Author |
Message |
thresholdpeople
Joined: Jul 17, 2014 Posts: 22 Location: NYC
|
Posted: Thu Jul 17, 2014 7:14 am Post subject:
Just got a Klee! Two issues, I think |
|
|
Hello everyone. I just bought a built Eurorack Klee. It has not yet been calibrated, I'll do that today, so apologies if that is causing one of my two problems. I have no problem doing diy, it's just a Klee came up for sale for a great price when I had the funds for it -- meaning I don't mind pulling the boards apart and fixing stuff.
The first problem seems simple: the manual step button does not work. Whether the clock is stopped, or running, no matter the state of the bits. I haven't taken anything apart yet, but I can sort of see where the button is soldered in and it looks solid.
The other problem is a bit more complex: I am running 8x2, pattern, normal. I have CV going out of A. The only bit on is step 1 (or any step, or any amount of steps in section A). I get no sequencing control from the faders unless I flip a bit on in section B. The exact same thing happens if I have CV going out of B, and no bits switched on in section A.
As far as I can tell this is not the way it should function, right? It is happening completely reliable, so maybe..yes?
Any help is greatly appreciated!
Thank you! |
|
Back to top
|
|
|
elmegil
Joined: Mar 20, 2012 Posts: 2177 Location: Chicago
Audio files: 16
|
Posted: Thu Jul 17, 2014 9:05 am Post subject:
|
|
|
That sounds like somehow A & B are swapped. Verify the cabling. |
|
Back to top
|
|
|
thresholdpeople
Joined: Jul 17, 2014 Posts: 22 Location: NYC
|
Posted: Thu Jul 17, 2014 10:34 am Post subject:
|
|
|
I thought that (haven't tried it yet) but what throws me off is that side A's switches and faders do end up corresponding to CV A's output. Same thing with B. It's just that until the opposing side's bit is switched on, there's no fluctuation in voltage.
This is from Scott's circuit theory and operation post:
http://electro-music.com/forum/phpbb-files/klee_encoder_8_2_105.png
Could it be that the CD4053's at U9b & U9c are either shorted together... or maybe they are the ones that are reversed?
Apologies since I haven't actually gone through all of the build documentation yet as I received the Klee yesterday, but do the A's and B's shift registers attach to U9 with wires, or is it only PCB traces? |
|
Back to top
|
|
|
elmegil
Joined: Mar 20, 2012 Posts: 2177 Location: Chicago
Audio files: 16
|
Posted: Fri Jul 18, 2014 5:12 pm Post subject:
|
|
|
I just tested my Klee, and no, there is no requirement for the "other side" to have an active bit for "this side" to have CV output.
I'm curious if you set the A bits to zero and one B bit, and take the A CV out, what happens.... |
|
Back to top
|
|
|
thresholdpeople
Joined: Jul 17, 2014 Posts: 22 Location: NYC
|
Posted: Fri Jul 18, 2014 8:14 pm Post subject:
|
|
|
Hi Elmegil. Thanks so much for your help and time.
You were absolutely correct -- the two shift register switch cables were backwards.
It became painfully obvious this was the case when I followed the testing procedure.
The person who I bought it from sold it to me uncalibrated, I just thought the trimmers weren't touched!
The manual step button is just broken, I believe I can't get any continuity when it's pressed, and I traced the leads further down the line and that all seems fine.
A major issue though, that I (think) I've found is during the trimmer calibration. The trimmer pots for range 2 and range 5 are completely unresponsive. And furthermore, the voltage reading for range 2 and range 5 are exactly what they are for range 6. Using the range 6 pot, I can dial in the voltage values for the other two ranges. I traced around all the leads going from the rotary switch to the diodes to the chip. It seems all fine. Not 100% what I should look for there, but it seems the way it should be. I checked all the pots with the connections to the chip itself, that seems fine too. Switching the rotary during a sequence definitely produces changes for all the positions, it's just that those two positions are tied in with range 6. Any ideas there? |
|
Back to top
|
|
|
thresholdpeople
Joined: Jul 17, 2014 Posts: 22 Location: NYC
|
Posted: Fri Jul 18, 2014 9:10 pm Post subject:
|
|
|
Oh, also, I'm assuming that it's normal that when 2x8 (or 1x16)/ Pattern / Normal and I stop the clock mid sequence, and I manually reload and restart the clock,the sequencer jumps a few steps ahead.. maybe to the step it was on when I stopped it. Sometimes it I think starts just fine at 1, although I'm trying to get it to do so now, and it's always jumping. |
|
Back to top
|
|
|
elmegil
Joined: Mar 20, 2012 Posts: 2177 Location: Chicago
Audio files: 16
|
Posted: Fri Jul 18, 2014 9:24 pm Post subject:
|
|
|
You should give me step by step how you produce the jumping, that doesn't sound right to me but I'm not sure I follow.
