I know the pinouts are quite a bit different than a standard 256k EPROM so I jumped the wiring to match but to no success.
I discovered with a scope on the drum machine that the CE input on the ROM has a clock pulse that is twice as fast as the A0 address line. Is it uncommon for the CE to require a clock? The ROM is cycled by a ripple counter and the data lines (only 6 bits are used) are latched and fed to a resister network acting as an digital to analog converter. The CE and CS are active low on 2 of the 3 chips (crash and ride). The hihat, not shown, is active high on CE and CS (not sure why Roland chose to do it that way - maybe to try and keep me out). Clock rate is approx. 31KHz (adjustable for tuning).
How would I go about clocking the CE using my GQ-4X to allow this chip to give up it's code without having to build another circuit? I thought about building a frequency doubler and use A0 line as the input but maybe somebody has better idea.
I have dumped so many ROMs with the GQ-4X programmer and I love it. I back up everything I own just in case.
If I take the CE pin and connect it to the A0 address line I get data on every other byte like this.
B7 FF 23 FF CE FF 4A FF 56 FF 1F FF ED FF 3D FF
And it will verify successfully every time.
Other than that if I keep the CE low I get all FF's and if I pull it high I get like 20 20 20...
I really think that if I can tap into the clock signal the drives the address lines, I can pull all the data off these chips instead of every other byte. ZLM, can this be done? If that is not doable then I will go ahead and build a sqaure wave doubler and double the pulse speed from A0 to the CE pin of the chip.
I have lots of faith in the GQ-4X. It has never failed me yet.
" But the GQ-4X is unable to output the clock on the CE pin."
I can rearrange any pin to any socket point on the programmer with a homemade adapter. I understand that I can't get the clock speed I'm looking for from the GQ-4X socket itself. I thought maybe I can take the unit apart and grab the clock rate I need from inside the programmer because something has to be timing the internal flip-flops/ripple counter etc.. Maybe I'll take it apart and use a scope to find out what's going on.
I loaded a generic 27256 device into the GQ-4X and connected the A0 socket from the GQ-4X to the input of the doubler. The output of the doubler went straight to the CE pin on the HN61256P ROM. I also used a homemade adapter to rearrange the pin configuration because these Mask ROM pin-outs are not directly compatible to standard EPROM pin-outs.
For ROMs C42 (Crash) and C44 (Ride) I grabbed the input of the last NAND gate for the CE because it needed an inverted clock signal and I tied the CS pin directly to ground. But for C43 (Hi Hat) I used the output of the last NAND gate and tied the CS pin to Vcc.
The doublers pulse width is sensitive to the frequency at it's input. I had to change the two resistors in the doubler from 10K to 33K. I also brought the speed of the GQ-4X down 1 notch to +1 and that is where it worked great.
It seems a lot of Roland drum machines of that time (mid 80's) use this same type of design like the TR-707 and the MKS-7 synth and probably a lot more.
Chip Enable Clock
Posted - 08/26/2010 : 15:01:50
