DSP Xover project (part 2)

Relay control

I'm trying to connect a wireless remote control transmitter to the "Relay control". See page 11 Najda Startup Guide.pdf.

The idea is to install a receiver with on/of function next to each speaker amplifier power supply.

The transmitter sends an "on" message when its two leads are shortcutted. So the amplifiers start. When disconnecting the leads an "of" message is sent and so the amplifiers are shut down.

Now the Najda Relay control does not produce a complete cut off in standby mode. So the transmitter doesn't understand that Najda is in standby mode. And therefore the amplifiers stay on all the time.

Can somebody take a look at the Najda board? What kind of circuit is the "Relay control". Can you see numbers or type? I have a boxed Najda and I'm trying to resist to open it.

I'm trying to solve this without an outside the box relay as the only problem seems to be able to "disconnect" the two leads from each other in the inactive state.
Did measure the voltage from the "Relay control" pins and I get 0,35V in Najda active state witch surprises me.
----
The transmitter and receiver are of Nexa.se brand. I have not found web pages but in Swedish. They have very handy transmitters and receivers to switch on and of lights and other electric consumers in your home.

I'm using:
Transmitter WBT-912
Receiver LCMR-1000
 
I think you need an outside the box relay. That is based on the information available on the following pages:
WAF-Audio
http://www.waf-audio.com/doc/Najda/Najda_Startup_Guide.pdf
Nexa | Sandare for inbyggnad
Dropbox - Svenska - Simplify your life

The relay control output on Najda may use a transistor switch. If so, it may not pull all the way down to ground. It only needs to go to a low enough voltage to make a relay turn off.

In addition, Nexa WBT-912 does not say that it can be used with transistor outputs to operate it. It appears to need a switch to control it. Only relay contacts would be appropriate for acting in place of a physical switch, and a relay should be able to operate from a transistor switched relay output terminal on the Najda board.

Based on the diagram in the Najda startup guide on page 11, it looks like you need a DC power supply to operate the relay. If you have a 5v power supply to power Najda, you could buy a 5v coil relay and use your existing 5volt power supply to power it. Otherwise, you probably should not use a relay and relay power supply of more than 24volts DC (just a guess to be safe), if for some reason you don't want to the 5v volt power supply for the relay coil.
 
@Speedysteve7 (#3316):

The post JLOP is writing about is #3215.
Works fine with Raspi 2 and Volumio V2.xx. Should as well do with Raspi 3.

Nounouchets solution seems rather ambitioned, a simpler approach to make one step beyond the above mentioned pure connection at a reasonable price ticket is the Kali reclocker.
Kali (i2s Reclocker)

(Experience with this welcome...)

:santa2:
I've got it all plumbed in, excellent sketch by Jlop, but as yet no sound.
Volumio / Spotify is there and send to be playing.

I've checked my wiring a few times, rebooted all but still nothing.
Got the loop wire in..

I have Volumio set to Generic I2S. Is that right?
Any other commands I need to give the Rpi directly?

I got it working fine with Hifiberry digi+ in to SPdif btw.. so I know Volumio/Spotify and Rpi are okay.

I have a Kali reclocker module on the way so I hope I can get it working..
 
Last edited:
Got it working this evening!

RPi3 into Najda I2S.

Checked all my connections, all fine.
Powered up, went to Spotify connected to Volumio, still have to increase the Volumio volume to max in Spotify (need to fix that) and then there was music!
But of a mystery why I could not get it to work yesterday. Settings all the same..

Anyway, it sounds quite different to the Hifiberry. A thick veil, more like blackout curtain has been lifted..

Incredible detail and mid / bass power.
Some soundstage change and a hint of high treble if not distortion, some artifacts that shouldn't be there.. (jitter?).

There's a little click / noise between tracks, not a pop though.
I saw a Linux line to cure that somewhere or other.

More listening needed.

Looking forward to Kali to see what improvements that brings..
 
