Any way to calibrate the pitch bend wheel?

Covers the Yamaha MODX6, MODX7 and MODX8

Moderators: Derek, parametric, Saul

User avatar
ITFlyer
Member
Member
Posts: 10
Joined: Wed Jun 05, 2019 4:33 pm

Re: Any way to calibrate the pitch bend wheel?

Unread post by ITFlyer »

YamahaUser wrote: Sat Feb 15, 2020 4:04 am Sorry, but not quite. What you've really got is a pair of nominally 10 KΩ pots with the end terminals wired in parallel. The giveaway is the 3.95 KΩ reading at center.


I don't have the time to get my CAD program up, so a description will have to do. Let's presume the pots are ideal, so exactly 10 KΩ. When measuring from one end to the wiper with the pot centered, you're actually measuring half of that pot's resistance (5 KΩ) in parallel with the series connection of the other half plus that of the entire resistance of the other pot (10 KΩ). In other words, 5 KΩ in parallel with 15 KΩ. That calculates to 3.75 KΩ, close enough to the 3.95 KΩ you measured with the real pots.
Aww crap. Of course, as soon as I read this, I thought "yup, he's right." Totally forgot about the parallel pot. :)
YamahaUser wrote: Sat Feb 15, 2020 4:04 am Most pots are manufactured with a small amount of lubricant on the resistance track. If that's removed with a "no-residue" contact cleaner, the pot's lifespan can be reduced. You might want to use a control cleaner that includes a small amount of lubricant.
Good point, I'll do that.


Shop Solid Top Guitars Starting at $295
User avatar
SysExJohn
Global Moderator
Global Moderator
Posts: 919
Joined: Tue Dec 02, 2008 1:00 am
Location.: Wellingborough, Northants, UK.
Contact:

VIP

Re: Any way to calibrate the pitch bend wheel?

Unread post by SysExJohn »

Sadly, I rather suspect that the issue has absolutely nothing to do with the pots, but with the, in effect, analogue to digital convertor.

Yamaha has decided (in its infinite wisdom) that, since 7 bit resolution is good enough for all the other CC's, then it shall be 7 bit for ALL controllers, including pitch bend.

That is, not all these keyboards are faulty, but Yamaha in its pursuit to drive down costs to has decided to do away with 14 bits (or even 10 bits) on its implementation of pitch bend.

Call me cynical if you wish, but why else?

JohnG.


UW500; 2 x MU1000 + PLGs AN, DX, VL, VH & PF; SC8850; EWI4000s + VL70m; Samson Graphite 49.
Privia PX-560; AN1x; Novation X-Station 25. E-MU 02 cardbus + 1616m x 2. Finale 26; Sonar 7 PE and CbB.
XGworks v4; SQ01; SOL2. Garritan GPO5; COMB2; JABB3; IO; Organs; Steinway; World; Harps; CFX Concert Grand.
Macke
Global Moderator
Global Moderator
Posts: 229
Joined: Thu May 30, 2019 8:03 pm
Location.: Sweden

VIP

Re: Any way to calibrate the pitch bend wheel?

Unread post by Macke »

ITFlyer wrote: Sat Feb 15, 2020 3:02 am So not seeing any obvious smoking gun, I figured the only thing left was dirt or dust in the pot itself. I flushed both pots out with a plastic-safe no-residue contact cleaner while working the pots continually. For good measure, I also disconnected the pot/mod connector from the circuit board and cleaned it as well.

I then put the keyboard back together, plugged it all in, and the verdict: it's fixed!!! No more spurious pitch bend MIDI jitter, and the pitch bend range has been restored.

For those who might be curious, I'll post some pictures of the guts of the keyboard that I took while it was open.
I'm glad you managed to solve the issue.

Thank you for also sharing your findings, the actual solution and pictures from the inside of the MODX. It will most likely be useful reference for others facing similar issue in the future. (Y)


User avatar
SysExJohn
Global Moderator
Global Moderator
Posts: 919
Joined: Tue Dec 02, 2008 1:00 am
Location.: Wellingborough, Northants, UK.
Contact:

VIP

Re: Any way to calibrate the pitch bend wheel?

