A VCA-Based Phaser

Where we discuss new analog design ideas for Pro Audio and modern spins on vintage ones.
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: A VCA-Based Phaser

Post by mediatechnology »

This may be an earlier citation for an SSM2164 based phaser from Osamu Hoshuyama. It's dated before Neil's Webbly World phaser. http://userdisk.webry.biglobe.ne.jp/000 ... ph0505.GIF

With some help from someone more skilled than I, I was able to modify the Hoshuyama circuit to use a THAT2180. Although I changed the Hoshuyama circuit to accommodate the overall non-inverting nature of the THAT VCA/I-V combination, what I didn't see in the above was the DC feedback path (R3 in the above sch) that wraps around the VCA. (R3 also provides AC feedback which is certainly more obvious.) The DC feedback provided by R3 permits the I-V stage - with Cfb made large so that it's an integrator - to run open loop without local DC feedback. Placing an Rfb in parallel with Cfb prevents the circuit from working properly. One of my errors was to return R3 to the left side of the input coupling cap (not shown) and this opened the loop at DC.

Since the THAT2180 doesn't invert overall, the sign of the DC (and AC) feedback provided by R3 requires additional inversion. This is done in the all-pass stage by reversing the inputs and taking R3 from the all-pass output. I've had a chance to breadboard one pole and it works quite well. Each pole requires a 2180 VCA and a dual BIFET op amp which is considerably simpler than the bootstrapped approach which requires three op amps. This Hoshuyama topology appears to be a far better approach.

I'll build up a few more poles and post a drawing. Since the "all-pass" stage is a differential amp - the integrator is the VCA/I-V combo - I may use a THAT1240 for summation. This will permit greater gain precision and deeper notches when summed with the direct input as well as eliminate 4 resistors per all-pass stage.
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: A VCA-Based Phaser

Post by mediatechnology »

OK, I've been playing with this for a few days and the Hoshuyama topology is the way to go. Sam Hoshuyama was very kind to allow me to modify and use his SSM2164-based circuit. With the help of Gary Hebert of THAT to transform it to a non-inverting VCA, I was able to protoboard a 4 pole phaser. I'll post sound files and some FFTs shortly.

This is a single all-pass section:

Image
A "printer friendly" copy of the schematic is here: http://www.proaudiodesignforum.com/Imag ... Pass_2.JPG

U2 needs to be a BIFET. U1 provides U2's bias current from input current sourced by R2. When U1 is in attenuation, (and reducing current) the amount of voltage required on the right-hand end of R2 to satisfy the bias requirements of a 5532-class op amp can be quite high. To put it a different way, as attenuation is increased, the effective value of R2 is multiplied by the attenuation. At 40 dB, the effective value of R2 is about 2M. A BIFET's low bias current reduces DC-shift to very low levels. With C2 at 220 pF, U1's attenuation range is 0 dB to about -33 dB to sweep the notch of a 4 pole section from approximately 13 kHz to <<100 Hz. Very large amounts of attenuation should be avoided because the VCA will "starve" U2 of bias current completely.

I like using the THAT1240 as the all-pass differential summing element rather than an op amp with 4X 10ks wrapped around it. Using a THAT1240 not only eliminates a bunch of resistors but permits very deep notches to be obtained due to the 50 ppm resistor ratio match. The notch depth and flatness will likely be limited by R1 and R2 matching. Although a little more costly, using a THAT1240 will provide an easier build and a more compact board layout. (One could also use SMT dual THAT2162 VCAs and THAT1280 dual line receivers as well.)
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: A VCA-Based Phaser

Post by mediatechnology »

I've tried values between 220 pF and 47 nF depending on the topology. 220 pF seems to offer the widest range using only VCA attenuation in the Hoshuyama topology. I determined this by looking at an FFT of the notch on white noise. Really large values, say 47 nF, compress most of the range in the subsonic but may provide a larger number of combs in the top octaves. I'd recommend that the layout be able to accommodate several lead spacings.

It just depends on how you want it to sound IMHO. There are so many combinations of sum/difference and tap settings that work, but work differently, which provide lots of choices.
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: A VCA-Based Phaser

Post by mediatechnology »

So far I've used identical C values per pole with them cascaded in series. The phase delay per section is the same and adds. It seems to sound better - deeper to me at least - than when the values are mixed. (During test I tried this.) Most of the phaser cites I see use identical values though it's never said why.

I think tuning depends on how one wanted to use it though.
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: A VCA-Based Phaser

Post by mediatechnology »

I just got around to recording a quick 4 pole demo using four cascaded all-pass sections shown below. The final output is direct + all-pass mono output also summed using a THAT1240.

The middle control range (and octaves) were used for this particular cut making it sound a little dark and thin. At about +0:28 sec you'll begin to hear some cheesy vibrato I manually introduced. The Cfb for each allpass is only 220 pF. Amazing you can get so much Vibrato with so little delay.

https://www.proaudiodesignforum.com/con ... Demo_2.mp3 (1.5 MB)

Noise demo with a slow sweep: http://www.proaudiodesignforum.com/Cont ... Demo_2.mp3 (1 MB)

Image
Bruno2000
Posts: 38
Joined: Fri Jan 04, 2008 5:32 pm

Re: A VCA-Based Phaser

Post by Bruno2000 »

Wow, great sound! More samples, please, like a Fender Rhodes.
Best,
Bruno2000
The large print giveth, and the small print taketh away - Tom Waits
Post Reply