JFET Buffer and buffer cable

[JohnH]

Yes there is indeed more going on. My point was just to show that the source cap could be smaller, if desired.

Lets do the tests, First we need a spice model and some JFET parameters. I usually use these:

* J201 with VGSoff=-0.8V and IDSS=0.6mA
.MODEL J201 NJF (VTO=-0.8 BETA=0.94M LAMBDA=2M IS=114.5F RD=1 RS=1
+ CGD=4.667P CGS=2.992P M=.2271 PB=.5 FC=.5 VTOTC=-2.5M BETATCE=-.5
+ KF=604.2E-18)


* 2N5457 with VGSoff=-1.6V and IDSS=3.3mA
.MODEL 2N5457 NJF (VTO=-1.6 BETA=1.29M LAMBDA=2M RD=1 RS=1 CGD=6E-12
+ CGS=2.25E-12 KF=6.5E-17 AF=0.5)


* MPF102 with VGSoff=-2.5V and IDSS=6mA
.MODEL MPF102 NJF (VTO=-2.5 BETA=0.96M LAMBDA=5M RD=1 RS=1 CGD=1.54248P
+ CGS=2.567P PB=1.49 KF=7.90591F AF=499.953M)

from this link:
JFETs, DIY stompboxes

I did some of this before, for my overdrive box:
BlueJuice Overdrive

I use a 150nF in the first stage, to get a bass roll off, then a 2.2uF in the second stage, for full range boost. My source resistor was a slightly larger value than here, at 3.9k

I will run a single stage, with parameters from this thread, using a range of source cap values.

[JohnH]

Have a look at this:



Its the first stage of my overdrive, adapted to be more similar to yours. I vary th esource cap, and the plots show the small signal gain of that stage. I varied from 220nF to 22uF, in log steps the red line is set at 2.8uF, with a small amount of bass roll off in the guitar range. Nothing wrong with the larger valeus though, if you want full bass range.

(my model also has a resistor in series, R16, but that is at 0 in this test)

[platon]

Mar 19, 2014 14:31:37 GMT -5 JohnH said:

But with a buffer in the guitar, the next thought is to not put a volume after it, but use the standard guitar volume before it. That way you get the full benefit of low impedance output. To add it some 'virtual cable' capacitance, you can wire a small cap just before the volume pot, ie across the outer lugs of the volume pot.

Hello to everyone!
I'm late to the party but since I'm having a guitar built right now, the idea of integrating a buffer came up and I found your nifty design, John. One question that sprang to my mind was about the 'virtual cable' cap. The capacitance of the cable normally interacts with the pickups and without I would expect the resonance frequency to raise significantly. Shouldn't a 'virtual cable' cap be installed in any case when using a buffer?

I'm planning to use a DPDT push/pull tone pot to switch between buffered and conventional, passive operation. One pole would toggle between the output signals, the other would switch the 'virtual cable' cab in and out of the circuit. However, I will have to go without the 'traditional' treble bleed because it would interact with the 'virtual cable' cap in buffered operation (unless I find a push/push pot with more poles). But since unbuffered operation only is planned to be a fallback in case the battery dies, that seems acceptable. Does this make any sense?

[JohnH]

Hello platon and welcome to GN2

Yes that makes sense. But if you want to have the virtual cable cap, switch it across the outer volume pot lugs. Then it will have a consistent effect at all volume levels and no treble bleed needed. But you might like the clear sound of just the buffer and no extra cap. It would be best to try it with and without. I have 2 buffered guitars, and dont the virtual cable caps. But my 2 best axes dont have the buffer, and I like their sounds through a 10'cable.

If you are trying a built-in buffer, I reccommend the version on page 6 of this thread.

[platon]

Thanks for your insights! Why would you rather use the version on page 6 (March 23, 2014)? What exactly is the purpose of C2/R11? C2 looks quite big.

Yes, I would replace the treble bleed by the 'virtual cable' cap. And I hope the buffer will help to reduce the noise level of the guitar (along with careful shielding) because I live right next to a railroad (all electric) and when I hold my guitars perpendicular to the tracks they all pickup noise.

[JohnH]

That version is the same as near the bottom of p1. It should give the best response, particularly if you use it to go into different inputs like mixers and pc sound cards. Its because it runs some more current through the transistor, but still not much, te battery will last a long time.

