Stereo/Panning Boss Tr2 Questions...

[Breadboard Warrior]

Hi everyone.

I've just designed a modded TR2 to run in stereo for my fender rhodes and I have a few questions for the more experienced guys around here.  

I have a recent TR2 (standard, mono) and I chose this as a base for the build mostly because the wave can be swept from square to triangle.  For the rhodes this is a really nice feature.  My real TR2 has a more modern VCA chip, but the schematics of the original are easily available on the net, plus it uses an M5207L01 which is actually a stereo chip.  And these are still available on good ol' ebay out of china for just a few bucks.

Anyway, after simulating the LFO and tweaking it to get the duty cycle to 50%, I created a 2nd control voltage with an inverting opamp as an exact mirror to the original.  All this was easy with spice and a little breadboarding.

The trouble was when I used the M5207L01 as per the old TR2 scematic, it had a very low output.  Not just slightly low, but really low - 1/10th of the input.



Notice the M5207 has an input resistor (R9) and is directly linked to the inverting input of the following opamp (IC2A).  This has a 22K feedback resistor (R12).

This pretty much follows the application note in the datasheet (http://www.datasheetarchive.com/dlmain/Datasheets-20/DSA-396453.pdf), which states that the output will have no attenuation if the opamp's feedback R is double the input R and the control voltage is 1 Volt above 1/2Vcc.  All of this checks out, but for me, it didn't work.

With a little playing around on the breadboard, I found I needed this instead.



As you can see the opamp is a standard inverting buffer with a 100K input Z.  Any lower than this and there was a drop in volume.

So what gives?  Would my new, "made in China", m5207L01 be different than back in the day?

(PS: I tried it with several opamps, not just the TL072.)

[armdnrdy]

The M5207L01 that you sourced IS probably from "back in the day"

It sounds to me like you are trying to turn the TR-2 circuit into a PN-2.

The TR-2 has a well known volume drop:
https://www.google.com/search?sourceid=navclient&aq=&oq=Boss+TR-2&ie=UTF-8&rlz=1T4GGHP_enUS450US450&q=boss+tr-2+mod&gs_l=hp..1.0l5.0.0.0.8522...........0.LxIt8ii8lRg

[PRR]

> Any lower than this and there was a drop in volume.

The datasheet you linked (which *appears* to be the right part) shows this output stage:



The output is Open-Collector.

It should be "infinite" impedance (or so near you could hardly tell a difference).

That your <100K into a current amplifier is causing a volume drop says there IS some non-infinite output impedance, defying the implication of that documentation.

i.e. gain should be only about the 22K NFB resistor. Resistance between the VCA and the opamp virtual earth pin should only cause errors (perhaps small), not change of gain.

Either your build is not as intended, or there's something odd about these chips.

[Breadboard Warrior]

The M5207L01 that you sourced IS probably from "back in the day"

Maybe.  As I mentioned the current TR2 that I have has a different (mono) VCA chip.  But I know the M5207L01 is still being used for tremolo pedals because my mate has one of these:  http://www.ebay.com.au/itm/New-Crossfire-Tremolo-Guitar-Effects-Pedal-/171116447828?pt=LH_DefaultDomain_15&hash=item27d755c854

It is not a clone, but a very close copy of the original TR2, complete with the M5207L01.  It's an excellent pedal and I was thinking about buying one and modding it, but it's SMT and pretty tricky to mod, plus a real TR2 came up 2nd hand for not much more.  I was disappointed to find they'd changed the circuit!  So I started from scratch.  I guess that cheap crossfire pedal could be just a company using up some huge N.O.S chip supply out of china, and I found an ebay guy with the same source...  Who knows...  But I'm actually thinking that the output Z of the chip has changed, but the only data sheet I could find is original.

It sounds to me like you are trying to turn the TR-2 circuit into a PN-2.

Yeah, I know about these.  They are pretty rare, and don't have the triangle to square sweep pot which I really wanted.  I also searched without luck for a schematic for them , so I'm not sure if the LFO is the same or not.  I'm fairly sure the behringer ultra tremolo is a PN2 copy, but it would also be surface mount of course, so I never bothered.  Who knows, maybe I could have just re-housed a behringer, and thrown a pot in to replace the switch...

