Adjustable Capacitance decade box 10p - 1u

[Diabolical Artificer]

How would you go about building such a box with a resolution of say 10p? I've tried working it out but got in a mess, also looked online for possible schematics.

Any idea's, Andy.

[Restoration73]

The main problem will be obtaining the rotary switches with adding wafers. For the lowest
ranges you may be better off using adjustable airspaced trimmers as there will be some
stray capacitance caused by the wiring. Although I have a commercial box by Jay-Jay
I use more a homebrew with 1 pole 22-way rotary switch with various preferred values.

[Studio263]

You can do each decade with only a few capacitors:

1, 2, 2, 5.

For example the 100nF decade:

100nF: just the 100nF one.
200nF: just one of the 200nF ones.
300nF: the 100nF one and one of the 200nF ones in parallel.
400nF: both of the 200nF ones in parallel.
500uF: just the 500nF one.
500nF+: the 500nF one and whatever you have to add from the above list.
1uF: all of the above in parallel.

All you need to buy is two values per decade, and one of those will be shared with the decade above (or below). Using the above example:

100nF makes 100nF (obviously), two in parallel makes 200nF.
Two 1uF in series makes 500nF

So just a few easy to buy values for each decade you want, plus four simple toggle switches to select them (you could do it with some simple digital logic and reed relays if you wanted to be really flash).

[AC/HL]

Something along these lines perhaps: https://www.ebay.co.uk/itm/Low-Cost-...OLN7:rk:1:pf:0

[Skywave]

Difficult to see how a minimum of 10 pF can be obtained. Any constructed & wired arrangement of capacitors will have a wiring capacitance in excess of that. However, here is an idea in principle.


Use several single-pole, 10-way rotary switches; a range of capacitors and one variable capacitor: 500 pF + 500 pF ganged & wired as 1000 pF. Its minimum capacitance will be approx. 20 pF. Add a scale + pointer and calibrate it in units of pF: 20 pF → 1,000 pF.

1. The selectors of all switches are joined and constitute terminal A. The 'ways' on each switch have one 'no contact'.
2. The first switch has capacitors in the range 1nF→ 9 nF. One lead from each cap. to each switch 'way', except the 'no contact' 'way'. The other leads are joined and taken to terminal B.
3. The next switch: 10 nF → 90 nF. Repeat capacitor connections as above.
4. Repeat until last switch has 1 uF as largest value.
5. Wire the (calibrated) variable capacitor from terminal A to terminal B.
6. Mark-up the scales for the knobs attached to each switch to show the capacitance thus selected.

The capacitance from terminal A to terminal B gives a range of capacitance from 20 pF to 1 uF. The accuracy will depend on the accuracy of the calibration of the variable capacitor and the accuracy of the fixed capacitors. Access to a calibrated capacitance meter will be very useful. The resolution will be a function of the variable capacitor. Its '20 pF' minimum may also be larger, depending on the total wiring capacitance.

In summary, all of that is similar to what constitutes a switched-resistance resistor box, but whereas Rs combine in series, capacitor combine in parallel - as we all know.

Easier to explain with a drawing, rather than words, but at present, my scanner is not working, so I cannot scan-in a hand-written sketch & add it as an attachment to this post. I hope you can follow the above: it does take a bit of digesting. I've reviewed it and edited it several times.

Al.

[julie_m]

If you can get BCD switches with the right action (made = 1) then you just need to arrange your capacitors in multiples of 1, 2, 4 and 8.

[Skywave]

Originally Posted by julie_m

If you can get BCD switches with the right action (made = 1) then you just need to arrange your capacitors in multiples of 1, 2, 4 and 8.

Fortunately, I've been 'doing electronics' for enough years to have accumulated a fair amount of salvaged (and once-they-were-new) components. And that includes a number of single-pole, 10-way rotary switches. I expect that some others here are in the same situation. That thought was contributory to my above post.

Alternatively, with only four capacitors in the range 1,2,2,5 (nF or uF) and a 4-pole, 11-way rotary switch, the entire range from 0 to 10 can be arranged for selection at will. That idea is described in detail in the Radio & Electronics Laboratory Handbook, 8th. ed., p. 208, by M.G. Scroggie. An analysis reveals that the switching arrangements give the sequence 0, 1, 2, 1+2, 2+2, 5, 1+5, 2+5, 1+2+5, 2+2+5 and 1+2+2+5.

Al.

[Diabolical Artificer]

Thanks all. I have shed loads of wafer switches (switch's/switchers/switches?) and it doesn't need to be too accurate, am using it to select the correct C in a NFB R/C. Still a bit confused about the switching, which always bends my mind, I'll do some sketches and report back.

That kit looks ok Bill, I could sub switches for the jumpers.

Andy.

[Radio Wrangler]

Seventeen miniature toggle switches and a binary series of capacitors?

Any circuit needing a 10pF capacitor is likely to object to the stray L and C of the connecting leads as well as the strays of the substitution box.

Tolerance and stability of larger values will make a joke of the theoretical resolution.