  • Like
Reactions: 1 user
Thanks
I've also get some very strange 'recessed' instruments on a couple of tracks only..
Hard to explain it, but say a guitar that is normally in your face, if now totally recessed and withdrawn.
Most tracks are fine.
Some remastered ones (but not all) are singing very odd..
Hi Steve
I have the same exact strange behaviour on some old Beatles songs, singer almost disappears ? I checked the tracks on other systems and everything is right.
Jean-Louis
 
Hi Steve
I have the same exact strange behaviour on some old Beatles songs, singer almost disappears ? I checked the tracks on other systems and everything is right.
Jean-Louis
Yes that's just the thing.

Can't you hear me knocking by The Rolling Stones - the guitar riff that should be in your face is as if in a box 30m behind the stage :)
That one's been remastered..

You And Your Friend - Dire Strait is also rather strange.

Almost all other tracks I've so far tried are not affected at all.

I wonder if it's the clocking. I have an Allo Kali reclocker coming very soon..
 
Listening more to this strange sounding Rpi I2S Najda set up, most of the music is coming from the left channel.
Putting the balance to the right a bit alleviates the general situation but not the weird echoey sound on those tracks that really are messed up.

I wonder what can be doing this?
It's a digital stream. It's either there or it's not?? Can't have analogue features like a low output channel, can it?
 
Hmm, there's no difference in the system with regards to physical phase, speaker wires etc

It's Rpi -> DAC and not there when using Spdif in. The Hifiberry is taking care of things.

It's as if all the info is there, just not where it should be, or how it should sound is messed up (must be digitally, like the info on where has slipped some).
I'm no I2S expert. I did a lot of work on PCM telephony back in the day.
Common Channel Signalling (CCITT7). Yellow/Red/Blue book.
 
Last edited:
Got the Allo Kali reclocker before Christmas!

IMG_20181222_115734980.jpeg

Plumbed it in as per instructions.

Made not the slightest difference to the weird instrument and vocal placement on a few tracks, or the slight left balance bias for all tracks I play.

So for now I'm abandoning RPi -> I2S! -> Najda.

I'll try it with the Hifiberry DAC and see what that sounds like and if no better, get a Allo Digi One and feed into the Spdif input!
 
Very interesting observations about the RPi i2s input - I have been using it for a while and always wondered whether my DIY amps or speakers have some (intermittent) gain / loudness problem as one speaker is a bit louder. Recently listened via the SPDIF input and did not have the problem of the voice being off center. So I should also look into the i2s input, not the rest of the chain. Thanks for sharing your observations!
 
I messed about with Kali reclocker a bit more.
It won't work with Hifiberry I2S to SPdif as Kali only talks to Slaves..
Multi armed Goddesses of time tend to be that way :)

I tried using Kali's master clock to the usual Najda I2S_Ext_CLK (pin 23).
No sound.
Then I tried Ext_CLK Pin 27 external clock input.
Again no sound.

Anyone ideas?
How does WaveIO do it so well?

Then I put it back on Kali's output BCLK / Najda's pin 23. That gives sound, the weird sound on a few tracks and a little off centre to the left balance..

I am feeding RPi and Kali separate 5V supplies now and it definitely sounds a lot better for it..
On many tracks it's sounding very good but not quite right mix as it were.
 
Looks like in Najda, I2S goes through the DSP function before the dac. Could it be that the DSP is somehow causing the problem? Can you change the symptoms by messing around with the DSP settings?

Also, looks like there is an ASRC between the I2S input and the DSP. The ASRC could be doing something odd. What input sample rate are you sending, 16/44?
 

Attachments

  • Najda Signal Path.jpg
    Najda Signal Path.jpg
    45.2 KB · Views: 350
Last edited:
Looks like in Najda, I2S goes through the DSP function before the dac. Could it be that the DSP is somehow causing the problem? Can you change the symptoms by messing around with the DSP settings?

Also, looks like there is an ASRC between the I2S input and the DSP. The ASRC could be doing something odd. What input sample rate are you sending, 16/44?

Thanks Mark, I'm not sure about that. It's the same route in for WaveIO using I2S and that performs fine.
I've contacted Allo directly about Kali use and connections - I've referred to the Najda blurb - hope they can see what's going / not going on, where my I2S numptiness cannot. I suspect a Master / Slave clocking incompatibility.

My other option is to get a Allo Digione or Signature and go in the SPdif route (= defeat). I've heard the Digione working just fine on Najda with a friends RPi and Allo Digione..