Roland Jet Phaser LFO issue

[armdnrdy]

I've posted previously about an issue that I was encountering at the bottom of the sweep of a Jet Phaser build. When the sweep gets to the bottom, it sort of falls apart, stays down there a bit before regrouping to work its way back up. The best description would be when the sweep gets to the bottom it gets confused.

Since that time I've tried many things to curb the problem with many great suggestions from Mark Hammer and R.G. (thank you both) with nothing working to remove the glitch beyond just masking it a bit.

I've tried five different types of JFETs (including the original 2SK30Y) matched with R.G's JFET matcher and the VP matcher that stemmed from STM and Rob Stand's posts. I've also tried many different JFET voltage ranges, but the problem remained. None of the different sets made any difference regarding the issue and there has always been excellent phasing.

Mark suggested that the problem might be in the feedback area and brought forth several things to try. I tried lowering the gain of the last phase stage by changing R78 to a lower value, C13 to a lower value to where the lower frequencies of the feedback path wouldn't come barreling through, and raising R19 which lowers the feedback signal. None of these adjustments really solved the problem.

Update:
I contacted a DIYstompboxes forum member, Tonemonger, from Tasmania.....yes Tasmania! Who owns an original Roland Jet Phaser. He was kind enough to verify that all of the components in the factory drawings are correct. I asked him about the bottom of the sweep of the original. This was his reply:

"Sadly I know exactly the glitch you are describing!
It sounds like it stumbles, looks around to get its bearings, realizes which way it was going and then continues on its way."

So...what we have here is a design flaw in the Jet Phaser! This flaw keeps one from adjusting the controls to certain settings. The problem is more prevalent when the Jet level, Depth, and Resonance are turned up, but even with those controls backed off, the problem remains. This issue also rears it's ugly head when strumming a bit heavier while in the regular phaser mode. The bottom of the sweep distorts.

I took voltage readings of the LFO area and other key area to see what I could find. I found that the LFO seems to swing more negative than positive. Here are some of the readings.

IC3 pin 8 LFO output (max +3.58 min -3.94)
IC3 pin 14                (max +7.24 min -7.72)
Q2  gate                  (max -1.19   min -1.46)
Q8 one of the phase stage JFETs (max -1.19 min -1.46)

Here is a more legible redrawn schematic:
http://www.aronnelson.com/gallery/main.php/v/diyuser/Roland+AP-7+Jet+Phaser+Redraw.jpg.html

I think that if there is a way to bring the negative voltage up a bit, that might solve the problem. I had the idea to cut the negative voltage trace going to IC3 pin 11, and add a schottky diode to drop the voltage 20-30 millivolts?

Any ideas how to solve this issue?

[slacker]

You could try messing with the bias section, making R40 bigger and or R38 smaller should move the voltage range the LFO sweeps up a bit.

[armdnrdy]

Hey Ian,

Thanks for the reply.

The Jet Phaser has a very distinct sound. If I adjust the bias to where the entire LFO is shifted up out of the problem area, then it doesn't sound like a Jet Phaser any longer. It changes the whole character of the sound. There is a sweet spot as far as getting the wah type talking sound.

http://www.youtube.com/watch?v=kaueqA4wNNg

Best examples from around 5:27 on.

[Fender3D]

You may change LFO's "bias".
Try connecting R28 to a trimmer wiper and feed it with a (little) more positive/negative voltage instead of GND.
This should raise or lower LFO's wavefom where you need it

[armdnrdy]

Hey Federico,

So lift the grounded side of R28, attach it to the wiper of a trimmer and feed one of the outer lugs positive or negative voltage? Is that right?

[slacker]

One outer lug to ground, the other to positive or negative supply. I think that will just do the same as adjusting the bias though. From what you said it sounds like if you move the bias up, the LFO ends up sweeping higher than you want.
What you might want to do is change the depth or range of the LFO so that it sweeps up to the same place but doesn't go quite as low. This will keep it in the area you want. You could try replacing R34 with a pot, lowering its resistance will reduce the range of voltages the LFO sweeps, by tweaking that and the bias you might get what you want.

[armdnrdy]

Thanks Slacker,

Well, I have a few more things to try. I just have to fix this glitch! When the sweep bottoms out, the phasing goes away as well. All you're left with is fuzz in the Jet mode. This just will not do! Roland should have addressed this back in the day!

[Fender3D]

One outer lug to ground, the other to positive or negative supply. I think that will just do the same as adjusting the bias though. From what you said it sounds like if you move the bias up, the LFO ends up sweeping higher than you want.
What you might want to do is change the depth or range of the LFO so that it sweeps up to the same place but doesn't go quite as low. This will keep it in the area you want. You could try replacing R34 with a pot, lowering its resistance will reduce the range of voltages the LFO sweeps, by tweaking that and the bias you might get what you want.

+1

[armdnrdy]

OK, good,

I have to go to work for a bit and when I return I'll give it a try.

