Enhancement for Roland's VC8E

[thunter0512]

I found that although Roland's version of the VC8E board set worked the X/Y deflection signals were very noisy and the VR14 display was less than perfect. This worried me as the expensive deflection amplifier transistors in the VR14 would be stressed by the noise.

At first I suspected that unlike in my LAB-8/e which has a built-in "high quality +/- 15V supply" supplying the DACs on the M885R, on my PDP-8/e the +/- 15V was derived directly from the Omnibus possibly causing noise in the deflection signals.

After some probing with my oscilloscope it turned out that instead the problem was caused by marginal timing of the FV2_LOAD_X_L and the FU2_LOAD_Y_L signals generated on the M869R board.

The fix was to replace U22 (a standard SN7400N) with a faster SN74F00N. With the SN74F00N the deflection signals are now nice and clean.

Previously with the marginal (slow) timing the 74193s on the M885R latched the contents of AC late and sometimes too late. It was right at the ragged edge of barely working. The effect varied possibly because of temperature. The SN74F00N brought the latching time forward by a few ns.

Anyway I strongly recommend that anyone who builds Roland's M869R to replace U22 with a SN74F00N.

Happy New Year.

Tom

 

[thunter0512]

While waiting for parts for my fried VR14 I have continued testing Roland's VC8E using the second VR14 in my LAB-8/e.

There is another minor tweak I recommend to lengthen the very short Z axis intensity pulse which resembles more a glitch then a well formed pulse.

The pulse starts out fine as it is generated by the M869R (U15 SN74122N and C40 470pF cap using the SN74122N's internal 10k resistor), but on the M885R the intensity pulse amplifier reshapes and shortens the pulse to a mere glitch.

The actual culprits are the slowish BC557 (Q12 and Q14) and BC547 (Q13) on the M885R. Faster transistors would work, but a simpler fix is to replace C40 on the M869R with a 1nF cap.
This will generate a longer intensity pulse on the M869R, which although it is subsequently shortened by the M885R's pulse amplifier, remains long enough (about 1 us) for the VR14 to work reliably.