Removing knob from switch on Roland JC77

[lukes]

A guitar signal is very low level and carried on an unbalanced cable (active/hotter pickups can help this). Because of this, we need lots of gain in order to be able to get it up to speaker level. It's nearly impossible to achieve that without also amplifying some noise along the way.

Hi Dude,
thanks for you message, but may I ask a question? What's the amplitude ratio in a pickup coil compared to a turntable cartridge? MC cartridges emit signals measured in hundreds of microvolts, and are amplified without noise (not a these levels anyways), and this over a full audio spectrum and at rather high quality levels, obviously. Why is it unreasonable to ask the front end of a guitar amplifier to do the same?

Sorry, I must be missing something because I really don't get it...

L

[g1]

Guitar amps are not expected to be quiet at a level like a hifi system is.
What I believe to be at issue here: are you trying to find a defective part, or a shortcoming in the design?
Consider that the amp is meant more for an on-stage environment than for home use, would the noise level be objectionable?
I'm not saying there is no problem with the amp, just trying to determine how noisy it is compared to other guitar amps.

[nosaj]

Guitar amps are not expected to be quiet at a level like a hifi system is.
What I believe to be at issue here: are you trying to find a defective part, or a shortcoming in the design?
Consider that the amp is meant more for an on-stage environment than for home use, would the noise level be objectionable?
I'm not saying there is no problem with the amp, just trying to determine how noisy it is compared to other guitar amps.

I would also like to add that a Peavey 212 duel I was working on had objectional hum coming from it till I put the chassis back in the cabinet then it was quiet as a mouse with the faraday cage replaced.

nosaj

[lukes]

Guitar amps are not expected to be quiet at a level like a hifi system is.

Agreed, except two points: first off it wouldn't be bad if they were quiet-ish, and second many normal amplifiers are a lot quieter than mine (random Fender or Vox from a shop I mean).

What I believe to be at issue here: are you trying to find a defective part, or a shortcoming in the design?

I have no direct evidence that any part is faulty (the reverb tank was, and I've replaced it). Btw, is the chorus 'rate' switch supposed to 'click' on the 0 side? I didn't notice it did that before.
And I haven't set out for a redesign of an amp that would in the obviously give me, well... another amp. It was more that when the amp was designed only certain parts were available to the designer: some hadn't been invented yet, some for cost/inventory/manufacturing reasons. You know like a '741/TL071/OPA27/OPA623 kind of historical path: processes and technologies improve and defects just get removed from designs.

The concept is more like: maybe the supply rails are noisy for a reason or another, maybe the first stage or two can benefit from a better jfet, things like that: small changes that would improve performance without affecting the design and spirit of the amp.

Example (which is arguably more about tidyness than anything, to be true): the ridiculous way in which the ground for the speaker is taken off the contact tab on the chorus and reverb foot switches, by means of a cold contact onto the *painted* chassis. Solder it to a ground tab (like the main one, which happens to be in the perfect place in the 77) and move on. Why did the designer choose that point? Likely to have a single wire per tab. Was that a good choice? Possibly yes given his constraints, but not one I like in my amplifier.

The 4558 opamps in this amp I don't believe were ever top performers, but are certainly poor performers by today's standards, would it be advantageous to replace them? I'm not sure, but it wouldn't be much work and it might be worth it. On the other hand, the final stage topology is minimalistic at best, but I can totally see that if you're driving a single cone it's likely to be all you need. Still a better output device (MJL3128, MJ21193) can only help, by reducing the demands on the linearization action of the Nfb circuit. You know we're talking spending a couple hours resoldering maybe 30$ worth of parts... No big deal.

Consider that the amp is meant more for an on-stage environment than for home use, would the noise level be objectionable?
I'm not saying there is no problem with the amp, just trying to determine how noisy it is compared to other guitar amps.

I thought there were tons of famous records made by miking amps like this one. I can't believe it would have been as noisy as mine is, the engineers would've had a hell of a time denoising it without making a mess of the guitar, especially in a pre-digital era...

In terms of comparing noise levels: when I play an amp in a shop "quiet room" (which is obviously louder than my house) I never heard any obvious white noise. I used to own a Fender Super Twin, which was louder than anybody in their sane mind would ever want to have, and I have no recollection whatsoever of noise (any pedal/effect attache to it was certainly far noisier than the amp itself).

