MUFF WIGGLER Forum Index
 FAQ & Terms of UseFAQ & Terms Of Use   Wiggler RadioMW Radio   Muff Wiggler TwitterTwitter   Support the site @ PatreonPatreon 
 SearchSearch   RegisterSign up   Log inLog in 
WIGGLING 'LITE' IN GUEST MODE

Roland TR-626 - DIY Slider Pot Mixer Idea
MUFF WIGGLER Forum Index -> Music Tech DIY  
Author Roland TR-626 - DIY Slider Pot Mixer Idea
Sunshine Jones
Hey,

I find myself with a Roland TR-626 on my desk. Haven't seen one of these since the 80's. The little brother, TR-505, was the drum machine I got beaten up over in the mid 80's a couple of times. Dudes shouting "Get a real drummer!" through the set (which actually used to happen all the time) but a couple of times conversation and understanding failed and it came to blows... 6'' 4" me with my satin scarf and false eyelashes dodging the ham fists of angry people who just felt that a person using sticks to make sounds was a better idea than me and my drum machine.

As if time hasn't more or less left that attitude in the dumpster, and people who respond this way along with it, I still feel that somehow something wonderful needs to be done with this arguably glorified remote control unit, and I'd like to try to put it into service.



I have seen the hacks and mods over at burnkit and while they're fun, I think what I'd like to do is to use Harry Axten's ROM expansion:
http://harryaxten.altervista.org/tr626expansion.html

And then enjoy a full flavored 12 bit/8bit sample world of the LINN and the DMX as well as arguably the very worst of Roland's sample library.

What's missing for me is the ability to at-a-glance adjust the volume of each individual drum.

I noticed this over at burnkit:



This is a TR-505 Mod from someone called Squelchbox from Ontario. I've tried to find them and ask them how they did this, but I haven't had any luck.

I think that there's no love lost in creating the channels for the slide potentiometers, and I might assume that the 707 style replacement slide potentiometers would electronically do the job:

https://www.mouser.com/ProductDetail/Bourns/PTA1543-2215CIB503?qs=sGAE piMZZMtC25l1F4XBU6KQRTmUDfzVQWaKO74%252bJhA%3d

Bourns 15mm throw, 50k Ohm

There are 8 individual outputs, so that suggests that there will need to be 8 potentiometers to control the outputs.

But my question at this point is where would or might one insert these pots?



Looking at the schematic I marvel and wonder.
I would not want pots following IC24 I don't think... that would then only be a volume control for the individual outputs themselves.

So would they insert before or after the transistors from IC18?
Would the follow the output of IC's 20 - 23?

Further, since I've never really just stuck a slider into a circuit before I wonder about the need for resistors between the insert and ext points?

If you were hoping to vindicate a device, as I am, by adding a 707/727 style mixer to it's front panel, where you would begin?

I hope this is as interesting to someone else as it suddenly has become to me.
Sunshine Jones
Well... sad, but it seems like either it's not interesting (and how could I blame ya?) or no one knows.

But the good news is that I think I've found Squelchbox!
I wrote to him on face-page and if he replies and has some insight, I will document and follow up here for the sake of anyone else with this sort of question in the mysterious future.
grenert
I think it's a pretty interesting idea, but perhaps most people would just adjust levels on a mixer. To add a mix slider to one of the individual outs, break the connection between the last filtering capacitor (C45, C46, etc.) and the output jack. Then add a 50K log slider in parallel with that capacitor. Be sure to orient correctly (slider terminal 1 at ground). Connect the slider terminal (terminal 2) to the output jack. It's pretty straightforward. If you look up designs for audio mixers, you should find similar circuits to learn from. I think the hardest part will be the physical side - modifying the case and mounting the sliders.

EDIT to add: Adding slider pots as I've described will also change their level of slgnal in the combined output as well, since that's taken right at the very end of the signal chain before the individual outs.

Have fun!
heapish
I'm interested. I think I may have some notes on it...
Ive seen it done but with a horrid looking box and pots on top. Sliders are preferable like a 707.
Sunshine Jones
grenert wrote:
I think it's a pretty interesting idea, but perhaps most people would just adjust levels on a mixer. To add a mix slider to one of the individual outs, break the connection between the last filtering capacitor (C45, C46, etc.) and the output jack. Then add a 50K log slider in parallel with that capacitor. Be sure to orient correctly (slider terminal 1 at ground). Connect the slider terminal to the output jack. It's pretty straightforward. If you look up designs for audio mixers, you should find similar circuits to learn from. I think the hardest part will be the physical side - modifying the case and mounting the sliders.

