swinkels

Yes it is related to clock source and firmware, so you need special MIDIbox firmware to get proper pitch. PAL C64 uses 0.985MHz clock, so the firmware can't work with both sources equally. Unfortunately there is no MB firmware for TDA1543A but i will try to compile it for you.

These extensions works only with special 'MIDIbox' SwinSID firmware version. Nevertheless your firmware is definitely much more close to real SID than the one reviewed by TK some time ago. It features e.g. ring modulation, improved waveforms and supports all filter modes.

It seems that you are using wrong DAC chip version. Please test firmware for TDA1543A. This should help.

Indeed, SwinSID emulates all SID filter modes without need of external capacitors.

You're right. R3 jumper also changes output biasing. This was needed because older C64 boards has bit different output stage and I decided to simplify desing by adding jumper instead of a more complex but universal solution.

@kristal=: Yes, because it is powered from 5V line only.

Hi, I am the creator of the SwinSID and I can answer your question directly. There are two jumpers. First is used to setup output stage and volume, second is for filter selection between 6581/8580 model. There was special firmware version specially designed for MIDIBox with extended filter range but this feature is not enabled in standard release.

Why not connect J14 to unused pin of J2? Anyone wants to build this new SwinSID PCB? I will help with programming and testing.

I'm using simplified STK200 with success but heard that it doesn't work with some mainboards. STK200 doesn't need any power source but you AVR should be running while programming. Please post all further technical questions in this thread: http://www.midibox.org/forum/index.php/topic,9386.90.html Here You will find my tip with setting fuse bits.

nILS, what is current status of your PCB layout? Is it finished yet? Small building tip: New AVRs are running with internal 1MHz oscilator. To use external 24MHz oscilator please set 'clock source' fuse bits to 0000 (all bits programmed).

Check your fuse bits configuration to enable external clock source - AVR have to run with 24MHz. SwinSID plays short 1KHz test tone during startup, when the frequency is lower then your clock frequency is wrong. Are you using TDA1543, not TDA1543A? (TDA1543A won't work). If you still have problems, attach sound output recordings here.

Yes they are swapped but no big deal.

The master AVR is actually slave and vice versa. Both are ATMEGA8515, not AT89S8515 , therefore change their names to avoid confusion.

There are in Poland for 1,93 euro http://tme.pl/katalog/index.phtml?f_szukaj=tda1543&idp=1

Thanks for the PCB. I haven't mentioned on my site that this schematic is for C64 version. It uses 74HCT138 for address decoding of each AVR from C64 bus. With MB we can remove this chip and connect CS signal directly to the MB core module as with MBSid - master CS goes to MBHP_SID:J2:SO and the second one to MBHP_CORE:J14.

ATMEGA8518 is really not so ancient as you think. Its memory is filled in a half only so there is no real need for switching now. This AVR has one advantage over new ones - three additional I/O pins used to transfer data with slave AVR in SwinSID X2. They are used in the latest SwinSID 1024 too. Switching to smaller AVR limits us to single AVR only. I can see only one real succesor of the ATMega8515 - ATMega162. It has the same pinout but twice the RAM and ROM memory.

Finally, my download page was updated with latest SwinSID firmware for MIDIBox and dual AVR schematics. There is source code too. I'm working now on the new SwinSID 1024 card with 512KB RAM and 512KB ROM for storing drum samples and new cool sampled sounds. Maybe some day it will work with midibox as a external sampler (when TK modify control firmware) but first, it is intended as a hardware MOD player for C64.

You should use ATMEGA8515 with 16PU or 16PI ending. 8PU is 8MHz version and may not work correctly.

I'm going to update info about new SwinSID X2 very soon. Currently I uploaded subpage with new demos recorded by Crisp and TK. The new firmware haven't change much since january, it has mostly SID and C64 compability fixes. You can download it from http://www.swinkels.tvtom.pl/swinsid/files/SwinSIDX2.hex

Yes you can do it by enabling double oscilator mode - in this mode anti aliasing is not supported. Now I'm only waiting for the new stereo demos from TK :)

SwinSID is still in development and the new version SwinSID X2 is almost ready now. It is designed to emulate two SID chip working with one MB core. In fact SwinSID X2 has two AVR connected with single DAC Latest firmware provides better sound quality (low aliasing), new waveforms and nice features: simple FM synth, two oscillators per voice and white noise generator. Firmware is quite stable now with all features working, but I need to design layout for new MB module. If you want to help with testing or designing layout then let me know. I will consider releasing the source code.

While playing high notes with full spectrum waveforms like sawtooth and pulse you can hear some harmonics. , but this is a standard effect with simple wavetable synthesizers. I am still working to improve sound quality and add new features to SwinSID with help and testing from TK. When I finish we will release music demos and comparisions, so please be patient

Here you can download and test the new firmware with features as above www.swinkels.tvtom.pl/swinsid/swinsid_20070727.hex TK already tested it and found some distortions with high frequencies but in most cases it should work fine.

PonyProg is a very nice program for flashing AVR. Yes, i will try to add more new features to standard SID emulation instead of pseudo-reverb effect. I will release new firmware this weekend with following changes: - Linear Attack support - Logarithmic Decay and Release - 2 x oversampling instead of interpolation - No reverb effect After some testing from you I will implement a lowpass filter soon.

I have used a simplified avr programmer like this and It works very well. http://elm-chan.org/works/avrx/stk200.png The new avrsyn filter is quite good, but too bad it is noisy ( probably 8 bit calculations ) I will try to port my own implementation of lowpass filter from SwinSID2 to original SwinSID soon.