The output cap and resistor get rid of dc at the output, and the highish value lets all the bass through, if driving lower impedance inputs. It could be smaller if wanted

[platon]

Thanks again, John! I think I'll give it a try in my new guitar.

[ashcatlt]

Aug 18, 2016 1:30:10 GMT -5 platon said:

However, I will have to go without the 'traditional' treble bleed because it would interact with the 'virtual cable' cap in buffered operation ...Does this make any sense?

Not to me, no. Won't the treble bleed interact with the virtual cable about the same way it would interact with an actual cable? Is this not what you want for some reason?

I say put the cap across the input of the buffer itself. Put in your treble bleed like usual. Use your DPDT to true bypass the buffer and everything should be cool.

[platon]

Aug 24, 2016 15:12:39 GMT -5 ashcatlt said:

Aug 18, 2016 1:30:10 GMT -5 platon said:

However, I will have to go without the 'traditional' treble bleed because it would interact with the 'virtual cable' cap in buffered operation ...Does this make any sense?

Not to me, no. Won't the treble bleed interact with the virtual cable about the same way it would interact with an actual cable? Is this not what you want for some reason?

I say put the cap across the input of the buffer itself. Put in your treble bleed like usual. Use your DPDT to true bypass the buffer and everything should be cool.

Yes, if the virtual cable cap is not put across the volume pot but across the buffer input, it should work. Thanks for pointing that out.

[JohnH]

On my buffered guitars, I never put in treble bleed, but sometimes try a small cap, not across the buffer but across volume outer lugs. Then you get the tone of having a bit of 'cable' capacitance, but perfect consistency at all volumes. That's because, the extra cap is always interacting with the pickup in the same way, unaffected by the volume pot (or the real cable after the buffer)


On passive guitars, a TB circuit is a good compromise but the buffer is better at preserving tone.

If you do cap across the buffer input plus treble bleed, its fine and it will work, but it negates some of the benefit of the buffer in terms of consistency

[platon]

Aug 24, 2016 16:13:56 GMT -5 JohnH said:

If you do cap across the buffer input plus treble bleed, its fine and it will work, but it negates some of the benefit of the buffer in terms of consistency

I agree. The reason why I am considering to keep the treble bleed is to be able to switch between 'passive' and buffered operation with a DPDT switch. The TB remains in the circuit for passive operation.

[matbard]

Hi again guys,

I've built some preamps using the Till and JohnH projects. Thanks to all, truly.

Recently I acquired two lovely pieces of gear: a G&L SB2 bass from late 80s and an Alembic IN-2 power supply/switcher. I was thinking about trying to make them "collaborate" using a particular preamp cable.

The Alembic is basically a 1R 18V power supply with a 5-pin XLR input. It has two pins for + an - Volt supply, a common ground and two signal pins(one for pickup). It was intended for use with Series instruments that have the 5-pin output receptacle. The SB-2 was a late Leo Fender project and had two volume controls(one for each pickup), but no tone. So it would be easy to rewire the controls with a stereo 1/4" jack and have separate outputs of the two pickups. So I have to do a stereo preamp cable with 1/4" stereo jack at one end and 5-pin male connector at the other end. I don't need gain, purely a buffer that can let use the gear. Also, maybe I will use two 250K "jackpot" potentiometers that lifts the ground on fully clockwise rotation and so let the pickups to get connected to the output without the volume load.

Which scheme would you suggest? The JohnH seems the better but I'm a little aware of the component to be put inside the connectors.

- Matteo

[JohnH]

The buffer cable i drew has x1 gain while Tillmans has some greater gain. Either should be fine but i wouldnt use a j201, not enough voltage swing availbale in these circuits and too much gain created. Almost anything else would be fine. I like 2n5457.

One key consideration may be that my version needs an extra wire core. His works in mono form using single core plus ground, mine needs 2 cores. A stereo version would need 3 cores for mine abd two for his. It could still be ok though. I have some 4-core screened cable that feels similar to standard guitar cord.

My version uses less battery power.

All that assunes you need to build it in a cable. You could build two buffers with a small battery into the guitar, or consider if you really need buffers?

[matbard]

I agree with you about the 2N5457. I have a bunch of these and they are my favorites for projects.

I don't mind about battery power, because the whole system would be powered by the Alembic IN-2 through the 5-pin connector, all running at 18 volts. If I could place some component inside the 1/4" stereo jack housing at one end and some other components at the 5-pin male XLR connector at the other, I could use whichever cable I want in the middle.

