[SOLVED] filter cutoff going down

Discuss x0x construction and related issues

Moderators: altitude, adafruit_support_bill, adafruit, phono, hamburgers

Please be positive and constructive with your questions and comments.
Locked
User avatar
antto
 
Posts: 1636
Joined: Thu Apr 15, 2010 3:21 pm

[SOLVED] filter cutoff going down

Post by antto »

after building the x0x, i had it calibrated to my other x0x, so when i set the CUTOFF knob to max on both of them - they resonate at the same frequency

last night i had the weird impression that the filter is going up on its own :shock:
actually, it does, only that when i compared the two x0xes - the new one had exactly half the cutoff frequency compared to the other one, and it was randomly going back up to where it should be, and then back down to half way

i measured the cutoff pot, but didn't see anything wrong there

it also noticed that setting the ACCENT to max sometimes makes the cutoff jump up to normal at first, and then it starts going back down to half way

looked in the voltage map, the first thing i decided to measure was the link which goes from the accent pot thru D24 to C13
and at D24 i get a lower voltage than on the other x0x
the voltage perfectly matches the behaviour of the cutoff frequency - it fluctuates up and down which shows on the multimeter too

so i immediately replaced D24
but the issue is still there :?
Last edited by antto on Tue Jan 14, 2014 4:02 pm, edited 1 time in total.

User avatar
aminoacid
 
Posts: 352
Joined: Tue Jun 20, 2006 5:27 am

Re: filter cutoff going down

Post by aminoacid »

If the cutoff goes up and down and that is a new build xox then perhaps its the new electrolytes that need some burn in time to settle...

User avatar
antto
 
Posts: 1636
Joined: Thu Apr 15, 2010 3:21 pm

Re: filter cutoff going down

Post by antto »

C13?

User avatar
aminoacid
 
Posts: 352
Joined: Tue Jun 20, 2006 5:27 am

Re: filter cutoff going down

Post by aminoacid »

Just speculating...

Theres a couple caps in the vcf and even in the psu that could have effect on the voltages in the vcf.

If you wanna know for sure, just let the box be on for an hour or perhaps a day, then measure again.

You also should Keep them on for some time from start to get everything settle every time you start your machine. I think my xox settles after one or two hours. Dont like shutting it off :)

User avatar
antto
 
Posts: 1636
Joined: Thu Apr 15, 2010 3:21 pm

Re: filter cutoff going down

Post by antto »

basically, i never really switch off the x0xes
only when i have to reboot to flash firmware, or when i have to solder/desolder stuff

the x0x has been running day and night since i built it
and i had it looping a pattern when i first noticed the cutoff changing on its own

left it this night playing a pattern with accents, and with the accent set to max (this is a bit like hunting ghosts) but it is still lower than where it should be and still goes up/down by itself

i don't think it's the power supply, it outputs stable 12/5.33V (otherwise i'd hear the vco detune and other weirdness)

if i short C14 - at first the circuit isn't happy, but when it relaxes after that (C14 is still shorted) the cutoff is right where it should be
maybe a clue

perhaps i should try replacing C14
and if that alone doesn't do it, replace all electrolytic caps in the VCF section?

User avatar
antto
 
Posts: 1636
Joined: Thu Apr 15, 2010 3:21 pm

Re: filter cutoff going down

Post by antto »

you can see it in the first one minute of this video: http://www.youtube.com/watch?v=cq9nkVWJqVw
the blue signal

near the end of the video, i short C15 for a moment (i also poke in other places)

User avatar
antto
 
Posts: 1636
Joined: Thu Apr 15, 2010 3:21 pm

Re: filter cutoff going down

Post by antto »

i socketed D24 and C13
i also desoldered one leg of R72 (the one which goes to the reso pot)

so, with R72's leg lifted - everything is normal, the filter cutoff is right where it should be, and there's no drift
that means, the CUTOFF and ENVMOD components work as they should, but the "accent wow" is gone, there's only the vca bump on accented notes, which comes from R119

i thought maybe something on the pcb is not good, so i put together the missing part of the circuit (the accent wow) on breadboard
took the signal from the middle pin of the accent pot, then diode, 100k resistor (to substitute R46+reso_pot), and a cap to ground
then hooked that to the lifted leg on R72
result: it worked at first... but after some time it started drifting too

so, even if i remove D24 and C13 from their sockets, and just connect R72's lifted leg to its hole - i get the drift :shock:
and the drift then changes instantly, like a s&h LFO, randomly.. sometimes short glitches, sometimes long glitches
when C13 is in its place - it smooths it out

so it's not the diode, it's not the pots, it's not the caps, it's not the ladder... i am completely out of ideas
:(

User avatar
aminoacid
 
Posts: 352
Joined: Tue Jun 20, 2006 5:27 am

Re: filter cutoff going down

Post by aminoacid »

If you have a scope you can perhaps check around where this fluctuation is and where it isnt.

Id check the current zinking fet in the envelope and parts around it, esp the switching ic.

Id have a look at the q9 and parts around it.

And even susspect cutoff, env or perhaps acc pots are bad.

Id start their at least. And try to investigate if the fluctuation is when notes are on or off, and other behaviors like pot settings.