As to absolute numbers, there are a few in my previous post, I notice I forgot to mention two point: my meter was set to 'fast' for the measurement (and I obviously waited for it to settle before I took the number down) and the base room noise at that time was about 33dBA. This means that at volume setting ~3-4 you'd have maybe 38dBA of white noise, quiet conversation kind of level. Would it be annoying for a gig in a cafe? Probably not ideal, but definitely workable. Would I record with it? No chance.

Oh, as to the material: I mostly play latin jazz, bossa nova and the like on archtops (at present I'm using an ES175 and an L5). I haven't tried my acoustics on it yet. I read some folks saying they do sound pretty good on these amps (I think they happened to be talking about 120's and 50's if I remember right).

Also, as I said, I'm also a curious person that asks a lot of questions, I hope that's not annoying people around here.

Cheers
L

[lukes]

I would also like to add that a Peavey 212 duel I was working on had objectional hum coming from it till I put the chassis back in the cabinet then it was quiet as a mouse with the faraday cage replaced.

nosaj

Yes indeed, same here. There is a half shield of tinfoil glued to the input side of the cabinet. It help a lot with hum. Seems to do near nothing with the white noise.

L

[Enzo]

yes, I think most recording studios at one time had a JC120 behind a door somewhere waiting to be used for a nice clean tone.

So let's assume you just have excess noise, rather than a design flaw. Track it down and isolate it. You say the tone controls affect the noise, good, that means the noise comes in before those controls. ANY control that has ANY effect on the noise is after the source. Likewise, controls having zero effect on the noise are either before the source or not in the same signal path.

One can use a scope and/or a signal tracer to follow the noise back through the amp. Usually hiss comes from a semiconductor.

We use the typical preamp out and power amp in jacks, or the FX loop jacks to isolate the front end or rear end of an amp. if the preamp has any inserts, we can use them the same way.

[lukes]

Hi Enzo,
thanks for your message.

let's assume you just have excess noise, rather than a design flaw. Track it down and isolate it. You say the tone controls affect the noise, good, that means the noise comes in before those controls. ANY control that has ANY effect on the noise is after the source. Likewise, controls having zero effect on the noise are either before the source or not in the same signal path.

One can use a scope and/or a signal tracer to follow the noise back through the amp. Usually hiss comes from a semiconductor.

Yes I do have a scope, I'm short of the correct schematic though. I think I have enough photographs now to put together a schematic of my board, I'll use the guidance of the 120 and 50 schematics I found to give it a similar layout, so that it'll look familiar. I don't know how long it'll take to get this done: I have two young daughters and rather full work days, it might be a week or two before I'm able to report progress with this.

We use the typical preamp out and power amp in jacks, or the FX loop jacks to isolate the front end or rear end of an amp. if the preamp has any inserts, we can use them the same way.

The only thing I have are "stereo" outputs. From the schematics I've seen these can be simple taps off the speakers outputs, so I'm guessing chasing those won't be much use.

I'll post the schematic once I have it done.

Btw, I would have thought there would be a small capacitor to ground near the entrance to shunt away RF garbage, but I haven't found one on the schematics I've seen. Is this not normally done on guitar circuits?

L

[Enzo]

Not then, more often now. But I see a 100pf cap at the input, shunting to ground.

The stereo outs are line outs from the preamps.

I thought I looked at Nick's site, but here it is:
http://bmamps.com/Schematics/Roland/...0Schematic.pdf

If the tone controls affect it, then I start to suspect IC2 is noisy.

You know there is another option. Roland. Call Roland and order the service manual for that model. They might charge you $10-15-20-30, but isn;t it worth it? Roland manuals are similar to those from Yamaha and Korg. Complete schematics, board layouts, parts lists, any special disassembly procedures (less of an issue in guitar amps than say keyboards), any diagnostics procedures, as well as any adjustments. I sure like to find schematics free online like anyone else, but for a complex amp like this, it seems like we wasted $100 of time wallowing in the fog already.

Just an option to consider.

[lukes]

Not then, more often now. But I see a 100pf cap at the input, shunting to ground.

The stereo outs are line outs from the preamps.

