Hi Folks,
I needed a switch for upcoming A/B double-blind testing in the coming weeks (more about this in an upcoming topic - stay tuned!...).
While such simple switches are easily available online, I couldn't find any that could switch sources and outputs at the same time.
Being a big fan of reusing/modifying stuff as opposed to buying new, I had the idea to repurpose an old LPT Printer switch, which allows to switch between 4 printers connected to a single PC. I saved that unit from the scrap (actually, I took it out of the trash) in Montreal about 15 years ago, knowing I would need it one day. Well, the day has come!
Here are a few pictures of the modified box. I added a 4mm banana-connector on the front to connect the metal enclosure to earth if needed. Earth and signal ground are or course isolated from each other.
It turned out to be a 4-way A/B/C/D switch instead of only 2 positions...even better! Now I can tests 4 devices at the same time!
For each of the 4 positions, the COM jack on the front is connected to one of the jacks on the back, and simultaneously the COM RCA is connected to the corresponding A/B/C/D RCA's as well. But the jacks and RCA's are completely independent. Therefore, I can use this box either as a jack switch only, as an RCA switch only, or both at the same time, which is perfect e.g. for testing different headphone amps by switching the signal between the source and the amp's input via the RCA's, and the outputs of the amp via the jacks (the headphone would be connected to the COM jack on the front).
The switch itself is of very, very high quality (in fact, I use the same ones on DIY preamps). It has silver-plated contacts, nice solid feel, extremely durable...). the switch alone costs way more than 30 EUR if you want to buy it new.
I wanted to share this, I hope some of you will find it useful!
Cheers
Denis
I needed a switch for upcoming A/B double-blind testing in the coming weeks (more about this in an upcoming topic - stay tuned!...).
While such simple switches are easily available online, I couldn't find any that could switch sources and outputs at the same time.
Being a big fan of reusing/modifying stuff as opposed to buying new, I had the idea to repurpose an old LPT Printer switch, which allows to switch between 4 printers connected to a single PC. I saved that unit from the scrap (actually, I took it out of the trash) in Montreal about 15 years ago, knowing I would need it one day. Well, the day has come!
Here are a few pictures of the modified box. I added a 4mm banana-connector on the front to connect the metal enclosure to earth if needed. Earth and signal ground are or course isolated from each other.
It turned out to be a 4-way A/B/C/D switch instead of only 2 positions...even better! Now I can tests 4 devices at the same time!
For each of the 4 positions, the COM jack on the front is connected to one of the jacks on the back, and simultaneously the COM RCA is connected to the corresponding A/B/C/D RCA's as well. But the jacks and RCA's are completely independent. Therefore, I can use this box either as a jack switch only, as an RCA switch only, or both at the same time, which is perfect e.g. for testing different headphone amps by switching the signal between the source and the amp's input via the RCA's, and the outputs of the amp via the jacks (the headphone would be connected to the COM jack on the front).
The switch itself is of very, very high quality (in fact, I use the same ones on DIY preamps). It has silver-plated contacts, nice solid feel, extremely durable...). the switch alone costs way more than 30 EUR if you want to buy it new.
I wanted to share this, I hope some of you will find it useful!
Cheers
Denis
This reminds me of a switch I did using guitar pedal foot button, it could swap both RCAs and Jack, here used to compare two headphones amplifiers (Whammy and JDS Atom)
Oh that's a nice idea too to use 3PDT stompbox switches!👍
Have you posted the results of your comparison in the Whammy thread?
Have you posted the results of your comparison in the Whammy thread?
This switch box is good example of the type of things people do wrong in amateur blind testing efforts.
The switch contacts are silver? How much current will be switched? Is the switch for line level signals? If so, then you should be using gold contacts. Silver oxidizes and needs a minimum current density to break down the initial high resistance.
Also, the pics show the RCA connectors are wired to produce a large number of ground loops. The ground loops probably wouldn't be so many in normal operation of the system. Of course ground loops don't just carry line frequency noise. They can carry EMI/RFI which can then cause veiling of the sound and obscure small differences.
However, if you switch grounds with gold switch contacts, the contacts will likely pit before too long due to arcing. It means you need a set of silver contacts to close across the gold contacts while the gold contacts are being switched (again, only needed for switched grounds).
Other than the switch box, there are probably a lot of other mistakes being made in amateur blind testing projects.
Last edited:
The contacts are certainly not pure silver, it is a marketing trick, as is for gold-plated contacts. Pure Ag20 is a p-type semiconductor and I bet no manufacturer would use that as a pure contact material (moreover, it's brittle).