EDIT to add: Adding slider pots as I've described will also change their level of slgnal in the combined output as well, since that's taken right at the very end of the signal chain before the individual outs.

Have fun!


This is GREAT information. Thank you.

It's strange, I hadn't thought to look at mixers for guidance. It seemed like all these potential sound output points didn't resemble a mixer, and that all seemed to happen inside the CPU with software, thus pages to control it.

I also imagined that if I were to tap just before the outputs I would only be creating a mixer for the individual outputs. And of course, I'd like a hardware representation of the "software" mixer.

But what you're suggesting seems like it would both interrupt the main out, and also control the level of the individual output?

I'm exited to try this.

I agree that the challenge will be how to securely mount the slide pots so that it's a robust and effective tool. I was imagining a sort of frame to mount them into, and then secure that to the inside chassis (plastic) I wouldn't like to drill into the panel more than in order to create the channels for the sliders.

It's not something to just bust out the jig saw and start cutting... even if it is only a "TR-626," it's still a vintage machine, and I would like to do it right.

Thanks.
I will follow up and report my progress.
Sunshine Jones
heapish wrote:
I'm interested. I think I may have some notes on it...
Ive seen it done but with a horrid looking box and pots on top. Sliders are preferable like a 707.


Yes, exactly. The 707/727 style mixer would change the nature of the 626 so deeply that it could really be a lot of fun.

I'd love to see any notes you can find.
grenert
Sunshine Jones wrote:

I also imagined that if I were to tap just before the outputs I would only be creating a mixer for the individual outputs. And of course, I'd like a hardware representation of the "software" mixer.

But what you're suggesting seems like it would both interrupt the main out, and also control the level of the individual output?


Yes, both would be affected. For example, the bass drum individual out passes to the left and right main outputs through resistors R164 and R165 (not sure which resistor goes to what side, since that part of the schematic is cut off). So changing the level before those resistors also changes the main outputs. As a side point, note that the resistors going to the left and right outputs are not the same for some sounds (hihat has 47K and 120K resistors, for example). This is the basis for the default right-left panning of some of the sounds in the main outs.

I don't know much about the 707. Does it allow you to fade every sound, or just some of them? On the 626, my suggestions would only allow you to change volume for sounds that are available from the individual outs.

Harry Axten's Linn/Oberheim EPROM mod is really cool and completely worth the effort. Don't forget to account for space for the added switches for his mod when you are putting in faders.
Sunshine Jones
grenert wrote:
Yes, both would be affected. For example, the bass drum individual out passes to the left and right main outputs through resistors R164 and R165 (not sure which resistor goes to what side, since that part of the schematic is cut off). So changing the level before those resistors also changes the main outputs.


Far out. Ok. I think I may completely understand. So the resistor value is what's producing the output, and to connect a 50k linear slide potentiometer will then attenuate the amount prior to its distribution, yes?

The individual output is then a cut in, where it is then removed from the main mix as it should be, but the attenuated resistor remains the volume control for the individual sound itself.

Very cool. Exactly what I wanted to accomplish. Thank you.

grenert wrote:
As a side point, note that the resistors going to the left and right outputs are not the same for some sounds (hihat has 47K and 120K resistors, for example). This is the basis for the default right-left panning of some of the sounds in the main outs.


That's a sharp and well noted point here. I believe that the outputs are not assignable, and thus, it is fixed in terms of which individual output the sounds come out. I believe it may even be slot based, if not software controlled in the CPU. Since it's PCM samples coming out, there is no sound source for the individual sounds apart from the ROM IC. So Roland have allotted output groups, and they are, I believe, fixed.

grenert wrote:
I don't know much about the 707. Does it allow you to fade every sound, or just some of them? On the 626, my suggestions would only allow you to change volume for sounds that are available from the individual outs.


The 707 and 727 only have one bank of sounds. So the sounds within it are all there is, there is no other layer or bank. So yes, each sound, and the master volume are given their own fader. It makes working with a rompler machine a lot more fun in that rather than making a bunch of patterns to make subtle changes, the user can just throw a few faders and it's a breakdown! Then bring back in the cowbell, add a little delay, and then throw the faders back up... endless fun.