I don't have any ideas off the top of my head about the trimmers. I have another one I'm building, if I find a bit of time this weekend I can try and see if there's something that jumps out at me. Have you pulled the analog board and examined it for shorts etc? That's pretty easy to do....
Good luck on the manual step switch though, that's going to be a real bear to get out of there. I think if I were to have to replace that I'd do everything I could to do so without removing the panel, because there are a lot of other things you'll have to get apart if you do that. I *think* if you clean the solder first, if you can get enough grip to unscrew the nut on the back side you ought to be able to lift the switch out, but I'm not sure about how feasible it is to do that unscrewing. |
|
Back to top
|
|
|
thresholdpeople
Joined: Jul 17, 2014 Posts: 22 Location: NYC
|
Posted: Fri Jul 18, 2014 10:31 pm Post subject:
|
|
|
I have a sequence running -- 2x8, pattern, and normal as the settings. I flip the switch to disconnect the clock, and the sequence stops (on whatever step it is on). Let's say I flip a switch to activate or deactivate a bit, and hit the load button. The LED will jump back to 1 on pattern A (I suppose I tend to have bit 1 on pattern A usually on, and I'm not sure I've tested this without it on....that'll be next, but I digress!). When I flip the switch to turn the clock back on, the LED will jump a few steps. I'm fairly sure that it jumps back to the step that it left off at, but honestly, it may not.. It's a little hard to follow with so many lights. It's mostly confusing to me, because once a sequence is going, I have no idea which LED 'is the 1 beat' so I'm assuming that it jumps to where it left off.
Sorry to be a bit unscientific about it. Just happy to have this running after spending a few hours taking it apart and putting it back together.
About the trimmers -- I traced a lot of the analog board and digital board and saw no shorts. Scott's scematics (the 8 page document, with everything broken down function by function) was really helping in assessing all of my issues, so I pretty much used that as a map starting from the IC and the pots, and working all the way back to the rotary switch, then back toward the IC tracing over the diodes, and into the 3 input selection pins on that IC. I check the wire connecting the rotary to the analog board. It seems fine. Perhaps the chip's gone bad?
I thought that the DC offset pot may have gotten everything out of range, so I tried re trimming everything many times, finding the 0V DC offset in different ranges of the DC offset pot, thinking that the trimmers that weren't working were just out of range, but they are totally unresponsive. The only thing that I didn't do is have a meter on them while trimming them, but I did get resistance values between them and the IC. Can't remember if it was unique for all three, or the same (as would be expected... I even reflowed the solder on most everything in the signal path.. nothing.
Thanks for the advice on the switch change over. Pulling it out may be doable, but from what saw fitting the new one in there may be a bit tough. Is it a pain to disassemble because there is a lot of hardware to remove, or are there a few precarious things that fit just so?
Thanks! Last edited by thresholdpeople on Sat Jul 19, 2014 11:56 am; edited 1 time in total |
|
Back to top
|
|
|
elmegil
Joined: Mar 20, 2012 Posts: 2177 Location: Chicago
Audio files: 16
|
Posted: Sat Jul 19, 2014 7:40 am Post subject:
|
|
|
It's a pain to disassemble because there are a few items (those switches in particular) that fit through the panel from the top and solder to the board. They will all have to be desoldered to actually remove the panel. Getting things to all line up to put on the panel is also a tricky business, but it's the desoldering that makes it really tough.
I find it doubtful that the 4051 could be bad in a way that affected two inputs and shorted a third into their place, yet still functions normally otherwise.