[hewo]

hi everyone, it's 2017 and here i am trying to optimally create stock telecaster optimal buffer using japanese n-channel jfet having smallest pinch 0.516vdc, 2sk170gr sanyo

so powering is 7809=1.1ampere capacity, clean power, R_drain=none, R_source=1k5, C_source=none, R_gate_tiedown=1m5, R_gate_powering=1m5, V_drain=9, V_source=4.57, V_gate=compromised_measurement_10M_ohm_multimeter_4.16, I_quiescent=3.146ma

my situation is confusion which way to align wet tantalum polarity for the n-channel jfet buffer input dc blocking cap.

you'd think the tantalum's anode (+ side) is faced toward the jfet's gate held steady midway of 7809 powering volts, and the jfet input tantalum's cathode pointing towards the stock telecaster connected to buffer by coaxial (low picofarad per linear foot).

intuitive gut feeling is that this is wrong orientation! but i am not certain, so i inquire here. i chose dc blocking wet tantalum to shrink the physical discrete welded buffer components integrated package into the tiniest my soldering skills could handle, the size of the raw 2sk170gr but in a three dimensional augmented package sporting extreme compactness. consideration of layout topology is key. provide adequate yet minimal physical separation of discrete components countering crosstalk between lowlevel signal and highers. provide minimal extraneous noise infiltration by physically shrinking integrated package as a noise target via reduction of welded antennae congestion (linear lead runs, component leads, powering supply can also bring in noise too if not carefully created)

please clarify if you can. normally non-polarized caps are employed here for dc blocking this centering gate topology, but i choose wet tantalum, those good ones from the prehistoric computer industry era (remember the microprocessor called 8088?) whereby conservatism was utmost owing to debut uncertainty of reliability and design. darn wet tantalum is pretty reliable and lasts forever? the ones i cannibalized off from computer boards have remarkable coveted performance (as in nil dissipation based upon meter quality measurement compared against modern electrolytic organics, panasonic oscon).

Attachments:

[JohnH]

Hi hewo, welcome to GN2.

Id agree that if you are using a polarised input cap, its + connection should go to the gate. But I dont see any benefit from using one since even a small value like 22nF will let all audio signals through down to sub-bass frequencies.

Also, I see you are using a 1.5k source resistor. I doubt you need this so low and it will draw considerable current. You can use 10k or 15k in most cases, with better linearity of output. What are you connecting the output to?

[hewo]

while truth dictates "muddiness" clouds the sound result were "overkill capacitance" signal dc blocking caps be employed, and also their "slower" rise fall charging cycles to the guitar signal frequency passed, if you purview the literature schematics in hifi, it does appear they use higher capacitance while guitarist utilize the least capacitance necessary, why, maybe because "muddiness" clouding is not wanted?

my understanding is lower capacitance has faster charge discharge characteristic, and, using tantalum furthers this.

why have quicker rise and fall charging? because it is more replicate. soundwise, is it better? of course!

schematics show 223 capacitor values, or so, within that hundreths of microfarads, to encompass the spectrum guitar frequencies.

my stock telecaster sports a 102cap for 1M_A volume pot treble bypass, and a 103cap for 1M_A tone pot. that's how factory made it long back in the 60's.

i don't want to re-weld those axial wet tants, it's a bh to prep preweld spatial orientation of component affix prior weld. it takes a lot of spatial jigs which position weld node before welding (soldering!)

question raised to what device (reception) this jfet 2sk170gr buffer sends buffered signal to. i built that schematic little gem mark II, a push pull bedroom guitar power amp with upstrm buffer, junk buffer, so i am improving substitue buffer (improve? let's see)

particular jfet 2sk170gr pinch only half volt! that's so wee. but telecaster can be manuvered because of it's passive energy annihilation via its pots. the remaining signal transported to substitute buffer created and forwarded into little gem mark II will be encompassed. while true intentionally striking telecaster to generatelarger swings of pickup output voltages can exceed particular 2sk170gr pinch envelope, it can be tempered by guitar onboard energy annihilation pots thereby reducing swing transported by coaxial cable into this substitute buffer and dwnstrm little gem mark II.

i have some concerns my selected choice of regulated power by 7809 ic ballast caps may be undervoltage for the 9v. i've tried to search the web for these tantalum caps (integral 1500uf but at what rating voltage operating?). maybe somebody knows these caps, looks european, they're surface mount miniatures cannibalized from a toshiba laptop logicboard, you know, the tiniest rectanguloid block, maybe i should pic it for viewing? we know computers use +-12, +-5, +3.3. their minuscule size is what scares me as to their voltage safety for applying 7809 powering.

