Got my Klee together (but stuck troubleshooting)

[softfin]

I've now finished building the Klee and managed to get it working for the most part.
I had some difficulties getting pattern A and B to work right, but exchanging the SCL4034s with CD4034s solved that. I actually had 5 pieces of SCL4034s and none of them worked
For the regulator I used a 7810 instead of the 78L10.

The situation is currently as follows:

All controls, leds etc. seem to work ok, except clock and load inputs. That means I'm able to clock and load the Klee manually, but the patterns don't move at all when I apply an LFO or other signal to clock and/or load inputs. I've been trying to solve this problem for days now and have made sure all connections between jacks and boards etc. work properly...

I was also unable to calibrate the Klee properly: Adjusting the trimmers for rotary switch positions from 1 to 6 were easy, but it seems I can't get the voltage trimmed higher than 3.2V on positions 7 and 8.

Feeling quite exhausted right now as I've already spent many nights troubleshooting the clock problem...

Any help would be appreciated.

Btw, printed and read 'Building the Klee sequencer', it was nice to read and there were lots of useful information there. Thanks a lot to Scott for all the efforts!

---

klee.jpg
Description:
Filesize:	306.98 KB
Viewed:	325 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

---

[Scott Stites]

We should be able to get that going. An oscilloscope and/or an logic probe will be very handy here. Do you have either?

The fact that you can get the thing to clock and load manually is a very good sign - those sections I consider more problematic to troubleshoot than the external inputs. It also eliminates a good chunk of the circuit as causing the problem.

[softfin]

[Scott Stites]

OK, even if you don't have an Oscope or a logic probe, there are some very important measurements we can make with a DMM. Since you've got the oscilloscope, that would still make it easier to see however. You can use either for this test.

The common theme between the clock input and the load input is that they both share the same IC with which to generate the pulses the Klee likes to see - U1 (LM358).

1. First of all, make sure that Bus 1 Load is OFF. This will interfere with external loads. Clock through manually to make sure everything is still clocking manually. Load manually to verify.

2. Once you've done all that, let's take a look at the clock input. Set the LFO to a very low rate and input it to the clock input. Preferably, use a square wave LFO (easier to see with DMM).

3. Ground your DMM minus lead to the Klee and use your +Lead to look at DCV.

4. U1 is serving as comparator. When the input voltage (your LFO) rises above a reference voltage, the output of the comparator should swing high to very close to the rail. This reference voltage is right around 2V. Because U1 is two comparators in one package, this reference voltage is applied to two different pins on U1:

Probe Pin 6 of U1 - the voltage should be around 2V. This is the reference to the clock in comparator.

Probe Pin 2 of U1 - the voltage should be around 2V. This is the reference to the External Load comparator.

5. If the voltages at pins 2 and 6 are OK, then we need to look at if the signal is getting to the opposite pin of each comparator input. We'll look at the clock input comparator. In step 2, we plugged in the LFO to the clock input. This signal must go through a protection diode, so that will drop it down around 0.7V, but if it is a nice modular level LFO, that is a pittance.

Probe Pin 5 of U1: You should see the voltage rise and fall between 0V and the positive peak voltage of the LFO. This signal will not go negative. Confirm that the level rises well above 2V and well below 2V. A slow moving square wave LFO will allow the DMM to settle on a reading before it cycles again.

6. If the voltage in step 5 is good, let's look at the clock input comparator's output pin. This voltage should snap up to close to the rail when the LFO is high, and snap down to ground when the LFO is low.

Probe Pin 7 of U1. Make sure that it snaps high and low, and is not railed constantly high or railed constantly low. The slow moving LFO should allow the DMM to settle on one reading before the LFO cycles to the next.

7. If Pin 7 of U1 still doesn't move at this point, make sure you have +15V at Pin 8 of U1 and Pin 4 is at 0V.

If everything checks out there, let me know and we'll move on.

[Scott Stites]

Sorry - for got to mention which board - that's U1 of the Digital Board.

[softfin]

[Scott Stites]

That's fairly interesting - your reference is way too low. Your LFO voltage is right on target. The interesting thing is that the output of the comparator should be railed high, not low.

But - first things first - this is certainly the source of the problem. We need to find out why your reference is so low.

1. The reference is supplied by a voltage divider formed by R25 (100K) and R29 (15K). Check to see if one end of R25 is at the voltage rail, and the other end is at around 2V. Check the value of R25 - make sure it is a 100K resistor. If, for example, it is a 1M resistor, that would make the reference voltage waaaay low (around 200 mV).