The jump doesn't sound right, I'll give a shot at testing that later today when I have a chance. |
|
Back to top
|
|
|
thresholdpeople
Joined: Jul 17, 2014 Posts: 22 Location: NYC
|
Posted: Sat Jul 19, 2014 8:47 am Post subject:
|
|
|
Later on today, I'll go back into it and also meter just the pots in question, see if maybe they're bad, but I'd also be surprised it they were.
I've been using the PCB schematics_122, page 2 of the analog board. It's a bit hard for me to tell when the switching is happening on pins 9, 10, and 11. There's no good place for me to alligator clip, so I have the probes on directly, and I can't really get a good grip on the rotary switch and see the meter all at the same time.
All this troubleshooting is really making up for the fact that I didn't put together the kit myself
It's been very fun playing around with the Klee! Getting to learn all of its quirks and functions has been real nice, especially now I'm at the point where I have confidence in its functioning that that strange thing that just happened is meant to happen, it's awesome. |
|
Back to top
|
|
|
elmegil
Joined: Mar 20, 2012 Posts: 2177 Location: Chicago
Audio files: 16
|
Posted: Sat Jul 19, 2014 9:43 am Post subject:
|
|
|
Here's what I would do with regard to the trimming... you say you've traced, but it's not clear if you've checked these things:
power off
test to see if there is continuity (a short) between pins 1, 5 and 14 of the 4051.
If so, pull the 4051 chip and check again.
if there is continuity with and not without, then I could see it being a chip fault.
if there is continuity either way, then you have a short somewhere and it's not the chips fault
if there isn't continuity, continue measuring.... (though it sounds like you may have already done these)
first verify continuity from pin 14 to the junction of R27 and trimmer 2
then continuity from pin 1 to the junction of R30 and trimmer 5
then continuity from pin 5 to the junction of R31 and trimmer 6
power on
check voltage at those pins. It's clear that you've measured the voltage *output* from the chip, but it's not clear whether you've checked the inputs. If the inputs are not all the same, and they don't all change in concert with trimmer 6, but rather change with their respective trimmers, then that could still be a (weird) problem with the chip.
You might also try that without the chip to see if that changes the behavior.
If any of this is stuff you've already done, I apologize for stating the obvious
Edit: oh wait, it could also be addressing issues...
check the diodes from the switch to the addressing pins to see if they have any shorts.
The easy thing to assume is that bits going to pins 11 & 9 are "stuck" but if that were the case you'd have problems with more than just trimmers 2 5 & 6.
You should measure the voltages on pins 11, 10 & 9, and see if the correspond to the correct addressing based on the switch position. Position 1 is all zeros, position 2 is 001 (but note that the order of the bits is opposite what you might think, so 001 is pin 11 high, pins 9 & 10 low) and so forth, basically position N is represented by (N-1) in binary on the address pins. |
|
Back to top
|
|
|
thresholdpeople
Joined: Jul 17, 2014 Posts: 22 Location: NYC
|
Posted: Sat Jul 19, 2014 11:36 am Post subject:
|
|
|
Ah ha! It is addressing issues.
I checked for continuity, and all is as you said it should be, no breaks. I checked for shorts on the IC pins, no shorts. No shorts on the diodes either.
So then I checked the voltage at pin 11, 10, and 9 and here are the results:
pos1- 000
pos2- 101
pos3- 010
pos4- 110
pos5- 101
pos6- 101
pos7- 011
pos8- 111
What's strange though is there doesn't seem to be a short in the diodes, and the diodes are not physically near each other either.
edit:
Now that I think about it, there was a bit of strangeness when I checked the rotary switch. I was a bit lazy about it, and didn't remove the board over it, so getting to the pins deeper was a bit hard, but there was some continuity that seemed strange to me, but I just attributed it to the rotary having a strange layout. I guess my next step is to find the make of the rotary and check out what's common and which pins are what.
Visually the pins of the switch look fine, no obvious shorts. Maybe I'll reflow the solder and see what happens.
