How To Repair Hammond Drawbar
A couple of weekends ago I took the two XB-2s that I had (at the fourth dimension) over to Ron'south store to take a little more room to spread out and test things. Similar real Hammond tonewheel organs, the XB-2 has "drawbars" that represent different harmonics (or subharmonics) of the fundamental frequency of each key being pressed; yous draw out the bars to mix different amounts of the different harmonics to get the timbre you want. This is condiment synthesis at its most visceral.
On the XB-2, the drawbar positions (either live or recalled from memory) are displayed on an LCD below the manual (keyboard). In alive mode, the bar graphs movement in and out in synchrony with the physical drawbars.
On ane of the two XB-2s, the LCD bar graphs didn't lucifer the drawbars — a couple of drawbars appeared to piece of work properly, merely some didn't piece of work at all and others moved multiple bar graphs on the display. Since the drawbar decoding is a relatively contained section of the organ, information technology seemed similar an like shooting fish in a barrel repair to tackle beginning.
The drawbar'due south wiring harness plugs into the principal lath on J121, at the left of this section from the service manual. Each drawbar is a detented slide potentiometer, then variable voltages are arriving on J121. The section enclosed in the dotted line and marked not used truly isn't populated on the circuit board, and then I omit it from discussion.
The 9 analog drawbar voltages are delivered to IC23 and IC24 (TC4051 analog multiplexers). The multiplexers receive their enable and select signals from the output of IC29 (74HC174 hex D flip-flop) which is latching signals previously delivered from the organisation data bus. (In other words, the 74HC174 is the drawbar select register; its own address is decoded elsewhere in the schematic.)
The chosen (enabled) TC4051 analog mux selects which input to pass to its output on pin 3, which is so op-amp buffered and delivered to the input pin of IC25 (BA9101 analog-digital converter). When selected (more system address passenger vehicle decoding), the ADC writes the digital value of the drawbar's position onto the organization data bus.
Side note: For the drawbars only having ix detents (0-viii), IC25 sure delivers a lot of $.25 of ADC resolution to the data bus.
I put a scope on IC23 (analog mux)'s output pin and I was able to view on the screen the time-division multiplexing of the drawbar positions (analog voltages) onto the unmarried line going to the ADC. It mostly matched what I saw on the LCD, although in that location were some quirks with a few of the drawbar time divisions appearing narrower than others. Ignoring the odd widths and recording which drawbar occupied which time division:
| Drawbar | 16 | viii | iv | ii | 1 | five 1/three | two 2/three | ane 1/3 | one 3/5 | |
|---|---|---|---|---|---|---|---|---|---|---|
| IC23 Pin | 13 | 14 | fifteen | 12 | one | five | ii | four | (IC24) | |
| IC23 Input | Good MB | 0 (000) | one (001) | two (010) | 3 (011) | 4 (100) | five (101) | half dozen (110) | 7 (111) | (IC24) |
| Bad MB | 0 (000) | 1 (001) | 7 (111) | 7 (111) | 0 (000) | 1 (001) | 7 (111) | vii (111) | (IC24) |
On the working XB-2 motherboard, the drawbars were selected and sampled in numerical order. On the broken motherboard, as you can see, whatever time the analog mux's select scrap A1 was enabled, the mux behaved as though bits A2 and A0 were enabled as well. Further, select scrap A2 didn't work on control as information technology should when drawbars 4-7 should have been chosen.
| Proper noun | C | B | A |
|---|---|---|---|
| Function | A2 | Aane | A0 |
| Pin | 9 | 10 | eleven |
It could be a bad 4051 mux; but equally nosotros had already replaced a leaky electrolytic capacitor in the neighborhood, it seemed worth another look at the circuit board first. The 4051′s select lines are on pins ix-xi, and what'southward this?!
I became suspicious of a damaged via on a trace that turned out to connect to pivot nine (A2). A continuity test showed that the via — even its top side — was no longer continued to IC23; the trace upwardly to the via had been eaten abroad past the leaking capacitor. The via — even its top side — did have continuity to its next stop on the PCB, and then the via itself was intact.
Ron heated the solder that had wicked into the via during reflow, inserted a piece of wire-wrap wire, and soldered the other stop directly to IC23 pivot 9. The drawbars now work perfectly. I suspect the floating select input on the CMOS mux was picking upwardly plenty signal from the PCB trace inductively coupled to its neighbor to trigger.
My hypothesis is that the previous possessor put the keyboard away considering of larger (ROM / CPU / Muse) issues; the capacitor leaked and damaged the drawbar multiplexer trace while it was sitting idle; and the possessor never even knew nigh the drawbar problem. At any rate, it was easily fixed and the troubleshooting was a rewarding mental do.
Source: http://www.neufeld.newton.ks.us/electronics/?p=1248
Posted by: daubdasked1981.blogspot.com
0 Response to "How To Repair Hammond Drawbar"
Post a Comment