(Originally I posted this in the 4U forum, but I figure the DIYers might be more active about this topic, so I'm reposting here:)
In the dreadfully immense amount of free time I've had, I've been going over all my Buchla clone builds that I've done and fixing things that are clearly not right. One thing that I didn't seem to get right was the time scaling on the 245r that I've built (and modified extensively as per Dave Brown's helpful webpage).
Has anyone been able to trim the Pulser timing knob to get times exactly as the legends describe without changing any of the resistors on the original BOM? ( https://cdn.shopify.com/s/files/1/0388/ ... M-v1.0.pdf ) With the default values for TR1 and a 10k linear potentiometer for the front panel control, the 10 second setting was more like 2 minutes. Measuring around on the board I found that the control goes through a big resistor balancing network with the trimmer into the base of one side of the one matched diff pair in the BOM. Measuring it out, it seems like it goes between a few millivolts for the shortest time setting up to .850 volts for the longest time. The other side goes into the oscillator circuit for the downwards sloping sawtooth, including what I presumed was the timing capacitor. I screwed around with the timing cap near the JFET (and also changed that JFET to something with more suitable parameters closer to an actual 2N4339) and changed the value down to 47nF from .15uF, which oddly didn't change much. Looking at that input balancing resistor network, I went after R42 next (R2 in the original Buchla schematics on Fluxmonkey). Since that becomes the top leg of a resistor divider on the max setting of 15V, I figured increasing that would lower the voltage from .850 down to something lower, and by changing it from 78k7 to 200k, I finally got a scale that I could trim with TR1 in order to match the front panel legends.
The 10 seconds is spot on, so is 0.5 seconds, and .002 seconds is right, but .005 is actually more like .017 seconds. I'm only assuming, but I think this scaling is due to a more logarithmic approach that Don got because, so I hear, he 'used logarithmic potentiometers', which might explain it.
Also my diff pair was an LM394 that had a beta of 650. The first thing I changed was taking that out for 2 cut tape 2N3904's that both had betas down near 170 or so instead. This barely changed the timebase. I had had an experience with the MFOS Soundlab Mk2 ages ago that led me to changing this.
ANYWAYS, tl;dr is: Did anyone else get a time scaling that matched the front panel legends without having to muck with the circuit? I had to change R42 to 200k and C4 to 47nF to get it mostly right.
In the dreadfully immense amount of free time I've had, I've been going over all my Buchla clone builds that I've done and fixing things that are clearly not right. One thing that I didn't seem to get right was the time scaling on the 245r that I've built (and modified extensively as per Dave Brown's helpful webpage).
Has anyone been able to trim the Pulser timing knob to get times exactly as the legends describe without changing any of the resistors on the original BOM? ( https://cdn.shopify.com/s/files/1/0388/ ... M-v1.0.pdf ) With the default values for TR1 and a 10k linear potentiometer for the front panel control, the 10 second setting was more like 2 minutes. Measuring around on the board I found that the control goes through a big resistor balancing network with the trimmer into the base of one side of the one matched diff pair in the BOM. Measuring it out, it seems like it goes between a few millivolts for the shortest time setting up to .850 volts for the longest time. The other side goes into the oscillator circuit for the downwards sloping sawtooth, including what I presumed was the timing capacitor. I screwed around with the timing cap near the JFET (and also changed that JFET to something with more suitable parameters closer to an actual 2N4339) and changed the value down to 47nF from .15uF, which oddly didn't change much. Looking at that input balancing resistor network, I went after R42 next (R2 in the original Buchla schematics on Fluxmonkey). Since that becomes the top leg of a resistor divider on the max setting of 15V, I figured increasing that would lower the voltage from .850 down to something lower, and by changing it from 78k7 to 200k, I finally got a scale that I could trim with TR1 in order to match the front panel legends.
The 10 seconds is spot on, so is 0.5 seconds, and .002 seconds is right, but .005 is actually more like .017 seconds. I'm only assuming, but I think this scaling is due to a more logarithmic approach that Don got because, so I hear, he 'used logarithmic potentiometers', which might explain it.
Also my diff pair was an LM394 that had a beta of 650. The first thing I changed was taking that out for 2 cut tape 2N3904's that both had betas down near 170 or so instead. This barely changed the timebase. I had had an experience with the MFOS Soundlab Mk2 ages ago that led me to changing this.
ANYWAYS, tl;dr is: Did anyone else get a time scaling that matched the front panel legends without having to muck with the circuit? I had to change R42 to 200k and C4 to 47nF to get it mostly right.
Statistics: Posted by Laughing — Wed Oct 16, 2024 3:48 pm — Replies 0 — Views 21