e-licktronic Forum

Finished the build this evening, after a slow several weeks and a few hours at the bench each night :^).

v 1.02, w/ 1.024 firmware on the atmega (accent appears to be working with the Mux Boards)

I am running into a few issues and they are detailed as follows --

1) SD
Getting a rather filtered Snare Drum out, as if it were just another TOM or BD voice -- it has no noise (makes me think that there is something wrong with the noise circuit)

SOLVED -- BAD/FAULTY SHIFT REGISTER CHIPS and 4070 IN THE NOISE SECTION.

2) HCL
Not getting an HCL out -- not from individual outs or from summed outs.
An issue with the Noise portion of the build would makes sense, as the HCL is generated through filtered noise?

SOLVED -- BAD QUAD VOLTAGE COMPARATOR IN THE HCL SECTION (AN6912). HAD AN INFINITE CLAP SOUND WHEN THE NOISE SECTION WAS FIXED, SWAPPING OUT A NEW AN6912 FIXED THE PROBLEM.

3) OH
Shortest decay value has a "secondary" tail that is noticeable — sounds like a sample is being played in reverse. If the decay time is increased, then this reappearance of sound or "secondary tail" is getting masked and is not perceptible.

UPDATE -- STILL WANT CONFIRMATION ON THIS ONE

4) Rather perceptible noise floor in the summed outputs? Individual outs seem quieter. With nothing programmed on an individual output voice, I can hear a low level clicking/thudding/muffled summed outputs on the indi out. I have to have the volume pot cranked rather high for this though.

SOLVED

Any suggestions would help!

Edit: Could someone let me know if this is their/the correct orientation of the pair of 2SD1469 transistors located at Q47 and Q48 of the SD voice. I looked at data sheets and compared mine to the build notes, but my transistors came with difficult to discern features, as in I don't know which side is which -- the dimple is too large to tell an orientation.

EDIT 2: Orientation is fine. Another user had a similar issue and posted images of their corrected orientation. Ours matches up and Vincent confirmed as well.

https://imgur.com/a/YJ6tzCn



best,
m

Anyone??

Been working on the snare voice and I have yet to identify the issue --

Snare does not respond to SNAPPY and TONE, just TUNE and LEVEL. It sounds like another TOM voice, as there is just low end, no high end frequencies; sounds like there is no noise getting to the circuit.

Probed all points in the SNARE section -- VCO-1 and VCO-2 both have periodic functions on the scope, but they look asymmetrical and not like perfect triangles, as Vincent told me that they should look like. They have more pointed bottoms, and smoothed/buffed/rounded tops.

Envelopes 1-5 are all there. Noise is there. Trig/Accent is there.

Vincent told me to swap out a fresh IC36 4069UBE. I swapped out IC36 with a working 4069UBE from the MID TOM section. No difference to the SNARE sound. The IC36 is not the issue, as I put the one from the SNARE section into the MID TOM section, and the MID TOM sounded fine.

My next hypothesis is either a bad pair of Q48/Q47 transistors, or some bad SIPs M5218 like IC40 because SNAPPY is not working — the low pass part of the routing is not working
or maybe IC39, as TONE via high pass filtering is not working either.
Any ideas?

UPDATE ON SNARE: I probed around and found a perfect triangle at R305. I lose the perfect triangle waveform in the signal path after R305.
Looking at the TR-909 schematics, the signal goes into the anti-parallel oriented diode clipping stage. On my build, VCO2 probe point displays an asymmetrical waveform, as I described in previous emails -- semi-pointed bottom, but buffed and smoothed top.
I ordered new M5218Ls and a pair of 2SD1469. Will try and swap some ICs for the SNARE section.

Ordered AN6912, BA6110 for a possible HCL fix.

HCL and SNARE solved

Still would like confirmation on HH transients if anyone else has experienced similar phenomenon.

If you find out about the OH "echo" thing, let me know. Same thing happening on my closed hihats. It sounds like its being played back in reverse from the end of the sample, and the length of the reverse echo gets shorter when I increase the length of the CH decay