MUFF WIGGLER Forum Index
 FAQ & Terms of UseFAQ & Terms Of Use   Wiggler RadioMW Radio   Muff Wiggler TwitterTwitter   Support the site @ PatreonPatreon 
 SearchSearch   RegisterSign up   Log inLog in 
WIGGLING 'LITE' IN GUEST MODE

[Kits Available!] Eurorack Klee Panel and Daughterboard
MUFF WIGGLER Forum Index -> Music Tech DIY Goto page Previous  1, 2, 3 ... 54, 55, 56 ... 58, 59, 60  Next [all]
Author [Kits Available!] Eurorack Klee Panel and Daughterboard
trigger303
Quote:

- Check if the opamp's decoupling capacitors have correct values and are correctly installed.
- Check if other parts around the opamps have correct values and are correctly installed.
- Check if swapping CD4034s changes which channel has noise.


I've checked above, but it doesn't seem problematic.
Now I'm clarifying the area where the noise is detected.


By the way, I noticed a strange behavior.

When both Random Level/Random Reference are fully CCW, Random LED is not lit.
But when I connect the external clock pulse to Clock input jack, Random LED is lit.

Is this the right or possible behavior ?
(I don't think so...)
NV
trigger303 wrote:

I've checked above, but it doesn't seem problematic.
Now I'm clarifying the area where the noise is detected.


To clarify, swapping the CD4034 ICs with each other did not change which channel was outputting noise? Which channel were you experiencing noise in both before and after the swap?

Quote:

By the way, I noticed a strange behavior.

When both Random Level/Random Reference are fully CCW, Random LED is not lit.
But when I connect the external clock pulse to Clock input jack, Random LED is lit.

Is this the right or possible behavior ?
(I don't think so...)


This is normal behavior, the thresholds of the CD4034 are slightly different when clocked than when not being clocked. The difference should be very slight, and with nothing plugged into the random jack turning the level knob up past reference should still set the LED fully on when the Random switch is flipped up.
trigger303
Quote:

To clarify, swapping the CD4034 ICs with each other did not change which channel was outputting noise? Which channel were you experiencing noise in both before and after the swap?


Sorry for lack of information. Swapping / exchanging CD4034 ICs didn't change anything, it was same result.
Now I rechecked which jack outputs a noise and how much amplitude it has.

CV A Output
2.0 Vpp

CV A Output (Glide)
0.2 Vpp

CV B Output
0.2 Vpp

CV B Output (Glide)
0.04 Vpp

CV A+B Output
2.0 Vpp

CV A+B Output (Glide)
0.08 Vpp
‚Äč
Consequently, all the jacks more or less output noises.
I may have missed the small noises when checked before...sorry if I was confusing you...

In this measuring, I set the Klee to these states :
===========================
All the programming pots are maximum, all the step switches are down, and no active stages.
16x1, Pattern, and Normal mode.
BUS assign swithces are all at BUS 2.
Range is set to M3rd.
GlideA and GlideB pots are fully C.C.W, and GlideA+B pot is fully C.W.
RandomLevel, RandomReference, and VariableRange pots are fully C.C.W.
Load Bus1 and Merge1-3 switches are all down.
External Random, Load, Range, Clock are all down.
===========================



And in addition, now I understand about the CD4034's threshold. Many thanks !

Quote:

with nothing plugged into the random jack turning the level knob up past reference should still set the LED fully on when the Random switch is flipped up.


Yes, when the Random switch is up, turning the RandomLevel knob up surely sets the Random LED on.
But while the clock is being supplied, Random LED is lit although the Random switch is DOWN.
Is it also OK?
NV
Sorry for the delay, I was called out of town for awhile.

The Random sounds like it's working fine, it's an analog circuit handling logic along with other features so at times it can behave a little different in one setting than another. The Random switch is intended to handle the input jack for Random while the thresholds still remain active, so it just sounds like your particular chips tend to have a threshold that's a bit more sensitive to changes. On my Klees for example, if I turn Random Reference all the way CCW the LED lights regardless of Random switch position. Moving the knob up a bit puts it into more predictable behavior.

