This is my new, highly improved, version of the 80017A chip.

This page is only temporary until the new website is online. For general info about these chips: Read the pages of the previous versions: D80017A and D80017.


The complete details with nice graphs and stuff will be placed on the new website, but these are among the improvements.
  • Full protection: The new AR80017A will withstand a potential difference of 36V between any 2 pins or a current of 500mA going in/out any pin. Whichever happens first. 36V is above the supply voltage of the Juno-106 and 500mA will blow its fuses. This means that you can short any number of pins, make wrong connections, have bad parts surrounding these chips, put them upside-down or wrongly aligned in sockets,... you name it, it can handle it. Why? From time to time I get returns and 99.9% of them died from shorting pins during installation or were obviously inserted wrong in a socket. Now I'm aiming for zero returns and I want to give peace of mind to those who are experimenting or less experienced.
  • Transient response: Apparently, in a rare case (only 2 people ever noticed it) the previous versions could sound a little bit different than the original 80017A's. To hear it, following 3 conditions have to be met. Resonance set to max so that the filter is self- oscillating, keyboard tracking set to a high level and you have to be playing notes that are far apart. When these conditions are met, an effect can be heard that can be best described as portamento on the self-oscillation of the filter: it takes a fraction of a second for the filter to reach its final frequency. The reason for this is that, depending on the filter's settings, different parts of the filter dissipate different amounts of power, which causes them to heat up or cool down. What you hear is this heating up or cooling down. The original chips have this effect much less. Reason for this is that the complete filter of the original chips is contained on a single piece (my design is built from many separate parts) of silicon. The net difference in power dissipation of all parts of the filter together is much smaller in total, so the achieved temperature difference is smaller and more important: all these parts heat up or cool down together. As you can't beat physics (only hippies with magnets on YouTube seem to be able to do that), full custom IC design is unaffordable for the quantities in which my AR80017A's are made and changing the operating power (and thus temperature shifts) by changing component values changes the sound, I thought this problem was unsolvable. This new design contains some analog wizardry that allows these physical effects to take place where they have less influence and this tuned so that it matches the original chips.
  • Out of spec noise generator immunity: a Juno-106's noise generator circuit can be out of spec and inject more noise power into the filter than the previous versions can handle. This can cause them to temporary stop sounding. Lowering the noise level, changing the filter cutoff or resonance brings them back to life. The new AR80017A can take all the noise you throw at it. This issue with the noise generator circuit is very rare (I'm aware of 3 Juno-106ís that had this problem) but can be caused by several reasons. The noise generating transistor, labeled TR21 (in the service manual) or Q21 (on the PCB), which should have been selected for its nice and flat noise spectrum can be an unsuitable one for the job. Maybe it degraded? Or more likely, it was 'selected' on Monday morning or Friday at beer-o-clock. Also, the capacitor (C41) in the RC filter after the noise VCA which should limit the noise bandwidth to about 5 KHz can be degraded. It's an old disc type ceramic capacitor and these are known to age. This will result in a too wide noise bandwidth. Finally, the Noise Level part of the adjustment procedure (see the Juno-106) can be done wrong. Doing this adjustment with a low-end digital oscilloscope or soundcard is asking for problems. Use an analog oscilloscope or high-end digital one with digital phosphor for this job. If you have D80017A's and have this problem, it can be properly cured by doing the noise adjustment correctly or by replacing C41 with a nice and new NP0/C0G ceramic cap when the noise bandwidth is too wide or by replacing TR21/Q21 with one that has a flat noise spectrum if the noise spectrum looks funky. Quick and dirty fix: turn VR32 anti-clockwise until the problem goes away. However, this will lower the maximum noise your Juno-106 can produce a bit.
  • Drifting along with the supply lines: In normal conditions you can expect your Juno-106 to have a stable power supply and the previous versions have been designed with this in mind. However, not all conditions can be expected to be normal. Extreme temperature changes (think indoor/outdoor stage use) will affect the operating point of the power supply. Also, given their age, it can be expected that not all Juno-106's will have a power supply that is in perfect condition. Its big electrolytic smoothing capacitors can be degraded, if beyond a certain level, this will introduce mains ripple on the supply lines. The pass transistors of the voltage regulators can be a bit toasted (small heat sink + little airflow inside the Juno-106) too. This will lower their forward gain, thus increase the impedance of the power supply and so allow dynamic loads to modulate the supply lines. Pretty much everything in the analog part of the Juno-106 is affected by what happens with the supply lines. The above has a myriad of effects on what you hear. The filter tuning changes a bit when your Juno-106 heats up or cools down. Mains hum will FM modulate the filter and this highly dependent on its cutoff frequency. Anything that modulates the supply lines will modulate the filters and VCA's and they will thus cross-modulate each other... The new AR80017A response to what happens to the supply lines now matches exactly the original 80017A chipsí response. Nobody ever noticed the difference, but it was there and I am a perfectionist.
  • High end 4-layer PCB: As addressing the above improvements added complexity and component count to the design it was no longer possible to make it on a double sided PCB without significantly increasing the size. So, I changed the design to a 4-layer one. With it I was able to even reduce the size and make them smaller than the original ones. These PCB's are made to the highest standard possible and frankly, I haven't seen anything like them.

You can buy them from me directly or from one of my distributors, see below.
I sell them at the price of 46 EUR each + 10 EUR shipping cost (international registered mail, worldwide with proof of shipment)
Shipping costs stay the same, no matter where and how many.

I have a buy 6 pay 5 policy.
I have a package price for 6 AR80017A's and 3 D5534A's: 280 EUR
I have discounts for large quantities and can provide invoices, full customs docs,... for registered companies.

Shipping usually within 14 days after payment.
Shipping only to the shipping address provided with the payment. No exceptions on this.

Accepted methods of payment:
  • Paypal (Preffered, use buttons below)
  • Bank transfer (IBAN-BIC)
  • Money Transfer
  • Cash
For 1 to 4 pieces of AR80017A: use this button
For a set of 6 AR80017A's (buy 6 pay 5): use this button
For the full set of 6 AR80017A's and 3 D5534A's (package price): use this button
For other payment methods and larger quantities: email me.
Do not mail for order status/tracking/whatever within 14 days after payment (or the date your E-check has cleared). All these emails are ignored. You will get a email notification to the email address you used for paypal when your package was shipped.