Hammond Drawbars- Lots of Them

[henrygr]

Hello all,

Just got my hands on the drawbar and switch assembly off of a Hammond H112 on ebay

http://feedback.ebay.ie/ws/eBayISAPI.dll?ViewFeedback&userid=noble_mutt&item=280006089649&iid=280006089649&frm=1883&ssPageName=STRK:MEWN:SID

I had never seen the drawbar assembley before off of a Hammond, and have read all the threads regarding their midification. The problem is most of the threads drew no conclusion or solution. Is their anyone about who can tell me whether I go either:

a- Scan Matrix, although I'm not sure if that end of the programming is fully complete

b- the 'resistive' route using increments of 1.2k resitors over all eight bus bars to a max of 10k, and set up that bridge every eight drawbars

c- The DIN route.

Any comments would be greatle appreciated.

MP

[Wisefire]

link doesnt work.. shows me the feedback an ebay member

[henrygr]

My apologies. Thanks for pointing it out wiswfire

Here is the link proper:

http://cgi.ebay.ie/ws/eBayISAPI.dll?ViewItem&ih=018&item=280006089649&rd=1&sspagename=STRK%3AMEWN%3AIT&rd=1

:)

[stryd_one]

50 bucks! Score!!

At a guess, I'd go with the 'resistive' route using AINs, just because it means less programming needed, and less IO used... But I've never seen the circuitry so I don't know how hard it would be to do...

[henrygr]

That was my main thought. It would also require less circuitry. And then, when I hear you say it stryd_one, I know it is right.... :)

[Wisefire]

arent these drawbars able to be set halfway..that would mean 1 ain for every drawbar..

[stryd_one]

I thought the drawbars had a few steps on each one? I just had a look at the pic to see if my memory serves and it looks like 8 stops each to me...

Hey henry once you take delivery, could you take a bunch of pics for us? Especially closeups of mechanics and circuitry and such... That would be really helpful.

Anyway, better to use one AIN for each, than one DIN and a stack of DOUTs, or a stack of DINs... Although I noticed that the wiring is still there, so it might be easier to solder a scan matrix.

Hey henry don't put tooo much weight in what I say... I'm still learning :) Don't be fooled by my post count, that doesn't mean that I know stuff, it means that I talk too much ;)

[henrygr]

I took delivery yesterday ;D ;D

Will post a few photos. Yes it is eight stops per drawbar- switch style, not potentiometer. Each stop in in a series of stops is fed by the same current. It is also split in two. So from above, it looks like a matrix of 8 by 22 switches. If I am not mistaken a scan matrix would require soldering a diode at each point. Iwant to do as little damage to the original piece as possible. So I was thinking of feeding each of the eight point, or stops, on the drawbars with the relative currnent of a 10k slide pot. Ie resistor of 1.2k at stop one, 2.4 at stop 2 etc.

I have seen somewhere on the web (http://www.geocities.com/JDPetkov/Hardware/B4ce/B4ce.htm, here in fact. It was a discussion around here a while back that put me onto him) that the correct wiring was 8 by 1.2 resistors in parallel. This is wrong. I think he made his assumpyions before seeing the unit. He assumed that as the drawbar was pulled out, contact made with previous stops was maintained, and that resistors in parallel would achieve a result. (This is a common misconception of Hamond drawbars)

Will get the Nokia busy and take some snaps.

FYI This also arrived yesterday-http://cgi.ebay.ie/ws/eBayISAPI.dll?ViewItem&ih=004&item=140005371147&rd=1&sspagename=STRK%3AMEWN%3AIT&rd=1

;D ;D

[stryd_one]

Next thing you'll be dancing around singing like Barry White "My darling, I.... Can't get enough of drawbars baaaby"

[henrygr]

Mmmm. Could you post an mp3 of that?!..... :-X

[stryd_one]

You don't know of barry white?  :'(

You gotta hear this guy man, his voice is like pure molten gold

[henrygr]

I know barry white- he saved my life!!!

I want the mp3 of you singing the drawbar song. Maybe we could choreograph a dance too. Could take off...

You heard it here first.

;D

[stryd_one]

Skinny white guy does barry white....hmmmm... I won't embarass myself ;)

[henrygr]

Not when you put it like that, no. That sounds like something that should belong on a porn site..... :P

[stryd_one]

That's it, I'm changing my sig  ;D

[henrygr]

OK No point taking photos as the unit is well sealed. I did a diagram though. The red lines with the red boxes indicate drawbars- just in case you were wondering, or being otherwise distracted by Barry White. The cross lines are where the contact busses run. I represents five drawbars, however they run for twenty two, The resistors in the diagram are my idea to midify it. I left out that the signal back to the DIN runs from the drawbar. I propose to break up the busses every eight drawbars and run a set of resistors again. Would this be necessary do you think, as Iwant to leave the unit as much as in tact as possible.

[henrygr]

ADDENDUM: I drew in the contact point, but note that only on bus has contact with a bar at any given time!!!!

