cvs from keyboard mode incorrect?

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
brandon daniel
 
Posts: 160
Joined: Sat Mar 12, 2005 12:46 am

cvs from keyboard mode incorrect?

Post by brandon daniel »

I haven't had a chance to check this from the seq or from midi yet, but I was doing a more precise osc tuning on my x0x when I discovered that the outputs of the digital section when in keyboard mode are putting out bad cv values.

low C when the keyboard is not up or down is the only one that is dead on (@ 2v). Low C when the keyboard is down is a little high (1.03v) though reasonable, high c when the keyboard is untransposed (or low C when it's up) is low (2.87v) and high c when the keyboard is up is low (don't recall exact voltage, but something similar in offset to the last one).

Of course I can "split the difference" in tuning it and get them roughly close to tuned, but it would be nice if I could find the source of the problem. Any ideas of things to check (other than the various power supply lines, which were fine the last time I looked but I will check again)?

Guest
 

Post by Guest »

some things to check

calibrate your multimeter
make measuremnnts with respect to analog ground
verify the powersupply is 5.333 volts
verify all outputs of the 74ac174 go high at 5.333 volts
verify that none of the 200k resistors are touching
check the solder joints and clean up the flux

its possible that the 1% resistors all drifted in the same direction
or that one of the resistors is bad
or that the 74ac174 is bad

User avatar
brandon daniel
 
Posts: 160
Joined: Sat Mar 12, 2005 12:46 am

Post by brandon daniel »

Anonymous wrote:some things to check

calibrate your multimeter
make measuremnnts with respect to analog ground
verify the powersupply is 5.333 volts
verify all outputs of the 74ac174 go high at 5.333 volts
verify that none of the 200k resistors are touching
check the solder joints and clean up the flux

its possible that the 1% resistors all drifted in the same direction
or that one of the resistors is bad
or that the 74ac174 is bad
Thanks for the clues. I trust the multimeter, since the reason I started checking the voltage was that in tuning the osc I could see a non-linear jump in tuning that happened right above A in the root octave, and as it turns out the voltages I saw jived with that.

I guess I've got some debugging to do this weekend, thanks!

User avatar
brandon daniel
 
Posts: 160
Joined: Sat Mar 12, 2005 12:46 am

Post by brandon daniel »

Anonymous wrote: verify that none of the 200k resistors are touching
Just to verify, you mean r74-r90 here, right?

User avatar
brandon daniel
 
Posts: 160
Joined: Sat Mar 12, 2005 12:46 am

Post by brandon daniel »

OK, just checked the IC, when each out is latched high it puts out 5.33 just fine.

I just double checked that none of the resistors are touching, but since the out of r90 is where things are wrong, things go bad somewhere in this divider.

User avatar
brandon daniel
 
Posts: 160
Joined: Sat Mar 12, 2005 12:46 am

Post by brandon daniel »

Another thought before I go pulling 16 resistors from the board.

Am I seeing this schem right? the flip flop IC spits out the tuning voltage reference on it's various outputs whenever they latch high, various outputs go through a voltage divider resistor network to end up with the proper keyboard voltage at r90, right?

All of my DMMs are missing the last digit, so I was only able to tune the voltage to 5.33(?), it seems that this add+divide (if I'm reading the schem right) method would certainly cause any deviations from the expected reference voltage to add up fast?

I'll borrow a better DMM from work today and tune to 5.333 to see if that helps at all before I go reflowing and/or pulling 16 resistors...

Guest
 

Post by Guest »

its also possible that ic11 or ic12 could be drawing more current
off of the resistor ladder than expected
but this is not likely

perhaps an easier way of figuring out which resistor is bad
would be to go through the scale and measure each voltage
if one of the resistors is bad
it should cause both positve and negative offsets in the voltages
from the expected values
if these offsets occur every other note you will then know its due to
a resistor connected closely to pin2
if its slightly down for two notes and then slightly up for the next two
then its probably a resistor close to pin15
etc

or it might be easier to remove ic9 and just measure all the
resistances of the resistors as they sit in the circuit

User avatar
brandon daniel
 
Posts: 160
Joined: Sat Mar 12, 2005 12:46 am

Post by brandon daniel »

Anonymous wrote: or it might be easier to remove ic9 and just measure all the
resistances of the resistors as they sit in the circuit
That looks like the easiest approach if tuning the reference to exactly 5.333 doesn't help me. Thanks!

Guest
 

Post by Guest »

you may have to remove one side of r78 also

out of curiousity
are you measuring the cv at the jack or at r90y

User avatar
brandon daniel
 
Posts: 160
Joined: Sat Mar 12, 2005 12:46 am

Post by brandon daniel »

Anonymous wrote:you may have to remove one side of r78 also

out of curiousity
are you measuring the cv at the jack or at r90y
R90, since my initial indicator was weird tuning problems.

Actually, for the most part things are fine from the bottom up until the A key in the middle octave of the keyboard, at which point the voltage goes about .025v lower than it should, and all the keys from there up are off by that much or more (increasing), high C in the high octave is down by .035v.

As you can imagine, this makes that octave and a half out of tune from the lower one in a way that can't be scaled out.

Guest
 

Post by Guest »

mid a is the first point at which r90 and pin10 become active
it stays active for the rest of the keyboard range
unfortunately pin10 also has the most influence over the cv
and it also sources the most current

reverify that pin10 puts out 5.333
i would expect the other voltages below mid a to be a bit high
if r90 r88 or r89 were at fault

User avatar
brandon daniel
 
Posts: 160
Joined: Sat Mar 12, 2005 12:46 am

Post by brandon daniel »

Anonymous wrote:mid a is the first point at which r90 and pin10 become active
it stays active for the rest of the keyboard range
unfortunately pin10 also has the most influence over the cv
and it also sources the most current

reverify that pin10 puts out 5.333
i would expect the other voltages below mid a to be a bit high
if r90 r88 or r89 were at fault
Well, the bottom half of the lowest octave is a *little* high.

I'll check pin10, though it looked like it was spitting out 5.33 last night when I checked it, and I'll replace r88-90 tomorrow.

Again, thanks for the help.

User avatar
brandon daniel
 
Posts: 160
Joined: Sat Mar 12, 2005 12:46 am

Post by brandon daniel »

Brandon Daniel wrote:Well, the bottom half of the lowest octave is a *little* high.
I'll check pin10, though it looked like it was spitting out 5.33 last night when I checked it, and I'll replace r88-90 tomorrow.
Again, thanks for the help.
I finally got around to looking into this (I'm a procrastinator, I know), so I'll update the thread for posterity in case anyone else runs into the problem.

I pulled r88-90, and none of them measured outside of the 1% tolerance spec'ed for these parts.

Fortunately I had bought 100ct 200k 1% awhile back just in case, so I figured I'd just go ahead and pull all the Rs in the DAC. I measured them all as I pulled them, and they were all within tolerance, but I noticed that I'd somehow gotten all the high ones on one end and all the low ones on another, so maybe this was compounding the problem.

Anyway, after I got them all out, I started hand-picking from my 100ct bag, and found 16 that were dead-on 200k, and put those in the DAC as replacements.

When I got everything back together, the tuning issue was gone, so something in there was the issue, now fixed by hand-picking correct 200Ks.

Just FYI.

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

Return to “Making x0x”