I was sent a 246 to modify similar to my 245 Sequential Voltage Source modifications.
I made reference designators from PCB images. The components cover the silk screen legends once populated which makes it near impossible to find components.
PCB1 Rear Reference Designators
PCB2 Rear Reference Designators
The original 246 used a 24V supply and the logic was a mix of DTL and early TTL that ran from +5V. This redesign uses CMOS with a 15V supply and the 24V circuitry is simply driven from 15V with no compensating changes. In addition, a number of resistors were changed or added.
I modified a 245 and 246 and my customer's email back was "So the report is - they work flawlessly. Null zone has been reduced to almost nil, adjacent stage pulse outputs now individually trigger the 281's - even three/four in a row. I encountered no problems with the stage trapping - either open ended on either side or between two mid-points."
These modifications are the same as for the 245 so see my 245 Sequential Voltage Source page for detailed information. This page simply consists of the component reference numbers for the modifications specific to the 246.
Stage CV Input
With 16 stages there are no unused gates with floating inputs.
Add a 100 pF capacitor across R65.
Stage CV Output Level
Remove R27, R28 R29, and R31 (note PCB1).
Change R294, R296, R298, and R317 to 330K.
Remove C57, C58, C59, C61.
Remove R279, R280, R281, R282, R283, R284, R285, R286, R287, R302, R303, R304.
Remove Q75, Q76, Q77, Q79.
Connect the base to the collector with a link on Q75, Q76, Q77, Q79.
Replace R288, R289, R290, R307 with a link (corrected).
Replace IC31, IC32, IC33, IC34 with LT1637.
Swap pins 2 and 3 of IC31, IC32, IC33, IC34.
Stage Pulse Output Level
Change R55, R61, R72, R78, R86, R92, R100, R106, R114, R120, R128, R134, R142, R148, R156, R162, R170, R176, R184, R190, R198, R204, R212, R218, R226, R232, R240, R246, R254, R260, R268, R275 to 6K8.
I adjusted the waveshape similar to the 245 but the baseline sits at 4V, not 2V (I believe) due to the number of stages. When I changed all the stage pulse output resistors to 6K8 the module would no longer advance. I simply adjusted the clock waveshape trimmers until it functioned reliably.
Change R301 to a 10K trimmer.
Change R306 & R315 to a 20K trimmer (wiper at the node connecting R306 & R315).
Change R30 to a link/wire.
Change R3 to a link/wire (note PCB1).
This image shows the clock and CV output modifications on PCB2.
This image shows all the modifications to PCB2.
PCB2 modifications photo (larger PDF)
Here are the four CV outputs with a full 0 to 15V range.
The clock I calibrate based on operation. These sometimes stall with all the outputs at maximum so I adjust my added clock trimmer for reliable operation.
The legends never match operation and it is important to use a 10K pot for the proper taper. I make sure it is never faster than 2 mS and sometimes adjust C4 a bit lower to make sure it doesn't stall on the longest setting.
The analog address circuitry will only pass ~13.5V signal so I calibrate measuring R271 setting TR2 for -13V which provides a 13.5V range to address 16 stages.
I had a 246 come in for service that would not advance with the clock at all. Individual stages could be addressed. In looking at the individual stage timing I noticed something quite odd. The green is the emitter of Q101/Q102 which resets the stage. Magenta is the collector of Q102 rising to turn off Q103. Yellow is the collector of Q103 which then falls as all stages are reset. Cyan is the base of Q102 which shows it reverse biased to turn it off. What is odd is the collector of Q103 drops to about 4V and then slopes down. It never gets low enough to turn of Q104 which creates the pulse to turn on the next stage. Thus is simply sits on the current stage. These reference numbers are for the original schematics. On the 246r Q11 is Q103 for stage 1.
For some reason, C103 is holding the collector voltage high as Q103 turns off. In this image I have removed Q102 and am simply driving the base of Q103 with an external source. Note the long RC decay once the collector drops to 4V. This is an asynchronous machine and the reset has to occur before Q104 turns off so it can set the next stage.
These modules are hard to work on and the board was not washed. I never could track down in a reasonable timeframe the root cause of this issue. One alternative was to significantly load the collector of Q3. A 1K resistor to ground pulled the waveform to 0V fast enough. However, this adds a 15 mA spike in current and these modules are finicky enough already. This image shows proper operation with these 1K resistors.
Another alternative would be to replace Q108 and R109 with 510R resistors. However, this did now work and I didn't want these current spikes so I did not pursue this further.
What I ended up doing was diode decoupling R108/R109/C103. I simply lifted the end of R109/C3 that connects to the collector and added a signal diode. This worked great. This image photo shows the yellow stage with the 1K resistor and the cyan stage with the diode. You can see the diode drop of about 0.4V in the gate level. The trigger is actually a bit higher because Q103 isn't loaded as much. Although the diode adds a drop, remember these were originally built with 5% resistors so the gate level could be off by 10% depending on tolerance build up. These levels are fine as-is.
Here's the diode modification I did to all 16 stages.
These modules are known to stall with all 48 controls turned to maximum. This one did as well. I did a slightly different clock modification. The R301 trimmer does very little so I left it as 15K. I instead used a 10K trimmer with the CCW and center for R306 and added a 1K resistor between CW and R315. This seemed to work better.
A 3/8" flat trimmer fits nicely over C62 if you bend it over. You can solder two trimmer leads directly to the pads and bend the remaining lead to the top to solder the 1K resistor to it.