Decade Boxes

Bricksnbeatles

Well-known member
Anyone else make decade boxes for variable parts substitution?

I’ve got parts on the way to make a 7-decade 2W resistance substitution box, and I’m still deciding on what max voltage rating and what decades I should do for my capacitance substitution box. I could do up to 7 decades for that as well— just need to figure if it makes more sense to do x1pF-x1μF (total range of 1pF-10μF) or x10pF-x10μF (total range of 10pF-100μF). I’m thinking the former since the upper μF range is typically just used for power filtering and other applications where you won’t really have a reason to test different values, and also for size reasons since those caps get pretty large and expensive when you increase the voltage rating.

I’m guessing I’ll have to fold my own enclosures for these things— time to build a hand brake?

Anyway, any of you build your own substitution boxes? Let’s see em! Or do you have a cool old one that you didn’t build? Let’s see those too!
 
  • Like
Reactions: fig

fig

Village Idiot
On the bucket list. Never thought I'd feel a need for one until recently. I've not used LTSpice to the point of running a sim and understanding the results enough to translate them into meaningful data. A recent project has me auditioning as many value samples as I have on-hand....so as is the economy of things......

Please post my instructions...er....I mean your progress. ;)
 

Bricksnbeatles

Well-known member
Will do! For the resistance box I’ve got a
On the bucket list. Never thought I'd feel a need for one until recently. I've not used LTSpice to the point of running a sim and understanding the results enough to translate them into meaningful data. A recent project has me auditioning as many value samples as I have on-hand....so as is the economy of things......

Please post my instructions...er....I mean your progress. ;)
Will do! The resistance box uses 1p12t rotary switches (each decade goes from 0-10, and then back to 0 for a quick reset) and the capacitance one uses 2p11t switches for 0-10 on each decade
 
  • Like
Reactions: fig

finebyfine

Well-known member
I like your capacitance range and would start there myself if you had just generally asked me before reading this because of your reasoning exactly with higher values being more associated with power filtering at least around here.

Seeing Parasit Studio's RC Box got me thinking about stuff like this recently if not exactly decade boxes per se.

s787426405301510_p55_i2_w1005.jpeg


One thing that starts to bother me when breadboarding is bench sprawl, and as tempted as I am to build a decade box or something like this RC box I know I'd have to get it WAY smaller to not feel like it was getting in the way. There's a really great documentation of a resistor box here: https://web.archive.org/web/2016061...nts.com/diy-resistor-substitution-decade-box/

If I was gonna do an RC box or resistance decade box - because my first goal would be "small enough to be something I'd use" over exact precision I'd personally start with smd resistors or with a wad of digital potentiometers and a microcontroller. Most digipots have a pretty wide tolerance of ±20% but there are some lower value ones ±1%, and theoretically tuning them might be reasonable. Plus then you throw the power supply limits from them and it might be more trouble than it's worth to go that route.
 
Last edited:

Stickman393

Well-known member
Something I've been considering for a little while here...

I've got a bunch of these little things:

51qIfonjquL._AC_SY355_.jpg

Single pole, 10 throw each.

Basically, buy a bunch of 1M, 100k, 10k, 1k, 100, 10, and (if you're a masochist) 1 ohm resistors. Would probably want to use 0.1% tolerance for something like this, cause the additive tolerance deviations could knock you wayyyyyyy off course.

Basic idea is to wire 9 of each decade in series, and using each corresponding numerical terminal as a shunt to that position in the series. For the 10k decade: 0=wire directly to 100k decade, 1=one 10k before the 100k decade, 2=two 10k's in series, etc.

Simple, intuitive...the idea appeals to me. But...one also needs to find a suitable enclosure and mill out a square hole...

The latter is easy enough for me...the former not so much. Eh.
 

Bricksnbeatles

Well-known member
Something I've been considering for a little while here...

I've got a bunch of these little things:

View attachment 19420

Single pole, 10 throw each.

