Author Topic: JFET FX loop, some questions (schem) (Read 8935 times)

Morocotopo · « **on:** April 15, 2009, 04:48:07 PM »

Hi people, I´ve made an external serial FX loop for my DIY tube combo

It´s based on a loop called Kleinulator I found on the web, modded by me. It´s a JFET buffer to adapt impedances for the send preceded by a level control to get the signal to a level suitable for pedals, and a JFET gain stage for the recovery section to get the signal level back up, followed by a level control. All powered by a voltage multiplier (taken from Geofex)to get 31V from a 9V battery or pedalboard power supply. This is to get as much headroom as possible. This is going to be mounted in a stompbox, connected between the amp´s last preamp gain stage (12AX7) and the phase splitter.
I´ve already made it, it works, but I have some things I´d like to improve, maybe you can help me...
- When the amp is at full volume, the loop can´t take the preamp´s signal without distorting. Right now I´m using it with the return level at full, to get maximum recovery, and the send level at about 1/4 open, with that settings I lose some final volume when it´s connected. At lower preamp levels it mantains "unity gain" when in use. Any ideas about how to get more headroom before clipping (it´s the buffer that clips, the first FET)?
- The loop thins the bass. I guess that´s a coupling cap(s) issue, can you help me identify wich could be the ones to change? Or is there something else to consider?

It´s a neat little circuit, and I´d like it to work as best as possible.

Any suggestions appreciated. Thanks for reading.
(By the way, if anyone wants to use it, by all means do so!)

newfish · « **Reply #1 on:** April 15, 2009, 11:02:28 PM »

Isn't C1 (47n) really tiny?

Maybe this is where all your bass disappears. That's the first place I would look to make a change.

Perhaps you could try a 100n or bigger here...

anchovie · « **Reply #2 on:** April 16, 2009, 12:14:47 AM »

Quote from: newfish on April 15, 2009, 11:02:28 PM

Isn't C1 (47n) really tiny?

Compared to what?

C1 forms a high-pass filter in conjunction with R2. With C1 = 47n the roll-off frequency is 3.4Hz, with C1 = 100n it is 1.6Hz. No noticeable difference to a guitar signal at all.

I'd suggest trying the circuit with R5 replaced by a direct connection from C4 to R6/Q2. This will send a tiny bit more FX output back from the loop too. As for the input buffer clipping, simplest thing I can think of is to attenuate the input signal even more so that when your preamp level is high it still keeps away from the rails of the buffer. Try putting the 10K resistor that used to be R5 from the input to the 5K pot.

alanlan · « **Reply #3 on:** April 16, 2009, 01:47:50 AM »

You've got a lot of gain in your output stage (Q2). That will probably be causing much of your distortion.

Morocotopo · « **Reply #4 on:** April 16, 2009, 02:30:09 PM »

People, thanks for the suggestions. Alanlan, the problem is that I have to attenuate the signal a lot before the send buffer to make it not distort, so I need as much recovery gain as I can get, so thats why the high gain structure there...
I´ll try bigger coupling caps and a higher voltage.

Thanks again

alanlan · « **Reply #5 on:** April 22, 2009, 01:47:22 AM »

I just realised what your problem is:
Q1 gate neds to be biased at roughly half supply (just under is ideal). As it stands, there isn't much swing available on Q1 source because it will be sat at a pretty low voltage (maybe less than 1V).

This should deal with your main woes.

CynicalMan · « **Reply #6 on:** April 22, 2009, 10:27:36 AM »

Quote from: alanlan on April 22, 2009, 01:47:22 AM

I just realised what your problem is:
Q1 gate neds to be biased at roughly half supply (just under is ideal). As it stands, there isn't much swing available on Q1 source because it will be sat at a pretty low voltage (maybe less than 1V).

This should deal with your main woes.

I'm pretty sure JFET gates don't need to biased like that.

alanlan · « **Reply #7 on:** April 22, 2009, 12:12:32 PM »

Quote from: CynicalMan on April 22, 2009, 10:27:36 AM

Quote from: alanlan on April 22, 2009, 01:47:22 AM
I just realised what your problem is:
Q1 gate neds to be biased at roughly half supply (just under is ideal). As it stands, there isn't much swing available on Q1 source because it will be sat at a pretty low voltage (maybe less than 1V).

This should deal with your main woes.

I'm pretty sure JFET gates don't need to biased like that.

They do if you want maximum swing on a source follower. The source needs to be roughly half supply to get a fairly symmetrical swing in either direction. By biasing the gate just under this, the source will end up roughly where it needs to be.

kiffou78 · « **Reply #8 on:** February 01, 2012, 09:50:10 AM »

Hi,

I would like to build your external FX loop. Very intersting for me since I have many DIY handwire tube amp without builtin FX loop (JTM45, Delux Reverb, Vox AC30).
I'd like to use one same FX loop for all and switch FX loop when I switch amp.
Has you said, I don't have space inside amp head to fit easly an FX loop (delux reverb is kind of busy (reverb, trem), and I made a small JTM45 combo 1x12' so no space).