While all of the jacks are outputting noise, it looks like output A has about 10x the amount. Since the two glide outputs have notably less I'm assuming they are essentially acting as a filter of the noise. What is the frequency of the noise you're detecting?

At this point I think you have to pick through the circuit path from the output A jack back until you no longer detect any noise. Make note of any considerable changes along the way, either a change in the frequency or amplitude of the noise. My first reaction to all of this is that one of the opamps is oscillating, and I'm assuming it's somewhere between the CD4034 and the output jacks. This isn't a problem I've encountered on another Klee before so it's likely there is a component somewhere along the way that is the problem - an incorrect resistor, faulty cap, solder bridge, etc. Fortunately there isn't a ton of circuitry there, and you should be able to find distinct points where the noise appears to be gone. Mark those points up on the schematic to make a map of where it might be located.

If lifting pin 1 of analog U8 caused the noise to vanish earlier, then that narrows it down a bit. However if I recall correctly you said you replaced U8 with a fresh TL074 and the problem persisted.

Another thing to try would be swapping the analog U1, U2, U3, and U4 ICs. I don't think this should be the problem, but at this point the easy solutions are out and we need to dig into the more unusual ones. Swap U1 and U3, and swap U2 and U4, then listen to the noise. Make sure to note the amplitude of the noise before and after the switch - if channel B becomes the louder of the two instead of channel A then you know it's one of those ICs.

If that changes nothing, then again it must be something between the output A jacks and the CD4034s. It'll just take digging through and finding where the noise starts and stops and where it changes, then making sure that every component along the way is the right value, installed properly (no bridges or cold joints) and functioning correctly.
trigger303
Thanks for replying !

Quote:
The Random sounds like it's working fine


I'm relieved to hear that.
I was wondering whether this majorly concerns noise problem or not,
but now I can forget about it.


And as always I appreciate your advice !
I feel that my investigation so far is desultory and discursive
because I didn't have enough knowledge about this circuit.
Now I think I can check things more organized and accurately
thanks to your many adivices !

I haven't rechecked things enough yet.
I will inform you if there is any progress.
trigger303
Quote:

What is the frequency of the noise you're detecting?


Sorry, I missed your question !
I attach two pictures which shows the waveform of the noise.
In both magnified and demagnified range.

Magnified

Demagnified

According to the magnified picture, it indicates the frequency of the noise is about 43.1 kHz.

As for the demagnified picture, the horde of the noises seem to occur periodically,
and there are silent intervals every 3.9ms.

I have no idea of this silent interval, but it may be important clue if you see this.
NV
Oscillation of any kind that isn't purposeful (running the clock at high frequency) is unusual in the Klee. This is the first build I've encountered it in, so it's somewhat uncharted territory.

I think the solution is still tracing the circuit back to determine where the noise stops and starts and then resoldering components, double checking values and IC/component orientations, essentially making sure that any potential error hasn't been overlooked. The fact that swapping various ICs didn't change the behavior suggests that there is a component error somewhere, and determining where the noise starts and stops will be the best way to narrow down your search.

After you've determined the points between which the error must be occurring in the schematics, finding the error means digging in and making sure all the components are the proper value and placed correctly, that there aren't any cold/missed joints or bridged traces/pads, that signals are being carried through the component, etc.
sduck
Almost finished building this up. Tonight I finished up the front panel assembly, and then had some fun with my lytro illum camera. This high tech camera is basically abandonware, so the promised click to refocus images aren't exportable, but I can make these little videos, and so here's one.

https://u12150305.dl.dropboxusercontent.com/u/12150305/IMG_0111.mp4

Edit: oh, duh. Figured out how to make animated gifs:

NV
Fantastic video sduck, thanks for sharing it! I've only barely heard of the Illum so it's great seeing more of it in action.
sduck
...and, it's alive! Finished it up, first testing seems to indicate most things working correctly. My top row of leds is not working, but hopefully that's not a big deal to sort out. Still have to calibrate things.