Unread post by SysExJohn »

ITFlyer wrote: Sat Feb 15, 2020 3:02 am So not seeing any obvious smoking gun, I figured the only thing left was dirt or dust in the pot itself. I flushed both pots out with a plastic-safe no-residue contact cleaner while working the pots continually. For good measure, I also disconnected the pot/mod connector from the circuit board and cleaned it as well.

I then put the keyboard back together, plugged it all in, and the verdict: it's fixed!!! No more spurious pitch bend MIDI jitter, and the pitch bend range has been restored.

For those who might be curious, I'll post some pictures of the guts of the keyboard that I took while it was open.
Restored to what, may I ask?
0 to 127?
0 to 1023?
0 to 16,384?

JohnG


UW500; 2 x MU1000 + PLGs AN, DX, VL, VH & PF; SC8850; EWI4000s + VL70m; Samson Graphite 49.
Privia PX-560; AN1x; Novation X-Station 25. E-MU 02 cardbus + 1616m x 2. Finale 26; Sonar 7 PE and CbB.
XGworks v4; SQ01; SOL2. Garritan GPO5; COMB2; JABB3; IO; Organs; Steinway; World; Harps; CFX Concert Grand.
User avatar
YamahaUser
Member
Member
Posts: 125
Joined: Tue Apr 23, 2019 8:43 am

Re: Any way to calibrate the pitch bend wheel?

Unread post by YamahaUser »

SysExJohn wrote: Sat Feb 15, 2020 10:39 am Sadly, I rather suspect that the issue has absolutely nothing to do with the pots, but with the, in effect, analogue to digital convertor.
The erratic behavior experienced by ITFlyer certainly seems to have been due to poor contact of the PB pot's wiper to the resistance track. It's not an uncommon problem.

 
Yamaha has decided (in its infinite wisdom) that, since 7 bit resolution is good enough for all the other CC's, then it shall be 7 bit for ALL controllers, including pitch bend.

That is, not all these keyboards are faulty, but Yamaha in its pursuit to drive down costs to has decided to do away with 14 bits (or even 10 bits) on its implementation of pitch bend.

Call me cynical if you wish, but why else?
The issue of 7-bit PB resolution is nothing new. Yamaha Motif owners have complained about the stepping effect of pitch bend for many years. It's not much of a concern when the bending range is just a few semitones, but becomes obvious with more extreme bends. It's somewhat surprising that Yamaha chose to continue the "tradition" on the MODX and Montage.

As to why...
I don't have schematics (yet) for the MODX and Montage to compare with the Motif XF and MOXF models, but I suspect that certain parts of the circuit design of the earlier models were incorporated without change in the current ones. That's probably especially likely for controllers such as PB, MW, and the keyboard. Indeed, not having to redesign things is a cost-cutting measure.

The only other thing I can think of is data density. Pitch Bend generates a lot of MIDI events, and 14-bit resolution of course would create many more of them than 7-bit. Perhaps under certain circumstances there was the potential for MIDI data overload (more likely on the Montage, since it has aftertouch), and Yamaha decided to limit the possibility. (I realize that's not a strong argument, but without design details, that's all I've got.  :-)  )
Last edited by YamahaUser on Sat Feb 15, 2020 3:32 pm, edited 1 time in total.


Yamaha A3000V2, A4000, MEP4, Motif XS6, Motif XF6, YS200
Fender Chroma Polaris | Korg T3EX, 05R/W | Roland M-GS64, U-220 | Etc.
Macke
Global Moderator
Global Moderator
Posts: 229
Joined: Thu May 30, 2019 8:03 pm
Location.: Sweden

VIP

Re: Any way to calibrate the pitch bend wheel?

