DSP Xover project (part 2)

Analogue input clipping

From the startup guide V1.1 board

1. Analogue Ins

Najda features one analogue stereo input, whose signal is sampled at the rate set as central

processing rate, which is one of 48, 96 or 192 kHz.

Nominal input level is 1 VRMS. A larger input signal will clip at AD conversion stage and the AD

Clip led will light on.

I'm looking at my old V1.1 board. No DIP switches.


From the V1.3 user guide.

1. Analogue Ins
Nominal input level is 1 VRMS or 2 VRMS , selectable with DIP switches located near the input jack.

Please refer to the table on the board silkscreen for setting those switches.

I am listening to my V1.3 board and it's rather plumbed in to take a look, but I reckon the switches are there[emoji6].
Somewhere on this thread there may have been a request to make Najda more input V flexible.

If you have the earlier version you will have to find a way to attenuate the input signal before it reaches Najda.


Re the 1 V to 6V setting.
The analogue output is variable 1V to 6V and that is done by changing the settings in the Najda in Control SW.
But that won't cure your input clipping.
 
Last edited:
Cheers Steve,

I do have the latest version - so do I just up to 2Vrms on the input jack? I still don't get what that means, but if that solves the problem and doesn't fry my AD chip, that will work for me.

Thank you for looking into the manuals, unfortunately it doesn't tell me what to do exactly.

Any help appreciated! Regards, Alex
 
Cheers Steve,



I do have the latest version - so do I just up to 2Vrms on the input jack? I still don't get what that means, but if that solves the problem and doesn't fry my AD chip, that will work for me.



Thank you for looking into the manuals, unfortunately it doesn't tell me what to do exactly.



Any help appreciated! Regards, Alex
NP.
Yes, if you go to 2V it's likely to irradiate the clipping.
Most line level devices, CD players analogue output, Tuners (remember them) and some phono stages are up to 2V.

I'm not sure why V1.1 was hard set to 1V.
Perhaps with low output phono stages in mind?
 
Hi Steve,

Great! Thank you, that makes perfectly sense - I will give it a try on the weekend, as the clipping has just been on the high hitting tones (kick drums, synth stabs etc.), I am sure it will do the trick.

Not sure why v1.1 was hard set to 1Vrms, but your explanation sounds at least feasible to me. There might be more to it from an audio engineering perspective I guess.
 
Not sure why v1.1 was hard set to 1Vrms.

2Vrms will have more noice if the input level from whatever you feed from is low.

It is easy to lower input signal level with a 10kohm potentiometer or two resistors / channal. Seek voltage devider it is basic elektroniks.


Some CDs have much more dynamics so these will clip where another CD wount. The clipping might take place alredy in your CD player.

You might also have the clipping comming from elswere. DSP is very sensitive to gain structure. There is not much headroom so never EQ +db. If you EQ gain say +3db at 200HZ then you have to lower the hole band pregain -3db.
 
If you stream try this track..
I could play everything I'd ever tried on Spotify and Qobuz but this track caused clipping on the I2S input!

Qobuz

open.qobuz.com/track/44622424

Spotify

Грустная сука - song by IC3PEAK | Spotify

Screenshot_20210522_085643_com.spotify.music~2.jpg

Warning it's a bit of an ordeal / horror track[emoji846]

In general I've found using resistors, L-Pads and attenuators (fancy bought ones even) in the line level saps sound quality/dynamics..

Worth trying all options though and seeing what works best.
 
Hi,
If anyone could explain what the difference is of 1V and 2V RMS nominal input level, that would be superb!
Cheers, Alex

Actually maximum input signal level for ADC chip used in Najda is 1V. For newer borads there is voltage divider/attenuator configured witch dip switches that allows 2V level also.
So if you are in 1V setting switch to 2V and see how it behaves. If still clipping you need to lower input voltage by some custom attenuator (or maybe your source has ability to attenuate output signal.)
 
I'm so sad that Najda is not support anymore for the reason we all know.

In these days I played with the FIR filter with the file generated from rephase ... ok the taps are not enough on the bass, but in my opinion is the most simple 4 way xover with simple interface on the market that can work with "FIR" file from Rephase, it's very sad that all this work will be lost, hypothetically if a new, more powerful version had come out that managed more taps per channel, especially on the bass channel, it would been a unique processor in the market. :-(