ARP 1047 SVF

From circuitbending to homebrew stompboxes & synths, keep the DIY spirit alive!

Moderators: luketeaford, Joe., lisa, Kent

Post Reply
User avatar
dksynth
Wiggling with Experience
Posts: 314
Joined: Wed Jul 01, 2015 12:56 pm

ARP 1047 SVF

Post by dksynth » Tue Jul 21, 2020 7:59 pm

I made my own version of the 1047 last year that I quite enjoyed.

But I don't completely understand some aspects of the circuit and I would like to. And maybe we can get some insight into how it works from some of the cleverer members here.... and maybe we'll be able to make something.

here is my redraw of the schematic:
1-47 schematic.png
I used my own little 4005 VCA chiclets:
4005.png
Which used different combinations of matched or single transistors depending on what type of 4005 it was.
4005 versions.png
source: https://www.guitarfool.com/ARP2500/4005.html

The 1047 is an SVF with discrete transistor VCAs. VCA 1P and 3P are driven by the expo onboard VCA 3P. VCA 2N has its own expo onboard. 1P and 3P are the filter VCAs, 2N is the resonance VCA.

What'd I'd like to know is the "tricks":

Whats the deal with the diodes to and from +15 and -15?
The use of double matched FETS for the integrators with the opamps... I take it that fet input opamps weren't a thing yet? Is this better than modern fet input opamps?

Is there a precedent for the kind of lead/lag compensation in the LP feedback path? I don't see it used in later filters with similar performance. serge VCFQ used interstage compensation and the UAF used feedback path compensation without a trimmer. (unless you count the snip of wires to be a trimmer).

here was my VCA cards assembled:

52434240_10218660159237615_9098918272757137408_n.jpg
My initial design used three of them in a serge voltage-controlled-triple-filter arrangement with automatic gain compensation between filters.
You do not have the required permissions to view the files attached to this post.

loki
Common Wiggler
Posts: 104
Joined: Fri Apr 25, 2014 4:20 pm

Re: ARP 1047 SVF

Post by loki » Wed Jul 22, 2020 11:30 am

Using an OPA2134 or OPA1678 is superior to the dual FET and LM301. You are correct that there were no FET input op-amps when the 1047 was designed. The configuration Dave Rossum used a few years later in the UAF is an improvement over the 1047. He used the FET in a source follower configuration and followed it with an LM318.

User avatar
teleport
Common Wiggler
Posts: 128
Joined: Sat Aug 12, 2017 2:43 pm
Location: Oakland

Re: ARP 1047 SVF

Post by teleport » Thu Jul 23, 2020 11:40 pm

This thread may be of interest :
viewtopic.php?t=188193

The correspondence between Colin and Ashley in the pdf in loki's post there discusses the lead/lag circuit.

Am also curious about those stacked diodes on the ACOM and XCOM nodes. If I had to hazard a guess it'd be that they're forming a reference that is a couple diode drops away from ground - in practice the actual voltage drop through a diode depends on the current to a certain degree, so activity in the 4005's may appear there. When it's driven is there signal on those nets?

loki
Common Wiggler
Posts: 104
Joined: Fri Apr 25, 2014 4:20 pm

Re: ARP 1047 SVF

Post by loki » Fri Jul 24, 2020 5:46 pm

For those interested, Dennis Colin's patent is: 3805091

User avatar
guest
Super Deluxe Wiggler
Posts: 4329
Joined: Mon Aug 19, 2013 11:49 am

Re: ARP 1047 SVF

Post by guest » Fri Jul 24, 2020 8:22 pm

the diodes are acting as voltage references. the VCAs will perform better if the collector voltages on the diffpair transistors are similar. the 2 diodes make up for the 2 Vbe drops in the current mirror in the VCA. this reference is also used for the exponential converter to put the collector of the exponentiating transistor above (or below depending upon NPN/PNP) the diffpair emitters. this has the downside of adding temperature dependency on the CV to the exponential converter.
openmusiclabs.com

User avatar
cygmu
Super Deluxe Wiggler
Posts: 1510
Joined: Mon Feb 15, 2016 3:33 pm
Location: UK

Re: ARP 1047 SVF

Post by cygmu » Sat Jul 25, 2020 7:50 am

To the extent that I understand the phase compensation stuff, I learned the theory from Lee Thomas's paper

The Biquad: Part I-Some practical design considerations
IEEE Transactions on Circuit Theory ( Volume: 18 , Issue: 3 , May 1971 )
DOI: 10.1109/TCT.1971.1083277

The paper is about the biquad rather than SVF but there is a lot in common. The big difference is that there is only one feedback loop in the biquad circuit. In the paper Thomas calculates the effect of finite op-amp gain on the phase around the loop and its impact on the location of the pole and the Q of the filter and works out how to compensate it with a phase-lead capacitor. It is stated that the cap can be placed in parallel with any of the resistors around the loop, which makes sense.

It always struck me that in an SVF, because there are two loops, to get the compensation exactly right would require phase-lead caps in several places. I have not calculated to verify this thought but I don't see how it can not be true, and presumably Colin, and people like Serge etc. who also designed SVFs, were aware of it too. The Serge filters have caps in a couple of places presumably to keep the phase correct at all points in the circuit. However, tuning them all individually to set everything just right must be essentially impossible, especially for a production line rather than a one-off DIY effort. So my guess is that Colin decided to put just one cap in and tune it by hand to get things working as well as possible. Once that decision is made it probably doesn't matter where around the loop you decide to add the compensation.