[stryd_one]

I'm a bit confused... I see the buss that runs to ground, and a lot of other busses running in parralel....

Does it work so that each bar contacts the ground always, and one of the other busses depending on it's position? or does it contact the ground always, and then each of the busses following?

So if the bar is set to '7' (drawbar slides out 2 'notches' from fully closed), does it contact only the 2nd buss, or does it contact buss 2 and buss 1?

[henrygr]

Thats where confusion has arisen before. The drawbar touches only one bus at any given time. The drawbar is in fact a large one pole 9 way switch, break before make, to put it into modern context.

See you kept the sig ;)

Mark

[stryd_one]

That might make it a bit easier :)

One thing about the resistive method that just occurred to me... The resistance of the buss itself will need to be taken into account... This might mean soldering the resistors onto each switch, rather than just the beginning of each buss. I'll leave that up to your multimeter, I doubt it will be much, so you probably can ignore it.

Hah no I just changed it ;)

[henrygr]

Good then I'll go ahead and try it. I would say the bus resistence is as negligable as normal copper wire connection pots together anyway.

Just a quick tought though. On the AIN multiplexer, the schematics say not to connect more that eight pots off of the 5v feed in a row. Is it possible to get away with longer. I might post a new thread on this to see if anyone else has succeeded.

Thanks stryd,

Mark.

[henrygr]

Good and bad news- they cancel each other out.

FACT: Hammond drawbars are 'make before break'- you heard it here first. Now that is good news if you have only one.

Why?

Because you would never need to worry about jitter on your AINS as you pull the drawbar-  jitter that would have been caused as the contact slipped between two busses on a 'break before make' switch. Only momentary, but jitter is litter!! (I like that one).

With me so far?

Now the bad news. When we have two bars (and i will limit the example to two to keep it simple) the situation changes. Lets say we have both our drawbars set at stop tree -go on, close your eyes and imagine playing a barry white song at the controls of a B3. Your solo is coming up. You need to pull out drawbar one to stop four. You hesitate slightly in anticipation of your applause, or because your grandmother is in the front row crying her eyes out, and you know it's really because you reversed over her cat getting to the gig. Even worse, you still have to tell her it's your fault.  ::)

So you pause. Drawbar one is in a limbo. It sits between stop three and stop four. But that's OK, because contact with bus four is completed, even though the contact with bus three is not yet broken. Good? ;D

Well, no actually. Total disaster. >:( What has in fact now happened is that drawbar one has completed a circuit between bus three and four, and as a resut, drawbar two, even though is sitting square on stop three, has also sent a signal to the AIN that it is at stop four.

Now, I have simplified the actual end result of this mess. What really would have happened is that the resistive value of busses three and four would have been read in parallel, and god only knows what the result would have been. Poetic justice for killing Marmalade? :o

I looked at the mechanism itself. Imagine the drawbar contact like that of a tram touching the electric lines above it. Most trams are equipped with a double contact at the end of, and perpendicular to, a long spring loaded pole. That mean the flow of current is never cut, even when the lines above are split momentarily. When one contact is sitting in the split, the other is still touching a live wire.

It would be possible to, if not remove, certainly isolate one of the contacts on the drawbar. The contact piece itself is just a small copper plate with two stud jutting up.

My problem is though, it brings me back to the 'break before make' jitter for litter. Is there someway of overcoming this? Would connecting the earth directly to the wie back to the multiplexer be a workaround???

Hope you enjoyed the story..

Mark.

[stryd_one]

LOL funny post ;)

Ok it might be easier to do a scan matrix then :( Darn!

Or.... You could just ignore all values which don't match your increments of 1.2k....

For eg if the drawbar is on stop 2 (2k4) AND stop 3 (3k6), you'll get 800ohms resistance, which isn't a multiple of 1k2, so it gets ignored... etc etc.

I've just fed all the values into MiscEl (a very cool calculator if you don't already have it) and it looks like it'll work.

Now back to bed for me. I have the flu, and no amount of troubleshooting can fix that ;)

[henrygr]

Very clever. Theoretically one could run into any amount of combination, but the likelyhood would be quite low. On further examination, a matrix is out os the question, as cutting the bus bars is not an option. I was thinking (last night tring to sleep no doubt) that if I painted one of the contact with an insulating paint, the problem might be solved. Removing oone of the contacts is not an option. Destroyed one of the drawbars last night trying that out.....Just as well I got the M 3 set as well to experiment on...

Get well soon stryd,

Mark

[henrygr]

Well. That was the M3 drawbars. To set them break before make, remove the drawbar, and apply a bit of superglue to the inner contact, then apply cling-film, or 'cling-wrap', depending on what side of the globe you reside. Replace the drawbar, and Voila, one perfectly baked break before make. Schematics to the AIN will take a while, but i'll post them....,.

Now, the H100 series drawbar. Bizaaaar!!! Different set-up altogether. Take the tram analogy again, but with one contact. Only there is 17 bus bars (1 for ground) set out like below. For the moment, I cannot make heads nor tails of it. Can someone please have a look???