Tascam RS-20 Spring Reverb (original version) Help

electric.wreck

New member
A little background, then onto my question --
I run an analog home studio with a Tascam M-520 and a Tascam TSR-8 at the center of it all. Although this is my first post (on any forum actually), I've found so much help and useful information from so many of you.
I'm posting this because a) I'm at my wits end, and b) there doesn't seem to be a good source of information/pictures of the RS-20 (besides what is in the manual).

This particular unit has a problem with (seemingly random) self-oscillation(?). I run it from the M-520 AUX send 1-2 (-10dbV setting on the RS-20), then back into channels 17 & 18. At any given moment, it starts feeding back into itself -- not a squeal, more like a cracking rumble. At times, this occurs when it has been turned on only for 5 minutes; and other times it can run for 3 hours before it starts oscillating (even with no signal going in).

Things I've tried, but haven't worked:
-- Racked and separated from all other gear while still powered through the same outlet as all other gear (plugging it into other outlets causes noise)

-- Adjusted internal trim (R42 & R46) as per manual

-- Cleaned what was the remains of deteriorated foam in and all around the springs (I thought the cause could have been tiny particles riding on the springs)

-- Secured the coils on the transducers (to which the springs attach) with a dot of superglue to ensure there is no movement/vibration

-- Checked wires and connections -- both internal and external. Nothing seems out of the ordinary

-- Opened the unit while it is feeding back. When I hold my finger down on the channel A springs, it reduces the noise, but the oscillation won't stop; and I still feel a charge moving through the springs (with no signal being sent from the console). Holding down the springs on channel B has no effect, just continued feedback

-- Also during feedback, disconnected inputs and outputs, but lights on the front panel still indicate signal (oscillation)


If any of you own one of these and have the patience to open it up and compare -- am I missing anything? Should there be shielding around the transformer? Any advice/experience?

Thanks in advance!
 

Attachments

  • IMG_2685.JPG
    IMG_2685.JPG
    441.5 KB · Views: 28
  • IMG_2684.JPG
    IMG_2684.JPG
    549.2 KB · Views: 30
  • IMG_2683.JPG
    IMG_2683.JPG
    723.2 KB · Views: 43
  • IMG_2682.JPG
    IMG_2682.JPG
    712.5 KB · Views: 32
Greetings and welcome. This is an odd problem. I have the ‘B’ version.

I don’t know if this helps, but the way I would attack this is use an oscilloscope to look at what the oscillation looks like at the output, and then probe backwards up the signal path until I don’t se it anymore. That should be the point at which the fault originates and then one could analyze what might be causing it at that point. It may be you have a failing component that is going into oscillation as it warms up. Another thing you could try is get some freeze spray and work your way up the signal path and spray the transistors in the signal path and any pianos to see if the oscillation stops when a particular component is chilled. Either way, it will be helpful to have the service manual. Do you have that or just the owner’s manual? If you need the service manual it is here: http://studionebula.com/docs/TASCAM_RS20_service_manual.pdf
 
Thanks for the reply; and thanks for that thorough M-520 thread from years ago... it's been incredibly helpful along the way.

I don't have access to a scope right now, but I'll surely take your advice with the freeze spray. I do have the service manual, and it'll be easy enough to follow that schematic. I have a hunch it's on the input side, so I'll start there. It may take a couple of days to get some spray (stay-home order where I am located), but I'll surely report back.

In the meantime, I'll see if anyone else chimes in with a similar experience -- albeit strange and rare.

I really like the sound of this unit. I put a limiter in front of it (not part of the equation involving the oscillation), and it smooths it out really well.
 
I tried out a can of air held upside down (not freeze spray, but it works) and took sweetbeats advice of freezing the signal path one component at a time while the unit was oscillating. I suspected it was coming from the input side, so I started from that side of the springs. It turns out that the culprit was the amp at U1 (an LM1877N-9). I'll get one coming today and will update once the new one is in place.
 
[MENTION=202021]electric.wreck[/MENTION] that’s great! I was thinking of suggesting regular canned air but didn’t know if it would be cold enough. Good job. My gut was the culprit was an opamp. Hope the replacement does the trick. Glad you were able to find one!

