Matriarch "Glide Bug"

Mother-32, DFAM, Subharmonicon, Grandmother, Matriarch
Post Reply
mrgnbrnl
Posts: 4
Joined: Mon Oct 12, 2020 12:18 pm

Matriarch "Glide Bug"

Post by mrgnbrnl » Mon Oct 12, 2020 12:54 pm

Hi,

The issue is quite simple to reproduce : With a single triangle wave, just play low C0 then C4 and it's quite evident there is a "weird glide twanky" noise artefact. I'm not the only one having this issue, you can find some people complaining about that on Gearslutz. I asked Moog, they tell me to recalibrate the machine and of course it did nothing. Now they are silent they didn't answer. So do you experience this as well ? And could it be fixed ?

Thanks.

User avatar
MrBlue
Posts: 47
Joined: Mon Dec 23, 2019 3:54 pm
Location: Burgundy France
Contact:

Re: Matriarch "Glide Bug"

Post by MrBlue » Tue Oct 13, 2020 2:18 pm

Nothing like that on mine...
Music is emotion suspended in space and time
illustrasons

mrgnbrnl
Posts: 4
Joined: Mon Oct 12, 2020 12:18 pm

Re: Matriarch "Glide Bug"

Post by mrgnbrnl » Thu Oct 15, 2020 4:03 am

I finally get contacted by moog support, they send me firmware 1.2.2 beta. But the issue is still here. I’ll try to send an audio file because it’s quite hard to notice it at first but I don’t think it’s a normal behavior. That’s not a big deal but some people seems to have it and others don’t...

mrgnbrnl
Posts: 4
Joined: Mon Oct 12, 2020 12:18 pm

Re: Matriarch "Glide Bug"

Post by mrgnbrnl » Thu Oct 15, 2020 4:32 am

https://www.gearslutz.com/board/electro ... st14645151

Check the post from user tricera he describe it very well and there is a sound exemple attached.

Synth.builder
Posts: 40
Joined: Fri May 31, 2013 4:16 am
Location: Cumbria, UK
Contact:

Re: Matriarch "Glide Bug"

Post by Synth.builder » Fri Dec 04, 2020 6:21 am

Mine has this issue and I've been trying to work out what is the problem exactly.

I think it's two things.

Firstly, the envelopes are triggered slightly before the VCO frequency has changed to the new note. That is, you get to briefly hear the old note's pitch just before the new note's pitch. Normally, the envelopes should only be triggered once the VCO has changed pitch. I looked at a recording I made of the Matriarch playing two notes four octaves apart (bottom C and top C). You can clearly see that the top note has a bit of the low note's waveform for short time before the pitch changes. This gives the start of the new note an audible blip in the sound. If the interval between the notes played is small then this blip is less noticeable.

Secondly, there's something else going on. This is because the blip is also heard, but perhaps to a lesser extent, when the VCA is put into drone mode. In drone mode the envelope timing has nothing to do with what we are now hearing. So what's happening? I'm really not sure, but I think the oscillators are taking time to stabilise each time the frequency changes. The bigger the jump between pitches the more time it takes for them to stabilise.

Now all VCOs will take some time to stabilise. But is there something in the Matriarch that makes for a slower response to pitch changes? It could be the glide circuit, but it could also be something else, perhaps in the VCO's exponential convertor circuitry.

I tried out a few things. Firstly, I found that the 'pitch in' socket to each of the VCOs is band limited in some way. That is, increasingly higher frequency modulations are filtered out. So fast changes are slewed - a bit like a glide circuit that you can't change. The linear FM input is not limited in the same way, although it is AC coupled which means it won't respond to slow moving or static CV inputs. Could the VCO's input circuitry be causing the unwanted glide? Perhaps.

I've no real answers sadly. The envelope timing could be fixed in a firmware update. In my view pitch CV must be established before the gate is activated. But the source of this other blip is still a mystery to me. It could be the inbuilt band limiting applied to the exponential convertor (hardware design), but it could be update rate of the CV control (hardware and/or firmware design).

I've sent Moog support an e-mail with my findings.

Synth.builder
Posts: 40
Joined: Fri May 31, 2013 4:16 am
Location: Cumbria, UK
Contact:

Re: Matriarch "Glide Bug"

Post by Synth.builder » Fri Dec 04, 2020 12:51 pm

So this afternoon I have spent some more time trying to work out what this slight blip is each time a note shifts from low to high. I have found that the pitch of the VCOs take a noticeable time to move from a low frequency to a high frequency. This is essentially a small but noticeable glide that you can't turn off.

I now think I know what is causing it.

