is more expensive than a decent timer IC with a crystal, but then again, the 555 has 3% error.

[free_electron]

3) Timing. Capacitors come in standard values as well so if you want a strange frequency or period then either the resistor or capacitor has to be a non-standard value. An example, a 1750 tone generator in a ham radio transmitter will need a 555 timer, a 100nF capacitor and a 3,958K resistor. http://www.changpuak.ch/electronics/calc_05.php

That is the worst example you could make. If you are using RC for timing, and your Capacitor has +/-5% tolerance (that,i f the temperature is nice, and you dont connect DC on it) there is really no point to use 3.958K instead of a 3.9K (E12 series). Unless you put there a capacitor which is more expensive than a decent timer IC with a crystal, but then again, the 555 has 3% error.

plus if you use a ceramic capacitor with BaTi as dielectricum ( which most ceramics use ) the mere fact you soldered it has reset the aging because you heated the crystal above its curie point ( 148 degrees c for BaTi ). so now it'll drift 1% in the first hour , another 1% the first day, another 1% the first week and so on ...  1750 hurts my ass..

[T3sl4co1l]

Yes, very stable capacitors are required as well.

Trimmer pots (or caps) can be used to adjust, in which case a stable resistor is desirable; metal film might be good, but anything tighter than 1%, probably useless.  There are low drift, moderate precision resistors out there for this sort of purpose.

Obviously, you'll be looking at polypropylene, polystyrene (or PPS apparently is good too), or C0G ceramic, type capacitors.  C0G is +/-30ppm, so it's quite good.

X7R is tolerable for simple RC filters and noncritical oscillators, but unusable for high order filters, precision oscillators and so on.

Tim

[Howardlong]

I do admit to having a current high volume design that uses two different 1.2uH 0603 inductors, but there is a reason. One is for a high current bias T that doesn't need to be very accurate, or have high Q; and the other is for an RF filter that needs decent accuracy, reasonable Q but no high current requirement. Typically you won't get a single part in that situation that would satisfy both scenarios.

When you're doing any volume, BOM cost becomes very significant and reducing your line item count is just one of many things to look at.

Amother aspect when doing any volume is to try to design out any alignment. The fewer manual tweaks the better, that will cost you time and money on the line. So trimpots are definitely out.

Anything that needs manual placing is another one.

Regarding the 1750Hz repeater tone generator, these days I'd do it in a PIC. Same applies to DTMF or CTCSS encoders, and decoders for that matter, but the decoders are a lot harder to make work then encoders.

I think HP and Agilent may use obscure values "because they can". I guess their margins can probbaly just about stand it.

[tszaboo]

I think HP and Agilent may use obscure values "because they can". I guess their margins can probbaly just about stand it.

Maybe they have a flat price at the assembly house for resistors? I've seen on older equipment, they were using 0.1% resistors everywhere, even for digital parts.

[SeanB]

Or they figured that stocking 2 or 3 versions of the same part, with only a tolerance variation, was more expensive than buying the highest tolerance in bulk. It reduces inventory cost, saves space and reduces the chance of having an error in the pulling of a reel for use in a line, you only have one value to pick, and that is a big time and error saver, plus they might leave the reel on for a few runs saving the loss of components in the tail and leaders.

[ion]

I've found this little program useful for weird resistance values (second tab):

http://koti.kapsi.fi/jahonen/Electronics/ResOptimizer/

You can even limit it to the values you have on hand and it will give you the parallel/series combinations closest to your target resistance.

[Zero999]

I've found this little program useful for weird resistance values (second tab):

http://koti.kapsi.fi/jahonen/Electronics/ResOptimizer/

You can even limit it to the values you have on hand and it will give you the parallel/series combinations closest to your target resistance.

Why download a program when you can use the following online calculators:

http://jansson.us/resistors.html
http://www.qsl.net/in3otd/parallr.html
http://www.random-science-tools.com/electronics/divider.htm

[Dr. Frank]

In Automotive Electronics, every 1/10 of Cent savings counts for resistors. ..
Therefore, cheap thick film technology, 5%, E 12 is preferred.
Using least possible different values and case sizes also saves money, due to lower storage cost.
Each additional magazine of the pick 'n place machine costs extra money.
Therefore, all of our engineers try to avoid special resistors or capacitors and try to reduce the variety, whenever possible. Design reviews support this cost reduction  process.

Anyhow, in some places, even 1% thin film are still required: At A/D inputs, for sensors for temperature, fuel, pressure, etc.
LCDs often need precise R divider ladders for the different phases.
Also LEDs require 1% resistors for exact initial luminosity because up to 20..30 LEDs in a cluster instrument need to give a totally even illumination. The LEDs themselves come sorted for equal forward voltage, and additionally for equal lumonisity classes, therefore each of these LED selections require one special resistor value, 1%.

Therefore, these "odd" components are sometimes required, even in such extremely low cost apps.
Frank

[zapta]

All those logistics issues are the main reason I switched early in my career from mostly hardware to mostly software.

[tszaboo]

In Automotive Electronics, every 1/10 of Cent savings counts for resistors. ..
Therefore, cheap thick film technology, 5%, E 12 is preferred.
Using least possible different values and case sizes also saves money, due to lower storage cost.
Each additional magazine of the pick 'n place machine costs extra money.
Therefore, all of our engineers try to avoid special resistors or capacitors and try to reduce the variety, whenever possible. Design reviews support this cost reduction  process.

The automotive was always a different industry with it's own weirdness. They sold 30.000.000 VW Golf, so every penny saving has a big effect of course.
I have an Agilent PSU nearby, I've checked the serial number, it was *000379. S oyou see, the number of units sold is ~5 magnitude less, making engineering cost 100000 times more. While an automotive electronics engineer (poor bastards) can spend a month to determine the correct resistor tolerance (I know for a fact this happens), if the guy at agilent spends 100.000 times less time on it, he only has 1 second. I think this could be a very good reason.
SeanB is probably on the right track. I usually go for the 1%, 100ppm resistor for common applications, maybe I should reach for the 0.1% instead? Digikey sais it is "only" 20 times more expensive.

[Zero999]

1) Voltage. You may be building a voltage regulator for (say) 13,8V and there's a voltage divider on the output that steps it down to the 1,25V reference. Well, if you use a 100 ohm resistor on one side you will need a 1004 ohm resistor on the other http://diyaudioprojects.com/Technical/Voltage-Regulator/

It's not worth going for 1004 Ohms. The LM317A has a tolerance of 1% (note that's the A suffix, for the plain one it's 4%) and 100R and 1k gives 13.75V which is within 0.3% of 13.8V.