I like the sound of my RS-20B quite a bit. It’s a permanent fixture for sure. It’s not an AKG BX series box, but I find it really useful for a lot of things. I think it’s underrated. I like using it for dub stuff, but also more conservatively to create a natural ambiance. Here is a piano solo piece called Snowfall performed by my eldest son some years ago when he was 13 or 14 I think. I need to refresh my memory on the composer. Anyway, I used the RS-20B exclusively on this. The instrument was a vintage Yamaha upright piano in my small carpeted living room; a very dry room. IMO the RS-20B afforded a very natural-sounding concert hall soundscape...note I varied the send level throughout the piece, but increased the level significantly for about the last 30 seconds of the piece, so that’s where you can mostly hear the RS-20B working:

 
What a beautiful piece -- and excellent playing too. Concerning the reverb, this track is a great example of the subtlety that can be maintained with these Tascam units -- which isn't the case for many spring reverbs in a comparable cost range.

My projects are almost exclusively centered around vocals and acoustic guitars, and I tend to under-use reverb. If I need an obvious effect, I usually lean on a 22-2 recorder for tape delay at either 7.5 or 15 ips mixed with reverb (I also EQ the stereo returns different from one another for inconsistency). But this RS-20 is just as you described. It makes a space and puts air in the room.

It is odd, though, how different the original and the B versions are designed. I realize yours has, among other things, a limiter, etc. But even the springs changed with the design (if I remember correctly, the B version has 2 per side, and they are noticeably bigger).

Anyhow, thanks again for your help. Will update soon.
And cheers to stuff that works.
 
Yeah really interesting how different the two versions are...different opamps, boards, springs, features...It looks like I posted pics of the guts of mine years ago in another thread here on this forum:

Tascam RS-20B Story...

It looks like the two sets of dual springs in the 'B' version are different wire size or coil pitch...so that would allow the effect to be more complex, maybe "dense" or diverse.

Do keep us posted. Really hoping the new opamp fixes your issue.
 
Update: I replaced the amp at U1 with a new LM1877N-9, and the unit has been turned on nearly all day with no oscillation.
However, this seems to have created (?) another issue: Channel A (left side) is now distorting (Channel B is clear as a bell -- no problem). I believe this is still an issue occurring on the input side, and it wasn't happening before I replaced the amp.
I'm still inspecting suspicious capacitors, etc. but coming up with nothing.
Faulty amp? Any thoughts?
 
It’s possible the IC might be faulty. I always order more than I need just in case. It doesn’t happen often.

I have also had experiences where replacing a faulty component takes another out that was on its last legs. But that’s rare too.

Bad caps don’t typically present as distortion. It’s usually, depending on the role of the cap, loss of signal altogether, but with signal getting through on peaks, or increased DC artifacts (more pops and clicks)...so it’s just distorted all the time?

Here again, aside from getting another IC and just trying it, having a scope where you can see the distortion and trace back to the point in the signal path where the distortion is no longer evident is the way I’d personally troubleshoot.

If I get a chance this evening I’ll look at the schematic again to see if I have other ideas.

[EDIT]

Also look very closely, like using a magnifying glass and good light to make certain you don’t have any cold solder joints, or stray solder causing energy flow to go somewhere it’s not supposed to go.

If you’re shotgunning components the next thing I would look at is the transistors Q1 and Q2...both 2N5461 parts. It’s possible they were stressed with the oscillation. Somebody smarter than me is going to have to identify the purpose of those transistors...they’re not pass transistors. None of the application notes on the datasheet show transistors in the circuit.
 
Thanks a ton for the tips, sweetbeats. It is indeed frustrating not having a scope at the moment, and I realize I'm just piddling around. Anyhow, I had to place another parts order for other projects, so I threw in a handful of those ICs (along with a socket) and transistors.

The solder joints are solid and contained, and I tested joining components with a multimeter for no resistance.

I wasn't too detailed in my previous post about the distortion on the left side. It is absolutely contingent on the material level coming in. With no input, it is silent as expected. When I do feed it signal, I can get the send from the console low enough to where it doesn't distort, but cranking the receiving channels on the console introduces too much noise.

Just for kicks, last night I switched the transistors (Q1 & Q2) to see if the effect followed the channel, and it did. So I'm pretty hopeful. I don't have much experience with these, but I'm curious how transistors would get stressed. The more I read, the more they seem to be often offenders. In fact, my next project (will certainly have the scope for this one) is the right channel of a A3300SX-2T machine which has a remarkably similar distortion sound happening somewhere between record mode and the playback head. Possibly transistors -- but that's just speculation, and I don't mean to stray from the thread at hand.

Will update when I can. Again, I appreciate your input.
 
