Looking for circuit to add expression to an audio signal on my Rodgers analog organ.

[Eric Mack]

Last year, I split out the String Celeste on my Rodgers 340 analog organ. For that project, I was able to repurpose an unused Rodgers Leslie interface to provide expression to that signal. From what I can tell, that interface tapped off the Light Dependent Resistor (LDR) used for expression on the main board and used that to drive the Leslie interface expression circuit. It worked just fine. (By the way, that is the depth of my current component circuit skills at present. I still have much to learn.)

Now, I'd like to split out another channel, perhaps the flutes, but I do not want to lose expression.

I could try and duplicate the Rodgers expression circuit on a breadboard - it has a handful of transistors, resistors, etc..

Before I attempt that, I wonder if there is an easier more modern expression circuit? Perhaps even one with an IC that would only need a proportional voltage input?

Open to ideas/suggestions.

This is as much for my education as it is for desire. I've learned so much from the people on this forum. I figure the smartest people who could help me can be found here.

[toodles]

One of the key issues with combining two different types of expression circuits is getting them to track equally. There was a Motorola volume control IC that used a voltage signal to control an analog signal, but it is obsolete, I think. There are probably others. But the LDRs used in the Rodgers are not linear, so their expression control circuit uses a voltage curve that is what works for the LDRs--and you'd probably find a volume mismatch which you might only be able to adjust for at one point.

Probably the easiest way to do what you want is to use a transistor as a current amplifier (i.e., an emitter-follower) to drive an additional lamp--package it up with an LDR (or several if you want more channels in the future) and copy the Rodgers expression circuit. The package should be light-tight. You could mimic Rodger's mechanical parts with a 35mm film plastic canister painted black inside & out. The lid could be glued to a circuit board or other thin stiff material, and the wires could go through small holes drilled in it.

[Eric Mack]

Thank you for your detailed response. It makes sense. Most of what you wrote, I understand.

Having opened up the similar black LDR module on my main circuit board (curiosity got to me) I can envision what you describe about a DIY version. Fortunately, I am old enough to know what a 35mm film plastic canister is, and I may still have one around somewhere. 😊

While I can probably duplicate the Rodgers expression circuit from the schematic and looking at the board, I’m certainly not at the point where I would know how to design a current amplifier circuit and I’m not sure I would even know where to connect it. (I’m going to guess that it might be driven from one leg of the light bulb that drives the existing LDR?)

Looking at the expression schematic, I was thinking I should try and find someone who can talk me through the circuit. That would not substitute for electronics theory and study, but it would help me understand why some of the components are where they are.

This is not a time critical project for me. I am enjoying learning more about analog electronics. I am thankful to have met several helpful people along the way.

[johnbeetem]

Looking at the expression schematic, I was thinking I should try and find someone who can talk me through the circuit. That would not substitute for electronics theory and study, but it would help me understand why some of the components are where they are.

Can you post the part of the Rogers 340 schematic that shows the expression pedal circuit and the amplifiers that use LDRs? I have a Rodgers Trio 321C and I think I understand how the LDRs work. I'll be happy to explain what I can. Otherwise I can explain the 321C circuit.

I think nowadays one would use a "programmable gain amplifier" (PGA) -- do an internet search. But I think LDRs were a good way to do it back in the 321C years.

[Eric Mack]

Hi John, thanks for your offer!

I will upload 5 files with a quick summary of each. The Two PDF files are probably the ones to look at but read my summary.
Happy to have a phone call or we can use email/forum. THANK YOU!!

[johnbeetem]

Hi Eric,

Thank you for posting. I'll take a look and report back here so that others can (I hope) benefit.

The Rodgers 340 is quite different from the 321C. The 321C has a very simple circuit for the light bulb and uses op amps with the LDR in the feedback path, so it's pretty easy to grok even with a minimal understanding of op amp circits. The 340 uses discrete transistors, but I'll do my best. I'm a computer engineer with lots of digital knowledge but weak in analog.

[Eric Mack]

Thanks, John. I, too, am weak on the analog side, but I am learning thanks to the web, old Heathkit courses from my teens and helpful people here.

[Eric Mack]

FILE #1. 19700118 - Rodgers 340 - 1365B Output Circuits (High Quality from SN 39192).pdf
This PDF is the mixer board and you will see the two expression circuits for the TIBIA (Top) and MAIN (Middle)

FILE #2. Rodgers 340 Leslie Adapter 6559-416 Schematic 1373 used for Expression Control.pdf
This PDF is the Leslie Interface Adapter, which I am now using to provide expression to the String Celeste channel which I split off from the strings. The leslie circuit has other features that I am not using. I am only using the expression circuit. It is probably cleaner and easier to explain this circuit than the mixer board but I'm guessing they are the same circuit.