The TR-2 has a well known volume drop:

When I started researching this project, this came up a lot, but as I mentioned this wasn't the subtle, "perceived" volume drop that players complain about, but a massive volume drop.  I didn't actually measure it accurately, but it would have been around 20dB.

That your <100K into a current amplifier is causing a volume drop says there IS some non-infinite output impedance, defying the implication of that documentation.

Good to know someone else thinks the same and I'm not crazy!

Either your build is not as intended, or there's something odd about these chips.

Well it happened with both the chips I bought, and both in the circuit and on the breadboard, so I think it's the latter.

Thanks heaps for the answers guys.

Is this common?  Where a chip is redesigned or is copied inaccurately?  For a while I wondered if mine were mis-labeled m5207L05's.  These need a 5V control voltage for zero attenuation.  But the breadboard proved wrong.  With a direct cap coupling at the output pins, the chip behaved as it should.  The scope probe has a 1Meg Z.  So when I added the opamp I started with 1Meg and worked down till it started to drop the signal. 

I guess you just can't really trust parts out of china these days.  Maybe I'll take another closer look at my mate's crossfire pedal and see if there's a load resistor between the VCA and it's opamp.

Anyway, the pedal works really well.  I used it on a gig last night...

[armdnrdy]

The PN-2 has a 220Ω resistor and a 100K pull up to VR (+4.5V) between the depth control wiper and pin 8 of the M5207L01.

The circuits are not exactly the same but...you may want to lower R8 (100K) and lose C4 (.1µf)

Maybe that will work for your lower gain IC.

Also..you might try sourcing another M5207L01 from a different supplier to see if it makes a difference.

[Mark Hammer]

To my mind, the value of stereo panning corresponds to the nature of the stereo pan.  And that, in turn, depends partly on the stereo setup.

If the speakers are spaced close together, a pure anti-phase pan (A is always the opposite in volume of B) provides nice shimmer and animation (as befits a Rhodes).  If the speakers are farther apart, though, one might prefer modulation by a quadrature LFO, with one channel always 90 degrees ahead of the other, instead of 180, such that the panning is of a less pronounced nature.

Just an idea.

If I recall correctly, the classic Rhodes setup was a cab right under the keyboard, such that the stereo panning didn't seem to swing wildly about.

[Breadboard Warrior]

you may want to lower R8 (100K) and lose C4 (.1µf)

Hmm OK, I might try that.  The control voltage(s) are pretty much perfectly at 1V above 1/2 Vcc and with the depth at full the lower swing is right on 1/2 Vcc.  On the breadboard, I carefully pushed the control voltage higher than 1V, and the chip started to boost for a half a volt or so then quickly started squaring off one side of the wave.  So I guess I could push that upper limit slightly.  I could also tweak the bias points (R28, R32) of the final opamp in the LFO.  This would take the whole wave up.  With my extra opamp that inverts the control voltage, I also inverted these two bias resistors to get a symmetrical mirror for the other channel.  This was one tweak I made on the prototype that was slightly different from the spice model.  If the control voltages aren't exactly centred and opposite, the panning is lopsided!

.you might try sourcing another M5207L01 from a different supplier to see if it makes a difference.

I'm afraid there's only one guy on ebay who sells them and I've looked at all the usual digikey, mouser, rs, etc etc... If I went ahead and made more of these or tried to turn it into a commercial product (very unlikely - there's two or three pedals already that have a wave control), I think I'd just use two of the chips that the current TR2 has.  It would use a little more current, but it has lower noise.

one might prefer modulation by a quadrature LFO, with one channel always 90 degrees ahead of the other

A really nice idea.  I'll definitely check it out.  Is that part of what these leslie sim circuits do?  

The gig on thursday night was in a small club, and I didn't use a keyboard amp, so when my two channels were mixed to my monitor send in equal measure, the panning disappeared of course.  I got the sound guy to drop one aux to half the other so my (mono) monitor would at least have some regular tremolo.  I was thinking, "I hope the audience appreciates this stereo tremolo I've spent so much time on!"...