Roland 80017a

The dreaded chip that takes the Juno-106 out. Title link takes you to shots via this auction. Docs with mp3s and shots also saved here. See this post for an alternative replacement chip.

Details:
Roland 80017a Replacement VCA/VCF chip for Juno-106/ MKS-30/ GR-700/ HS60
Problems Typically Solved by Replacing the Failed 80017a Chip Include:
* Erratic or sudden loud noise from synthesizer output, particularly after 10 minutes of powering up.
* Note hang-ups or long sustain.
* One synth voice will not respond to filter or amplifier changes.
* Unstable operation.
* Juno-106 or MKS-30: Notes will not play, most typically every 6th note will not play due to failed 80017a chip.
* GR-700: One particular voice or string will not sound, "D" string, "A" string etc.

About the 80017a VCA/VCF chip:

The 80017a chip was a real workhorse for Roland synthesizer engineers in the early 1980s. This one chip contained both the classic, driving -24 dB voltage controlled low-pass filter with a smooth voltage controlled amplifier. By placing all the components on one chip, Roland was able to keep the classic sound of their analog synthesizers across various products, with low production costs. Almost immediately Roland knew they had a problem on their hands, and changed production methods.

When these chips fail, sometimes voice will stop working completely. On a keyboard synth, this shows up as every sixth voice not working. On a GR-700 guitar synthesizer, this shows up as one string’s synthesizer voice failing. But the chips may not fail completely, resulting in a lack of control over the filter quality for one voice, or by sudden, loud erratic sounds from the synthesizer. The only solution to these problems is to replace the existing 80017a with a new chip.

This chip is from lot 44, produced after Roland solved the manufacturing problems. Probably 20 years after the original production date, this chip is still working well without failure."

Demos:
Failed Chip
Good Chip

"There are four audio samples of the open string "D." The first audio sample has both resonance and filter cutoff tracking the envelope generator. The VCA is set to "gate," no VCA modulation. This sample highlights the resonance effect. The next sample has the resonance turned off, and only the filter cutoff is modulated. Again the VCA is set to "gate," no VCA modulation. The third sample has the filter wide open, with the VCA being modulated by the envelope. This is to highlight VCA modulation. The last patch is a complex modulation patch of the filter with sample and hold qualities. The mp3 file is a little over one minute in length. There is silence at the beginning, and the sample was normalized to make it as loud as possible."

Update via the comments:
"The statement that there are good lots (date codes) for these chips is completely untrue. I've seen all date codes from 40A to 61A fail."

Update: Looks like there is another one up for auction.