Nice to know. In my case 0.5 rms is enough as i use differential and then get 1 v rms, using two channels. Thanks !
Hello guys,
First, thanks a lot for all your very usefull sharings on this topic.
I'm having trouble with the lowcost chinese combo ADAU1452 + AD1938
After reading the datasheets multiple times and spending hours configuring the registers in different manners, I still can't get it to work properly.
C/D arrives on SDATA_IN0 (slot 0 & 1) => OK
A/B should arrive on SDATA_IN1 (slot 0 & 1) => Nothing
On the scope, the SDATA_IN1 pin has no sign of life, despite having the DSP clocking on BLCK_IN1
The schematics provided somewhere on github seems outdated (at least partly), some jumpers and connectors missing, etc...(I've already sent a message to the aliexpress seller but for now, no reply ...)
Also, I noticed there was a STC15W101 MCU on the AD1938 board wich I can't figure out the purpose ...
I suspect a hardware problem ... ? Or maybe configuration ?
Thanks in advance 🙏
First, thanks a lot for all your very usefull sharings on this topic.
I'm having trouble with the lowcost chinese combo ADAU1452 + AD1938
After reading the datasheets multiple times and spending hours configuring the registers in different manners, I still can't get it to work properly.
- USBi is working, registers and eeprom R/W OK
- All outputs working, but 1/2 inputs( only C/D and not A/B)
C/D arrives on SDATA_IN0 (slot 0 & 1) => OK
A/B should arrive on SDATA_IN1 (slot 0 & 1) => Nothing
On the scope, the SDATA_IN1 pin has no sign of life, despite having the DSP clocking on BLCK_IN1
The schematics provided somewhere on github seems outdated (at least partly), some jumpers and connectors missing, etc...(I've already sent a message to the aliexpress seller but for now, no reply ...)
Also, I noticed there was a STC15W101 MCU on the AD1938 board wich I can't figure out the purpose ...
I suspect a hardware problem ... ? Or maybe configuration ?
Thanks in advance 🙏
Attachments
Well, after replacment of the opamps (opa1612 is replaced with opa2340) THD in all channels dropped by orders of magnitude. Although in two channels the THD is still too high (probably, the problem is in the ceramic SMD capacitors). When buying a cheap chineese adau1452+ad1938 DSP, be sure to inquire about what opamps are soldered on the PCB.
I have experimented with measuring the ADAU1452 and the CS board quite a lit lately.
ADC is a ASUS U7
Reference DAC is a ESS 9039 pro based DAC from SMSL
The CS board uses 0 and 1 as differential output. Goes into a analog HP-LP filter and then into two stages of -6dB differential analog lineamps fom BB.
Then into the analog in on U7
So above 700 Hz the U7 is the limiting factor.
There are also a reference line of a -110 dB 3.rd harmonic that is made syntetic in the ADAU.
Also an experimental measurement of cancelling the fundamental from the CS board with the reference dac so that ADC distorts way less. I am not sure at what HF frequency it can not be trusted anymore. But it looks like the CS codec distorts a lot less than the ADC, And surely a lot less than the analog components on theese series of boards. Both CS codec board and AD codec board is affected by the underspec of the analog component in the filters.
ADC is a ASUS U7
Reference DAC is a ESS 9039 pro based DAC from SMSL
The CS board uses 0 and 1 as differential output. Goes into a analog HP-LP filter and then into two stages of -6dB differential analog lineamps fom BB.
Then into the analog in on U7
So above 700 Hz the U7 is the limiting factor.
There are also a reference line of a -110 dB 3.rd harmonic that is made syntetic in the ADAU.
Also an experimental measurement of cancelling the fundamental from the CS board with the reference dac so that ADC distorts way less. I am not sure at what HF frequency it can not be trusted anymore. But it looks like the CS codec distorts a lot less than the ADC, And surely a lot less than the analog components on theese series of boards. Both CS codec board and AD codec board is affected by the underspec of the analog component in the filters.
I was about to buy one of these boards thinking that the Cirrus CS42448 version didn't have any opamps for them to muntz or downgrade, but it seems like all of these boards require a lot of part replacement and hacking to get the noise down to acceptable levels. I had similar experiences with a MidiWorx board from AliExpress, which either had noise, non rail-to-rail opamps (and it would oscillate if you put "better" opamps on it), etc.
The Cirrus CS42448 board measure well enough used spdif inn and balanced dacs out.
But the analog output passive filter must be removed.
Analog in requires the 4 HP filters on the non signal inputs.
But with the original caps, the 3.rd harmonic is high. Useable under 20dbfs
But the analog output passive filter must be removed.
Analog in requires the 4 HP filters on the non signal inputs.
But with the original caps, the 3.rd harmonic is high. Useable under 20dbfs
But are there other options out there? Really like the TDM between ADAU and codec/DACs
Then I mean low price and minimal analog filtering so original spec of codec can be achived
Then I mean low price and minimal analog filtering so original spec of codec can be achived
Just tested my unit for the first time, got it configured with a simple input->output program. It works fine, but it seems too low. Am I supposed to boost it with a node before the output? Or is the signal coming from the codec board just low by itself?
The codec board is the CS42448.
The codec board is the CS42448.
Hello there, after reading the whole thread I decided to go directly for AD EVAL-ADAU1467Z. It works like a charm from the first plugging in. The codec is supposed to be quite usable and the price is not so higher than the china board including codec.
It was around 200euros. Comparing to the china boards with codek and USBi, the price difference was less than half. Plus zero hassle. I manage to tune AutoEQ and parametric equalizers and crossover in less than hour without any previous knowlenge of SigmaStudio.
Now I want to attach TPA Dclass from TI to one of the I2S buses and configure it as sub.out.
I have to admit, thank you guys. You motivated me a lot to get the board and UMIK and play with the audio once again.
Now I want to attach TPA Dclass from TI to one of the I2S buses and configure it as sub.out.
I have to admit, thank you guys. You motivated me a lot to get the board and UMIK and play with the audio once again.
I have an ADAU 1452 board together with an AD1938. The AD1938 has 4 in and 8 out. Is it correct to understand that out3 and out4 are only pass through? I want a 3-way, + sub.
I don't know if I undarstood you correctly but you can use it for you project. There are available 8 output channels (4 stereo), currently I'm finishing car audio processor with SPDiF and analog input (automatically multiplexed when SPDiF is on/off) and 3 way output (woofer-midwoofer-tweeter) plus subwoofer.
Ok. Then I should be able to DSP all outputs. Do the AD1938 and ADAU 1452 both have to be powered or does one of them have to be powered? I don't have the USBi yet, is this essential?
5V rails are connected so you can supply only one board as far as 3V3 rails are also connected together. In my project I split supply lines and built external power supply 2x5V and 3V3, than I obtain quite good distortion level. See below. Measurement made by REW using Motu M2 interface, balanced output and internal (DSP) 1kHz generator. The USBi is essential but you can go the cheap way:
https://www.diyaudio.com/community/...rammer-using-cypress-cy7c68013a-board.269111/
https://www.diyaudio.com/community/...rammer-using-cypress-cy7c68013a-board.269111/