It's not at all the same when you need to press menu, and shift, and then tap the level up or down for each sound in the dark... not the same at all.

grenert wrote:
Harry Axten's Linn/Oberheim EPROM mod is really cool and completely worth the effort. Don't forget to account for space for the added switches for his mod when you are putting in faders.


Totally. Those switches are vital. Really looking forward to Harry's expander. DMX and LINN and a few 707 sounds in this machine would make it a sort of pan-80's explosion of goodness...

Adding the manually controlled mixer will make this thing wonderfully useful.
I really really appreciate your notes, and help.

Thank you.
Sunshine Jones
A little news:
I heard back from SquelchBox and they directed me back to burnkit2600 because that's what they did to create their mixer modification:

It reads:

Quote:
Individual Outputs

C28 – C35 – You can wire up jacks to these points for individual outputs. Decreasing the value of the resistors right near the caps (R63, etc) will lower the volume of that drum in the mix. Decreasing the value of the resistors near the transistors (R64, etc) will raise the gain of that drum to a slight crunch. Instead of doing individual outputs, I was able to wire up 100k pots in parallel to the pair of resistors for each channel. This lets you bring the volume of that drum from nothing, to normal, to crunch.


Here's the link to that:
http://www.burnkit2600.com/tr-505/#outputs

Lots of other fun things to do (like trigger outs etc.)

But bear in mind that this is all for the TR-505
They are very similar devices, but they are not the same, so R63 isn't likely to be the same thing on both devices.

It's great to read a success story, and this makes the project look way more doable.

Now to find the correct 50k linear slide potentiometers.
I think the style of the 626 asks for black top, white line sliders. So the Alps faders used for the 707 would not be best. They would need a cap.

Unless there's a source for black and white 707 style caps out there somewhere...

I will continue to post progress...
grenert
For audio mixer purposes, you will want to use audio (log) potentiometers, not linear. The reason has to to with psychoacoustics and the non-linear way we hear sound volume. I personally don't know this field, but this particular aspect is well established.
Sunshine Jones
grenert wrote:
For audio mixer purposes, you will want to use audio (log) potentiometers, not linear. The reason has to to with psychoacoustics and the non-linear way we hear sound volume. I personally don't know this field, but this particular aspect is well established.


* DANG *

Starts an email to mouser...
Sunshine Jones
Well, I opened up the TR-626 and had a look around. Based on the conversation here and going from the TR-505 Mod for output volume pots, I metered and found the last two resistors before the output.

Here is the exploration study:



Based on the TR-505 Modification to add potentiometers for the volume:



I extrapolated the same configuration for the TR-626 output mods using a B50K slide potentiometer with a 15mm throw.



Here is a schematic view of it



Sadly it doesn't work. The volume of the kick is not effected in the mix, nor from the individual output.

Back to the drawing board.
Sunshine Jones
Well, I took another look at the volume mod diagram for the TR-505 and compared it with the schematic and board of the 626 and moved away from the outputs and back a step to the transistor and resistor relationship to see if this might create a physical volume fader.

There is an array of transistor and three resistors for each of the 8 output groups. I applied the same potentiometer to one of these groups to see if it would adjust the volume:



Here is the hypothesis in diagram



Result: FAIL

No influence whatsoever upon the output volume.

* dang *
Sunshine Jones
Well, I took another look at the schematic, and got out the DMM and followed the trail of connections and spotted where the previous idea had not worked:
I was tapping into the resistors too soon.

Still following the Burnkit2600 modification idea for the TR-505 output volume mod, I re framed, and applied the theory again.

This time with the two resistors before the transistor at the sound generation before the output, and specifically at the BD point.



Here is the note/diagram of the modification:


Result: FAIL

Still no control over the output volume.
I sense that I am close, but perhaps there is something fundamental that I am missing which might be obvious to the seasoned schematic reader... I said in the beginning here that I understood the concept, but did not understand where to apply this mod. So I'm learning a lot, but it would be super cool to have something, really anything, actually happen when I move the fader.

I believe I may need an adult, or a more experienced set of eyes to help me figure this out.
Sunshine Jones
Tried another theory, this one provided from the Face-Page Synth DIY group as a reply from someone who has successfully performed this "easy mod" on a TR-505.

The suggestion was to loop in between R105 and R97
This seemed strange to me, as these points don't have continuity, but because the suggester had done the mod, and gotten it working, I decided to give it a try.