Thanks (once again) for creating this Much Easier version of the klee! It almost felt like cheating when I did what was the moral equivalent of the wiring in one evening, instead of the 2 weeks it took previously.
sduck
Finally got around to replacing the stage leds to ones that worked, and calibrating. Everything seems to work perfectly!

Calibration: I couldn't get range 3 to get all the way to -.583v - I had to add on another resistor to R28 to lower it's value a bit. I think I used a 270K. Perhaps R28 should be 100K instead of the 150K? Also, I couldn't get range 5 to sit right at -1v - it seemed to want to jump between -.992 and -1.008 or so. There was quite a bit of play between the 2 places where it would jump, so I put it right between those 2 places, presumably it'll be ok.
NV
sduck wrote:
Calibration: I couldn't get range 3 to get all the way to -.583v - I had to add on another resistor to R28 to lower it's value a bit. I think I used a 270K. Perhaps R28 should be 100K instead of the 150K? Also, I couldn't get range 5 to sit right at -1v - it seemed to want to jump between -.992 and -1.008 or so. There was quite a bit of play between the 2 places where it would jump, so I put it right between those 2 places, presumably it'll be ok.


The trimpots should allow for some space around the target voltage. Each of the trimpots pull from the same 10V reference, so if one is coming up short that's unusual. It might be an issue with the resistance in the pot itself. As you already did, swapping R28 to a different value should give it a wider berth.

With range 5 also being jumpy it seems like there might be an issue with the batch of trimpots you're using. If the imprecision of the 1V interval starts to bug you I'd try switching that trimpot out to a different brand/batch as a first step.
sduck
For the range 3 thing, it was so close (I think it got to -.570v), I suspect the resistor or the trimpot were just a bit on the edge of their tolerance range. Not a big deal. Range 5, again not a big deal - in actual practice it works fine.

Now I'm deciding if I want to mess with the LEDs any more or not - I got the top leds working, by switching them all out to a different kind, but now the fader leds don't work. Not really a big deal either. Bottom line - the whole thing works!
Sleipnir
sduck wrote:
Now I'm deciding if I want to mess with the LEDs any more or not - I got the top leds working, by switching them all out to a different kind, but now the fader leds don't work. Not really a big deal either. Bottom line - the whole thing works!

I just got the full Synthcube kit, and you are making me think about testing all the LEDs before I turn on the soldering iron. MY ASS IS BLEEDING
sduck
I'm pretty sure synthcube has tested their led choices. I didn't - it is just a matter of voltage requirements. The first ones I had for the top row worked fine, just needed more juice that the klee provides. The new ones are probably a bit too efficient, and are sucking up the juice that the fader leds need. It's just a matter of balance.
NV
sduck wrote:
I'm pretty sure synthcube has tested their led choices. I didn't - it is just a matter of voltage requirements. The first ones I had for the top row worked fine, just needed more juice that the klee provides. The new ones are probably a bit too efficient, and are sucking up the juice that the fader leds need. It's just a matter of balance.


Are you using the slider LEDs as constantly on or as lighting in sync with the bits? If they're constantly on then they're just pulled from the +12V rail, so using the 4.7k resistors in the "Con." spaces should light them all up fine regardless of what the Klee is doing.

If you're using them as bit synced then they're pulling from the same CD4034s as the bit LEDs, so the slider LEDs will run a bit dim depending on the efficiency of the bit LEDs. You can improve this by swapping the LED resistors on the main board. These are Digital R32 - R47.

The CD4034s have an absolute maximum power dissipation of 500mW, but ideally you'd give that some wiggle room. If you want to really maximize the brightness then here's a formula you can use. Note that you should have your two LEDs (bit and slider) at around the same forward voltage, which they should be if you're using standard diffused and not intermixing with blues/whites/UVs or ultrabrights:

(((PSU voltage - lowest LED forward voltage) / LED resistor value) x PSU voltage) x 8 channels = mW through the CD4034

For example, using 2V forward voltage LEDs (around what the included green slider LEDs run) with the standard 4.7k resistors and running all 8 channels at once, we get this:

