Analog servomotors

[JohnRoberts]

Lots of moving faders use simple motors with a DC voltage on the sense winding.

I did a bunch or work with digital pots, for a microprocessor controlled analog, automatic mixer. I designed the prototype to accept either DBX VCAs, or DPOTs. I fired up the DPOTs first and they worked so well I never bothered to fire up the VCAs. In theory DPOTs can have zipper distortion, but I found in use they sounded adequately transparent (on speech). I even designed in zero crossing synchronization so the gain changes could be forced to occur only at zero crossing (to eliminate clicks). I didn't need to implement that either.

FWIW the Automatic mixer application is dealing with speech and modest gentle gain changes. This SKU never got evaluated by critical customers in the marketplace so my opinions are purely subjective.

Sadly the company I designed this for (still around just barely) got spanked by low cost digital mixing platforms. The market for a digitally controlled analog (automatic) mixer would be vanishingly small.

In hindsight I could have made my microprocessor controlled automatic mixer even better by adding some relatively inexpensive A/D convertors but that opens another can of worms...

JR

 

[interested human]

Lots of moving faders use simple motors with a DC voltage on the sense winding.

I did a bunch or work with digital pots, for a microprocessor controlled analog, automatic mixer. I designed the prototype to accept either DBX VCAs, or DPOTs. I fired up the DPOTs first and they worked so well I never bothered to fire up the VCAs. In theory DPOTs can have zipper distortion, but I found in use they sounded adequately transparent (on speech). I even designed in zero crossing synchronization so the gain changes could be forced to occur only at zero crossing (to eliminate clicks). I didn't need to implement that either.

FWIW the Automatic mixer application is dealing with speech and modest gentle gain changes. This SKU never got evaluated by critical customers in the marketplace so my opinions are purely subjective.

Sadly the company I designed this for (still around just barely) got spanked by low cost digital mixing platforms. The market for a digitally controlled analog (automatic) mixer would be vanishingly small.

In hindsight I could have made my microprocessor controlled automatic mixer even better by adding some relatively inexpensive A/D convertors but that opens another can of worms...

JR

I've looked everywhere and can't find the solutions I need to do this exact thing! Could I hire you to do something similar?

 

[JohnRoberts]

I am not taking new consulting gigs these days but i will offer free advice shared here publicly.

The analog side is pretty straightforward. You start by doing a survey of what DPOTs are available. The technology available was still improving years ago when I did this exercise. Important specs are working voltage which impacts max signal swing. Then step size is important.

Digital control is again straightforward with multiple serial digital communication options (SPI, I2S). IIRC that last design I did used dual DPOTs and I controlled them both at the same time for the stereo channels.

I anticipated incorporating zero cross switching so gain changes were coordinated to reduce step change glitches, but i found for my application I didn't need the added complexity because gain updates were so frequent the step sizes were small and didn't sound bad.

JR

 

[gyraf]

An interesting recent development is closed loop stepper motors based on simply gluing a small permanent magnet onto the stepper shaft and taking position through a magnetic sensor, e.g.:

https://www.aliexpress.com/item/4001064683300.html
..position precision 0.1125 degrees - and it works well indeed

 

[abbey road d enfer]

I anticipated incorporating zero cross switching so gain changes were coordinated to reduce step change glitches, but i found for my application I didn't need the added complexity because gain updates were so frequent the step sizes were small and didn't sound bad.

Zipper noise reduction is well documented e.g. in the AD5291/5292 datasheet.

 

[JohnRoberts]

Zipper noise reduction is well documented e.g. in the AD5291/5292 datasheet.

yup, zipper noise is pretty old news.

When I was coding up an automatic mixer using DPOTs I incorporated zero cross synchronization hooks so gain changes could be coordinated to only occur coincident with zero crossings (x times zero=zero). Another old school trick that can be used to reduce audibility of gain change steps is HF pre/de-emphasis. I was pleasantly surprised by how good the DPOTs I used sounded without zero crossing, by my application was mainly for speech.

JR

PS: After that "only analog" mixer company abandoned our AM project, I was very tempted to drop in some cheap A/D convertors and just do it all in the digital domain but in my judgement the world didn't need a cheap AM digital or otherwise, now that most premium digital mixers can provide AM as a free software feature.

 

[mhelin]

Digitally controlled audio volume controls like Cirrus CS3318 have built-in zero crossing detectors:
https://www.cirrus.com/products/cs3308-18/TI have similar solutions (PGA chips like PGA2311).

 

[JohnRoberts]

Synchronizing gain change steps with zero crossings definitely reduces the energy and therefore audibility of artifacts while large gain steps can still be audible, just not clicky when using zero crossing sync.

I suspect in my Automatic Mixer design I was pushing the gain changes, based on my computed gain sharing algorithm, so frequently that the gain changes were very tiny and so less audible. Plus processing speech that is more percussive doesn't have smooth LF content to highlight interruptions like music does.

JR

PS: Sadly I can ocassionally hear perturbations from pitch sifting processors time compressing or expanding movies to fit smaller or (rarely) larger time slots. The technology for splicing pitch shifted samples together is quite good but I can sometimes hear a burbling in symphonic background music. I suspect there is less motivation to stretch or shrink presentation for commercial free viewing.

 

[FIX]

I did a remote controlled mic preamp once that used the clarostat motorized pot, It would sit in the stage or studio and then the remote unit in the truck or control room. It could be manually overridden, and if the power went out, it went to min gain, then back up when stable, and sent data over 2 balanced mic cables (redundant) and could get data if only one sig wire of one cable was hooked up and any other wire from either cable. Bullet proof...

 

[Tubetec]

Alps motor pot with a remote dual pole changover momentary up/down switch and a battery box might suffice if repeatability isnt an issue , KISS ,'keep it simple stupid' works .