Here is the schematic



Result: FAIL

I didn't think that this wold work, but it did get me rooting around a little more and looking at the schematic and making comparisons with the DMM and I decided to try this same technique, but move the connections for the unrelated resistor (R105) back into the loop and tried R104, but the result was the same.

Fail.
Sunshine Jones
Finally, I tried something suggested by a user in PM here from Muff's

A totally different approach: Rather than work with the transistor at the VCA, instead take a look at the output jacks and inset a potentiometer between the tip of the jack and the last capacitor before the jack. The instructions were also to cut the trace between the tip and the capacitor so that the slide potentiometer would take the place of the resistor.

I don't have the heart to cut the trace. Not without some indication that this will work. I realize I am really just fumbling around inside the TR-626 with a DMM and a hot soldering iron, and in my experience this is how I tend to fry things and break them.

So I wired it up without cutting anything:



Here is the schematic



I love this approach because it's all happening on the underside of the PCB, and the board is very clearly marked, so it was a very clean, and simple hook up.
Unfortunately without cutting the trace the result was FAIL.

I hate to fail and fail and fail, but that's a huge part of this exploration. I'm a little discouraged, but I still think that adding a tactile mixer to the front panel of the 626 is time well spent, and will deeply improve the device.

I'm going to put it back together and give this a bit more thought, perhaps consult an adult and see if another set of more experienced eyes can assist me.
Houtson
Sunshine Jones wrote:
Finally, I tried something suggested by a user in PM here from Muff's

A totally different approach: Rather than work with the transistor at the VCA, instead take a look at the output jacks and inset a potentiometer between the tip of the jack and the last capacitor before the jack. The instructions were also to cut the trace between the tip and the capacitor so that the slide potentiometer would take the place of the resistor.

I don't have the heart to cut the trace. Not without some indication that this will work. I realize I am really just fumbling around inside the TR-626 with a DMM and a hot soldering iron, and in my experience this is how I tend to fry things and break them.

So I wired it up without cutting anything:



Here is the schematic



I love this approach because it's all happening on the underside of the PCB, and the board is very clearly marked, so it was a very clean, and simple hook up.
Unfortunately without cutting the trace the result was FAIL.

I hate to fail and fail and fail, but that's a huge part of this exploration. I'm a little discouraged, but I still think that adding a tactile mixer to the front panel of the 626 is time well spent, and will deeply improve the device.

I'm going to put it back together and give this a bit more thought, perhaps consult an adult and see if another set of more experienced eyes can assist me.



Hi Sunshine Jones - you’ll need to cut the trace for that last one to have an effect (without doing that your slider doesn’t really have any effect as the trace shorts between the signal and the wiper).

If you want to try it without cutting the trace you could wire up another headphone jack and connect that to your slider and the 626, or desolder/remove the 626 jack and connect it with wires to the slider and 626.

Cheers
Sunshine Jones
Houtson wrote:
You’ll need to cut the trace for that last one to have an effect (without doing that your slider doesn’t really have any effect as the trace shorts between the signal and the wiper).

If you want to try it without cutting the trace you could wire up another headphone jack and connect that to your slider and the 626, or desolder/remove the 626 jack and connect it with wires to the slider and 626.

Cheers


Far out. But I'm confused...
It sounds like you're saying that a trace within the jack itself must be cut?
I assumed it was the trace between the tip connector of JK3 (yellow wire) and the rest of the board.

What am I missing?
Houtson
Sunshine Jones wrote:
Houtson wrote:
You’ll need to cut the trace for that last one to have an effect (without doing that your slider doesn’t really have any effect as the trace shorts between the signal and the wiper).

If you want to try it without cutting the trace you could wire up another headphone jack and connect that to your slider and the 626, or desolder/remove the 626 jack and connect it with wires to the slider and 626.

Cheers


Far out. But I'm confused...
It sounds like you're saying that a trace within the jack itself must be cut?
I assumed it was the trace between the tip connector of JK3 (yellow wire) and the rest of the board.

What am I missing?


That’s right, it is the trace between the tip connector and the rest of the board, just like your picture

however I understand you don’t want to cut the board so an alternative is to get a new connector (either new or desolder one off the 626) and then wire that directly to the board (missing out the trace that you have marked to be cut) - that will let you test it without cutting the board. Hope that makes sense
Sunshine Jones
Well, I've had some success at last.