Unread post by Macke »

Here are some earlier threads about this issue as well as promised:

viewtopic.php?f=186&t=13690&p=80238#p80238
https://www.yamahasynth.com/ask-a-quest ... -pitchbend


User avatar
SysExJohn
Global Moderator
Global Moderator
Posts: 919
Joined: Tue Dec 02, 2008 1:00 am
Location.: Wellingborough, Northants, UK.
Contact:

VIP

Re: Any way to calibrate the pitch bend wheel?

Unread post by SysExJohn »

YamahaUser wrote: Sat Feb 15, 2020 2:19 pm The issue of 7-bit PB resolution is nothing new. Yamaha Motif owners have complained about the stepping effect of pitch bend for many years. It's not much of a concern when the bending range is just a few semitones, but becomes obvious with more extreme bends. It's somewhat surprising that Yamaha chose to continue the "tradition" on the MODX and Montage.

As to why...
I don't have schematics (yet) for the MODX and Montage to compare with the Motif XF and MOXF models, but I suspect that certain parts of the circuit design of the earlier models were incorporated without change in the current ones. That's probably especially likely for controllers such as PB, MW, and the keyboard. Indeed, not having to redesign things is a cost-cutting measure.

The only other thing I can think of is data density. Pitch Bend generates a lot of MIDI events, and 14-bit resolution of course would create many more of them than 7-bit. Perhaps under certain circumstances there was the potential for MIDI data overload (more likely on the Montage, since it has aftertouch), and Yamaha decided to limit the possibility. (I realize that's not a strong argument, but without design details, that's all I've got.  :-)  )
Pitch change is the one thing that the human ear is most sensitive to. Massively more so than e.g. volume.

This is why, when the MIDI 1 standards were being devised, Pitch Bend was given its very own message type, rather than being bundled in with all the other 32 CC's, with LSB and MSB transmitted separately. It needs more than 7 bits to sound smooth, even when used with the standard 2 semitone up/down range.

To relegate it to just 7 bits is stingy in the extreme by whatever company implements it, just to save a few pence in manufacturing costs.
Shame on you Yamaha!


UW500; 2 x MU1000 + PLGs AN, DX, VL, VH & PF; SC8850; EWI4000s + VL70m; Samson Graphite 49.
Privia PX-560; AN1x; Novation X-Station 25. E-MU 02 cardbus + 1616m x 2. Finale 26; Sonar 7 PE and CbB.
XGworks v4; SQ01; SOL2. Garritan GPO5; COMB2; JABB3; IO; Organs; Steinway; World; Harps; CFX Concert Grand.
User avatar
ITFlyer
Member
Member
Posts: 10
Joined: Wed Jun 05, 2019 4:33 pm

Re: Any way to calibrate the pitch bend wheel?

Unread post by ITFlyer »

So I was curious, and did a bit of research to see just what my various synths are sending for pitch bend data. I ran a MIDI monitor on each. Here's the results I found for resolution of pitch bend data.

ASM Hydrasynth: 14 bits. Both MSB and LSB populated.

Roland System-8: 8 bits. LSB alternates between 0 and 64, so going up I see MSB 64 LSB 0, then MSB 64 LSB 64, then MSB 65 LSB 0, MSB 65 LSB 64, MSB 66 LSB 0, and so on. The most significant bit of the LSB is the 8th bit (as you would expect)

Yamaha Reface CS: 8 bits. Identical to the Roland System-8

Yamaha MODX: 7 bits. But very odd behavior. With pitchbend from all the way down to center, LSB is zero, and MSB is the value being sent (from 0-64). Once it goes above 64, it also starts sending LSB data, with the formula apparently LSB = (MSB-64) * 2 - so the LSB increments at twice the rate as the MSB starting when it goes above 64, and continues until both MSB and LSB get to 127. It still is only 7 bit resolution however, despite this wacky behavior. I'm guessing they use this to make sure that at full deflection either way, you get either 0 or 16383 (otherwise the upper value would be limited to 2^14-127 instead of 2^14-1).

I'm very surprised that the MODX (and Montage) are limited to only 7 bit pitch bend data, whereas even the diminutive Reface sends at least 8 bit data, effectively doubling the pitch bend resolution of the MODX.


Post Reply