I have 2 questions :

-1- I know about electronic but not so well (enough to build tube amp all the same...). I understand your schematic except one point : Why do you have a stereo input jack ? One is for signal (taken from amp before PI stage), and the other one goes to Q3 base... why ? I figure out that is to switch ON/OFF the power (maybe wrong) but where does it comes from the amp ? At first, I've planed to use a stereo jack as SEND/RETURN jack on the amp (only one hole to drill), then I saw this stereo jack on your input.... I'm confused.
-2- Did you draw the circuit board ? If yes could you send it to me, or post it on this forum ?

Thanks a lot for your work

R.G. · « **Reply #9 on:** February 01, 2012, 10:26:48 AM »

Quote from: alanlan on April 22, 2009, 12:12:32 PM

They do if you want maximum swing on a source follower. The source needs to be roughly half supply to get a fairly symmetrical swing in either direction. By biasing the gate just under this, the source will end up roughly where it needs to be.

Yep. That source needs to be sitting at about the middle of the 31V (!) power supply for best swing.

However, there are some issues with the use of a J201.

The J201 is an unusual JFET. It's useful for some stuff, but you have to watch its peculiarities.

For starters, Idss is 1ma max and may be as little as 200uA. 200uA through 10K is 2V. A minimum-Idss J201 cannot raise the source resistor of 10K higher than 2V. A maximum-Idss device can't raise it more than 10V. A JFET with a minimum Idss of 15ma is needed to ensure you can raise the source to 15V with a 10K source resistor. Worse, you want that point to be able to swing up about 10 more volts; otherwise why have such a big power supply that you can't use? So you need to either (1) get a JFET with an Idss of 25ma minimum or (2) change the source resistor to be much higher to cut the current needs. If you go the high-current route (and can find a suitable JFET, which is not a sure bet) then you'll need to start worrying about the 625mW maximum dissipation limit of a TO-92.

At the risk of being old school and redundant, is there a reason that this can't be our old friend the TL072, biased to half supply? Same number of resistors, I think. And it will happily work on 31V and swing 27 of those volts into loads greater than 2K ohms.

PRR · « **Reply #10 on:** February 01, 2012, 04:24:59 PM »

> Why do you have a stereo input jack ? One is for signal (taken from amp before PI stage), and the other one goes to Q3 base... why ? I figure out that is to switch ON/OFF the power (maybe wrong) but where does it comes from the amp ?

Re-draw it logically.

When MONO plug is inserted, the jack's Ring contact is shorted to Sleeve. This pulls a small current through R12 R13 to jack ground. Meanwhile the R12 voltage turns-on Q3. Q3 then passes battery to the rest of the circuit. OTOH when Ring is not shorted to Sleeve, no current in R13 or R12, thus 'no' current through Q3, circuit is dead and battery is not drained. (There is a small leakage in Q3 but with a healthy Silicon transistor this is much less than battery self-leakage.)

> I've planed to use a stereo jack as SEND/RETURN jack on the amp (only one hole to drill)

You may be able to find stereo jacks with isolated switches. This could be used to trigger the transistor power switch, or simply switch directly. Or if inside an amp, you might tap power somewhere (self-bias power tube cathode is good) and not need any switch. Or if you use a 1-Spot or similar pedalboard wall-power supply you do not need any battery switch (no battery).

Morocotopo · « **Reply #11 on:** February 02, 2012, 03:22:48 PM »

OK, update to this. I tried the Jfet loop extensively, and eventually was not satisfied with the performance. I´m not saying that it´s not possible to make a good Jfet loop, but it clearly is beyond my knowledge. So searched for other schems and finally settled in the one called Paralooper, modded it for my needs and came up with this:

Schem
http://www.aronnelson.com/gallery/main.php/v/Morocotopo/Op+Amp+External+FX+Loop+1_1+SCHEM.tif.html

Layout
http://www.aronnelson.com/gallery/main.php/v/Morocotopo/Op+Amp+External+FX+Loop+1_1+LAYOUT+.tif.html

PCB
http://www.aronnelson.com/gallery/main.php/v/Morocotopo/Op+Amp+External+FX+Loop+1_1+PCB.tif.html

So, R.G., your suggestion is clearly on the right track, since this is an opamp one. Guess wich opamps I used?

This works really well, with the high voltage power it has headroom for days, and the opamp makes a really good driver for the phase inverter/power tubes, the amp sounds "healthier"! Doesn´t color the sound at all. Only thing is, it draws quite a lot of power, 75 mA! Battery power won´t be very efficient, money wise.
So Kiffou, I recommend you make this instead of the Jfet one.

Thanks to all for the help!

kiffou78 · « **Reply #12 on:** February 18, 2012, 11:47:25 PM »

Thanks a lot !

Did you use a tl072 as op amp ?