Without cutting a trace, and at the suggestion of Tomash Ghz I moved down a Cap and resistor. I guess I wasn't reading the schematic correctly, and working on the wrong trace, so no wonder nothing worked.

Dropping down to C25 and R105 back in the transistor section before the OP Amps I have successfully managed to connect a physical B50K slide potentiometer to the Bass Drum and it works!

No traces to cut.

Here's the note card



And here are the practical images of the hookup





Notes:
It's a little crunchy. Not exactly Bass Drum to silence as it probably should be. Personally, I think the crunch is a little bit of an improvement over the standard sample rom, but that wasn't what I set out to do.

So I am going to test this hookup on a different instrument and be sure that it's actually working, and then consider the most popular cut trace solution, as I feel that this will produce mixer quality/post op amp / post transistor results which will probably be the quietest and more elegant solution.

However, it works, it works! And it got done to one degree of success or another without cutting a trace.

More as it comes together.
grenert
I implemented hypothesis five on my TR-626. You can see the cut trace between where the green and yellow wires are soldered:

Good and maybe bad news:
I didn't have a 50k log pot on hand, only 100k and 25k log pots. The good part is that it does completely shut off the sound and open it up completely from both the individual out and the mix out. The possibly bad news is that for the mix out, it seemed like most of the effect of the pot seemed to be in the first half of the pot range, rather than across the whole range. It seemed worse with the 100k pot. The reason why I say "possibly bad" is that I only did a fairly cursory listening test, didn't actually look at the output on a meter to really see how it was changing. But definitely this will work to fade in/out sounds from the mix out. It's just that it may be more sensitive than desired. The individual out seemed to work normally. Note that assessment might be just because I *know* that it will work normally, since it's really just adding a volume control at the output. It may be that the mix out is similar sounding, but I haven't really measured it and am not certain from the schematic if it may behave differently from the individual out.

I look forward to hearing about your progress! I wasn't initially thinking I'd want to do this on my own TR, but having played around with sliders fading sounds in and out on my Minipops, I really like that ability!
Sunshine Jones
grenert wrote:
I implemented hypothesis five on my TR-626. You can see the cut trace between where the green and yellow wires are soldered:
Good and maybe bad news:
I didn't have a 50k log pot on hand, only 100k and 25k log pots. The good part is that it does completely shut off the sound and open it up completely from both the individual out and the mix out. The possibly bad news is that for the mix out, it seemed like most of the effect of the pot seemed to be in the first half of the pot range, rather than across the whole range. It seemed worse with the 100k pot. The reason why I say "possibly bad" is that I only did a fairly cursory listening test, didn't actually look at the output on a meter to really see how it was changing. But definitely this will work to fade in/out sounds from the mix out. It's just that it may be more sensitive than desired. The individual out seemed to work normally. Note that assessment might be just because I *know* that it will work normally, since it's really just adding a volume control at the output. It may be that the mix out is similar sounding, but I haven't really measured it and am not certain from the schematic if it may behave differently from the individual out.

I look forward to hearing about your progress! I wasn't initially thinking I'd want to do this on my own TR, but having played around with sliders fading sounds in and out on my Minipops, I really like that ability!



I think at the output a 10k pot makes the most sense. So it stands to reason that the 25k pot worked better. Not sure what's going on with the mix output.

My explorations have been interesting as well:
The B50K slide pots back before Q3 between C25 and R105 works great. Excellent response at the individual out as well as the mix.
However:
The 0-detent is all of the action I'm looking for. Silence to 100%. The detent to fully open produced an overdrive. That could be cool for benders and digital heads, but that's really not what I was after.

Not entirely sure how to meet the capacitance or voltage that's present at that insert point, and match it with a 15mm throw slide pot. There are many 10k this size, but I found various other measurements challenging to score.

The real hassle been when I flipped the PCB and attempted to instal it on the front (for convenient and easy mounting to the front panel) and I couldn't get it to work.

I have to spend a little more time, measure, and be sure I'm not inverting the circuit. Seems the likely culprit.

I think having tactile faders for these digital 80's machines is a must for sure. Love it.

Thanks for your help @grenert . I'll report back when I've got more progress.
MUFF WIGGLER Forum Index -> Music Tech DIY  
Page 1 of 1
Powered by phpBB © phpBB Group