I had the need to decide which of three incarnations of SLAPS for bass I would go for.
I also wanted to compare the three with my Trimodal amplifier.
With an Arduino Nano, some relays and potentiometers I built a switchbox:
according to this schematic:
At the listening position the test is conducted using this control box:
This is how it done:
First the amplifiers to test are selected. It could be that you don't have all four amplifiers connected or that you already have ruled out one in a previous test.
Then you can calibrate the input levels to each amplifier. I did it by using a True RMS voltage meter on the common loudspeaker connection.
Next the button to DUT relationship is randomized. Press the buttons a couple of times to ensure that.
Then it is time for the blind test. At this point no one except the program knows the button to amplifier relationship. Keep notes.
Next the button to amplifier relationship is revealed.
It could be good to repeat the above a couple of times.
You might find that one or two of the amplifiers can be removed from the "contest" so the blind test can be easier.
Here's a youtube video on my first test.
Later, when I have found the winner, I will re-work the hardware so that I can test which op-amp is the better one for my preamplifier; is it LM4562, OPA1612 or OPA1642.
I also wanted to compare the three with my Trimodal amplifier.
With an Arduino Nano, some relays and potentiometers I built a switchbox:
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
according to this schematic:
An externally hosted image should be here but it was not working when we last tested it.
At the listening position the test is conducted using this control box:
An externally hosted image should be here but it was not working when we last tested it.
This is how it done:
First the amplifiers to test are selected. It could be that you don't have all four amplifiers connected or that you already have ruled out one in a previous test.
Then you can calibrate the input levels to each amplifier. I did it by using a True RMS voltage meter on the common loudspeaker connection.
Next the button to DUT relationship is randomized. Press the buttons a couple of times to ensure that.
Then it is time for the blind test. At this point no one except the program knows the button to amplifier relationship. Keep notes.
Next the button to amplifier relationship is revealed.
It could be good to repeat the above a couple of times.
You might find that one or two of the amplifiers can be removed from the "contest" so the blind test can be easier.
Here's a youtube video on my first test.
Later, when I have found the winner, I will re-work the hardware so that I can test which op-amp is the better one for my preamplifier; is it LM4562, OPA1612 or OPA1642.
Nice looking box🙂 - how do you address the issue that the sounds of relays switching could be slightly different for different switching combinations and be subconciously registered? Is there a way to double-blind test for being able to discriminate the sounds of the relays? Just a thought.
Do you mean the actual clicks of the relays or the sound one can year through the loudspeakers?
Anyway, I just did a test with just the relay switching, in blind test mode - no difference between switching the different amplifiers.
And then with a 200 Hz test tone, also in blind test mode - still no difference between the amplifiers.
In not blind test mode I first imagined that there was a difference, but that went away after some "clicks".
Note that I have a break before make delay time of about 12 ms in this manner when switching:
Break all relays
Wait 12 ms
Make relay for input
Wait 12 ms
Make relay for output
For every input there are two relays and for every output there are four relays.
So that's why you can here a cascade of relay clicks when the amplifiers switches.
But it all sounds the same.
Of course, with time some relays might deteriorate and thus can be noticeable.
I don't plan to use them for that long though, but if I did, a relay board can easily be replaced.
All things considered, one must start a blind test session with testing this to make sure there are no audible differences when the relays switches.
Good point!
Anyway, I just did a test with just the relay switching, in blind test mode - no difference between switching the different amplifiers.
And then with a 200 Hz test tone, also in blind test mode - still no difference between the amplifiers.
In not blind test mode I first imagined that there was a difference, but that went away after some "clicks".
Note that I have a break before make delay time of about 12 ms in this manner when switching:
Break all relays
Wait 12 ms
Make relay for input
Wait 12 ms
Make relay for output
For every input there are two relays and for every output there are four relays.
So that's why you can here a cascade of relay clicks when the amplifiers switches.
But it all sounds the same.
Of course, with time some relays might deteriorate and thus can be noticeable.
I don't plan to use them for that long though, but if I did, a relay board can easily be replaced.
All things considered, one must start a blind test session with testing this to make sure there are no audible differences when the relays switches.
Good point!
Just did a double blind test at full listening volume - I could barely hear the switch between the amplifiers.
At low volume levels relay clicking is certainly audible. Also, there is different sound when relay goes NC-NO or NO-NC.
In my AB box I cure this by switching auxiliary relays. They switch nothing, only make clicks when it is appropriate.
My solution has two independent section, line-level and power-level.
The former main function is to match the volume levels, the latter to switch the amplifiers outputs.
When switching path is A->A (or B->B) power relays are not engaged and clicking sound is generated by auxilary relays.
In my AB box I cure this by switching auxiliary relays. They switch nothing, only make clicks when it is appropriate.
My solution has two independent section, line-level and power-level.
The former main function is to match the volume levels, the latter to switch the amplifiers outputs.
When switching path is A->A (or B->B) power relays are not engaged and clicking sound is generated by auxilary relays.
Attachments
Yes, but I don't plan to do the bass blind tests at low volume.
There's no different sound as in my solution every input and output have their own relays.
When all relays are NC, no inputs or outputs are selected.
Same here. At NC, the input is shorted to ground and the output has a 4 ohm resistor across its terminals.
So there are three relays for every amplifier, 32 in all.
Ingenious. And great looking box too.
I guess yours is a permanent solution while mine only shall last a couple of weeks.
The preliminary testing results is that I couldn't hear any difference.
So I need better test music, any suggestions?
Using Audacity I can make a blind test mix.
So I need better test music, any suggestions?
Using Audacity I can make a blind test mix.
Kudos to you and LKA. I suspect you've expended far more time, money and effort on this than many people who feel confident in their views (either way - and including myself) on this topic. 🙂
Switching unused relays to ensure A->A clicks can't be differentiated from A->B is some serious attention to detail!
Switching unused relays to ensure A->A clicks can't be differentiated from A->B is some serious attention to detail!
Thank you, Earl Grey.
Not that much time actually and the hardware was lying around.
It'll be re-used in other projects I reckon.
Yes, that was an out the box solution.
Oh sorry I was being sincere, I like the test track in your video 🙂
And interesting result. I guess it is the expected result, but interesting data point nonetheless.
No offense taken.
Yes, the result is perhaps somewhat expected given the type of music in the test and how I listened to it.
But I thought Brian Bromberg would have brought up any differences being an acoustic piece of music.
So what to expect then?
The SLAPS amplifiers have positive feedback and distortion cancelling through their bifilar coils that are expected to give fast response without any TIM.
Well, the fast response must mean percussion then.
TIM is piano solos or choirs?
I don't think that I can pinpoint any of these properties directly.
So it must be how big a sound stage I get, the number of individual voices I can hear in a choir and so on.
It is also good if I'm fairly acquainted with the music piece in question.
This calls for longer sessions with each amplifier, still using the double blind test approach of course.
On the other hand, it's then harder to remember the impressions from one amplifier from the other.
A third type of test is to check how long I can listen to an amplifier before I get listening fatigue, if ever.
That can be even harder given that the tests then perhaps must span over several days and how perceptive I am a particular day.