(((12V - 2V) / 4700 ohm) x 12V) x 8 = 204 mW

Keeping it under 350 mW should be safe - the CMOS will draw more current during transition states and the LEDs also present a capacitive load to the IC, meaning you're not quite as far from the top as it seems. So with that in mind I'd say you could go with 3k resistor LEDs and still be solid (320mW). If you wanted to push things a bit further you could experiment with lower values but the more mW the more it edges towards burning a CD4034.
sduck
Yes, using them as bit synced. I figured it was just a matter of adjusting those resistors, but wasn't aware there was a possibility of blowing the 4034s - thanks for the warning! I'll probably pull one of them and throw a temp 10K pot in there, with a 3K resistor in line, and see what value will work best before switching them all.
romanb
Hi All! Hi NV!

I felt the urge to tell you how thankful i am for your work on making this sequencer a relatively easy build.

I have been making music on computer with Buzz for almost 2 decades now and only got started with modular after demoing the TINRS Tuesday at Superbooth 2017 for three days at my friends booth. He showed me the essentials of modular, which I never really tried to understand before and then launched me to the public demonstrating his Tuesday procedural sequencer. I had so much fun, after Superbooth I was infected with the modular virus and I NEEDED to start building my own rack. It's just too much fun to pass on! It's going downhill, fast! (note: downhill as in moutainbike downhill FUN!)

After building a dozen or so diy full kits from various manufacturers, I went to look for a decent sequencer, only to be disappointed of what I saw was available. Of course, completely unjustified, considering the virtually limitless possibilities I have in Buzz and me having no real clue as to what was available for modular and in hardware generally. I stumbled upon a video of the Klee on youtube and after watching a couple more of the rare Klee videos, I was convinced that I want one.

For quite some time I looked around, found no full kit anywhere, so I went for the partial kit from modularaddict, only to discover a few days later that synthcube restocked on the full diy kit just after I had ordered the partial kit...

"yeah, whatever... parts sourcing practice!"

After days of searching for parts, I started to get frustrated and finally settled for cheating and took some mouser carts from this thread, updated with available parts and non-minimum order quantity of 1000s of pieces for some parts, and finally ordered the parts after many checks and rechecks to make sure all things are proper...

I received the parts last saturday and started building the analog/digital board, which to my surprise was pretty easy to build.

That said, I want to heartfully thank everyone who had their hands on developing this beauty and for the time you spent on making this easy build a reality.

THANK YOU! we're not worthy

as of this moment, I only need to get some LEDs, the knobs and I will probably have to drill the holes a bit wider for the buttons to 16mm, so that my shiny, heavy duty buttons that have been laying around for about two years without purpose finally have a worthy device to be installed on. smile

below, have some pictures of the not quite finished Klee, hopefully soon to sport some shiny buttons... smile
(beware: big sizes!)
synthcube
glad you found your way to the klee- in our experience its among the best DIY kits available!

enjoy!
NV
Great to hear you're enjoying the build romanb, thank you for the kind words! If those buttons don't line up quite right with the solder pads you can solder some bare wire to the lugs and guide them in that way. This also makes it easier to pull the panel if you need to do some troubleshooting, so I tend to recommend it either way. Let me know if you run into any problems and we'll get your build up and running!
romanb
Build Update: IT'S ALIVE!!!

The Klee Sequencer has been on the operation table for a while now. It's been power shorting the test rack and on further analysis and probing, found a diode that was reversed and letting power go straight to ground.
I ordered a few replacement diodes (2x BAT85 along some 100x 1N4148 and 100x 1N914 because, you know, can't have enough of these). The replacements arrived, replaced on the board aaaaaand... it just kept shorting. o____o

In my desperation to revive the Klee, I simply thought I will just have to recheck -EVERYTHING- again and see what I find out.

After some hours of multimetering the shit out of every connection, following all power trails, finding some shorts which I had found before too, I was a bit at a loss first.

And then I thought about how the power cycling started... I plugged in and out of the random jack a couple of times, and when unplugging it had started shorting.