FILE #3.
The Leslie Expression Circuit board gets its expression resistance from an LDR on the main mixer board. It does not have its own LDR but rather parallels off the existing two LDRs on the mixer board. So the LDR to the Leslie expression board is wired in parallel with the MAIN LDR 66 on the Output Circuits Schematic 1365B.

John, this may not be useful to you, but for my own record, I dodcumented everything I touched. This photo shows how the MAIN LDR is tapped from the back and sent to teh LDR as noted on the LESLIE interface.

FILE #4.
This is a photo of the Leslie Adapter Interface that I turned into my Celeste Expression Interface. For reference only.

FILE #5. 20180217 - Summary of splitting String Celeste using Rodgers 340 Leslie Interface.docx
No action for you. I am including my summary of how I split the celeste and reused the leslie as an expression circuit in case anyone is curious. I want to share with the OrganForum as much as people have generously shared with me.

[johnbeetem]

OK, I looked at the Output Circuits (aka "mixer board") and Leslie Adaptor schematics. Wow, it's been a while since I've seen "adapter" spelled that way I had to remember a bunch of transistor amplifier theory from my undergrad days, but Wikipedia was quite helpful for refreshing my memory.

First, let's take a look at the Output Circuits Board (OCB) schematic. Take a look at the circuit that's surrounded by a dashed line. It's labeled "Standup Preamp 6464-132" and there are four copies. I'm guessing that if you look at the OCB you'll see four small boards soldered into the OCB, probably at right angles.

The Standard Preamp Board (SPB) is a three-stage transistor amplifier. The first two stages (Q1 and Q2) are "Common Emitter" amplifiers (https://en.wikipedia.org/wiki/Common_emitter). A common emitter amplifier has an input connected to the base of the transistor. The emitter follows the signal at the base. The collector inverts the signal at the base and amplifies the voltage according to the resistors connected to the collector and emitter. The SPB input signal is AC-coupled through C1 and amplified twice through Q1 and Q2. The output of the second stage is Q2's collector and is an amplified, twice-inverted version of the input signal at Q1's base.

R2 is interesting. It couples Q2's emitter back to Q1's base through a 15K resistor. This is a feedback resistor, and I think it's used to produce a fixed gain through Q1 and Q2. Transistors produce a lot of gain, but the actual gain can vary widely. Transistor manufacturers specify a minimum gain, but not the max. So if you want a well-defined gain, you use feedback and set the gain using a resistor ratio. Resistors are much more precise than transistors and don't vary as much over temperature.

Q3 is an "Emitter Follower" (https://en.wikipedia.org/wiki/Common_collector). Q3's emitter follows its base (connected to Q2's collector) with a fixed base-emitter voltage drop, usually 0.6-0.7 volts. The AC portion of the signal doesn't change voltage, but Q3 gives the signal far more current.

Note that there's another feedback resistor, in this case between Q3's emitter and Q1's emitter. This resistor is not on the SPB -- it's on the OCB. This way each instance of the SPB can have a different feedback resistor. The top three SPBs all have 15K resistors, but the pedal's SPB has a different value. I'm guessing that this resistor sets the overall voltage gain of the SPB instance.

The SPB has another AC-coupling capacitor C13 at the output. The input and output capitors allow designers to use the SPB without worrying about the DC level at the input or output.

Now let's look at the light-dependent resistors (LDRs). First, check out the "Main Expression Pedal" circuit in the lower left part of the OCB schematic. R519 is a variable resistor controlled by the pedal. The variable resistor acts as a "Voltage Divider" (https://en.wikipedia.org/wiki/Voltage_divider) to produce a voltage controlled by the pedal. This voltage is amplified by common emitter amplifier Q523 and emitter follower Q524, which produces the current needed to light the bulb L524. As before, there's a feedback resistor R524.

Following the dashed line on the OCB schematic, the light from bulb L524 is detected by LDR 46, which is connected to an SPB output. LDR 46 works with R47 as a voltage divider: the voltage at R47 is the SPB's output times the ratio R47 / (LDR 46 + R47). If the light is weak, LDR 46 will have high resistance and the SPB output will be attenuated and low volume. If the light is strong, LDR 46 will have small resistance and SPB output will remain strong and loud. The signal at R47 is amplified by emitter follower Q49. Q49's emitter is the output of the circuit, which is AC-coupled through C49.