Substitution boxes are a lot less useful than they appear. I can't remember using one since my university days.

David

[ionburn]

Substitution boxes are a lot less useful than they appear. I can't remember using one since my university days.

I can see the issues with a capacitance box and making one. Low values will be all over the place with the connection lead capacity effects, and high value tolerance unless expensive non- electrolytics are used (there would be issues with polarisation and bias voltage / leakage with these anyway).

That said, although I do not use a capacitance box. I find that absolute values of capacitor are seldom required, except for very low values (below 500pF with <10pF resolution), at which point strays get problematic anyway. I use switched resistance boxes all the time for valve circuit testing / optimisation. My valve test setup is distributed rather than one box, which enables easy setup of specific test circuitry.

[Argus25]

Probably everyone would know about this one, very popular in the Antipodes, easy to operate thumb wheel switches. It started out a as Silicon Chip kit and is supplied in full form by Altronics, 100pF to 9.99 uF, it is the bees knees. Try to encourage somebody to buy you one as a Christmas present, or maybe Easter

https://www.altronics.com.au/p/k7520...ecade-box-kit/

...notice the "lovely" K2575 ESR meter advertised on the same page makes a great stocking stuffer.

[Diabolical Artificer]

Maybe a 10 puff resolution is a bit silly and unobtainable with any precision. Thinking about the intended application it may be better to just bung in the standard line of silver mica caps, E12? So maybe 100, 150, 220, 330, 470, 560, 680 puff and 1n etc. The lower decade might not be needed at all. I'll have a look at my stash and see what I have.

Nice box Argus and probably not too dear what with the Oz $ being funny money.

Andy.

[Argus25]

Nice box Argus and probably not too dear what with the Oz $ being funny money.

The internal design & construction of that unit is well done too, I remember the article that came out on it, it is all very orderly inside and well thought out.

[Skywave]

Ref. post #5.
Attached to this post is my idea of a switched-value capacitor.
A few explanatory comments about it.

S1 and S2 are s.p , 10-way rotary switches.
VC = a 1000 pF variable capacitor.
C = the resultant capacitance.

Illustrative example.
As drawn, C = 44 nF if VC is set for zero capacitance. (see Note below).
The next smallest switched step for C is with S1 at 5 nF and S2 at remaining at 40 nF. Then C = 45 nF.
If VC is calibrated in units of capacitance 0 → 1,000 pF (1,000 pF = 1 nF), then each step in C can be smaller than 1 nF.
For example, if VC is calibrated in units of 100 pF, each step in C will be 100 pF = 0.1 nF.
Therefore, in the above example, with S1 at 4 nF and S2 at 40 nF, as VC is varied over its range, C = 44 nF → 45 nF
For VC in steps of 100 pF (0.1 nF), result is C = 44 nF; 44.1 nF; 44.2 nF . . . . . 44.9 nF; 45 nF.

Note: The reference to VC at zero capacitance is theoretical, since it will have a minimum value, typically 5 to 10 pF.
VC = 1,000 pF can be obtained by wiring a dual-gang 500 pF + 500 pF variable capacitor and joining both gangs.
With the scheme as drawn, max. value of C will be 100 nF = 0.1 uF; min. value of C = VC min. - typically 10 pF
► Therefore, range of C is from < 100 pF to 0.1 uF; resolution = 100 pF (or less - depending on VC calibration)

This switching scheme can obviously be extended to another decade by adding another s.p, 10-way, rotary switch.
The switched capacitors for that additional range would be, in uF:
0.1, 0.2, 0.3, 0.4, 0.5, 0.6. 0.7, 0.8. 0.9
The 'resolution' of 100 pF would remain, so a typical value of C would be: 300 nF + 60 nF + 2 nF + 300 pF = 362.3 nF
Subsequent 'next values' would be (nF): 362.4; 362.5, 362.6 . . . . 363.0 (VC at 'zero' for that last value)

Note 1: The capacitance values stated are for illustration only. In practice, preferred values could be used, depending on the unit's required purpose.
Note 2: Overall, this idea is the same as we would find in a switched resistance box which included a calibrated variable resistor. The essential difference with this idea is that unlike resistors, capacitors are 'added' by connecting them in parallel.

Al.

[m0cemdave]

If you wire up some banks of 10 or 11-way switches to provide a BCD sequence, you can make a substitution box using standard values.

1
1 + 1 in parallel = 2
3.3 + 0.68 in parallel = 3.98 (close enough to 4)
3.3 + 4.7 in parallel = 8

Repeat for each decade required.

About 30 years ago I built a box like that for prototyping loudspeaker crossovers. I think it went from 0.1 to 100 uF. Not sure if I still have a circuit diagram but it's standard stuff, must be in a textbook somewhere.

[Diabolical Artificer]

Thanks Al for that but Jeffrey over at Goldbourne has offered me a capacitance decade box cheap. I was rather looking forward to building this too and using some of the big multiway switches I have.

Thanks to all too who contributed, Andy.