PGA2310 and Bascom and ATMega8535

[simingx]

Hi, I've been cracking my head over this for the past 2 nights and just *cannot* get the simple circuit to work!

The PGA2310 is connected to a ATMega8535 as follows:
SDI : PB0
SCLK : PB1
MUTE : PB2
CS : PA1

ZCEN is permanently connected to VD+.

The source code is as follows:

$crystal = 1000000
$regfile = "m8535.dat"

Dim Totalvol As Word
Dim Volvarr As Byte
Dim Volvarl As Byte
Dim Testbit As Byte


Cs2310 Alias Porta.1
Mute2310 Alias Portb.2
Source1 Alias Portc.3

Const Wait1 = 1
Const Wait2 = 10

Config Portb = Output
Config Portc = Output
Portb = 1

Mute2310 = 1
Portc = 0

Source1 = 1
Cs2310 = 1

Totalvol = &H1111

Cs2310 = 0
Waitms 10
Shiftout Portb.0 , Portb.1 , Totalvol , 0 , Null , 1000
Waitms 10
Cs2310 = 1

When the code is executed, a quick probe with a 'scope shows the correct signals at the pins of the PGA, but no reaction! The PGA does not change it's output in any way.

Even if I do the very primitive method of bit-banging, it doesn't work too.
What could be wrong? I've already replaced a chip to no avail.
If anyone has any code, please send it...

Thanks!

[JochenH]

Hi,

i have to say that i don't know BASCOM, but from looking at the code i see no statement where PORT A (CS-PGA2310) is set to output. Perhaps thats the problem.

Jochen

[Bee]

Hi

I've successfully driven a CS3310 with an Atmel, and written a fair bit of Bascom code, and it does look as though your code is OK. I would set mute high, CS low, Data in high, and toggle Clock 16 times high to low and see if it works. You should also be able to watch the data go through the chip by looking at data out. Are you saying you've tried that, and checked its happening with a scope? If there's still no output, then have you double checked your implementation of the analog side?

Brian

[Audiopower]

Hi ,try to use the hardware SPI which is implemented in ATMega8535. I've used this metod in my project and it works. Everything is writen in bascom help.

[simingx]

D'oh!

JochenH was right...

The problem was my totally forgetting to define Port A as an output... the "correct" waveform seen by the 'scope was actually caused by the internal pullup!

The code using Shiftout works now... much more compact!

Thanks to all who looked...

me <-

[simingx]

Here are the pictures of the work in progress:

CPU board (left) and analog board (right)


Close up of analog board


'Scope output with output set to -6 dB (0.5x attenuation)
Both traces are set to 500mV/div.

[Bee]

The problem was my totally forgetting to define Port A as an output...

It usually is something like that!


Nice looking boards - you should be under way now.

Have fun -


Brian

[JochenH]

Yes, looks really nice.

When doing the listening test i would also try to bypass / replace the 2 buffer opamps with a piece of wire. Most sources have a low enough output impedance for the PGA2310, and a piece of wire sounds always more neutral (but not always better) than an opamp.

Jochen

[gmarsh]

That's a sexy VFD... where'd you get it?

[MCLL]

How much 20 x 2 VFD display?