Thanks guys!

[armdnrdy]

Well...I tried the suggested fixes and neither seemed to do the trick.

I lifted the ground side of R28, connected it to a 20K pot, (wiper) connected lug one to ground and lug three to +v and then to -V. Both voltages produce a shift in the whole LFO. I could move the LFO out of the problem area but just like adjusting the bias, it would lose depth, and tone, and it would not sound like a Jet Phaser any longer.

Replacing R34 with a pot adjusted the rate of the sweep. It seems that C16, R31, and R34 around IC3C all control rate.

I don't really understand what role D4 and Q2 are playing in this LFO circuit. If someone could enlighten me, maybe I could come up with a fix.

Any other suggestions are extremely welcome.

[Mark Hammer]

Um, perhaps Ian meant R38?

As for the non-phasing, imagine what you would hear if a triangle or sinusoidal wave was veaily clipped at one end of the excursion, such that it was essentially "square" on that end.  It would appear to sweep in one direction, come back a bit, and then sort of "freeze" for a little while before continuing sweep in the opposite direction.  At slow enough speeds, there will appear to be no phasing, because there is no real sweep.  Remember it is the constant movement that creates the effect.

I'm suspecting that if you scope the LFO, you might see that.  On the other hand, if the JFETs are all matched and their range is uniformly mismatched to the parameters of the sweep generator, then the LFO signal might look good, but not have any audible impact on JFET resistance for a portion of the sweep.  Again, the reason why we match JFETs in phasers is so that none of them reach that "This is as far as I go and I'm not budging" point before any of the others; i.e., all JFETs are always changing in response to the sweep generator.  If they all uniformly reach that no-more-change point (and heaven only knows how Roland selected them back in the day), then you're LFO can look good, but be mismatched to the JFETS (or vice versa).

As we may have discussed off-line previously, the range of sweep can be dickered with via the control element (JFET), the caps in each allpass section (C20/22/24...etc), OR the resistor that parallels the JFET in each section (R44/49/54...etc).  So, once you get the bias set to where the sweep seems to be continuous, without freezing at any point, perhaps it's time to dicker with those resistors or caps.  Making either or both of them larger in value will lower the point of the first notch....which may get you closer to the desired sound.  Obviously, of the two options, swapping resistors is the safer and cheaper.

One thing I would suggest for testing, though, is to somehow isolate the phase shifted signal for, say, two or maybe 4 stages, make some experimental changes, and listen for any possible improvements to phasing, before going to town on the whole thing.

[slacker]

Sorry my bad, it's R84 not 34, the 12k attached to the depth switch.

[armdnrdy]

Hey slacker,

I had previously tried changing the value of R84. Lowering it lessens the problem at the bottom of the sweep but also kills the depth considerably which is essential for this phaser's unique sound.

Mark,

When I scope the LFO I see something a bit unusual. (At least to me) I see some sort of wave that almost looks like DC with irregular sawtooth waves spaced sparsely across the vertical graticule. The whole line will drop down the screen (drops in amplitude) and then pulse into a huge, tight sine wave very quickly and then back to the line I previously described. After that display it heads back up in amplitude and pulses into a huge, tight sine again.

As far as JFETs go, I ran across a post of someone receiving help from R.G. with a Jet Phaser that wasn't phasing. This gentleman pulled the 2SK30s out of his phaser and measured them with R.G.s matcher. The values didn't seem to be very close matches, but it was established that they averaged around -0.80.

After trying many different types and ranges matched with R.G.s matcher and a vp matcher with no apparent improvement, I settled on a closely matched set matched with R.Gs matcher, the vp matcher and matched by resistance off. That's a heap of matching!

I've printed out all of your off-line suggestions and have been checking them off as I try them out. It was yesterday that I entertained the idea of changing the phase resistors/caps. I think I'll try pulling the resistors first and socket  them to try different values.

[Mark Hammer]

When I scope the LFO I see something a bit unusual. (At least to me) I see some sort of wave that almost looks like DC with irregular sawtooth waves spaced sparsely across the vertical graticule. The whole line will drop down the screen (drops in amplitude) and then pulse into a huge, tight sine wave very quickly and then back to the line I previously described. After that display it heads back up in amplitude and pulses into a huge, tight sine again.

THAT is weird. Of course, what we don't really know much about is whether there is any particular spec to shoot for with respect to the JFET in the sweep generator.

[armdnrdy]

what we don't really know much about is whether there is any particular spec to shoot for with respect to the JFET in the sweep generator.

After receiving the info about the issue at the bottom of the sweep in an original Jet Phaser from Tonemonger, (outline in blue writing above) I'm now fairly convinced that my build is acting "normal". With that being said, "normal" will not do! As I stated earlier, this issue limits the usability of a lot of settings. ( It's a bit ridiculous to have a phaser that distorts while strumming in the lower part of the sweep) At this point I think that I'm trying to fix a design flaw.