Maestro MP-1 clone - help for choosing the FET

[gigimarga]

Hi all!

I've just finished a MP-1 clone using this schematic (page 9): http://www.jedistar.com/pdf/maestro_phaser.pdf (and a project called "Der Meister Phaser" from another site).
The clear signal passes OK the circuit, but I have no phasing effect.

I obtained a pulsing voltage from 3 V to 6.2 V, i obtained 3.9V in the junction point of Q1A/Q1B/Q3, but, even after I 've tried over 20 BF245A/BF245B/2N5457, I wasn't able to obtain 0.6V from the LFO.

At the page 4 it's described a procedure about how to set the trimpot, but I haven't the required tools.

In the list of parts, the FET is E101. Which FET is this, because I can't find any infopmation about it?
It's a simpler way to choose the FET?

Thx all!

[isildur100]

According to datasheetarchive.com, a J201 is equivalent to E101. You could try that.

[gigimarga]

Thx a lot isildur100!
I've tested the only two J201 that I have with no luck 

[gigimarga]

I've tested over 40 pieces of J201/2N3819/BF245A/2N5457 and only 2 of 2N3819 seems to make a little change in the sound when the trimpot is maximized (the source voltage becomes 3.9V)...it seems very strange to me that no FET worked...I changed the trimpot from 10K to 100K, but no luck too

[isildur100]

I do not think it is because of the jfet, unless you are using the wrong pin-out. A j201 should work. The trimmer is there to center the phase sweep and on my MP-1, it is set near the middle position. This is a complex build and you probably have another problem somewhere in there. Use a J201, make sure you are using the correct pin out and then verify your build for errors. I'm almost sure it has nothing to do with the jfet.

[gigimarga]

I do not think it is because of the jfet, unless you are using the wrong pin-out. A j201 should work. The trimmer is there to center the phase sweep and on my MP-1, it is set near the middle position. This is a complex build and you probably have another problem somewhere in there. Use a J201, make sure you are using the correct pin out and then verify your build for errors. I'm almost sure it has nothing to do with the jfet.

Thx a lot for your kindness isildur100!
The PCB it's very tight (you can find it on another site if you search for "der meister phaser"...) and I found a lot of soldering errors, but I hope that I removed all
Anyway, excusing me for bothering you, can you post the voltages from the original unit?

[isildur100]

I will try to get you voltages but I can't guarantee I will have time this week-end Tell me what voltages you are interested in.

[gigimarga]

Thx a lot isildur and don't be hurried
I mainly need the voltages from the LFO, because the clean signall passes the phaser and it's a little amplified, so I think that the problem is from the LFO

Thx again!

[PRR]

Replace the JFET and trimmer with a 20K resistor. This will work: it will sweep the 3094 chips with a nearly exponential current. The FET just makes a more-exact sweep, and as John says, the trimmer sets the center pitch. But a 20K resistor will give a very obvious sweep. And it avoids selecting FETs or trimmers, or possible pin-out confusion.

I agree that this is probably not the problem.

The matching of Q1A and Q1B is important. They are two different devices, and may not be "matched" close enough for this odd circuit.

Disconnect the collector of Q1B and connect it to ground through a 5K (or 4.7K) resistor. Short R49. Ideally you want 1.0V across the 5K resistor. Anything 0.5V to 2v will surely "work", just not at the center-pitch that Norlin intended. If you get that, un-short R49. The voltage on the 5K resistor should sweep to near zero and up to about 5V. If it sweeps too fast to read on your meter, tack 220uFd caps across C15 and C16 to slow it way down.

If that works, put a 20K (or 22K) resistor from +9V to the pad that you took Q1B's collector out of. This biases the 3094 chips to a medium current so they will pass signal. Lift one end of R37, the direct signal path. Now the output should be the naked phase-shift signal. It probably won't have any strong "effect" without the direct signal to cancel against. It may be muffled. It may hiss. But you should get the input signal. If nothing comes out, look for trouble in that long chain of 3094s and associated parts.

[gigimarga]

Thx a lot PRR!

I will try tonight all that you wrote (and I hope I will make it alive!).

What I forgot to say is that in "Der Meister Phaser" the CA3094s are replaced by halves of LM13700. I think this could affect the biasing.

I used a 2N3904 with a gain of 282 and a 2N3906 with a gain of 286. It's enough good? May I match something else?

Anyway, two peoples said that built it succesfully, one using a 2N5457 and one using a MPF102, so it's verified.

[gigimarga]

Replace the JFET and trimmer with a 20K resistor. This will work: it will sweep the 3094 chips with a nearly exponential current. The FET just makes a more-exact sweep, and as John says, the trimmer sets the center pitch. But a 20K resistor will give a very obvious sweep. And it avoids selecting FETs or trimmers, or possible pin-out confusion.

I agree that this is probably not the problem.

The matching of Q1A and Q1B is important. They are two different devices, and may not be "matched" close enough for this odd circuit.

Disconnect the collector of Q1B and connect it to ground through a 5K (or 4.7K) resistor. Short R49. Ideally you want 1.0V across the 5K resistor. Anything 0.5V to 2v will surely "work", just not at the center-pitch that Norlin intended. If you get that, un-short R49. The voltage on the 5K resistor should sweep to near zero and up to about 5V. If it sweeps too fast to read on your meter, tack 220uFd caps across C15 and C16 to slow it way down.

If that works, put a 20K (or 22K) resistor from +9V to the pad that you took Q1B's collector out of. This biases the 3094 chips to a medium current so they will pass signal. Lift one end of R37, the direct signal path. Now the output should be the naked phase-shift signal. It probably won't have any strong "effect" without the direct signal to cancel against. It may be muffled. It may hiss. But you should get the input signal. If nothing comes out, look for trouble in that long chain of 3094s and associated parts.

Tonight I made the tests above (thx a lot PRR again) with the following results:

1. "Replace the JFET and trimmer with a 20K resistor." - no change...I heard only the clean signal
2. "Disconnect the collector of Q1B and connect it to ground through a 5K (or 4.7K) resistor. Short R49. Ideally you want 1.0V across the 5K resistor." - I used a 4.7K resistor, but I got only 0.42V across it (with a fresh battery - 9.2V)
3. "If you get that, un-short R49. The voltage on the 5K resistor should sweep to near zero and up to about 5V. If it sweeps too fast to read on your meter, tack 220uFd caps across C15 and C16 to slow it way down." - I didn't get, but I un-shorted R49. The voltage on the 4.7K resistor didn't sweep: it was stucked to 0.41V!

Because this didn't worked, I didn't made the last step ("If that works, put a 20K (or 22K) resistor from +9V to the pad that you took Q1B's collector out of."), because I was afraid that I can damage the LM13700s.

I think that the problems are from bads soldering of mine (the PCB is very tight, with a ground plane, and my solder iron is very poor). I will try to remake a larger PCB for it, but it's hard to believe that I will have time soon for that

Thanks a lot all for your help! 

[PRR]

> I didn't made the last step

You should.

[gigimarga]

> I didn't made the last step

You should.

OK...I will try as soon as possible.

Thx a lot!

[gigimarga]

Sorry for replying with a so long delay, but I have no time for DIYing last and present week.

I made the last step of testing ("If that works, put a 20K (or 22K) resistor from +9V to the pad that you took Q1B's collector out of."), but I didn't heard any sound. I'm almost sure that problems are from the poor PCB and soldering.

I hope that I will have some time to made a bigger PCB next year

Thx a lot all for your help, with a special mention to PRR.