Nice sleuthing [MENTION=202021]electric.wreck[/MENTION] ... swapping the transistors. I was going to suggest that but didn't want to spur you to more poking around...more exploratory measures which might invoke more risk. But my hesitancy shouldn't and didn't prevent you from doing it and I think that was a very good test based on my theory.

Looking more at the schematic and the components, Q1 and Q2 are P-channel J-FET parts. There is an application where the transistor serves as a voltage controlled resistor...as voltage increases at the gate, the resistance increases. I'm not an expert in these things but, looking at the schematic for channel A for instance, you'll notice there are essentially two feedback loops, one involving C5 and R11, and the other involving R17 in series with Q1. As the output of U1A increases, the voltage present at the gate of Q1 also increases which increases the resistance at Q1 arresting the gain of U1A. I might be totally wrong about this, but I think this is like a simple voltage controlled gain limiter to reduce the potential *poink* of the spring reverb. In other words, the gain of U1 in this circuit would not be linear, but rather the output gain factor would level off as the input level increases.

Components can go bad. They have limits. Which came first, the chicken or the egg...did your transistor go bad allowing high gain conditions which stressed the opamp and part of its failure state was to crawl into oscillation? Did the opamp fail of natural causes sending harsh signal voltage to the transistor which drove that into early failure? Maybe so...If I was a betting man I'd wager the transistor failed and that brought an early demise to the opamp.

If your experience with the new opamp feels like the gain kind of runs away as you increase the input level, like its good, good, good and suddenly kind of wham-o too hot, that is consistent with what I'm understanding Q1 and Q2 do in the circuit.

Do keep us posted. Interesting problem on a rare piece of equipment.
 
Update:
I replaced the amp at U1 again with a new LM1877N-9 and went ahead and replaced Q1 and Q2 with two new 2N5461 transistors. Everything with this unit now functions as it should. I'm a happy camper.

The first replacement of the U1 amp was labeled "LM1877N-9A." I didn't think anything of it at the time, but when my second order came in, they were labeled "LM1877N-9" (no A). So far, I haven't found any data sheets on these that detail what the "A" means. But my going theory (again, speculation) is that the first IC replacement (A version) didn't play nice with the transistor and somehow crashed it. Before that, my only issue was the amp failing into oscillation as it heated up. Something about that first IC replacement junked the transistor.

All that to say -- the proper LM1877N-9 matched with (2) 2N4561 JFETs did the trick. Of course, other unknown factors may be involved, but I figured I'd share.

I've attached some pictures of the RS-20 schematic and the LM1877 pin diagram just to shed some light on our discussion. Note that the small numbers around "U1A" and "U1B" correspond to the pins on the IC, and that explains the interaction of the input and feedback loop that sweetbeats mentioned.

Concerning the increase of input levels -- coming from the oscillator on the M-520 console at 1K through the AUX 1-2 outs and driving it well into the red (about +6 on the meters), I couldn't get it to sound "too hot." Passing actual musical material sure did make it sound springy and sprongy, though. But not out of line with what springs typically do (to my ears anyway). I think your theory is spot on, but it would take a lot more signal than I'm putting into it to have a sudden wham-o. Perhaps on a circuit that is not for reverb, that effect would be much more obvious.

Thanks again for your input, and I hope this helps any passers-by understand the RS-20 a bit more.

Screen Shot 2020-04-03 at 8.31.12 PM.pngScreen Shot 2020-04-03 at 8.32.53 PM.png
 
Yes!

Glad it’s working. That’s awesome.

I still maintain the original issue was probably related to a failed transistor. And I don’t think your ‘A’ vs non-‘A’ opamp has anything to do with it. Typically those suffix designators simply have to do with the packaging, whether reel, or bulk, or ammo pack, or sometimes it’s simply related to a specific run of the part as ordered in mass quantity by a particular company. 1877 is 1877, regardless of what alphanumeric bits follow that. Did you end up socketing the U1 location in the PCB? If so it would be interesting to put the ‘A’ part in and see if it is problematic. And even at that you won’t know if it was a bad part from the get-go or if it failed on install due to a bad transistor...and you don’t have additional new ‘A’ parts to test. But it would be interesting.

Anyway, I’m very pleased to have assisted in getting your RS-20 working right. I’m glad for that.
 
Ah, ok. Thanks for clarifying the suffix designation.
Just out of curiosity, and because the unit was still opened up on the bench, I installed the "A" amp again (yep, I ended up putting a socket at U1). The distortion came back. When I put the "non-A" back in, it cleared right up again. I realize this doesn't necessarily prove anything about the "how," but I figured I'd document this finding as well.

