Relay-controlled A/B switch box, part 2
Author: Dmitry Nizhegorodov (dmitrynizh@hotmail.com). My other projects and articles
|
|
|
1. Intro
This article is a continuation of [1], where a simple yet versatile relay-controlled A/B switch box was described. This part shows some useful additional gizmos greatly enhance versatile of the box, refine some of the basic hookups described in [1] and list several more.
2. Add-ons
A small collection of useful gizmos can greatly enhance versatility of a simple A/B box.
2.1 Attenuators
Frequently, gain stages lack necessary built-in attenuators that become neded to level the gain of A and B. In other cases, the built-in gain attenuators are present but use of then for some reasons is not desired (see "elaborate applications" section). What todo?
One option is to build a pair of attenuators into a box. For simplicity, only one Channel, say A, can be equipped with an attenuators. That was originally done. You can see the holes in the box where a pair of (originally present on the "minolta" chassis that was cannibalized) excellent precision 1Meg potentiometers were located. Very soon those were replaced by 100K audio-taper pots then by 47K pots then by 10K pots and then all were gone. Why? Because different applications demand different range of shunt resistance and no value seems matching all. If you must have built-in attenuators, and tube amps are used, then choose 47K as the best tradeoff.
Instead of using fixed-value built-in attenuators, I use external ones, as shown on the picture. Sometimes a stereo attenuator with two pairs of RCA jacks can be used, In most cases, mono adapters with an RCA jack, a potentiometer, and an RCA plug are the best. These are very versatile, can be hooked to the box or directly to the gain stages. The latter is much preferred. The capacitance in the wire after the potentiometer and before a gain stage must be as small as possible.
I recommend to build an assortment of these, with potentiometers of different values. Most useful is to have a pair of 10K, a pair of 47K, a pair of 100K and maybe a pair of 1M. CD players and SS preamps can feed 10K ones, and these are preferred whenever they can be used, since pretty long wires can be connected afterwards. Tube preamps can feed 47K, but 100K is a safer bet; yet cable capacitance starts showing its ugly head, that is why it is the best to plug the attenuators directly onto tube amps.
Each mono attenuator should be made with an input RCA socket and an output RCA jack. The jack is on short cable. The socket is the input, and the jack is the output. This arrangement will help to always keep the attenuators as close to the receiver as possible, to minimize influence of cables.
2.2 Y Adapters
Y-connectors extend versatility of the A/B switchbox. 1 male to 2 female ones are useful to split outputs. Two-male-one-female are useful for gathering signals (merging) and thus have somewhat limited use, yet occasionally are useful to grab output signals from switched channels (see below). Note that it is a very bad practice to use Y connectors to merge live signals!. Many devices have low output impedance and would not like competting for voltage; even if they would, mergin the outputs almost always results in distortion. Unlike that, most devices have sufficiently low output impedance and can drive several loads. Thus, contrary to what some audiophiles may believe, a mainstream CD player has output impedance in tens or at most hundreds of ohms, and can drive up to 10 tube amplifiers, which input impedance often in tens of kiloohms and more. Which Y connectors to use? The connectors are sold by many stores including Radio Shack, DigiKey, PartsExpress, Percy, etc. Choose according to your liking. As it holds with other connectors, choice of wendor dictates the price more than materials do.
2.3 RCA to Binding post adapters.
It's a gizmo, or several variants of a gizmo, which purpose is to hijack the switching section attached to binding posts and use them to switch RCA cables - up to 4 3-plet groups. See the pictures on the right. These are not sold in stores but can be easily done DIY way. Each gizmo is nothing but an RCA jack soldered onto a banana plug or a spade. The spade version is very useful when heavy RCA cables are involved. In both cases only the signal (center) pin is connected; the ground is not. Ground is provided by the gear to which RCA cables are connected. In very unusual case when it is needed (most often it is not) ground can be provided with alligator clips.
For these gizmos, select RCA sockets with sturdy center posts (shown two teflon sockets with machined metal parts).
2.4 Shunts
Another set of goodies useful when swithing power amps is on the right. These are shunt resistors that can be inserted in parallel to the amplifier outputs. These can be useful to prevent unvanted spikes in transformers when load is disconnected. Note that since all the wiring arrangements presented here never run signal into unloaded amp, these are mainly safety add-ons. Resistance needs be on the order of 10x of load impedance. 100 ohm is recommended. Shown are quite powerful resistors (5 and 10W) but even 2W shunts can be used in many setups. The shunts are inserted in the 3-way binding posts and shown are all 3 variants: resistors terminals are to go into the holes in the posts; spades are titened and bananas are inserted from the top.
3. Advanced Rigs
in this section we alaborate the basic configuration, mostly by adding several adapters and attenuators. This opens up some surprising possibilities.
3.1 Basic applications augmented with volume controls.
Here we reiterate the basic applications discussed earlier with volume controls inserted in low-power signal path. The volume controls are the attenuators described above, and on the wiring diagrams represented as circles with a diagonal arrow inside.
|
|
Only one pair of controls is added; it is inserted into the path that sounds louder - that's the one that needs attenuation. On most pictures, we show that in position A.
The Source/Source setup now allows to compare all kinds of gear - CD plays of various brands, CD plays vs SACD, SACD vs DVDA, LP players.
