Earlier in the month I posted an article about a mod for the SY85 that allows the player to get higher velocity values from the keybed: SY85 Keyboard Velocity Modification
Recently I picked up a floppy emulator unit – a Gotek SFRM72-FU-DL. More about that in a future article. Whilst the SY85 was open during the installation I was able to run some experiments on the PKS CPU. The Velocity Modification reduces the normal clock speed of the PKS CPU to about 92% of the normal 8Mhz. At this lower speed the SY still functions normally but what would happen if this was reduced further?
The PKS CPU looks after these functions:
- Scanning the Keybed for pressed keys and calculating the velocity
- Scanning the Front Panel buttons
- Monitoring the Sustain pedal
Using the little function generator in the top photo I was able to clock the PKS CPU to these frequencies:
- 8MHz – the standard rate in the SY85
The function generator was hooked up to the SY’s PKS CPU with a hastily constructed link using a 1x scope lead and an IC Pin-Hook lead. Not really ideal for sending signals but it worked well enough for this experiment. I used MIDI-OX to monitor MIDI activity from the SY85
Nothing to report. As this is the standard frequency of the PKS everything was normal including the failure to get anywhere near 127 velocity from the keybed.
MIDI-OX now reports 127 velocity from the keybed. The keybed still reliably registers all keypresses ON and OFF. Large sections of keys were stabbed concurrently – no missing keys reported in MIDI-OX. Sustain pedal still detected reliably – I was hitting this on and off quickly with my hand. The front panel buttons still functioned but I did detect an occasional ‘miss’. If I very quickly pressed Program 1, Program 2, Program 1 in a loop it would miss about 1 press in 10. Pressing and holding the ‘+1’ button to auto-repeat to the next Program was down to half-speed.
MIDI-OX shows 127 velocity is easy to achieve. Keybed detection is still reliable with no latency in the sound. The Sustain pedal is now missing the occasional hit, and the front panel buttons are sluggish giving the impression that some are faulty. If you are used to quickly selecting programs & menus with a quick stab-stab-stab action then this is annoying. Keeping a finger on the buttons a little longer still works.
MIDI-OX shows 127 velocity is very easy to achieve. Keybed detection is still reliable with no latency in the sound. Sustain detection is sluggish. The front panel buttons are getting tedious to use – a finger needs kept on each one for a duration before detection. Pressing and holding the ‘+1’ button to auto-repeat to the next Program takes an eternity.
Even at 1MHz the keybed detection was still 100% reliable. My little function generator would not generate a lower frequency unless I used a different mode but the highest frequency in that mode is about 65Khz which is too low for this experiment. I guess I need to dig out the other one.
1Mhz and below:
After hooking up this unit it appeared that the lowest clock speed the PKS would take was around the 620KHz mark. Any lower and it stopped responding to keypresses and panel button pushes. Either the code inside the PKS crashed at this speed, or the main CPU gave up waiting for data, or the signal at the end of my hastily constructed lead was degrading enough to be of no use. Increasing the speed did not bring the PKS functions back again – the SY needed to be power-cycled.
At 620KHz the keybed detection was still very reliable. MIDI-OX showed all the Note-On & Note Offs for every key I hit, even when I quickly tramped an arms-length of them up and down. Obviously the keybed detection code is the priority, and the sustain pedal and panel switch detection in second place.
Driving the PKS at 4MHz clock was still usable and I was able to get 127 velocity without hammering the keys. A frequency between 4MHz and 8Mhz would be an ideal value – the panel switches would still be responsive, and those higher velocity values easier to reach without key-hammering. Something around the 6MHz would be useful – but where will I find that..?
Seek and ye shall find..
Whilst looking at the SY85’s Block Diagram I noticed something down at the bottom right corner – a Clock Generator diagram showing an output of 6.114Mhz
After a bit of tracking-down I found this comes from IC21 on Pins 12 & 13. It appears be used for the effects ICs and other things. Scoping these pins shows this signal comes on immediately the SY is powered on, and is a square 5V 50% pulse. I just need to link this to the PKS CPU’s clock input.
The Velocity Mod – Rev. B
For this modification the trace indicated by the arrow on the top left of the photo needs to be cut. This is the same trace cut detailed in the SY85 Keyboard Velocity Modification
A link wire needs attached to the signal source at IC43 Pin 10 and run to Pin 3 on IC12. A 120R resistor was put in series at the IC12 end of the link to kill some signal ringing that appeared there. The mod worked fine without the resistor but it is worth cleaning it up if it is easily done. Any value between 68R & 120R will work. Scoping IC43 Pin 10 before and after the mod showed no difference in signal quality.
The link was routed away from other high-speed signals as much as possible.
Signal Source – IC43 Pin 10
Signal End – 120R Resistor then IC12 Pin 3
The SY85 is now closed up again and over time I will see if anything out of the ordinary appears. This mod can be easily undone so nothing is permanent.
If you choose to do this modification be aware that it is at your own risk – I take no responsibility for any damage you may do to your SY85. It is assumed that the person doing the modification is experienced with working with PCBs and that full antistatic precautions are followed.