Ok! Back in the rack. Thanks again for your time, sweetbeats.
 
Hey thanks for following up on that. I was curious. But, right, we don’t know why the ‘A’ part doesn’t work, but at least we know it doesn’t work. Was the oscillation on one side or both?
 
The initial oscillation was on both sides. It was difficult to catch it when it started because it was quite random, but I seem to remember seeing it happen one time where Channel A started, and in a few seconds Channel B followed. Cooling the IC with the can of air stopped it immediately on both sides.
The distortion after the first IC change was on Channel A, and it followed to Channel B when I switched those transistors around.
 
I do want to clarify -- if I had a scope on this, all the answers would be apparent. Through this project, I've only been equipped with my ears, so take these anecdotes with a grain of salt.
 
So your previous post spurs something I was starting to ponder about this whole thing...like, we know you had a bad transistor in there. And we also know something was wonky with the original opamp. But with the fact it oscillates on both sides with the new (but maybe of questionable integrity at this point) ‘A’ opamp, I’ll offer this:

With opamps, as I stated earlier, alphanumeric suffix designators on the part number *usually* have to do with the packaging of the part or something insignificant having to do with a specific customer. I suppose there could be a grading reference like with many transistors (A, B, C...A being the highest performing of a particular type), and if this *is* the case then maybe there are better bandwidth specs or higher slew rate or something on the ‘A’. I don’t know if this is the case with 1877...usually it’s a different part number when there are performance changes like the 4558 vs 4559. But I suppose it’s possible, and if that were the case, and the design of the circuit was tenuous enough, maybe a higher grade part might go into oscillation. Many times when you “upgrade” an opamp to a faster part it is necessary to strap the power inputs to ground using a small value ceramic cap to avoid the opamp oscillating. That seems like a stretch here because you are using the stock part, and it’s not a “fast” part by modern standards. But I suppose if the circuit with the in-built gain limiting feature with Q1 and Q2, who knows? Maybe it’s a sensitive circuit. I’ve had in issue in the past where I had an amp stage that started oscillating with stock parts and I decided it was a design flaw. So you could always, if you’re curious enough, solder in a small value ceramic cap from the + input to ground and the - input to ground of U1 as close as possible to those input pins, and then put the ‘A’ part back in and see if it still oscillates.
 
There may be some confusion as to what I am calling "oscillation" and what I am calling "distortion."

As for the original issue, I described it as oscillation because it was more of an unstoppable feedback (kind of a low-pitched crackling that drove the channel indicators into the red) that was not contingent upon any signal being passed through.

After IC "A" (I suppose we'll call it that!) was installed, the oscillation never came back. However, when (and only when) signal was sent through, Channel A came through distorted/overdriven -- the material was still decipherable, but was fuzzed out. That's when our attention turned to the transistors. The cause/effect relationship between these components gets tricky for me because of those loops. I do see your logic about the cap across the ground/IC ground though.

I think for me to experiment anymore, I would like to get more familiar with this stuff in general (namely, transistor functions) and also have an oscilloscope available. I sure did learn a lot through this and will no doubt pick it back up as other projects demand.
 
I think the fellow who recommended the use of a scope to trace the weird sound is right on. I have done enough troubleshooting in electronics to know that these intermittent problems can be the most difficult to troubleshoot. If you get access to a scope, you should look at the power supply when the oscillation is occurring. One function sometimes overlooked about a power supply is its decoupling function; that is the function of providing isolation between stages on the same power supply. I can envision a case where unintended coupling between two stages is happening because a decoupling capacitor is failing at times. That idea of using circuit cooler is a great one, too. I have used that trick to diagnose intermittent problems on circuit boards. There is, though, the possibility that the spray of the cooler might "blow" the suspected component; but if that happens, you know for sure of one part that you will have to replace, and that part might be the culprit in your case. I know of such a case because I was troubleshooting a distortion in one of those Wurlitzer portable electronic pianos. I sprayed a few parts without any change; but when I sprayed one of the output transistors, there was a pop and the sound went dead. We fixed the piano by replacing the two output transistors in the push-pull circuit. Now, it's possible that I "unbalanced" that push-pull circuit to blow it completely; but the fact is that replacement of both transistors with a matched pair permanently cured the distortion. So I have to believe that the fundamental problem was with one of the original transistors in the amp.
 
Back
Top