Fet Phaser Idea

[jambers]

Any one care's to check these PCB's of it?

I've made 2, in consideration the pinout of the fet's or bjt's.

LDR are not on the picture, because I have no library of them. But you can clearly see the where they have to be put; right across the led's.

Suggestions of comments are welcome, since this is my first attempt on making a PCB that's more then a few resistors, capacitors en trannies.
I also used a dual gang pot for the 'Regen', since I think that's what is being used. But I could be wrong

[gaussmarkov]

LDR are not on the picture, because I have no library of them. But you can clearly see the where they have to be put; right across the led's.

o.k.  i just updated gm-resistors.lib to include a 5mm LDR like the old Mouser 338-54C series or a Selco 8P series component.  i haven't double-checked but it should work fine.

[mac]

Please, before making a PCB, I guess it is better to wait for users breadboarding it, and hear their suggestions and mods.
Also, I´m still experimenting: single big LED and BJTs.

jambers,
Did you bradboard it?

mac

[dschwartz]

mac..what´s the current state of this?
i´m very interested in this..like someone said before, is so nice to have something different than a fuzz or a TS variation to discuss.

the circuit is beautiful in its simplicity and tweakability..thanks for sharing!!!

[mac]

I have a deadline in a couple of weeks so I did not have much time for experiments.
Actually I have a both a BJT and FET version on breadboards.
- BOTH: too many regen path. I guess just one going to Q3 base is enough.
- BJT: I want to increase the BJT impedance to have a deeper effect, so I guess I have to use darlingtons. Anyway the regen path restores any loos caused by Z-in.
- FET: I tried placing the LDRs at the drain with the 100k, and the 103 cap at the source in 2 fets, and leave the other 2 unchanged. Just for fun.
I found a little annoying effect at very low speed, when the LEDs go out the signal level increases noticeable. I have to work on that. The BJT version seems a little more consistent about this.

mac

[dschwartz]

you mean when the voltage on the oscliator goes lower than 0.7V and the led goes out, then there´s a loud sound..or that the lower the light, the higher the volume, like a tremolo thing?

if it is the 1rst, then adding a little dc bias to the led would soften the response (like adding a 1V bias voltage)

if second...maybe the balance between the inverted and non inverted parts of the filters is not ok..
or maybe, as somebody said before, i don´t see the clean signal bypassing the filters, as all the other phase designs i have seen..

you can couple A FET and a bjt to get low output Z and high input Z..

[Mick Bailey]

You could drive a lamp with a suitable power transistor and get some bulb latency in there!

[mac]

if second...maybe the balance between the inverted and non inverted parts of the filters is not ok..
or maybe, as somebody said before, i don´t see the clean signal bypassing the filters, as all the other phase designs i have seen..

Yes it is more like a tremolo.
Most likely it is a balance thing, but i am going to connect the mix pot in the usual way: wiper to Q5 gate, other two to Q1 gate and last 104 cap which must be disconnected of Q5 gate.

you can couple A FET and a bjt to get low output Z and high input Z..

I thought about it, like in the booster I posted long ago: a fet buffer direct coupled to a bjt. But I have to add 4 more bjts :S

[dschwartz]

you can couple A FET and a bjt to get low output Z and high input Z..

I thought about it, like in the booster I posted long ago: a fet buffer direct coupled to a bjt. But I have to add 4 more bjts :S

yup, that wil kill the simplicity beauty of it..

it´s possible tu make a phase shifter with a cmos phase inverter?

[TELEFUNKON]

you can even use a transformer!

[mike_a]

Hey mac,

any news on this lovely project? 
did you decide which way to go (i.e BJTs or FETs.)?

Cheers,
Mike.

[Gus]

You can try the univibe bootstrapped darlington stages, make the transistors a single modern darlington

[dellamorte]

this looks like a good candidate to hook up to a vanishing point

[mac]

this looks like a good candidate to hook up to a vanishing point

Hopefully I'll resume soon.

You can try the univibe bootstrapped darlington stages, make the transistors a single modern darlington

Gus, I thought about darlingtons.
BJTs work and their bias are more consistent, and darlingtons can easily replace those univibe bootstrapped transistors. Making Zin over 1M will ensure the same results I got with fets... teoricamente
Maybe 10Ks at C and E, and a >4M7 from B to C. Even 15ks to reduce current draw.

Besides I want it to work with a single big ultra bright led as Miguel suggested. I was wondering which color to use. Human eyes are more sensitive to some freq than others, yellow IIRC, at the same intensity. LDRs might behave the same.

mac

[bioroids]

You should check the specification of the LDRs you are using. The ones I have here, I don't remember how many nm were, but I remember they matched the best with Orange Leds, and second with yellow Leds. I'm using Red leds anyway, because that's what I have in hand. As a matter of fact I'm using one regular 5mm diffused led for 4 LDRs, and it works just fine.

Regards!

Miguel

[mac]

Gracias Miguel.

mac

[mike_a]

To my knowledge, all LDRs sensitivity peaks at around 590nm-630nm (wavelength).
That would make orange and yellow LEDs the best candidates. 
I would go with ultra-brights as they have a larger light intensity swing, which would in turn lead to a larger resistance range of the LDRs.

Regarding the darlingtons, I would suggest trying MPSA14.
They should have very high hfe.

[DougH]

Looks nice! (Somehow I missed this thread.)

Sounds good too! Nice job.

[petemoore]

  This is now the little added 'Accolades' section of the smart thread...
  Nice schematic ! Nice Soundclips ! A departure from what I've seen phaser schematics look like from before...making headway...'less finicky' about components is a good feature.
 
 

[mac]

I'm considering to separate fet and bjt versions on different linked threads.

I'm breadboarding the BJT version using bc549c hfe=500. 10Ks at E and C, 4M7 from B to C. Vc is near 6.3V. 2n5088/9 can be used too.
If a single LED works then I'll test it on the fet version as well.

mac