I guess (just a guess) they have some built-in contact self-cleaning feature (I've seen similar switches that have some sort of mini-brush that wipes the contacts clean). These switches were designed to cope with a lot of switching cycles
The contacts are for line-level signals and headphone out indeed, so I'm not stressing the contacts at all (a few thens mA's tops and less than 5V). Furthermore, as this was initially meant for LPT connectors, it has a lot of independent 4-position switches stacked on top of each other. Since the time I posted the picture, I paralleled several switches to even lower the contact resistance.
Good point regarding ground loops. In my case however, they are not much of an issue. I use this switch as the signal ground reference (banana connector in the front), and the HPA's I'm comparing have separate signal ground/earth.
I guess (just a guess) they have some built-in contact self-cleaning feature (I've seen similar switches that have some sort of mini-brush that wipes the contacts clean). These switches were designed to cope with a lot of switching cycles
The contacts are for line-level signals and headphone out indeed, so I'm not stressing the contacts at all (a few thens mA's tops and less than 5V). Furthermore, as this was initially meant for LPT connectors, it has a lot of independent 4-position switches stacked on top of each other. Since the time I posted the picture, I paralleled several switches to even lower the contact resistance.
Good point regarding ground loops. In my case however, they are not much of an issue. I use this switch as the signal ground reference (banana connector in the front), and the HPA's I'm comparing have separate signal ground/earth.
This is exactly why I'm doing double-blind testing instead of biased amateur tests. Even if this switch added some distortion/noise/etc, it would be added to all systems in the comparison (A/B/C/D), hence it would still be possible to compare these devices.
Gold over silver is probably the most common contact material for small-signal relays. A brief summary by another forum member:
https://www.diyaudio.com/community/threads/audio-switching-relay-recommendations.348883/post-7336353
--------------------------------------------------------
One small signal relay datasheet shows the contacts use gold-alloy material:
https://components.omron.com/us-en/sites/components.omron.com.us/files/datasheet_pdf/K106-E1.pdf
-------------------------------------------------------------
Also, gold plating is used for contact material precision audio attenuators:
https://www.goldpt.com/info.html
"All current models employ thick, hard-gold plated switch contacts"
------------------------------------
OTOH, sliver oxide is used on power relays because of its tolerance to arcing. However, a minimum current density is needed to breakdown resistivity of the surface layer.
https://www.diyaudio.com/community/threads/audio-switching-relay-recommendations.348883/post-7336353
--------------------------------------------------------
One small signal relay datasheet shows the contacts use gold-alloy material:
https://components.omron.com/us-en/sites/components.omron.com.us/files/datasheet_pdf/K106-E1.pdf
-------------------------------------------------------------
Also, gold plating is used for contact material precision audio attenuators:
https://www.goldpt.com/info.html
"All current models employ thick, hard-gold plated switch contacts"
------------------------------------
OTOH, sliver oxide is used on power relays because of its tolerance to arcing. However, a minimum current density is needed to breakdown resistivity of the surface layer.
Last edited:
My subjective impressions are scattered on Diyaudio thread, and were summarised on this blog: subjective
I posted my measurements done with Audio Precision in #23 .
I also measured some output power (#3,093), and measured the response of the various Op Amps on a 10kHz square wave: #2,324
Interesting thread there as well:
https://forum.headphones.com/t/pass-labs-whammy-solid-state-class-a-headphone-amp-kit/5272
I also have created mine AB tester, but used relays, and a RaspberryPi controlls them. I have written a simple Python program for that.
This approach is better for me, because I can switch from the listening position.
This approach is better for me, because I can switch from the listening position.
And if its computer controlled it can be double blind without needing someone else, by using random generator.
IIRC Scott Wurcer described being able to beat a relay controlled ABX tester solely by ear. He listened to the acoustical noise pattern of relays switching to determine which source was playing.
The switching itself is not easy to hear, but (in case of ABX testing) I made several switching quickly to make my mind confused.
By the way, I not use the ABX test too often. I just listen to both sources by switching forth and back between them and try to be honest.
I mean, if there is no difference, then I don't try to find one. The ABX testing is just a statistical method to indicate, that you can't clearly distinguish the sources.
By the way, I not use the ABX test too often. I just listen to both sources by switching forth and back between them and try to be honest.
I mean, if there is no difference, then I don't try to find one. The ABX testing is just a statistical method to indicate, that you can't clearly distinguish the sources.
Properly implemented ABX tester has dummy relays and dummy switching to hide any audible cues from relays. Also a "true" random number generator should be used so that switching sequence is not repeated.
Thanks Denis. I made something similar just switched between 4 headphones. The only issue that bugs me is the attenuation is higher compared to headphones connected directly to the headphone amp. Not an engineer but could this be related to resistance in the A/B box? If anyone has had that experience or can offer suggestions I thank you.
most probably, yes. That's the first thing to check. do a check with a DMM, what resistance do you have between input and output of your switcher? it should be negligible, only a tiny fraction of an ohm.