Best capacitors for decade cap box??

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

hymentoptera

Well-known member
Joined
Jan 5, 2013
Messages
298
Location
USA
I'm sourcing parts for a custom decade capacitance box.. I have rotary switches, banana jacks, etc, on the way, and it's time to start buying lots of precision capacitors so I can make this project happen.

I wanted to ask you fine folks if there is any particular types of capacitors I should be looking into (or avoiding?).

I need better than 5%, so my options for larger values are slim I think, and will require me to buy large quantities and hand select the ones closest to nominal value. I'd like at least 1% precision across the entire band, except for at the picofarad range where I'll likely trim them out with variable capacitors I think.

I'm thinking polystyrene for most of the nanofarad range, as they can be found in low tolerances, but most of them aren't rated for more than 63v (I'd like 100v to be my target minimum voltage rating here, no reason why really), but I'd have to buy much fewer of them as opposed to if I went with big foil types, since I'd have to sort through dozens just to find what I needed in 5% parts.

I've got a metric boat-tonne of micas coming with 1% and 2% tolerances, as I've read that they are pretty temperature stable, and stability is important for what I'm trying to do. Modern MLCC's have TERRIBLE voltage coefficients, and as for temperature, regular NP0/C0G ceramics are hard to find in 1% tolerances. Is there a certain type I should be looking for in any given value range which meet my needs?

I'm willing to order from Mouser, DigiKey, etc, or even ebay sellers if needed. Price is not really a concern. Tolerance and stability (volt and temp) are most important.

One thing to note: If I can scrap together enough capacitors (and if the rack chassis space allows!) I'm going to stack every single switch position with it's own bank of capacitors in parallel, as to avoid the common configuration in commerical decade cap boxes where only the E12 values are available, i.e. 10pF, 12pf, 15pf, 18pf, 22pf, 27pf, etc... instead I want a TRUE decade box, even if it's means 100+ capacitors in a 20lb box of madness with hundreds of solder connections. I realize I'll be limited to around 1uF max like this, but this is what I really want. I know this is crazy, but this is DIY, and this is how I'm going to do it. Many of my values (E48 values) are almost available as micas and styrenes, but only go up to a few tens of nanofarads and then become to bulky to consider.

So my question is, what types of capacitors should I be looking for for any given range? And also, what other considerations are there in building a decade cap box that I may have missed?

And one more question; I plan to have a switch on the front which shorts the jacks together through a resistor to drain off any built-up charges. What value resistor should I consider (I'm thinking 10Meg?), and any other advice on this section as well.

So I guess that's four or five questions really...

The planned specs I'm aiming for include:
*19" rack mount decade capacitance box
*1% or better tolerance
*ranges from single digit picofarads (i know this isn't easy!), up to single digit microfarads.
*high stability
*banana jacks and rotary switches, probably in a 19" rack enclosure
*switchable bleeder resistor(s) strapped across banana jacks
*???
 
A precision diddle box sounds like a oxymoron, (yes i have both an old resistor and cap diddle box), but these are crude instruments at best for design.  The cap box has a single 12 pos sw, the Resistor box has a high and low range switch to cover more ground. 

you probably need to stop worrying about precision and single digit pico Farads... Your wiring could easily swamp the lowest cap positions.

I don't follow how a single bleeder resistor could bleed all the caps that aren't selected.

In general make the resistors higher power (low values often release their smoke in use), and the caps higher voltage.  Source good quality film caps when you can.

What do you anticipate using this for?  For tube work you'll need decent high voltage caps.

JR   
 
I have to agree with John here.  Anyone doing DIY has no need for 1% tolerances.  These boxes are for swapping rough values to get an idea of what will work/do what you want.  The 1% boxes are for laboratory work that (usually) has nothing to do with audio.  If you really DO need that, find a used NIST calibrated unit.
 
John, you're right about the single resistor not cutting it, I don't know what I was thinking there. I'll have to come up with something else there.

I've already considered dropping the single pF switch as preliminary tests are giving me 4 or 5pF just off the breadboard I'm playing with..

As for 1%, it's just a target, not a necessity. I don't expect anything unrealistic here, it's just another development tool for the bench.

The need for the box comes from both the irritation of popping caps in and out of the breadboard fooling round with DIY guitar pedal mods. I'm building several resistance boxes (both a rackmount one, then several smaller ones in project boxes) for the same reason, and also because it's just what I want.

Micas and styrenes come in E96 values, and they're cheap and plentiful, besides I already have piles of them in my trays here, so being able to tweak filter cutoffs by selecting different resistor/capacitor values sounds like fun, maybe even a real time saver, and I've always wanted to build these anyways. The need isn't so much important and the journey getting there ;)
 
instead of different values, you could maske a box with different types for the same value,

like a .022 in Wima, Jensen, ERO, Spraque, Vitamin Q ,Teflon, Poly, wax, oil

have you tried these Russian caps that everybody is raving about>

http://www.ebay.com/itm/Pair-1uf-200V-Russian-Vintage-K40Y-9-PIO-Capacitor-Great-Tone-Caps-/191600893293?pt=LH_DefaultDomain_0&hash=item2c9c4d9d6d
 
CJ said:
instead of different values, you could maske a box with different types for the same value,

like a .022 in Wima, Jensen, ERO, Spraque, Vitamin Q ,Teflon, Poly, wax, oil

have you tried these Russian caps that everybody is raving about>

Haha, no. I don't think so. Can't tell if you're being serious. But I have some of those caps around here somewhere. I believe they are .039uF if I remem correctly.