I thought I looked at Nick's site, but here it is:
http://bmamps.com/Schematics/Roland/...0Schematic.pdf

Indeed! And it's the right one: PM code matches the one on my board.

If the tone controls affect it, then I start to suspect IC2 is noisy.

You know there is another option. Roland. Call Roland and order the service manual for that model. They might charge you $10-15-20-30, but isn;t it worth it? Roland manuals are similar to those from Yamaha and Korg. Complete schematics, board layouts, parts lists, any special disassembly procedures (less of an issue in guitar amps than say keyboards), any diagnostics procedures, as well as any adjustments. I sure like to find schematics free online like anyone else, but for a complex amp like this, it seems like we wasted $100 of time wallowing in the fog already.

Just an option to consider.

It would be worth indeed. I do live in New Zealand, I'm guessing email might be more the way to go here, and then finding a way to convince them to post me the booklet, or email a pdf. But yes you're right, it would totally be worth it.

I'll check on their website what they say about support.

L

[lukes]

Hi Enzo,
sorry another question: I was looking at this family of schematics for a few days now, and I am a bit puzzled by the amplifier on the input side of the reverb.
I don't understand the feedback structure very well: on the one side the output buffer is mysteriously outside of the feedback loop (per-se, at least: and yes I did notice its power supply is GND/-33V, all DC-lifted with capacitors), but then the ground of the return path from the tank is lifted with a 4.7 Ohm resistor (which I'm suspecting means I need a tank with higher impedance than the 8Ohm I bought, darn...) effectively making a tap at "2/3rds@1kHz" in some sort of decay-with-frequency sense which is then fed back into the inverting input? Thus achieving what? A rise-with-frequency amplitude response? (Seems to be mild, it's taken off a 39k/2.2k partitor, earlier 120's have 680/680 and the bottom resistor to ground on the side of the tank is 0.5Ohm).

What's going on there, do you know?

L

[Enzo]

Didn't know you were in NZ, have no idea what Roland does down there.

I rarely look at circuits like that, I just assume they were designed correctly if th amps run properly as a group. But what I see as I look is a current feedback circuit. Perhaps it is a poor analysis, but if I draw 10ma through that 4.7 ohm resistor, I get about 0.05v, then the 20/1 voltage divider that the 39k and 2.2k represent knocks that down to about 0.002v feeding back. If I change that to 0.5 ohms and a pair of 680 ohms, my 10ma current now drops 0.005v, but is only divided 2/1, which gets me the same 0.002v feedback. In other words, we dropped the sensing resistor by a factor of ten,. and at the same time lowered the voltage division by a factor of 10. So the overall result ought to be similar. If I did that wrong, someone will correct me. I don;t see it as frequency compensating.

By the way, this looks to be a whole Roland manual, not for yours, but for the JC120/160, just to show the information included.
http://bmamps.com/Schematics/Roland/...e%20Manual.pdf

[lukes]

Didn't know you were in NZ, have no idea what Roland does down there.

Apparently nothing: the email goes to their australian site. I wrote, we'll see if they answer.
There is also an Auckland number, I'll in a day or two if the email does not do much useful.
Thanks for the tip, though

I rarely look at circuits like that, I just assume they were designed correctly if th amps run properly as a group. But what I see as I look is a current feedback circuit. Perhaps it is a poor analysis, but if I draw 10ma through that 4.7 ohm resistor, I get about 0.05v, then the 20/1 voltage divider that the 39k and 2.2k represent knocks that down to about 0.002v feeding back. If I change that to 0.5 ohms and a pair of 680 ohms, my 10ma current now drops 0.005v, but is only divided 2/1, which gets me the same 0.002v feedback. In other words, we dropped the sensing resistor by a factor of ten,. and at the same time lowered the voltage division by a factor of 10. So the overall result ought to be similar. If I did that wrong, someone will correct me. I don;t see it as frequency compensating.

