Why specifically a polyester capacitor in the feedback of this H-bridge driver?

[Icchan]

I'm refurbishing/repairing older Yamaha DM1000 mixing console and while I was at it, I noticed that they've used specifically a Polyester (or as they say in the Japanese repair manual: Mylar) capacitors for the dc-blocking in the feedback loop of the operational amplifiers that are driving the H_bridge of the DC_motors of the moving faders.

That somehow caught my eye and I can't come up with any specific reason why they'd choose relatively expensive and large polyester capacitor instead of basic ceramic?

The service manual can be found here: https://www.manualslib.com/manual/896121/Yamaha-Dm1000.html

And on the PDF page 266 you can see the motorized fader driving circuits.

Here's a snip from one of them:

So I was thinking, if I ask from the community, this might be a very good learning opportunity :)

So why do you think they chose specifically a polyester capacitor for this location?

[KE5FX]

Given that it's audio equipment, they may have wanted to avoid acoustic noise from a ceramic capacitor. 

[Gyro]

47nF isn't big for a film capacitor, and polyester is the cheapest of the film types. Maybe they wanted closer tolerance than a ceramic, or maybe just habit - an audio company would be used to fitting film capacitors in opamp feedback loops so they may have just done it by default. It may even be that 47n polyester was a company stock item and 47n ceramic wasn't.

[TimFox]

Assuming these are all TH (leaded) parts, I doubt the polyester/Mylar capacitor was more expensive than an X7R ceramic at the time of manufacture.
Even though this is not in the signal path for the audio, the polyester capacitor will not suffer from the voltage co-efficient of X7R (or worse) ceramic dielectrics.
A quick look for current-production 0.047 uF leaded parts in Mouser (at quantity 1) shows PE film at roughly $0.32 ea and X7R MLCC leaded at $0.35 to $0.40.

[Icchan]

...maybe just habit... maybe 47n polyester was a company stock item and 47n ceramic wasn't....

As boring of an explanation might be, it could be true
I was just wondering exactly because it's not in audio path what so ever and since these are SMD-parts, these are somewhat large.. though all the boards are HUGE in this device anyway so there might not be any reason to skimp on size.

It was just so specifically stated in the schematics, that these are polyester, it felt weird.
We'll see if anyone comes up with some insight as to why or is the most boring explanation the correct one

[SiliconWizard]

47nF isn't big for a film capacitor, and polyester is the cheapest of the film types. Maybe they wanted closer tolerance than a ceramic, or maybe just habit - an audio company would be used to fitting film capacitors in opamp feedback loops so they may have just done it by default. It may even be that 47n polyester was a company stock item and 47n ceramic wasn't.

Yep. Very likely. Tolerance issues are doubtful here, this doesn't look like a critical value in this driver. The BOM factor is much more likely, along with just habits.

[AndersJ]

Ceramic caps are often microphonic.

[BrianHG]

In the old days, when I used to use 1000pf caps for analog C band video de-emphasis circuit, Polyester caps gave me a clean flat filter response.  Ceramic would smear the picture and the effect was clearly visible both onscreen and on the oscilloscope.  However, the motor controller filter you illustrated here does not require such a precision wide-band, consistent smooth slope filter demanding circuit.  I guess the engineer behind the circuit just may have had some experience in the past and just said, lets use the good quality caps.

Another thing is that in a high temperature environment, a ceramic cap would drift and it's high frequency impedance completely goes way off in the 47nf range compared to a 47nf Polyester cap.

Again, the circuit you have shown us does not require such demanding performance as it just drives the motors for the faders unless that cap is fighting a high current drive from jfet 'FT800'.

[bdunham7]

It was just so specifically stated in the schematics, that these are polyester, it felt weird.
We'll see if anyone comes up with some insight as to why or is the most boring explanation the correct one

Where does it say polyester--I don't see in in the schematic.  Is it in the BOM?  What is the voltage rating of the capacitor?  Are these the maroon rectangular-ish gumdrop-sized ones or do they appear different?  I don't see anything unusual in using those, I actually have an assortment of 630V and some 1000V of these that I keep around for random projects, like tube radios, snubbers and anything where I want low leakage.  AFAIK they're very, very reliable (at least the type I have). 

[MathWizard]

I have lots of sub-1uF green mylar caps, the kind rated over 100V. Next time I build an oscillator, I'll have to compare them to ceramic cap's.

I hardly have any variety of cap types, and most are just cheap ebay ones.

[TimFox]

I have lots of sub-1uF green mylar caps, the kind rated over 100V. Next time I build an oscillator, I'll have to compare them to ceramic cap's.

I hardly have any variety of cap types, and most are just cheap ebay ones.

Polyester/Mylar capacitors will compare well against X7R and other Class-II dielectrics, but are inferior to NP0/C0G Class-I dielectrics.
However, 0.1 to 1 uF C0Gs are not to be found:  Kyocera quotes a range of  0.5 pF to 0.1 uF for surface-mount C0G devices.
Polypropylene film capacitors compare well to NP0/C0G and are available well past 1 uF (although large).

[floobydust]

[...] So why do you think they chose specifically a polyester capacitor for this location?

Many times it's just because the part is already used on the board somewhere else and it's easier to use it again. Fewer unique parts on a BoM, the better.

[TimFox]

I remember addenda and updates to -hp- manuals where it was obvious that a decree had come down to reduce the number of discrete resistor values required in a BOM, to simplify production.

[tom66]

I second/third the statement about microphonic capacitors.  If I'm reading the circuit right you have FT800 possibly acting as a muting transistor and two channels summed in via R809/810.  So we can ignore the FET for now, the gain of the circuit is (with R827 @ 160k) around 80x. This would make it particularly sensitive to microphonic effects. Any current injected into the virtual ground via the capacitor would need to be offset by the amplifier causing a large output voltage.  The film cap eliminates this risk, and it's a mixing console, so manufacturers are going to want to consider this as sensitive ears might notice effects like that.