I ve got some 4558 and 4559, can i use it ?

kiffou78 · « **Reply #13 on:** February 19, 2012, 12:15:50 AM »

Hi,

Other thing :

In this last opamp version, you have 2 inputs and two outputs :

As i understand things : one In/out is for pre preamp fx loop and other is for post preamp fx loop right ?

That means 2 way for using your fx loop :

1- pre preamp : guitar - guitar input fx loop - send fx - fx - return fx - fx guitar pre amp output - guitar amp guitar input
2- post preamp : guitar - guitar amp input - guitar amp post preamp output - fx loop guitar preamp input - send fx - fx - return fx - fx guitar power amp output - guitar amp post preamp input

I guess those two mode are exclusive ?

Am I right ?

Morocotopo · « **Reply #14 on:** February 19, 2012, 06:25:52 AM »

Kiffou, you can use 4558 - 4559, it will work OK, or almost any dual opamp with standard pinout, the advantage of the TL072 is that it has FET inputs, that are better suited to the guitar signal, low noise, etc. You might not even notice the difference in sound, or might like the sound more, just try them all. Use sockets to be able to switch the opamps easily.

Regarding the ins/outs, yes, you can use it with the FX´s divided into pre and post preamp ones (that´s what the in from gtr/pre fx out - out to amp input jacks are for, they are just a patch point), or just guitar straight into the amp and all FX into the loop itself.
Some of the jacks have switching functions. For example, if you don´t plug a cable into the ret from post FX out, the signal is not interrupted, it turns into a signal splitter. Of course you lose the FX routing. Other jacks are switching too, but now i can´t remember the possible uses I planned when I made the schem!

I´ll have to think about those...

Hope I make myself clear. Later I´ll try to make some diagrams about all the possibilities.

kiffou78 · « **Reply #15 on:** February 20, 2012, 10:44:40 AM »

Hi, thanks for your answer.
After quick analysis, I've found 4 way to plug it (well 3, cause one is like a raw cable)
Am I right ? Is there any more left ? :

RawCable.jpg :

PrePreampFXLoopBooster10dB.jpg :

PrePreampBooster10dB.jpg :

PostPreampFXLoop.jpg :

Morocotopo · « **Reply #16 on:** February 20, 2012, 12:52:26 PM »

Wow Kiffou, you made some work! Great way to analize signal flow.

The first drawing is, of course, not very useful.

The second one I haven´t thought about, but it´s another possible use, yes. In that case I would set the send level to max, because at that point the guitar signal hasn´t been buffered, so if you turn that down you´ll be attenuating the raw guitar signal, not good for the signal to noise ratio. I would control gain with the ret level. Keep in mind that the actual voltage gain of the last opamp is 11, but in dB that´s 20,8 dB, not 10.

The third one is the same as the second but without effects.

The fourth one is the actual use I made it for, the only difference is that I use the non used jacks as a convenient patch point with pre-preamp effects, but functionally the same.

Here´s my examples of possible connections:

http://www.aronnelson.com/gallery/main.php/v/Morocotopo/Opamp_FX_loop-1.jpg.html

http://www.aronnelson.com/gallery/main.php/v/Morocotopo/Opamp_FX_loop-2.jpg.html

http://www.aronnelson.com/gallery/main.php/v/Morocotopo/Opamp_FX_loop-3.jpg.html

http://www.aronnelson.com/gallery/main.php/v/Morocotopo/Opamp_FX_loop-4.jpg.html

Hope it´s of use!

kiffou78 · « **Reply #17 on:** February 21, 2012, 10:47:14 AM »

Ok Morocotopo !

I start to think it will fit lots of my needs

!

Now, next step before building this is to figure out where to split signal chain into my amps.

I've got two amp I'd like to Mod this way : JTM45 HW and Deluxe Reverb HW.

I'm quite sure with JTM45, but not so sure with Deluxe Reverb.....
The way I designed it for Deluxe Reverb will add FX Loop on Vibrato channel only since I want the FX Loop to be before the reverb Unit.

Now this is my thoughs :

JTM45 :

Deluxe Reverb :

PRR · « **Reply #18 on:** February 21, 2012, 03:37:22 PM »

> not so sure with Deluxe Reverb.....

The way you tapped, the +180V DC at the tube plate is sent to the Send jack. The thing you are Sending-to may explode.

If you fully understand the "why" of that, you'll see that you can tap _after_ the coupling cap.

For Return, all the other paths to the V4B node are cap-blocked or grounded, so this node is zero volts DC.

In nearly all cases there should be 220K-1Meg bleed resistor from jack to ground, to drain any leakage or stored-charge on coupling caps.

kiffou78 · « **Reply #19 on:** February 21, 2012, 11:54:14 PM »

Thanks PRR, for sure I had dust in eyes yesterday evening !!

I modified Previous post deluxe reverb schematics cause they were very dangerous

Now comes the new schematics :

JTM45 :

Deluxe Reverb :

News:

Author Topic: JFET FX loop, some questions (schem) (Read 8935 times)