There are some additional resistors and capacitors in the LDR circuit and SPB. They're probably for stability or something like that. We're hitting my analog limits. The additional LDR resistors are a lot larger than R47 so I think we can ignore them for understanding the LDR 46 / R47 voltage divider.

There are three copies of this LDR circuit on the OCB and one more on the Leslie Adaptor (LA). There's a note on the LA schematic that says LDR 21 is actually located on the OCB. It's probably next to one of the LDRs on the OCB schematic, and shares one of the two light bulbs. However, the LDR shouldn't be connected electrically to anything on the OCB, though there may be ground shield on the cable connecting the two boards.

The LA doesn't use an SPB as a preamplifier. Instead it has a "Common Base" amplifier Q18 (https://en.wikipedia.org/wiki/Common_base) followed by emitter follower Q20. Between the two schematics we've covered all the basic transistor amplifier types. Isn't this fun?

[Eric Mack]

Wow thank you. Such detail. I’m going to have to print this out and reflect on it for a few days. But I do have a question upfront. So the Wire pair that goes from the main LDR over to the Leslie interface LDR circuit I suppose it’s not really inserting resistance into the Leslie circuit but rather it’s just supplying a voltage to the Leslie circuit. Is that how it works? I knew that they were supposed to be isolated and so I thought it was odd that they would parallel one LDR over in place of another but if they’re simply providing a signal voltage then that would make sense to me. Like I said, I'll need to print study what you wrote before I get it. Thank you so much

[johnbeetem]

Eric asked: "So the Wire pair that goes from the main LDR over to the Leslie interface LDR circuit I suppose it’s not really inserting resistance into the Leslie circuit but rather it’s just supplying a voltage to the Leslie circuit. Is that how it works?"

According to my understanding of the two schematics, the wire pair is indeed inserting resistance. Specifically, LDR 21 on the Leslie Adaptor schematic is "Located on the Output Circuits Board". LDR 21 shares the light from an OCB light bulb with another OCB LDR. If the Leslie is controlled by the Main expression pedal then it should share L524 with LDR 46. But there should be no electrical connection between the leads of LDR 21 and any other components on the OCB.

I'm guessing that LDR 21 is connected to the Leslie Adaptor through that gray cable in File #3. It looks like there's a ground shield, which is good since you don't want a those LDR 21 wires picking up noise.

This is my understanding from the schematics -- you can take a look at the OCB to see if I've guessed correctly.

[Ben Madison]

Ack, Wikipeadia . I admit I read that website but using it in "Real" work and is not look at nicely by the academic world. I think though it is nice to get a general overview of a topic but to get to the substance I consult books or other websites i would not write my papers with it.
Speaking of papers i need to get hopping on a pipe organ related paper.

[Eric Mack]

John, toodles wrote at the top of this thread, “Probably the easiest way to do what you want is to use a transistor as a current amplifier (i.e., an emitter-follower) to drive an additional lamp--package it up with an LDR (or several if you want more channels in the future) and copy the Rodgers expression circuit.”


Is that current amplifier lamp circuit one you’d feel comfortable specifying?

There is no urgency, however since I have the help available I will explore this project to learn.

[johnbeetem]

As with all comments and advice I give at the Organ Forum, this comment is offered with No Warranty. Now that that's of the way...

I like Toodles' suggestion. Copying the Rodgers circuit improves the likelihood of success.

Let's take another look at the "Main" expression pedal circuit on the Output Circuits schematic. Q523 is a common emitter amplifier. Its collector inverts the voltage at the expression pedal pot. Q523's collector is connected to the base of emitter follower Q524. The voltage at Q524's emitter follows Q524's base with a 0.6 to 0.7 voltage drop. Q524 amplifies the base current so the emitter can drive light bulb L524. R524 provides feedback which I assume is for improving stability and linearity.

Now if you want to drive a second light bulb you can follow Toodle's suggestion of adding another emitter follower. One way to do this is to add a second transistor (call it Q2) wired in parallel with Q524: Connect Q2's collector to +12V, Q2's base to Q524's base, and a second light bulb L2 between Q524's emitter and -12V.

