Hello. I'm currently making a 2015 smoothie and the phase is barely there. I've listened to a phase 45 on youtube and the difference is night and day. I'm using SMD "MMDF5457"s which I bought as a matched pair. I've checked the pinout and I'm 99% sure it's correct, especially after voltages.
I haven't started audio probing because I'm not sure exactly where I should be probing. The schematic is in the build doc here:
https://www.madbeanpedals.com/projects/archives/1590A/Smoothie.zip (2015 version)
Here are the voltages:
IC1 IC2 g s d
1 3.52 3.52 Q1 1.4-1.7 3.51 3.51
2 3.52 3.52
3 3.35 3.52 Q2 1.42-1.7 3.52 3.52
4 0 0
5 8.17 8.17
6 3.52 1-7.5
7 3.49 3.6-4.7
8 3.51 3.1-5.2
I've attached some pictures too, thank you.
Does the trimmer do anything helpful? Go back and reflow your solder and add some where the pads are still visible. That may help, it's usually step one in troubleshooting.
Sent from my iPhone using Tapatalk
Thanks for your suggestions.
I just added some solder to pads that looked like they needed it and re-flowed some. The trim pot has a very small sweet spot that gives a bit of phase, but it's very faint. It seems to be the same as before.
This seems to me as a issue of LFO gener. and details around. There must be some changes at C8 (10uF) up and down in proper range. Set it at slow freq. and measure how much are changes at that cond. And the same do at gates of those J-FETs. There must be changes at gates cca from cca 3V - 0V. Those values you found there seem to me too low.
BTW, in manual you can read NOTE:
The Phase 45 originally used 2N5952 JFETS. The Smoothie lists 2N5457 for Q1 and Q2 because they are more widely available and work just as well as the 2N5952. You may use either, but keep in mind that the pinout for the 2N5952 is opposite the 2N5457. This means you will need to rotate the transistors 180° on the PCB if you use 2N5952. The moothie also requires the JFET transistors to be matched. While you may be able to stick a couple of random JFETS in there and get lucky, the best result comes from matching the cutoff voltages for the two transistors as closely as possible. This is actually very easy to do, provided you have a quantity of transistors available to work with.
The process involves placing individual transistors in a test circuit on a breadboard. Each device is then measured with a multimeter to read its gate/source cutoff voltage. This is the measurement you want to match with another transistor. The goal is to find two transistors whose Vgs are within just a couple of percent of one another. The closer the match, the better response one gets in the phase circuit. This process takes all of five minutes once you have the circuit ready to test the devices, so do not be discouraged...it is actually very straightforward. The only caveat is that you generally want to have a decent pile of transistors available to match because the Vgs values can range widely between individual transistors. 2550 transistors should be enough to find a match. You should be able to purchase that quantity relatively inexpensively. Plus, you will be left with many extra transistors for future projects!
At C8 I'm getting 3.25~ to 5.1~ roughly.
Gates are about 1.45 to 1.75.
The JFETs I bought were sold to me as 1.305v and 1.311v.
Thank you.
That's kind of low for a phaser, are they 5457s?
Sent from my iPhone using Tapatalk
That zener diode D2 in power source has to be 4V7, so at VB terminal you should have cca 4,5V, so as at Drain of jfets. You have there 3,5V. Try power level at that zener diode D2.
d2 is giving me 3.5v, should I change that out for one that I know is 4.7v?
Thanks.
Yes change it, but try that one, you have inside, when power is not connected with machine. If there is not some shortage, bending power source to low value. This gives me hint, to controll power value of supply after D1. There must be something about power source voltage.
Here is something how to test zener diode:
So, the 5818 is there to prevent you from throwing more than 9 volts into the circuit, it doesn't drop the power down. You can measure your power supply at the input, both sides of the 5818 diode and pin 8 of IC 2 and r22 on the side closest to the IC should all match. Do you have any updated pics? Also, your voltages look backwards. the square pad is pin one, it goes like this:
1 8
2 7
3 6
4 5
Sent from my iPhone using Tapatalk
Primarly, D1 protects circuit from reverted polarity. And little (Uf) brings down power value. But it doesn't matter for our problem, if it is good. It is low schottky diode for 1A / Uf=0,5V. Only, if circuit has some shortage or too high consumption for power source, it will lower current and voltage more, than we need. But now measure that zener diode.
Quote from: Zerro on July 15, 2021, 11:46:14 AM
Primarly, D1 protects circuit from reverted polarity. And little (Uf) brings down power value. But it doesn't matter for our problem, if it is good. It is low schottky diode for 1A / Uf=0,5V. Only, if circuit has some shortage or too high consumption for power source, it will lower current and voltage more, than we need. But now measure that zener diode.
Yeh, sorry, I'm all turned around backwards with the zener. I'll hide behind the current and continue to watch.[emoji106]
Sent from my iPhone using Tapatalk
No problem, it's just that slacjs is probably newbie, so it could little jam his mind :@)
Thanks for all the suggestions and help. I put in a new diode and it seems have improved a lot. It's definitely more phaser-y, though not as much as I would expect, but I wonder if the phase 45 is just a bit less phasery than something like a phase 90?
New diode is giving 4.3 volts now. Here's the new voltages:
ic1 ic2 q1 q2
1 4.37 4.36 g 2.1-2.45v 2.1-2.45v
2 4.37 4.36 s 4.36 4.36
3 4.15 4.35 d 4.36 4.36
4 0 0
5 4.36 4.5 to 5v
6 4.36 4.5 to 5v
7 4.36 1.33 to 8.7v
8 9.38 9.36
input 9.6
d1 9.33 0
r22 9.33
Great! So 4V7 zener gives 4,3V with current consumption. Now, if you have a some more patience, I shoul try here zener 5V1. Little higher voltage at center bias doesn't harm. Try this and we will see. If nothing happens, you should little lower that R19 - 3M9, at cca 3M3 or 2M7. Every j-fet has slightly another thrashold for good functionality. But, maybe that with those 2 phase stages you can't await something more. So, try those tricks, fiddle about bias trimmer, and refer us about. Here are some more tricks, we can still try.
Yep, the phase 45 is exact half of a phase 90. It's very subtle at its deepest.
Sent from my iPhone using Tapatalk
For distorted guitar it would be enough, to hear. Something like wah pedal with low resonancy. Very soft phasing with fuzz or similar fx.
My 2¢ worth:
I had a pair of fets that were matched down to 0.003 volt and used them in the Smoothie and the effect is pretty pronounced for a two stage phaser. JimiLee is correct, it's a fairly subtle phaser but if you have a bunch of fets to get a really tight matched pair, that will help. Also, the trimmer to bias the fets is super sensitive. A 10 turn trimmer will help get the best biasing
Quote from: danfrank on July 26, 2021, 04:46:13 PM
My 2¢ worth:
I had a pair of fets that were matched down to 0.003 volt and used them in the Smoothie and the effect is pretty pronounced for a two stage phaser. JimiLee is correct, it's a fairly subtle phaser but if you have a bunch of fets to get a really tight matched pair, that will help. Also, the trimmer to bias the fets is super sensitive. A 10 turn trimmer will help get the best biasing
Truth. I put j201s in the last phaser I built. Works really well. I've never heard of a 10 turn trimmer. Sounds handy though.
Sent from my iPhone using Tapatalk
These are actually 3 turn but better than a single turn. Mouser has actual 10 turn trimmers
Quote from: danfrank on July 26, 2021, 04:59:13 PM
These are actually 3 turn but better than a single turn. Mouser has actual 10 turn trimmers
Gotcha, thanks.