Thanks so much elmegil! |
|
Back to top
|
|
|
thresholdpeople
Joined: Jul 17, 2014 Posts: 22 Location: NYC
|
Posted: Sat Jul 19, 2014 11:54 am Post subject:
|
|
|
Ok, I just took the digital board off. Probed the switch, it's a bit strange, but I haven't looked up the datasheet yet. I reflowed the solder on all the pins, reassembled, same thing. It seems like a few of the switch pins are common. I'd assume there'd be two common, and 8 positions, but i'm getting a few others that show continuity, and a few that have some inherent resistance between them, but not direct continuity. Some are completely not connected to each other. |
|
Back to top
|
|
|
blue hell
Site Admin
Joined: Apr 03, 2004 Posts: 24079 Location: The Netherlands, Enschede
Audio files: 278
G2 patch files: 320
|
Posted: Sat Jul 19, 2014 12:46 pm Post subject:
|
|
|
Reordered the data a bit ... dont havve the schematics handy, but its strange that when bit 0 and 2 are the only bit to be one they will affect each other ... something with a pullup or pulldown resistor not connecting to Vcc or ground?
Code: |
pos num code seen conclusion
pos1 - 0 - 000 is 000 - 0
pos2 - 1 - 001 is 101 - 5 error
pos3 - 2 - 010 is 010 - 2
pos4 - 3 - 011 is 011 - 3
pos5 - 4 - 100 is 101 - 5 error
pos6 - 5 - 101 is 101 - 5
pos7 - 6 - 110 is 110 - 6
pos8 - 7 - 111 is 111 - 7
|
_________________ Jan
also .. could someone please turn down the thermostat a bit.
|
|
Back to top
|
|
|
elmegil
Joined: Mar 20, 2012 Posts: 2177 Location: Chicago
Audio files: 16
|
Posted: Sat Jul 19, 2014 5:20 pm Post subject:
|
|
|
Rotary switches can be very heat sensitive. I haven't had much trouble with the particular model in the Klee, but there's one in the Dual Mirror core that I think I melted three while building two modules. That kind of damage is definitely going to behave strangely.
Unfortunately I don't see anything you can do to fix a dead rotary without taking the panel off The big switches at least, you could conceivably pull through the front of the panel since that's how they mount.
If it's going to come to that, probably a good idea to go ahead and get the switches too (and I'd get more than the one you KNOW you need, because you don't want to have trouble desoldering cause you more grief.
You might be able to talk to Synthcube and see if he'd be willing to sell you just the parts you need. No promises there, but it can't hurt to ask. He's a pretty reasonable guy. |
|
Back to top
|
|
|
thresholdpeople
Joined: Jul 17, 2014 Posts: 22 Location: NYC
|
Posted: Sun Jul 20, 2014 8:45 am Post subject:
|
|
|
Ha, damn, I already ordered the switches from Mouser.
Blue Hell, I don't think it's a pull down resistor issue. Each diode ends up at one of three inputs for the IC, and there is a pull down for each pin, a 22k resistor R46, R47, and R48. The way the inputs have organized themselves into bits, each position of the switch that works ends up with a working bit from each row.
http://electro-music.com/forum/download.php?id=10350 page 6 (or 2 of the analog board)
I guess the other place to recheck is where the switch traces go to the header pins. After that it's new rotary time.
If it isn't that, it has to be the switch, right? The circuit's signal path is straight forward, there's nowhere else for it to short. |
|
Back to top
|
|
|
thresholdpeople
Joined: Jul 17, 2014 Posts: 22 Location: NYC
|
Posted: Sun Jul 20, 2014 9:07 am Post subject:
|
|
|
http://www.taiwanalpha.com/english/p_e_141.htm
It's the one pole version.
Too bad it's impossible to get this without mail order. Such a shame to have missed out on at the very least combining this with the switches order. Seems like it may be my third mouser order this weekend.. |
|
Back to top
|
|
|
elmegil
Joined: Mar 20, 2012 Posts: 2177 Location: Chicago
Audio files: 16
|
Posted: Sun Jul 20, 2014 11:05 pm Post subject:
|
|
|
So no, there is no jump forward in normal operation.
I just set mine up with a single bit (first bit), 16x1, pattern.
clocked it, paused with the clock disable, hit load switch which reset back to that single bit, enabled the clock again, and it proceeded from the start per normal.