[warmstrat]

Firstly apologies for not catching up on the entire thread since I was last here - it's too detailed for me to skim through at the moment - so if I'm being dumb or redundant please ignore/flame at will.

Five years on and my trusty buffer cable is only on its second 9V battery (!!) but the glued-up mess of components and solder crammed into the jack plug by Capt. Hamfists (yours truly) has started to make banging and crackling noises.


Got to thinking about making a tiny little PCB (with surface-mount components obviously) that could be soldered into the plug and sealed up very easily. Could even do the battery box side of the circuit also and make quite a few of each within the minimum order of most PCB shops.

Does that tickle anyone? Board layout files would obviously be free to use and I can probably send anyone a couple boards for free+shipping costs.

[JohnH]

Hi ws
Sorry i missed this last week. Nice to see you again. Good luck with the buffer cable mk2. Do you have the gear needed to work with smd parts? Too small for me!

[warmstrat]

Hi John

I had a tutor in university show me how to do SMD work with regular soldering equipment... If you have a fairly fine pair of tweezers to place the parts, and the PCBs are solder masked, you can rely on the surface tension of solder to cling only to the tinned metal pads and component legs (provided you don't apply too much solder). It's not hard once you get the knack...

This video looks like most of the techniques explained, with the exception that I usually have the whole lot swimming in a drop of liquid flux which burns off as the solder fuses - supposed to give better heat distribution so less risk of overheating delicate components (this could be hearsay though).

[roadtonever]

Mar 21, 2014 22:38:40 GMT -5 JohnH said:

This is what I did with the button-cell buffer:



The cells are two CR2025, mounted in a holder, which was taped to the back of a pot. Hot glue might have been better.  The buffer was assembled point to point around the jack. I left off he 2.2M input resistor for this application, and added 1M (could be higher)to tie the output of the cap down to zero. Looks like I used a 39k source resistor, and the Jfet was MPF102.

As to how long these things run for, it's gonna depend....

This circuit, as described above, if you used a 33k resistor, and if the JFET ended up with 2.5V at the source, will draw 0.076mA.

Here is a data sheet for CR2025:

CR2025 data

It seems to be good for at least 100mAhours, at low current, still at nearly full voltage, which gives about 1300 hours run time. But it will be also limited by the battery shelf life, and 10 years, maybe not! It worked fine for me for about a year, then I took it out to build in something more complicated. The buffer and batteries have been in my spares box for the last 6 years, and are now reading about 5.5V total.

If you could find space for a 9v lithium pp3, it would have 10x that capacity, and a larger headroom for discharge, so could last several further years.

Is the entire buffer inside the guitar or are you combining with more parts in an external box?

EDIT: Looks like it. How neat.

Let me ask another question. How suitable is the 2N5457 for 6v operation? I was thinking about using your stripboard layout from here: guitarnuts2.proboards.com/thread/3773/lp-maximiser-humbuckers-max
Except substituting the input cap for 22nF. And the 2.2M resistor for 1.5M. It's destined for a five string bass with no space for a 9v battery.

[JohnH]

Yes that's all it is! With the x1 no gain buffer, then 6V and 2n5457 should be fine The strip board layout is ok. Probably can omit the 10uF cap I'd reduce the 100k to say 22k, and adjust the bias resistors (shown as 2.2M and 3.3M( to suit your specific jfet and set about 3V at the source

[roadtonever]

Jan 9, 2022 4:11:48 GMT -5 JohnH said:

Yes that's all it is! With the x1 no gain buffer, then 6V and 2n5457 should be fine The strip board layout is ok. Probably can omit the 10uF cap I'd reduce the 100k to say 22k, and adjust the bias resistors (shown as 2.2M and 3.3M( to suit your specific jfet and set about 3V at the source

The averaged values in post #9 only apply for 9v operation then? I'm not certain I will be set up to do an individual biasing without delaying my project substantially.

Oct 3, 2006 15:45:29 GMT -5 JohnH said:

Chris and Channelman. Thanks, I agree with what you said.