I'm really just wondering what caps will be the most stable over time/temp/volts, and if my choice of micas, styrenes and films is could be improved on.
 
I'd consider ceramic NP0/C0G for anything up to 100 nF. These are quite easily sourced at 1% (at least up to some nF) if that precision is required (I agree with previous post), offer the best temperature stability, are cheap, small, etc...

Samuel
 
SO I've found a ready-made PCB for this, here http://www.siliconchip.com.au/Shop/8 (about halfway down, just Ctrl+F for "capacitance") and I bought both the PCB and the front panel.

Also bought the box this guy used in his 23 April 2014 post
http://tronixstuff.com/category/kit-review/
and 6 rotary switches from Altronics.
http://www.altronics.com.au/p/h0151-ub1-157lx95wx53hmm-grey-abs-jiffy-box/
http://www.altronics.com.au/p/s3021-1-pole-2-12-position-pcb-mount-rotary-switch/

Everything's coming from AU so it should take a few weeks I would think.

Anyways, the point is that looking at the schematic shown in Silicon Chips magazine the solution they came up with for a true decade cap box is much better than what I was planning to do. They're using only 19 caps per decade where they fall something like this... this would, for instance, be for the 100pF decade:

100pF = 100
200pF = 100 + 100
300pF = 150 + 150
400pF = 180 + 220
500pF = 100 + 180 + 220
600pF = 270 + 330
700pF = 100 + 270 + 330
800pF = 330 + 470
900pF = 220 + 680

So you'd need a total of 19 caps per decade. Where multiples are needed of course you'd just run them in parallel.

Parts count looks like this:
5 x 100
2 x 150
2 x 180
3 x 220
2 x 270
3 x 330
1 x 470
1 x 680

I thought I'd post this in case anyone else finds this searching in the future and they're looking to build something similar. I still plan to build my 7 decade rackmount box since I've pretty much got all the parts now, but first I'll build this little 6 decade one and see how useful that is.  :)


 
> using only 19 caps per decade

A dum-dum way would be to buy forty-five 100pFd caps. 1*100, 2*100, 3*100... 9*100.

More caps but the cost-per-cap falls WAY off when you buy more than one or two per value.

That 5X 2X 2X 3X.... order will drive the parts-pickers nuts. (And increase the chance of error.) All will be at per-each price, no quantity discount (that usually starts at the 10-pack).

DigiKey, Kemet C317C101G2G5TA
100PF 200V 2%
1 = $0.85
50 = $24.95 ($0.50 each)

I'm too lazy to run your 19-cap series through Digikey; but if 100p to 680p costs $0.85 to $1.67 each, then that 19-cap series might cost OTOO $23.94. Hardly much different from 50*100p for $24.95, and you get 5 spares.
 
PRR said:
> using only 19 caps per decade

A dum-dum way would be to buy forty-five 100pFd caps. 1*100, 2*100, 3*100... 9*100.

...

Very true.

Best part is I'm sure I already have just about everything I need in my bins, except for a few larger values like 6.8uF is one I'm looking for right now. Not too common in film 100v, but there are some on ebay pretty cheap.

The 45 x 100pF would be insane. I was originally going with something in the middle and would have needed about 30 caps per decade, so 19 is still a relief! It never occured to me before how easy it was to come up with just about any value using only the standard preferred numbers.
 
> just about any value using only the standard preferred numbers.

Yes, that was interesting.

> something in the middle

After I posted, I thought: if you could get 200pFd at small tolerance, then five 100p and twenty-one 200p (26 per decade) gets you there. However this misses the 10-pack discount. (Perhaps two 10-packs of 200p and one 10-pack of 100p, make a 200p out of two excess 100p.)

Relevant to my "annoyance" point-- I just ordered a grocery-list of terminals, breakers, and relays. I double-checked it thrice. The order was confirmed. But when I got the order, it was one relay short. (OTOH, last week I came out 27 hose-clamps ahead: instead of 3 clamps they shipped 3 *boxes*.) Long lists of odd quantities always baffle the pickers.
 
Ha it's so true. I love spending hours on (insert name of any major distributor) only to have my order arrive 1 pc short, or several pcs over. It's kinda like playing scratchoffs. :)

I never report it, maybe because i've never been shorted any large dollar value, and besides I've been the guy picking the orders and I know that a record is kept of exactly who screws up what and how much money it "cost" the company. Pick too many wrong locations and you'll be filling out applications instead of orders ;)
 
I just recieved a WWII decade box full of mica caps. It goes upto 1uF. Buying all these caps today would be an insane amount of money. Happy camper here  8)
 
I have a capacitor substitution box in the works. It's to tune a motor run cap value for a synchronous motor. DC link caps were suggested as motor run caps don't come in the small values.
 
related tangent:

Eico decade resistor using only four R per SP11T switch.  0-9 increments per switch.  1/2/3/6 R increments for resistors.  Tricky.
20980563619_37ae7896a5_h.jpg

 
Beautiful! I've bid on a few of those Eico units on the 'bay, but never won one.

I love the random 10% 100r CC thrown in amongst all the .5% precision parts. I'm sure it was carefully selected for match, but it's just great to see. Then again, maybe the factory part went kaboom and that's all the owner had to replace it with!  :D
 
Back
Top