Basically, buy a bunch of 1M, 100k, 10k, 1k, 100, 10, and (if you're a masochist) 1 ohm resistors. Would probably want to use 0.1% tolerance for something like this, cause the additive tolerance deviations could knock you wayyyyyyy off course.

Basic idea is to wire 9 of each decade in series, and using each corresponding numerical terminal as a shunt to that position in the series. For the 10k decade: 0=wire directly to 100k decade, 1=one 10k before the 100k decade, 2=two 10k's in series, etc.

Simple, intuitive...the idea appeals to me. But...one also needs to find a suitable enclosure and mill out a square hole...

The latter is easy enough for me...the former not so much. Eh.
Go for it! I’ve seen people make homemade decade boxes with those switches and they came out awesome! I just went with 1% on the resistors since I couldn’t find cheap 0.1% resistors that were higher than a 1/2W rating, so I’ll just test and sort the resistors to make sure they’re all close to spec (and I’ll also alternate resistors that are above spec or below spec so there’s not to much of a deviation from spec over the course of a decade).

I’ll have like 6 or 7 more of these giant industrial-type rotary switches after I’m done with the resistance box, so I thought of doing an inductor decade box, but that would be too expensive and of fairly limited use.
 

Stickman393

Well-known member
It's not really the deviation over a decade that I worry about so much as the additive deviation from adjacent decades...

After all, say you have two 100 ohm resistors with a 1% tolerance...and they both measure at 101, the upper limit.

Well, if you put those two in series they add up to 202. Twice the difference of each (individually speaking)...but as a percentage, they're still within 1 percent. Schweet. But...let's say you want to make a 680k resistance...

8 10ks that measure out as 8080, 6 100ks that measure as 60600...now you've got an additive resistance of 68680 ohms. Which is STILL within 1%. Huh.

Well...what about a 5.6 Megaohm?!?!?!? You could potentially measure as high as 5,656,000 ohms. STILL within 1%.

...my point...ah yes...my point is that my brain forgot how percentages work for a brief moment. Yeah, 1%'s are probably just fine for 95% of stuff...but do you realize how big that five percent is? It could be like...five thousand. Out of one million.

Eh...I should go to sleep...
 

Bricksnbeatles

Well-known member
It's not really the deviation over a decade that I worry about so much as the additive deviation from adjacent decades...

After all, say you have two 100 ohm resistors with a 1% tolerance...and they both measure at 101, the upper limit.

Well, if you put those two in series they add up to 202. Twice the difference of each (individually speaking)...but as a percentage, they're still within 1 percent. Schweet. But...let's say you want to make a 680k resistance...

8 10ks that measure out as 8080, 6 100ks that measure as 60600...now you've got an additive resistance of 68680 ohms. Which is STILL within 1%. Huh.

Well...what about a 5.6 Megaohm?!?!?!? You could potentially measure as high as 5,656,000 ohms. STILL within 1%.

...my point...ah yes...my point is that my brain forgot how percentages work for a brief moment. Yeah, 1%'s are probably just fine for 95% of stuff...but do you realize how big that five percent is? It could be like...five thousand. Out of one million.

Eh...I should go to sleep...
Ah yeah I see where your head was haha

Didn’t realize that’s what you were thinking— I was just thinking along the lines of: if the first resistor in a decade is 1% higher, and then every resistor after is 100% to true spec, then the first position is 1% out of spec, the second is .5% out of spec, the third is .33% out of spec, fourth is .25% out spec, fifth is .2% out of spec etc until you reach the tenth and final, which is .1% out of spec; but if the last is 1% off spec and all the prior resistors are true to spec, then you’ve got 0% deviation in all positions of the switch except the last, which only is .1% out of spec, so it’s better to use the closest-to-spec resistors first in the chain, and the furthest from spec last, but since 0% tolerances are pretty much impossible, it’s best to alternate +/- tolerances:

Say:
R(1) + R(2) + R(3) + R(4) + R(5) + …….

100.1 + 99.8 + 100.2 + 99.6 + 100.3 ……

= 100.1; 199.9 ; 300.1; 399.7; 500 …………..