The advantage of this scheme is that L524 and L2 should track exactly the same way, assuming identical light bulbs. The disadvantage is that the presence of Q2 and L2 will draw more current through R523 which will affect the voltage across L524. How much depends on how much Q2 amplifies: if it has very high gain the current used will be small and it won't have much effect. If I were to try this I'd hook up a voltmeter across L524 and see what happens when you add in Q2 and L2. You might need to reduce R523 so that its voltage drop is the same as if Q2 were absent.

Another way to add our emitter follower Q2 is to connect Q2's base to Q524's emitter. The advantage is that Q524's emitter has a lot of current so adding Q2 won't affect it as much. The disadvantage is that you get a second 0.6 - 0.7V voltage drop, so L2 won't be as bright as L524. This configuration of transistors is like a Darlington Pair (https://en.wikipedia.org/wiki/Darlington_transistor) which is a great way to get lots of gain from bipolar transistors.

If you adopt this scheme, you may be able to adjust for the dimmer L2 by adjusting the 4.7K pull-down resistor in the LDR circuit.

I'd try both configurations and see how the bulbs look. Measuring the voltage across lit bulbs is probably more accurate than looking at them.

[Eric Mack]

John, I’m working through yours and toodles explanations of the expression circuit.

So, the computer guy in me is breaking this down into subroutines/functions (i.e. “circuits”). Defined by function, I see four independent circuits:

1. Expression Pedal Circuit which causes a lamp to shine according to the swell shoe position
2. Standard Standup Preampwhich is the same for all channels, but appears to have an external resistor which we suspect establishes the gain
3. LDR Expression Circuit made up of an LDR, a transistor, two capacitors and six resistors. This is the key circuit that we agree needs to be duplicated to maintain the Rodgers design
4. Level & Conditioner Circuit that takes the post-expression circuit and allows fine level control before it sent to the S-100 Amplifier for each speaker channel. There is also some other mystery stuff but I’m guessing that is conditioning of some kind, perhaps to set output voltage levels…

Then, on the Leslie Adapter board, there is another circuit that allows the ECHO ON stop to control what goes into the expression circuit. It begins with R33 and ends with C16 which joins the audio signal into the “common base” amplifier you described on the Leslie interface.

Q1. Is a preamp a preamp? Meaning, can I take a modern IC audio preamp and use that?


Q2a. The ECHO ON Stop control circuit that I described above, is it effectively using Q29 to shunt the audio signal to ground? Is that how it is done?

I would have thought that it would have disconnected the audio into the common base amplifier (preamp) but I guess shunting to ground would kill the signal, too. Assuming I am reading that right?

Q2b. So, if I have understood how they use a stop tab to switch mute an audio signal, is there any reason I could not use that same circuit in other situations where I might want to mute a signal? (No application in mind for that, yet. Just trying to think about what you have shared.)

[johnbeetem]

Hi Eric,

I'll have to get back to later on your questions, seeing as you said there was no urgency. I have to deal with some tax stuff first.

Since you're a computer guy, I recommend checking out the Cypress Semiconductor PSoC -- Programmable System on a Chip. These are microcontrollers that also have programmable digital and analog sections. I'm most familiar with the PSoC5LP (with ARM Cortex-M3) and PSoC4 (with ARM Cortex-M0). You can get development boards for cheap: US$10 for the PCoC5LP and US$4 for the PSoC4. I've only played with the parts a little myself, because you normally need to use Windows-based tools to design with them. The tools are free-as-in-beer, but I generally refuse to use Windows unless someone is paying me

https://www.cypress.com/products/32-...ex-m3-psoc-5lp
https://www.cypress.com/products/32-...rtex-m0-psoc-4

[johnbeetem]

I like the way you broke the functions down into those four independent blocks. You have discovered that in analog design an amplifier has two primary purposes:

1. To increase the power of a signal so it can drive a heavy load like a speaker, a relay, a motor, or a light bulb.
2. To provide directionality, so that your system can be broken into nice, clean modules with a well-defined flow.

Each of the four modules you have identified has a high-impedance input, i.e., it doesn't present much of a load to whatever is driving it. Each of the modules has a low-impedance output, i.e., it has plenty of power to drive the next module(s) and/or components.

Now to your questions...

Q1: Is a preamp a preamp? My answer would be NO. To replace a preamp (such as Rodgers' Standard Preamp board) you have to have compatible supply voltages and compatible input/output signal ranges. Rodgers' Standard Preamp has AC coupling capacitors on the input and output, which means you don't have to worry as much about the DC signal level. However, they're electrolytic capacitors so you need to be sure the levels are compatible with the direction of the capacitor.