When I build with 2N5457, I usually test each one on a bread board to choose the best biasing resistors, For this buffer circuit, it tends to put the gate biased at about 1/3 supply voltage, so R3 and R4 are 1.5M and 3.3M. Ive found these values wotk in practice with the JFETs that I have tested - and I guess I am lucky not to have come across the extreme range. For circuits using gain where Im using the JFET for overdrive, I prefer the sound of the 2N5457 over the J201, but then it is definately necesary to set each one up individually.

The Tillman circuit did not look like it was optimum, and CM's analysis confims it. I had one working better with a 10M resistor from gate to positive - and a larger one would have been a bit better still, of about the size CM suggests.

John

[roadtonever]

What's the basic rundown of steps and tools needed for biasing? If it's not too hard I'll try. On hand I already have the JFET, a stero jack, 2cr2032 batter holder and metal film resistors. And for the remaining stuff I'm visiting my electronics supplier soon.

[JohnH]

The idea of biasing is to set the dc conditions at the source (output of the jfet in this circuit) so that it is just above half the supply. Then when a signal is applied, it can swing up and down cleanly without distorting by reaching the power supply or ground voltages. (6V or 0V).

The ideal range for a 6V supply is between 3V and 3.5V at the source.

Looking at the strip board diagram, the jfet is the black shape. The source is its middle wire.

So measure dc Volts using a multimeter set to a 20V dc range. Put the leads to the source and the -ve battery = ground.

The jfets are often not very consistent. But if you have the parts shown, with 2.2M and 3.3M resistors, it will probably bias a bit too high. A more likely ideal is to change the 2.2M to about 1.2M or 1.5M.

And change the 100k to 22k.

[roadtonever]

Here we go. Populating the board.

[ozboomer]

Not trying to be a smarty... but for other readers who aren't so au fait with some of the issues when modifiying the electronics...

Just t'be aware that if you try to 'piggyback' the stripboard+battery-packed CR2025s on the back of a pot, realize the whole arrangement may not fit inside the guitar cavity under the scratchplate; I found this out the hard way when trying such a mod with Squier Bullet and Affinity strats.

As always, YMMV...

[roadtonever]

So I made to last minute decision to acquire a breadboard so I could properly bias the JFET. I'm getting an abnormal readings? 6.37v instead of ~4.5v that I expected with a 9v battery I tried for a quick test?





EDIT: Removed everything expect JFET and 1.2M and 3.3M resistors. 2.73v reading this time using a 9V battery as a power source. Is that expected with a 9V? Where do I go from here?

[JohnH]

Good idea getting a breadboard. I found them to be very useful for this testing.

Just to check the jfet; does it look like the one in this link?

www.theengineeringprojects.com/2018/03/introduction-to-2n5457.html

The source is the middle lead, and in our circuit it will go to the 22k resistor.

The Junction of the two high-value resistors goes to the gate. The drain goes to the positive battery. The reading is from source to ground.

When you got that 6.37V reading, what resistors were in? and which was going to negative ground?

A nice test when it's set up with the bias test: Lick your figures and touch ground and also the gate connection to the resistors. The source voltage should swing. (This is quite safe with these jfets, but not good in circuits that use mosfets.)

[roadtonever]

Jan 14, 2022 3:47:41 GMT -5 JohnH said:

When you got that 6.37V reading, what resistors were in? and which was going to negative ground?

Basically while reversing my test leads I probed the wrong terminal on the battery. Anyway I'm paying attention now and I'm getting readings that are more expected. Between source and ground I'm reading:

2.7v with a 9.4v power source
2.6v with a 8.8v power source
2.3v with a 8.1v power source
1.9v with a 6.5v power source


I verified that my part looks like the pictures in your link, and the pinout was helpful.

Values I've tested with are 22kOhm, 1.2MOhm and 3.3MOhm. How do I proceed from here?

EDIT: Quickly swapped 1.2MOhm for 1.5MOhm. Now Reading 2.9v with 8.9v power.
EDIT2: 3.7v with 2.2MOhm. And now 2.7v with 6.5v power.
EDIT3: Kept trying different combinations of resistors. 2x3.3MOhm and 2x2.2MOhm both read 3.3v between source and ground with 6.5v. Shall I consider the 2x2.2MOhm optimal for my JFET?