Yes, I see what you mean. Makes perfect sense to me. The question then is: the positive in to this opamp is ~300mV (per quote near where the input is taken o the line above). Shouldn't this voltage on the inverting input be comparable? 200mv seem to imply there'd be 1A through the coil, which seems excessive, or is it? At the quoted 1.4V is would be dissipating ~ 1.5W which looking at how tiny it is in practice maybe it's a lot? Arguably most of the signal ought to come from the feedback resistor R15, so maybe this could be 50mA (6:1) ratio, but that' still 250mA into the coil... I don't know at all if that's too much or not. I have a vague memory of seeing a page at one point with recommended current/voltage ratings for Accutronics tanks, I'll see if I find it again.

The frequency bit occurred to me as follows: the buffer is a voltage source, and it's in series with an inductor (8Ohm@1kHz is maybe 1.275mH?) being the winding in the input transducer of the tank.

(I got the inductance playing with this thing Inductor Impedance Calculator)

Now, given there's a moving magnet in the middle of its field, I'd imagine the impedance growing faster with frequency that if it was simply a coil in air, because of the effect of the mass being moved around becoming harder and harder as frequency grows, but still, it ought to be a coarse approximation of some kind

So, compared with your typical potential divider, I thought the top "resistor" here, being the tank, would effectively make the bottom one going to ground smaller and smaller in comparison, no?

By the way, this looks to be a whole Roland manual, not for yours, but for the JC120/160, just to show the information included.
http://bmamps.com/Schematics/Roland/...e%20Manual.pdf

Wow, thanks! I notice this one (page 3) has that .5->680/680 scheme I was telling you about, for example.

Whereas this 1979 JC50 here http://music-electronics-forum.com/a...hematic-79.jpg has a slightly different structure: no buffer, the supply is right off a TA9200P (which I'm guessing is a buffer, but I haven't found it on the web). The tank model seems different and the feedback if 0.5->680/220. The fact that it has 10R in series with the load might indicate the load impedance to be a fair bit higher.
Kinda cool to see that little zener regulator down to 15V right next to the reverb, but actually serving all the effects to its right... (gets to 14.5 due to a resistor in the collector circuit of Q24, center of the bottom row).


I want to think about what you said some more. Thanks for taking the time, really appreciated!

Cheers,
L

[lukes]

Alright. I think I'm starting to see it. I think it's a scheme trying to compensate for the rise in impedance in the coil driving the input transducer. As the impedance rises, the feedback from the coil drops in voltage because the coil is a higher impedance at higher frequency. The feedback dropping means more gain, a higher voltage from the buffer, and a new equilibrium established a bit higher. It's not obvious to me how this blend of the linear response of the opamp on its own (well with R15) and just a bit of the coil will be enough though: per your calculations, with a drive of 28mA on the coil, we'd have about 6mV on the feedback line, which is 2% of the 300mV we have on the non-inverting input, or about 35dB down if I'm not mistaken. Isn't this too much trouble to be worth the fuss?

Rod Elliott has some notes on this here Spring Reverb (Tables 1 and 2). His circuit (being some 30years younger and all) feels much more natural to me: a 5532 opamp with a BC639/640 buffer. Few caps missing here and there, but as a concept it seems right on the money.

Also, here there are a few schematics from other manufacturers https://www.amplifiedparts.com/tech_...d_and_compared all around similar concepts. The blues junior is simpler but very similar to Rod's, and the '63 Reverb uses the feedback on the screen, kinda smells a bit like an ultralinear in that case.

Thanks again for your help Enzo,
L

[lukes]

Roland. Call Roland and order the service manual for that model.

And good advice this was indeed! The service guy has already sent me the pdf of the manual, just asking in return not to share it on the web.
I am amazed at their promptness, I have to say.

FWIW, it contains front and back photograph, board layout and details of the bottom mount components, schematic and a full parts list.
It's very similar to the one for the JC120/160 Enzo posted earlier in the thread

Thanks for the tip, Enzo!
L

[Enzo]

Might be a typo, I think the reverb driver in those certain versions was a TA7200 rather than 9200. I have a few of those ICs in fact, I just inventoried them a day or two ago. The TA7200 is a 3.5w power amp. After all, that is what a reverb driver is: a small power amp. Google TA7200 for a data sheet, if you are interested.

The other versions used an op amp with a basic push pull transistor output as the little power amp.

Come to think of it, I also have a couple of the DPDT lever switches, probably like the one your knob was on. Long flat shaft blade. Straight out at rest, and flips down to one side, but not the other side.