The value of the potentiometers depends on the sources. For most CD players it can be as low as 10k. Keeping it that low will reducse sensitivity to capacitance of interconnects. Yet preferred positon for the controls - as close to the amp as possible, in our case - directly attached to the A RCA sockets on the box.
CD players or DACs with homebrew or boutique tube stages may require load of higher resistance - up to 100k in some cases. Same is true for some tube phono stages and preams.
On the right is a configuration which performs double switching of RCA cables. Tree more switching RCA pairs are accomplished with the RCA-binding-post adapters described in the previous section. The gain states can be any RCA-to-RCA devices to compare such as preamps. It is neat to use this configuration to compare some preamp with a straight wire (short RCA cable or jumpers), or to compare two pairs of RCA cables.
In the latter case the attenuators can be removed since cables should have equal gain.
Yet there is a variation of interconnect cable comparison setup based on this configuration where the the attenuators are not removed. Instead, attenuators are placed in each channel after the first switch and before the cables. The attenuators are set to mid-turn positions. Then, the cables can be compared with respect to influence of their capacitance on high-frequency response.
.
.
.
A little twist to the "one amp, one pair of loads" configuration is when one wishes to compare a pair of loads connected to different output taps of the same amp. The attenuators are needed to compensate the louder path. Usually, the output taps designed for higher output impedance are louder, since on the picture the attenuators are in the channel A which corresponds to 8 ohm taps, and no attenuators in the 4 ohm path.
Why to connect speakers to different taps or use "wrong" taps?
Often, speakers prefer particular output impedance. Depending on many factors, particular output impedance may be preferable.
.
.
.
.
Next is a configuration allowing to compare loads of different sensitivity. A pair of Y-connectors is used to split the output of the source.
Attenuators are inserted in the path that feeds more sensitive speakers (in other words, those with higher efficiency).
Note that this is a "system" test rather than "component test". It is so because speakers of vastly different sensitivity running at the same SPL stress the amplifier that feeds them very differently.
It is not unusual to find a pair of 81dB/W/m and a pair of 103 dB/W/M speakers in the same room. Comparing them will mean a very serious test to the amp's versatility.
.
.
.
.
Another twist to the above is when one wishes to hear some additional impedance connected between the amp and the load, and compare two variants of that. The setup is the same except only one pair of loads is shown:
Z an be any two-pin device. As long as it a two-pin device, two pairs of that can hang off the amp without causing any problem. Z can a resistor, a cap, a coil, an RL filer or any combination of the above. Z can also be a short-circuit (short wire) in which case it is compared against a Z. it is useful to evaluate notch/shelving filters etc. Z can also be connected on both sides (+ and -).
One specific (and very interesting in terms of consequences) variant of the latter is when Z+ and Z- are speaker wire. You can compare a long cable against very short one, against a reference cable, or compare two sets of cables.
One very common use of this setup is when comparing two pairs of caps that are in series with the tweeters. Do PIOs sound better than mylars? When comparing to pairs of equal-value Zs, you can skip the attenuation and wire the source directly into the amp.
When path A contains a pair of shorts and B contains resistors, evaluation of interaction between the speaker and varying output impedance can be conducted. For starters, it is interesting to add R that is equal to speaker's impedance.
Related hint: when comparing two pairs of amps, and one is tube SET and another is solid state, makes sense to experiment connecting R =~ 1..6ohm in series with the SS amp's outputs.
Next is a variant where all the bananas are in the + path.
This setup offers great convenience and flexibility. The two pairs of bananas are sockets where all sorts of components can be insrted. and because these are isolated on both sides, not only 2-pin components, but those that refer to the ground can be used. This includes various complex filters.
.
.
.
.
.
.
.
.
.
.
One example is a comparison of a straigh connection of a load (which can be tweeter etc) vs connection via a 2nd order CL filter. Various circuits incluing low/highpass filters or zobels can be tested this way.
The attenuators are in the path containing straigh wire.
.
.
.
.
.
.
.
.
.
Next we discuss a variant that is a bit similar to the above. Yet its goal is to compare components inserted in a different part of signal path. Namely, after preamp and before the amp.
Z can be any device, as simple as a 2-pin passive and as complex as any 3-pin box.
We show 12 RCA-to-BP connectors assuming the devices can interact via RCA cables (or be RCA cables). In those cases when Z are passive parts such as wire, caps, etc, only 2 pairs of RCA-to-BP connectors needed.
.
.
.
.
.
Here is an example, where we compare 2 pairs of signal caps. Note as the gain will be strictly the same, no need for Y-connectors and attenuation here.
.
.
.
.
.
.
.
.
.
.
.
As a conclusion, a configuration that requires 2 AB boxes. This permits performing serious comparison of speaker cables. The one mentioned above had limitatin that both cables are connected to the speakers, at one side. This was a cable that has impedance that shunts teh speakers can alter the sound of the other one. The configuration with two AB boxes if free from that since cables are completely disconnect from the amp and from the speakers. You can compare a long cable against very short one, against a reference cable, or compare two sets of cables. On the picture the "A" pair of cables has grey-swamp color, and the "B" pair is grey-yellow.
.
.
.
4. Part three has even more!
There is yet more to say - part three contains additional thoughts, ideas, variants, photos.
5. References
[1] Part One abbox.htm
Author: Dmitry Nizhegorodov (dmitrynizh@hotmail.com). My other projects and articles