The ARP Quadra phaser circuit - stage ladder

[jessenator]

I attempted this circuit once before and would like to understand a bit (or many, many bits) more about how certain aspects of the circuit are designed so that suitable changes/substitutions can be made possible. My eventual end goal is to have another go at it.

The first hurdle I want to get myself over is understanding a bit more about how the actual stages (all 14 of them) are using the transistors; more specifically, if one wanted to revert to discrete components* for space concerns—I know I could be waaay off here, but here's a stupid test i did around the ladder with 4 stages: monolithic, TO-92, SOT-23, and another with smaller MLCC caps


Just looking at the schematics (since I'm a hack and no proper EE), they don't appear to be under the same prescriptions as something like a Phase 90, so I wondered:

1. since the original calls for the 3086 array (spec sheet here), if I was looking to find a suitable sub, what characteristics should I be most concerned about when matching to a discrete component? I'm all for doing the legwork, but I don't even know what to compare :S

2. Again, since the original asks for these monolithic jobbies, am I going to plan on matching for pairs? quads? more? I've read that BJTs are a bit easier to match than the JFETs of the Phase 90-like circuits, but figured I'd ask.

3. is this part of the exercise stupid and should I own up to the reality that this thing is monstrous and should resign myself to having to source and pony up for some 3086/46 chips, or just build something else?  : P

I ask, because if a comparably functioning/characteristic BJT can be found and the only real expense is time in matching them, I could see that as more viable than shelling out $10 / 14-DIP array, since there are small amounts of real estate gains to be had, especially if one goes with SOT-23 pieces. E.g., again I have no idea what I'm looking to compare, but the MPSA18 is only pennies in some places and *I could get ~80 for the price of a single CA3086, loads of headroom to find matches.

Anyway, thanks for the advice/pointers.


edit, oops, forgot to include the actual circuit in question; upper right:

[alanp]

BCM847DS is another option, and one I've used myself. Matched dual NPN transistors in a single six pin SMD package. (Swine of a job to solder, but they're already matched, and they're in current production.)

The matching isn't amazing, but it's better than bothering to do it yourself.

[jessenator]

Thanks for the reply!

BCM847DS is another option, and one I've used myself. Matched dual NPN transistors in a single six pin SMD package.

Nice! Here's what I managed to shrink it to (far right):

The matching isn't amazing, but it's better than bothering to do it yourself.

Any noticeable distortion happening?


So you think this circuit is based then around pairs vs quads? or is that what you meant by the matching not being amazing?

[alanp]

Hadn't noticed the quad versus pair difference, but given that it's a transistor ladder, I'd be surprised if you need more than matched pairs.

Also, some people say that matching the capacitors used in transistor ladders is just as important. Haven't looked into it myself, but you might want to.

[jessenator]

Hadn't noticed the quad versus pair difference, but given that it's a transistor ladder, I'd be surprised if you need more than matched pairs.
Also, some people say that matching the capacitors used in transistor ladders is just as important. Haven't looked into it myself, but you might want to.

Ah, good to know! Thanks for your help on this.

[mousefart]

I'm a bit late to the party but, I've been working with the same Juergen Haible schematic over the winter. The component values are brutal to translate from the handwritten schematic. Random Source implemented BCM847DS's on their Eurorack version of the JH Tau Phaser. Any luck getting this circuit to function?

[jessenator]

I'm a bit late to the party but, I've been working with the same Juergen Haible schematic over the winter. The component values are brutal to translate from the handwritten schematic. Random Source implemented BCM847DS's on their Eurorack version of the JH Tau Phaser. Any luck getting this circuit to function?

This one's on the back burner now :/ I may get back one day, but I've had to shift focus.