Firstly, I controlled an external VCO and an external envelope with the Matriarch's keyboard control voltage (CV) and gate output. No significant blips were heard. This means that the problem is not the glide circuitry of the Matriarch. When the Matriarch's glide control is fully counter clockwise the pitch CV moves fast enough so as not to cause noticeable slides.

Secondly, I connected the Matriarch's VCOs and envelopes to an external midi-CV convertor (Mutable's Yarns). This time playing a four octave spread of two notes on my controller keyboard produced the blips from the Matriarch's output. This suggests that these tiny glides are being generated by the Matriarch's VCOs.

I previously speculated that the envelopes are being triggered too early and that the pitch CVs were being applied too late. I now believe that this is not the case. It is the VCOs that are lagging behind because of their inherent sluggishness to the CV controlling the pitch.

I believe that these glides are either caused by slew limiting in the VCO control circuitry, or that there is some sort of thermal catch up going on in the exponential convertors. I hope it is the former since this could be fixed with some component changes. Sadly I don't think this is something that could be fixed with a firmware update.

Maybe it is time to just embrace it as a feature and live with it. However, I do feel a need to take a peek inside and see what can be done.

hudri
Posts: 29
Joined: Tue Sep 08, 2020 7:40 am
Location: EU

Re: Matriarch "Glide Bug"

Post by hudri » Sat Dec 05, 2020 5:53 am

Nice findings. I have also looked at it with the oscilloscope when I heard about the issue (but left it at that as it did not perturb my workflow).
What I found out:
* The glide circuit itself is fine - measured at KB CV OUT, rise time for 4 octaves is around 200μs with little overshoot.
* VCOs (measured at WAVE OUT) take around 2ms to adapt to the new frequency, which is enough to be audible.
* Envelopes have nothing to do with it, as it also occurs when playing legato (with Multi Trig off).
* Neither do mixer, filter, delay or VCAs, as you can connect the audio output directly to VCO WAVE OUT and it will still be audible.

Another thing I wondered is why the Grandmother supposedly does not have the issue although it has (probably) very similar VCOs. I think the main difference is that Matriarch has additional gates at the VCO outputs, used for Paraphonic modes.
And that could also be a partial mitigation in firmware: If the gate is closed before triggering a new note with big frequency difference, open the gate 2ms after setting the Osc CV. That means the osc has stabilized at the new frequency when the gate opens. But it doesn't help when playing legato, or with long envelope Release.
Synth.builder wrote: Fri Dec 04, 2020 6:21 am I've sent Moog support an e-mail with my findings.
Good luck with that, I have sent them around 10 bug reports, partly with solutions or workarounds. I have NEVER received an answer from them, so I gave up sending them anything.

Synth.builder
Posts: 40
Joined: Fri May 31, 2013 4:16 am
Location: Cumbria, UK
Contact:

Re: Matriarch "Glide Bug"

Post by Synth.builder » Sun Dec 06, 2020 7:16 am

hudri wrote: Sat Dec 05, 2020 5:53 amGood luck with that, I have sent them around 10 bug reports, partly with solutions or workarounds. I have NEVER received an answer from them, so I gave up sending them anything.
That is hugely disappointing. My first e-mail to tech support was replied to within 24 hours and with a usable solution to my problem. But my last two requests have apparently fallen on deaf ears, so what you say is not surprising. What a shame. Small companies like Moog can learn a lot from listening to customers, many of whom will have a wealth of knowledge to share.

Part of me wants to go inside again and take a peek at the circuitry that drives the VCO core. It could just be a poor choice of feedback capacitor in the CV summing circuit. My worry is that with all this CPU control inside, particularly regarding VCO tuning and scaling, that any changes I do make may mess with the VCO calibration routine.

That said, I may just to sell the Matriarch and build something similar myself.

Synth.builder
Posts: 40
Joined: Fri May 31, 2013 4:16 am
Location: Cumbria, UK
Contact:

Re: Matriarch "Glide Bug"

Post by Synth.builder » Mon Dec 07, 2020 12:33 pm

Just an update on this.

Moog did get back to me today and confirmed that this is indeed a hardware issue and not something that can easily be fixed.

User avatar
ummagumma
Posts: 827
Joined: Sun Aug 21, 2016 8:37 pm
Location: Canada

Re: Matriarch "Glide Bug"

Post by ummagumma » Mon Dec 07, 2020 4:25 pm

that is interesting, thanks for posting the R&D!

well, I guess it would just be the R... 8)

mrgnbrnl
Posts: 4
Joined: Mon Oct 12, 2020 12:18 pm

Re: Matriarch "Glide Bug"

Post by mrgnbrnl » Tue Dec 15, 2020 5:58 pm

Hardware ? That means every matriarch should behave like that... Moog must find a solution. The « glide » is really too obvious with high pitched notes. I don’t understand why Matriarch designers/testers didn’t notice it.

Post Reply