2. If you don't get positive rail voltage at least one end of it, it's likely not connected to the positive rail for some reason. If the other end of it is 2V, then that signal is not making it to U1.

3. If you do have rail voltage on one end of R25, but extremely low voltage on the other end, check the value of R29. Make *sure* it is 15K. If it were 1.5K, for example, you would again get something around 200 mV. Make sure one end of R29 is 0V (connected to ground).

4. If the resistor values are correct, and R25 is connected to the rail, and the path to the comparator reference pins (pins 2 and 6) have continuity, pull out the LM358 from the socket and measure the reference voltage at pin 6 of the socket. If the voltage is correct now, your LM358 is likely bad.

Let me know what you find, and we'll move on from there.

Edit: Adjusted value of too low ref caused by order of magnitude error in resistor value - in that case, the voltage would be 200 mV, not 2 mV.

[softfin]

[Scott Stites]

Woo-hoo! It's clocking and...

>DOH!< I still didn't fix the voltage levels in the build doc yet?

The voltages at the test points indicated will be negative - sorry about that.

Adjust them for the negative equivalent, or switch your DMM leads around and adjust them for positive. Man, sorry about that!

Does that work?

[Scott Stites]

Kwik Kwestion: Are you using variable or fixed voltage for range 8?

[softfin]

[Scott Stites]

R32/trimmer R40 determine the value of Range 7. If the top of R32 is 10V, maxing out the trimmer should give you 5V (R32 and R40 are both 10K, so that would divide the voltage in half). Adjusting the trimmer should drop it down below 5V through to 0V. (Of course, at the test point, it would be -5V).

R33/trimmer R41 determine the value of Range 8. Standard value for R33 is 1K, and R41 is 10K, so maxing out R41 should put the voltage to 9V (that's -9V at the test point). It should adjust on down to 0V.

You mention using a 7810 - is the voltage coming out of it a solid 10V or so?

Cheers,
Scott

[Scott Stites]

Check the regulator voltage on both sides of R33.

[softfin]

[Scott Stites]

You're not using variable range - right? IE, position 8 is fixed range - no front panel variable range control, etc.?

[Scott Stites]

Are you sure R41 (range 8 trimpot) isn't 1K instead of 10K? That would give you around that kind of drop across R33 with the trimmer maxed out. It should be 10K.

Assuming R33 is 1K, a 1K resistor in series would divide the voltage roughly in half (given the tolerances of R33 and R41) at the junction between them, which is what you're measuring/adjusting.

[softfin]

[Scott Stites]

Ah, OK. Somethings not kosher - the voltage you see on the R33 should be pretty much the voltage you see at the TP (only TP voltage would be inverted). Something is happening between R33 and the TP.

Crank R41 all the way up, so you get that 6.45V reading on R33 again. Then measure the voltage on pin 3 of U7 (the CD4051). It should be very close to that. U7 is just an IC switch that selects that voltage and puts it out on its pin 3.

If CD4051 pin 3 is not close to that, check the voltage at pin 4 - this connects directly with R33 (the 6.45V side).

If pin 3 of U7 is very close to the 6.45V reading, read the voltage on pin 3 of U6 (TL072). It should be the same reading. If that's OK, check the reading on pin 1 of U6 - again, it should be as close to identical as possible as pin 3 of both U7 and U6.

Let me know what you see.

[Scott Stites]

Make sure you're in range 8 when you do that....

[softfin]

[Scott Stites]

OK - and do you have 8.9V on pin 1 of U6, too?

If so, let me know which bit you have on for calibration, and which pot is cranked for max.

[softfin]

Btw, when I powered the Klee up after changing the R41 for a new trimmer, half of the leds were continuously on, so I suspected another of the 4034 had blown up. I ended up having to change both 4034s to get it to work again. They must be very delicate parts

[softfin]

[Scott Stites]

You should have 8.9V on pin 3 of U7 AND on pin 3 of U6.

If you don't have 8.9V on pin 3 of U6, remove it from the socket and check to see if socket pin 3 has 8.9V. If it does, I suspect U6 may have problems. If you have a spare, try it out. If a spare does the same thing, then one of the CD4066s (U1 through U4) could be pulling the output down....hard. You could remove them one by one to find the bad one.

If you still do not have 8.9V on the socket pin 3, then suspect the connection between pin 3 of U7 and pin 3 of U6.

The CD4034s blowing up like that is unusual. Make sure you're well grounded.

[Scott Stites]

Oh yeah - check U6 for +15V on pin 8 and -15V on pin 4.