Minuscule and barely worth doing honestly, but it’ll make the average variation from spec much smaller than it already would be, and also reduces the maximum variation from spec so that even a resistor a full 1% from spec never sets the overall resistance read from that decade is even near 1%.
 

JetFixxxer

Well-known member
Something I've been considering for a little while here...

I've got a bunch of these little things:

View attachment 19420

Single pole, 10 throw each.

Basically, buy a bunch of 1M, 100k, 10k, 1k, 100, 10, and (if you're a masochist) 1 ohm resistors. Would probably want to use 0.1% tolerance for something like this, cause the additive tolerance deviations could knock you wayyyyyyy off course.

Basic idea is to wire 9 of each decade in series, and using each corresponding numerical terminal as a shunt to that position in the series. For the 10k decade: 0=wire directly to 100k decade, 1=one 10k before the 100k decade, 2=two 10k's in series, etc.

Simple, intuitive...the idea appeals to me. But...one also needs to find a suitable enclosure and mill out a square hole...

The latter is easy enough for me...the former not so much. Eh.
I tried this one, but realized that I F*** it up after I solder all the resistors on and so now it's been sitting my "junk" box for the past 3 years.
 

Stickman393

Well-known member
@Bricksnbeatles

Yeah, that was me overthinking it a bit and rabbit-holing without checking myself. Seeing as how 99% of what we end up using for pedals falls into the e24 series, implying a maximum tolerance of 5%, 1%s are JUST FINE.

But...all that resolution could cause one to be like 'ohhhh, I can use this to experiment with subbing out e192 resistors!'. In which case...like...come on. You're not using a decade box on a circuit that requires that kind of precision. Come on. Stop it. This is just silly.
 

Bricksnbeatles

Well-known member
@Bricksnbeatles

Yeah, that was me overthinking it a bit and rabbit-holing without checking myself. Seeing as how 99% of what we end up using for pedals falls into the e24 series, implying a maximum tolerance of 5%, 1%s are JUST FINE.

But...all that resolution could cause one to be like 'ohhhh, I can use this to experiment with subbing out e192 resistors!'. In which case...like...come on. You're not using a decade box on a circuit that requires that kind of precision. Come on. Stop it. This is just silly.
Time for .001% tolerance resistors. 🤫
 

fig

Village Idiot
Circling back as I often do....it takes me a little longer to compile ;)
After reading, I decided to pick up a dead cheap solution and see if it fills the need.
For $16, I got 2 7-decade, 1Ω step, 1% tolerance resistance boards.

agHJZQF.jpg


You have to set jumpers, so it's not quite as convenient as turning knobs, but they are cheap, take up very little space, and I believe will work well for most pedal building purposes.

Mine will be used extensively with my protoboard. The other will be headed to NY for a second opinion. ;)
 

finebyfine

Well-known member
@fig I almost bought one of those too before deciding a decade box isn't really something I need! The size looks much better on that than I was expecting and is the exact size I would want one to be to be something I'd actually grab.
 
  • Like
Reactions: fig

fig

Village Idiot
There's also a small capacitance decade box, but overall there seems to be just a handful of values used in pedals and even fewer that might need tweaking. It's simple enough to build one out using those needed values with the preferred package, or just use the TSC.
 

JetFixxxer

Well-known member
Circling back as I often do....it takes me a little longer to compile ;)
After reading, I decided to pick up a dead cheap solution and see if it fills the need.
For $16, I got 2 7-decade, 1Ω step, 1% tolerance resistance boards.

agHJZQF.jpg


You have to set jumpers, so it's not quite as convenient as turning knobs, but they are cheap, take up very little space, and I believe will work well for most pedal building purposes.

Mine will be used extensively with my protoboard. The other will be headed to NY for a second opinion. ;)
I got one of those and I ended up letting the smoke out. I even printed a nice cover for it and that melted a little..... too much power over a fairly long period and I lost :)
 
Top