I'd suspect something noisy creating extra clocks when you first start up. |
|
Back to top
|
|
|
thresholdpeople
Joined: Jul 17, 2014 Posts: 22 Location: NYC
|
Posted: Mon Jul 21, 2014 6:13 am Post subject:
|
|
|
Ok, I think it might have been me and not the Klee -- just tried to recreate the problem now, and it didn't happen. I know the last time I went around the back to mess with the rotary switch I noticed that the bit select ribbon cable was off by 1 position. It was probably that, with the combination of having multiple bits on.
I ordered a new rotary last night, just in case. Going to go in and see if there's a short at or near the header pins after work today. |
|
Back to top
|
|
|
elmegil
Joined: Mar 20, 2012 Posts: 2177 Location: Chicago
Audio files: 16
|
Posted: Mon Jul 21, 2014 7:14 am Post subject:
|
|
|
Good, cable problems are much easier to solve than replacing components |
|
Back to top
|
|
|
thresholdpeople
Joined: Jul 17, 2014 Posts: 22 Location: NYC
|
Posted: Thu Jul 24, 2014 9:43 am Post subject:
|
|
|
Just got the buttons and rotary switch in the mail today. Will try doing the replacement button and tracking doing the rotary issues tonight when I get home. Hopefully the switch replacement will go smoothly! |
|
Back to top
|
|
|
thresholdpeople
Joined: Jul 17, 2014 Posts: 22 Location: NYC
|
Posted: Thu Jul 24, 2014 5:09 pm Post subject:
|
|
|
I've got the panel off! Didn't realize both switches would have to get desoldered to be able to remove it!
Strangely enough the step switch now seems to work under the probes of my meter...
I can't seem to find an obvious short anywhere for the rotary. I think it's a bad rotary as I get continuity between pin 3 (pos 2) and 6 (pos 5), but only when it's on one of those two corresponding states. What's strange about this, though, is that neither are continuous with pin 7 (pos 6), and they should be based on the trimmer pots defaults to pos 6's trimmer, unless there's another problem else where (which can very well be the case).
I just thought of another thing --and I'm not an electric engineer so maybe the answer lies there, but-- it's strange that if there's a short between those three positions, I can't use all three positions to dial in settings.. theoretically there should be 3 trimmer pots in parallel because of the short, and yet there's only one that's able to adjust the resistance.
I need to get some solder braid and will be removing this rotary switch!
I hope putting a new one in just fixes it, however based on what I'm seeing and thinking, it may not. Any ideas? |
|
Back to top
|
|
|
elmegil
Joined: Mar 20, 2012 Posts: 2177 Location: Chicago
Audio files: 16
|
Posted: Thu Jul 24, 2014 5:42 pm Post subject:
|
|
|
Rotaries that are messed up behave strangely. They aren't necessarily hard shorts or hard opens, they can be scrambled inside.
Once you have it out, re-measure and see if you still have shorts in weird places.
BTW I'd recommend getting a solder sucker at Radio Shack or somewhere and using that as much as possible before the braid. I've pulled pads more times than I care to recall trying to use braid only. For some things I would just suck it up (no pun intended) but on the Klee, on a PCB that's not being produced any longer....I'd want to take it safe. |
|
Back to top
|
|
|
thresholdpeople
Joined: Jul 17, 2014 Posts: 22 Location: NYC
|
Posted: Thu Jul 24, 2014 7:01 pm Post subject:
|
|
|
Do you use the turkey baster solder sucker from Radio Shack?
Also, do you have a good method for removing flux off a pcb? |
|
Back to top
|
|
|
elmegil
Joined: Mar 20, 2012 Posts: 2177 Location: Chicago
Audio files: 16
|
Posted: Thu Jul 24, 2014 8:21 pm Post subject:
|
|
|
I don't know about turkey baster
The one I have is an aluminum cylinder with a plunger. You press the plunger down and it cocks it. When you're ready for it to suck the solder up, you click a button on the side, the plunger goes up and creates a vacuum, and voila!
As for flux.....
I use water soluble for when I can wash it (before pots and trimmers and things that shouldn't get wet at all), and no-wash for everything else.
You can clean rosin flux off a board with alcohol and a medium brush (think like a toothbrush, but hopefully bigger ), but you need to be sure you get pretty pure stuff so you don't leave behind anything that could cause other problems. I've heard of people using acetone, but then others warn that acetone can damage some types of components. |
|
Back to top
|
|
|
|