Q2a. The ECHO ON Stop control circuit that I described above, is it effectively using Q29 to shunt the audio signal to ground? Is that how it is done?

Yes, that what the ECHO ON circuit is doing. I assume "ECHO ON" means to turn on the Leslie speaker, which would be placed at the far end of the room (hence ECHO). If the ECHO ON stop is off, I'm guessing that nothing is connected to the node labelled "ECHO ON STOP". This means that R30 pulls the base of PNP transistor Q29 towards -12V, which switches it on, effectively shorting the emitter to ground. If there's any audio signal on the + terminal of C16 it gets coupled through to the - terminal and immediately shorted to ground. Capacitors cannot change voltage instantaneously, so if one side is grounded the AC signal on the other side is effectively grounded as well.

When "ECHO ON STOP" is set to +12V, the base of Q29 is raised to 0V and Q29 shuts off. The signal at C16+ still couples through to C16-, but instead of being shorted to ground it can freely swing up and down since R29 is a large resistance. This means C16 has virtually no effect on the signal going to Q18 and so it goes through the rest of the Leslie amplifier.

I think C31 is there to make switching Echo On and Off more gradual, and D32 is for discharging C31 if you turn power off, so that Echo isn't suddenly on when you switch power on.

Q2b. So, if I have understood how they use a stop tab to switch mute an audio signal, is there any reason I could not use that same circuit in other situations where I might want to mute a signal? (No application in mind for that, yet. Just trying to think about what you have shared.)

Nowadays I would use an "analog switch" (https://en.wikipedia.org/wiki/Analogue_switch), which lets you switch an analog signal on and off using a digital control signal. Analog switches use MOS transistors, which weren't available back in the Rodgers 340 and Trio days. There are lots of analog switch parts available nowadays, including multiplexers.

[Eric Mack]

John my microprocessor skills are limited. I’ve done some arduino stuff but limited mostly to experiments with simple circuits.

While I’m eager, there is no urgency. I’ll wait to see what you can share as you have time.

Best of success on your taxes.

[Eric Mack]

For those interested in this thread, now or in the future, here is what an a Light Dependent Resistor Assembly looks like


Here is a photo of the inside of the MAIN the LDR Container :



As you can see, there is a light bulb in the center and then one or more LDR's in a circle around the perimeter, each isolated from the other. Then, there is a black plastic cover. In my Main Swell shoe LDR assembly, I think there will be at least three LDR's - one for the main, one for another channel.

My organ has a third LDR, which was apparently added by Rodgers for the Leslie expression. And, yes, these LDRs are all isolated from one another.

There is room for up to 4 LDR's inside of the container. So, to add another expression circuit, it may be as simple as just soldering in another LDR into an unused position in the existing socket. No need for a separate light bulb assembly. This means all LDRs would have the same bulb driving them. (Yes, I understand we would still need to match LDRs as closely or have a compensating resistor).

UPDATE:For those following this thread, I have also added a PDF of a single LDR assembly that might be helpful LDR Assembly Schematic.pdf

So, if that works, it might not be necessary to add an emitter follower to drive a second light bulb and LDR assembly, as toodles suggested in post #7 above.


That said, if you or toodles would care to suggest the values to "create an emitter-follower to drive an additional lamp--package" I would be very interested to learn from you which transistor and support components you would recommend.

[jbird604]

Excellent idea! Adding another LDR is certainly the first step, then copying the rest of the Rodgers circuit. It "might" just work without any additional circuitry, if you use the LDR as a simple series resistor in your audio path for the additional signal source. I can't recall whether the lamp gets BRIGHTER when you open the shoe (more volume) or brighter when you CLOSE the shoe (to decrease the volume). If it does the former, then you just need to pipe your audio through the LED and it will act like a volume control. If it does the latter, you would use the LDR as a parallel "shunt" across the audio and ground, so that the brightening of the bulb would shunt more audio to ground, lowering the volume.

The action might not be as smooth though as it will be if you actually copy the complete Rodgers circuit.

[twnelson]

The LDR is actually in the feedback path of the expression amp circuit, so depressing the shoe changes the lamp from bright to dark to progressively decrease the negative feedback, increasing the audio level. That's why when an expression lamp goes out for whatever reason, the associated channels go to full volume. It also changes the amp response curve so that the high frequency response is reduced at low volume levels, imitating what happens when pipe chamber shutters are closed. There's also a RC network that slows the voltage change from the expression pot, again simulating the mechanical response of shutters.