Stijn, who remote supported me with the troubleshooting on skype, first thing he told me back then was "check the random plug socket...". By then I had already spent quite some time doing some checks and stuff, so at the time he told me that, I was tired and under the impression that I already did so and I did not want to repeat the procedure of checking stuff I already checked.

Now that I had the replacement installed and the Klee keeps shorting, checking almost everything again, finding some shorting power trails but not finding the actual cause, I went on to check the random plug alone. And indeed, 3 of it's 4 feet were shorting each other!!!!!! (Rng+Tip+Gnd).

Thinking back when I was building it, I remembered when applying solder, on one plug it seemed like an endless pit, sucking up solder and not closing the hole.

I went on to remove the solder until the hole was back, and then just reflowed it with what was left, applied a little bit more solder, and let it cool down. Multimeter did not beep anymore. Also checked all the previous random power shorts on the boards, no beeps... So I connected power from the test rack and turned it on... No power cycling. I had found the error now and fixed it. YAY!

I went on and built it back together, soldered the pots' wires back on, soldered the buttons' wires back on, put it back together, all the while in skype, Lauri, who was helping me before sort of text-yelling at me "DON'T DO THAT! Test it extensively first! NOOOOOOOOO...!!!"

I did not listen. I was desperate. I finished putting it together. I mounted it on the test rack. I was euphoric, because I finally found the problem, which was what Stijn had told me to check from the very beginning, but I was too tired to do again... and I decided to leave it be for now... Lauri was right. I should properly test it before I declare it fixed. My noobness needs to learn. And I have two masters of their fields behind my back helping me with this stuff, I really shoud listen...

I went back to the test rack and decided to properly test the random jack socket. I connected power, no cycling. I connected a cable into the jack socket, *POOOOFF* rack powercycling!

"SEEE!!??!?!?" Unicorn told me, "I told you to look closer! You did not listen! But nooooooooo... You want to do it your own way! You want to do the same mistakes THEY have already done and you want to just repeat them..."

Bunny was cowering in the corner, bandages splattered with dark brown blood and dirt, visibly starting to fall off all by itself... Eyes in tears and terror, looking around with a nervous gaze... Hedgehog was nowhere to be seen.

Chicken, as usual, was completely oblivious of everything that had happened. Chicken just clucked happily, hopping from grain to grain, picking them up one by one...

"You have to listen to them! If not, at least take their words into consideration! They know what they are talking about."

There was nothing visibly wrong with the random socket. I decided to reflow it again. Multimeter not beeping. Rack power normal. Cable into random socket *POOOOFF* rack powercycling.

Clearly, there is still waaaay too much solder in there somewhere, shorting Gnd+Tip. Removed some more. Reflowed. Multimeter no beep. Rack power, cable into socket... nothing... Power still on. Pressed cable some more, and power cycles again. Huh!

"I'm telling you, you have to look closer! Listen to them!" - "Shut up, Unicorn!" Hedgehog grumbled. "I heard what they said! I know! But I will learn my own way! I always did! Shut up!"

More solder removed. The wick felt like it weighted 35kg by now, but I kept removing solder, until there was none anymore. Multimeter no beep. Rack power no cycle. Cable into socket with pressure, no cycle... Apparently there is not even any contact anymore. Applied a limited amount of solder, reflowed all the other sockets just in case. Tried to repeat the problem, but it did not happen anymore.

I brought the test rack to the big rack and started the bring up procedure from the excellently written manuals. After a while I repeated some steps, simply because it was fun to see the Klee working again, all safe and alive, blinking lights reflecting off the shiny buttons and sexy knobs.

After a while, I finally started calibrating the ranges, which took me just a couple of minutes and... Well... It is done.

The Klee is finally finished, tested and calibrated!

WOOOOOHOOOOOO!!!

Oh! What's that? I have an unopened package somewhere...

Time to play the Klee for a while, then sleep and then on to the drum modules!

NV
Thanks for the update romanb, glad you got it all sorted and working! Fantastic aesthetic choices as well, the silver look is really sharp. Enjoy your successful build!
underthebigtree
Hi all -

I finished and calibrated my Klee some time back, and to the best of my knowledge, it is functioning perfectly. What an incredible instrument!