User avatar
antto
 
Posts: 1636
Joined: Thu Apr 15, 2010 3:21 pm

Re: filter cutoff going down

Post by antto »

the very first thing i suspected was pots being bad (since it was new, and the knobs required a not-very-nice amount of push to get on the pots)
but i tested the pots immediately, compared the measurements agains the working x0x - no difference there
i measured the cutoff, reso, envmod and accent pots

the only place i find the fluctuation is in the path from R72 (one side only) all the way thru r46 till one side of D24
so this means:
R72 (the point connected to the middle pin of the reso pot), C13, R46, D24

the reso pot is okay

also, when i remove C13 and D24 - what's left is the hole on R72 (i've lifted the leg) which is connected thru the resonance pot, which goes to nowhere
yet, if i connect that hole to R72 on the other x0x - the other x0x starts to drift exactly the same way

pot position doesn't matter, and this drift is there no matter if playing notes or not (appears randomly)

ehm, when R72 is with the lifted leg - no drift (but also, no accent "wow" of course)
when R72's leg is connected, there is the drift, even if C13 and D24 are removed :shock:


i'll socket the FET in the envelopes section, i got some spare J201 which should work there

User avatar
antto
 
Posts: 1636
Joined: Thu Apr 15, 2010 3:21 pm

Re: filter cutoff going down

Post by antto »

oh, and something interesting
with R72 with the lifted leg - if i connect it to R72 on the other x0x - i get the "accent wow" from the other x0x on both
and, no drift, everything works

http://www.youtube.com/watch?v=wTdES8vjO2A

User avatar
aminoacid
 
Posts: 352
Joined: Tue Jun 20, 2006 5:27 am

Re: filter cutoff going down

Post by aminoacid »

So then perhaps its some kind of short on the pcb...

Have checked thoose leads from r72 to d24, if they are intermitting with other leads close-by and/or that runs under thoose components? För ex under the res pot?

How correct english was that..? :)

User avatar
antto
 
Posts: 1636
Joined: Thu Apr 15, 2010 3:21 pm

Re: filter cutoff going down

Post by antto »

yes.. after i reduced the suspects down to the connection from resonance pot to one side of R72 (the hole under the lifted leg) and since C13 and D24 are removed
i probed from the hole under R72 for connectivity to to ground, to the 12 and 5.33V lines, and to random places around there and thru the whole board - no "beep"
i switched to measure resistance (to the highest range my multimeter can go - 2000K) again from the hole under R72 towards ground, the power lines, and random places - didn't find anything

that hole is only connected to the middle pin on the reso pot, as per the circuit

i socketed Q39, placed one J201 in there - didn't fix the problem, actually, the envelopes work regardless if Q39 is in or not :shock:
i wonder what its role is

i changed (again) Q9,10,11 - still nope

the only thing that stops the drift is lifting R72's leg up and not touching that hole (which means, i got cutoff and envmod control working, but no accent "wow" ..)

User avatar
aminoacid
 
Posts: 352
Joined: Tue Jun 20, 2006 5:27 am

Re: filter cutoff going down

Post by aminoacid »

Did you remove r46 aswell by the way?

Id check for shorts between r72 and the housing of the res pot, c14, cutoff pot, r61, all pins of res pot.

Id check for low ohms , not high. Or perhaps both.

The fet is draining the env constantly. There should be 0V there if no notes are playd.

User avatar
antto
 
Posts: 1636
Joined: Thu Apr 15, 2010 3:21 pm

Re: [SOLVED] filter cutoff going down

Post by antto »

after i re-did the test with re-routing the circuit between R72 and the middle-pin of the accent pot on breadboard - it turned out that it doesn't drift
in my first test, i don't know what i've done wrong, i was too frustrated or something

so the only component that was a suspect in that situation was the reso pot

now i got a replacement, soldered it in - the cutoff freq is where it should be, and so far i haven't heard any drift

User avatar
antto
 
Posts: 1636
Joined: Thu Apr 15, 2010 3:21 pm

Re: [SOLVED] filter cutoff going down

Post by antto »

update:
even tho the issue was solved - it reoccured... even on my old x0x which is not built by me, nor have i touched the reso pot with my iron
now the result is.. i got 5 of these "ALPHA" dual 50k linear pots, 4 of them are busted in this same way - the two halves are not isolated, there is a connection between them with varying resistance (from +inf down to ~60kOhms)
not only that, but it took just 1 week for one of them to get "busted" after it was carefully soldered and not even used (didn't even put the knob on)

so at the end, i found other pots which are pin-compatible, from another brand, their shafts are 5mm longer, and have a center detent (which is very easy to remove)
those are from ALPS "RK12L123000E"
soldered one of them in today, and it works
if anyone decides to go for these pots, have in mind that you'll probably have to trim the shaft 5mm down, and also, if you're not gonna use the BOM knobs - do notice that the shaft on that alps pot is round on the lower half, and this sticks about 4mm above the front panel, so make sure this isn't going to be an issue if you use some other knobs

comparison:
Image
green == alpha, blue == alps

Locked
Please be positive and constructive with your questions and comments.

Return to “Making x0x”