User avatar
The Real MC
Veteran Wiggler
Posts: 670
Joined: Sun Jun 02, 2013 8:36 pm
Location: In front of computer

Re: ARP 1047 SVF

Post by The Real MC » Sat Jul 25, 2020 6:30 pm

cygmu wrote:
Sat Jul 25, 2020 7:50 am
To the extent that I understand the phase compensation stuff, I learned the theory from Lee Thomas's paper

The Biquad: Part I-Some practical design considerations
IEEE Transactions on Circuit Theory ( Volume: 18 , Issue: 3 , May 1971 )
DOI: 10.1109/TCT.1971.1083277

The paper is about the biquad rather than SVF but there is a lot in common. The big difference is that there is only one feedback loop in the biquad circuit. In the paper Thomas calculates the effect of finite op-amp gain on the phase around the loop and its impact on the location of the pole and the Q of the filter and works out how to compensate it with a phase-lead capacitor. It is stated that the cap can be placed in parallel with any of the resistors around the loop, which makes sense.

It always struck me that in an SVF, because there are two loops, to get the compensation exactly right would require phase-lead caps in several places. I have not calculated to verify this thought but I don't see how it can not be true, and presumably Colin, and people like Serge etc. who also designed SVFs, were aware of it too. The Serge filters have caps in a couple of places presumably to keep the phase correct at all points in the circuit. However, tuning them all individually to set everything just right must be essentially impossible, especially for a production line rather than a one-off DIY effort. So my guess is that Colin decided to put just one cap in and tune it by hand to get things working as well as possible. Once that decision is made it probably doesn't matter where around the loop you decide to add the compensation.
One of the first publications of a SVF design is IEEE 1968 Vol 56 pp 1137-1139 by J. Tow of Bell Labs. While not a complete analysis, it does have brief discussion of the sensitivity of the system.

I collected quite a few discussions on SVF analysis, including ElectroNotes. I have yet to see any discussion on phase compensation in dual loop SVF designs.
I haz a wrench, to fix da synth.
http://www.analoguediehard.com/

loki
Common Wiggler
Posts: 104
Joined: Fri Apr 25, 2014 4:20 pm

Re: ARP 1047 SVF

Post by loki » Sat Jul 25, 2020 8:42 pm

The Real MC wrote:
Sat Jul 25, 2020 6:30 pm

One of the first publications of a SVF design is IEEE 1968 Vol 56 pp 1137-1139 by J. Tow of Bell Labs. While not a complete analysis, it does have brief discussion of the sensitivity of the system.

I collected quite a few discussions on SVF analysis, including ElectroNotes. I have yet to see any discussion on phase compensation in dual loop SVF designs.
Here is a copy of Active Compensation of Op-Amps. It is no specifically aimed at the SVF but it should be easy enough to apply. Especially because there are now inexpensive dual op-amps that are well matched.
You do not have the required permissions to view the files attached to this post.

User avatar
The Real MC
Veteran Wiggler
Posts: 670
Joined: Sun Jun 02, 2013 8:36 pm
Location: In front of computer

Re: ARP 1047 SVF

Post by The Real MC » Sun Jul 26, 2020 1:48 pm

loki wrote:
Sat Jul 25, 2020 8:42 pm
The Real MC wrote:
Sat Jul 25, 2020 6:30 pm

One of the first publications of a SVF design is IEEE 1968 Vol 56 pp 1137-1139 by J. Tow of Bell Labs. While not a complete analysis, it does have brief discussion of the sensitivity of the system.

I collected quite a few discussions on SVF analysis, including ElectroNotes. I have yet to see any discussion on phase compensation in dual loop SVF designs.
Here is a copy of Active Compensation of Op-Amps. It is no specifically aimed at the SVF but it should be easy enough to apply. Especially because there are now inexpensive dual op-amps that are well matched.
Had not seen that paper, thanks.
I haz a wrench, to fix da synth.
http://www.analoguediehard.com/

Bernie Hutchins
Learning to Wiggle
Posts: 23
Joined: Sat Jun 27, 2020 3:19 pm

Re: ARP 1047 SVF

Post by Bernie Hutchins » Mon Jul 27, 2020 5:52 pm

I tried (in 2013) actively compensating the three SV blocks (summer, two integrators, all three separately) and reported the very mixed results:

http://electronotes.netfirms.com/EN215.pdf

No Magic Bullet! Since passive comp works quite well, and since it is unclear if the active comp can even be tweaked (with duals? perhaps), why not leave well-enough alone?

-Bernie

User avatar
J3RK
Super Deluxe Wiggler
Posts: 6620
Joined: Tue Dec 22, 2009 12:46 pm
Location: Seattle

Re: ARP 1047 SVF

Post by J3RK » Mon Jul 27, 2020 9:12 pm

Bernie Hutchins wrote:
Mon Jul 27, 2020 5:52 pm
I tried (in 2013) actively compensating the three SV blocks (summer, two integrators, all three separately) and reported the very mixed results:

http://electronotes.netfirms.com/EN215.pdf

No Magic Bullet! Since passive comp works quite well, and since it is unclear if the active comp can even be tweaked (with duals? perhaps), why not leave well-enough alone?

-Bernie
I’ve been off the site for a bit, and never saw posts from you before... this might not be the exact right place for this, but can I just say you’ve been a big inspiration for a lot of the work I do! Thank you!

Continue the 1047 discussion here:

:mrgreen:
Are you lemon? Does your head come to a nub?

Post Reply

Return to “Music Tech DIY”