I do have one question, that I hope someone here can help with. I am trying to use the load in jack to generate odd time signatures with the Klee. I can get certain devices to trigger it properly, but the ones that I want to use won't work - the Klee ignores the load pulse.

It must have something to do with either the amplitude of the pulse, the shape of the leading edge of the pulse, or the length of the pulse.

Here are pulses that successfully trigger the load in:
Pamela's workout
Bus out from the Klee itself
A very short AD pulse from Maths
Doepfer A-155 trigger out
Pulse out from Music Things Turing Machine

Here are pulses that do not successfully trigger the load in:
Trigger out from Synthrotek Sequence8 (which is what I'd really like to use)
Trigger out from Doepfer clock divider/clock sequencer (A160-A161)
Any LFO square wave

Any help would be greatly appreciated!

-Nick
NV
underthebigtree wrote:

Here are pulses that do not successfully trigger the load in:
Trigger out from Synthrotek Sequence8 (which is what I'd really like to use)
Trigger out from Doepfer clock divider/clock sequencer (A160-A161)
Any LFO square wave


The Load input doesn't require a sharp edge, it just needs a voltage that clears a +2V threshold in a positive-going direction. This means even a slow LFO sine wave or envelope should trigger it. I just checked mine with a 1Hz sine offset to oscillate between +1.5V and +2.5V and it triggered consistently.

It's especially odd that the A-160 won't trigger it, since Doepfer modules put out very strong gates and the A-160 holds them for much longer than most triggers, which should rule out other sensitivities. I'm reading 11.3V gates on my A-160, held for the entirety that the LED is illuminated.

Is the behavior consistent with the Klee never loading at all with any of the A-160 outputs, or is it intermittent or with particular outputs?

Do the successful triggers always work consistently, or do they display intermittent behavior?

Do you have a multimeter to compare voltages? Try running one of the successful square pulses at a low rate and through a slew into the Load jack. It should function just fine even at sub-Hz rates and with enough of slew to turn it into a saw, but see if there's a point where it stops triggering. If the successful pulses still work even when you mess them up to that extent, then something must be different about the other pulses.

For the pulses that don't work, try experimenting with different levels of offset and attenuation to see if there's something that causes them to work, likely slowly attenuating the Doepfer A-160 outputs to see if the load starts triggering and at what voltage. If so then we might be able to figure out where the circuit isn't up to snuff in the build. The load signal doesn't go through much (Digital U1, U2, U3, and U5, along with a handful of resistors/caps) so it's possible one of the chips it's running through is doing something to the signal.
underthebigtree
OK, mystery solved - but still need to find a solution for my problem. But at least I know what the problem is now.

So I dug my ancient oscilloscope out of mothballs, and found a tiny space to set it up. Proceeded to measure the pulses from the various systems. Hey they are, for your entertainment:

Pamela's Workout: 5v 5ms square wave
Klee bus out: 10v 1ms square wave
Maths AD envelope (fastest setting): 10v 2ms triangle wave
A-155 trigger: 8v 6ms near square wave (trailing edge slightly higher than leading edge)
Turing Machine pulse: 7v 5ms square wave

So all within a reasonable distance of each other.

Now we turn to the Sequence8, and the A161. The problem is immediate and obvious: They are both gates, not triggers. The period of a voltage change was in the hundreds or thousands of milliseconds - couldn't get a clear reading on my old scope, but that is the deal. The Sequence8 was also only putting out about 3V.

Now all I have to do is figure out how to convert a gate to a trigger, and amplify it, or find some other piece of gear that is a clock divider that deals in odd numbers, that can output a true trigger pulse. I'll start looking at Modular Addict and Synthcube to see if any small, inexpensive kits are available to do this.

Many thanks for pointing me in the right direction,

-Nick
MUFF WIGGLER Forum Index -> Music Tech DIY Goto page Previous  1, 2, 3 ... 54, 55, 56 ... 58, 59, 60  Next [all]
Page 55 of 60
Powered by phpBB © phpBB Group