Bass guitar FET preamp

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Well hi,

I have this Sadowsky/Aguilar based discrete bass guitar preamp schematic. It is designed for 9V but i want to run it on 18V for more headroom. Now, after i've built it and tested it, i have two questions:

1. Cause it's designed for 9V, it obviously distorts on 18V. What modifications i'd need to make to make it run clean with as much headroom as possible on 18V?

2. This circuit acts as radio receiver too for some reason and has some high pitched noise when Treble is turned up. I can hear stations on it with a lot of noise when signal is not fed with Treble boosted. Am i missing something? Proper shielding/grounding?

Thanks!
 

Attachments

  • FET bass preamp, Fender tone control.jpg
    FET bass preamp, Fender tone control.jpg
    182.5 KB · Views: 1,701
What voltage do you read on Fet Drains?
What Fets did you use?
Are they matched? (they should).

They are most probably mis biased but need existing voltages to suggest corrections.

As of RF interference:
1) build preamp inside a metal grounded box or inside a Bass grounded shielded cavity (graphite paint or aluminum foil)Will you buiild it in a box pedal stykeor mount it inside a Bass Guitar?

2) What do you mean with "signal not fed"?
Are pickups connected or input is left open?
 
Haven't measured the voltage in the circuit, they are matched, used two same as in the schematic - 2N5457, measured them while matching, both with -3.8V and 1.7mA if I remember correctly when matched them, but they are definitely matched.

It's going to be inside the cavity of a bass, will definitely shield it, just found it unusual to pick up radio signal that much, you can almost recognize what's being said.

By "signal not fed" I meant no input signal but everything connected. As of now, I'm testing it with a signal from a headphones output on a smart phone, the bass isn't ready yet to test with it. I just play a bass clip on YouTube till I finish the bass.
 
Last edited:

PRR

Member
Joined 2003
Paid Member
> both resistors should be either 10k or 100k; equal, not 10k AND 100k.

Look again. This is a gate bias, NOT an op-amp center bias.

1/11 bias is not wrong, but may not be right, depending how many volts are bound up inside the JFET at a reasonable drain current.

For a single JFET, one simple fix is to make the 10k bias resistor a 22k trimmer and trim for a happy drain voltage.

> make it run clean with as much headroom as possible

Then DON'T jack-up the supply voltage!! The gain in a multi-stage amplifier increases at least as fast as the supply. For same input, it will distort the same. Except if you keep jacking-up, you will smoke your amplifier input (or some JFETs).

Define JUST how much gain you need for a sensible output level, and design to that.
 
It's not mine, but yes as PRR mentioned center is required for opamps.

OK i'll make the 10k bias resistor a 22k trimmer and try that. I want to run it on 18V cause the Aguilar OBP-1 is basically this circuit (same guy designed both) just modified to fit 9V and 18V to get more headroom.

That's what I want to achieve, so jacking up the voltage apparently makes a difference, just I'm not sure what needs to be modified to go up to 18V while getting clean sound. Not sure if different FETs have been used in the Aguilar though.
 
Couple details.

1) I do not see a Volume control in your schematic.
You can add it at 3 points, best is after tone control .
I suggest 100k Log/Audio.

2) as suggested above, replace bias divider 10k resistor by 22 k trimmer, and adjust until Drain voltages are about 9V (halfway of +V supply)

I asked about matching because if not, both may be at wildly different voltages.

3) to kill RF place 100pF across GS junctions.
 
Last edited:
Ok, so i did the modifications and got some interesting results.

Making the bias 10k resistor to 22k trim is working very nice. Just one interesting thing i noticed... with the 10k resistor it distorts, with the 22k trim it doesn't distort when around 10k value, just changes the overall gain of the circuit. The lower resistance i set on the trimmer the lower the gain. It's a bit weird that it doesn't distort with the trim but it does with the resistor at about the same value.
Didn't measure the voltage though, adjusted by ear when it sounded optimized with the signal input with both tone controlls at max.

The 100p caps on GS junctions lowered the radio reception by a lot, the high pitched noise is the same but that is very likely due to bad shielding atm. The noise is only present with boosted Treble.

Ok, i might add a Volume pot after the tone control or i'll just use passive Volume before the circuit cause i already have this setup with push pull pot.

Thanks for all the tips PRR and JMFahey, very appreciated.
 
Last edited:
Are you sure the 10k resistor is actually 10k and not a misread 1k for instance. Sometimes it is difficult to choose between red and orange as they sometimes look the same.
This is the blue one, metal film and the rings are tricky to read, but took it from a 10k labeled pack, so it should be 10k. I'll measure it when I'm home just to make sure what I've used
 
Ok so i have connected this circuit connected to a high output pickup and it distorts a lot. When i bypass the circuit and feed the signal directly to the amp is MUCH lower in volume and sounds normal. Clipping gradually goes away as i lower the volume on the bass guitar. I used to test the circuit with my mobile phone but now the pickup has arrived and clips with it. It's a passive pickup btw.

Same thing with 9 or 18V, on 18V is just louder. Tried adjusting the drain resistors and the bias and there little difference in clipping, mainly only the gain changes overall. When i try to adjust to unity gain it barely pases signal from the bass, BUT works nicely when i plug in audio from my phone and set the volume to about 25% on the phone.

How do i fix this? I like the tone this circuit produces cause of the "colored" EQ it has and gives very nice crisp and warm touch to the bass sound. Thanks
 
The circuit is very non-linear I'm afraid. If the input FET is overloaded it will distort heavily as there is no source degeneration nor GNFB global feedback. Hence the high output distorts a lot, only low levels are decent.


If you want identical performance at different signal levels the circuit needs to be linear.
 
It does sound like the gain is way too high, adjusting the drain and the bias only affect the overall gain but the distortion very little, it's still present.

Funny thing is, some that have built this have really nice results, and some same thing as me, all built with same component values. I'm suspecting that the jfets could be fake or with different pinout, as I've read that is also a thing today, I got the jfets from eBay from China so I wouldn't be surprised.

I don't have an oscilloscope if that's what you mean by scope and post waveforms :/
 
Hi Schotky,

Got your PM - good thing you sent it - sorry I didn't see this project sooner. Figure it's better to answer here where others can 'check my work'.

Others have mentioned, so this isn't news: This circuit has a crap-ton of gain designed into it. One or both Source bypass caps will need removal. Is there any chance that you copied two instances of a circuit originally designed as a single-stage amp?

I doubt if it has anything to do with different pin-out or fake parts -- JFETs often vary over 4:1 or 6:1 range (and are still within specs per the PDF!) in the current they conduct at Vgs=0! That alone likely explains the widely varying results. Your PM mentioned 'matching at 2.2mA'; don't understand that, since a 10k resistor (the Drain loads) conducting that current would have 22 Volts across it.

Misters Tillotson and Fahey have already hit this nail on the head -- you're getting rail-slamming clipping.

Then post the Drain and Source voltages and the one from the divider, when operating on 9V, since that was the 'original design'. Then try lifting one leg of C7 and extracting the signal from the wiper of the treble pot. It'll be much lower amplitude (gain will only be about 6 dB) but maybe less-clipped; should give us a starting point.

I may be kicking myself for weeks for doing anything to discourage your high ambitions, but you've just got to buy a 'scope if you're going to continue at this level ..

Cheers

edit: Forgot from a couple posts back -- this circuit cannot be 'adjusted to unity gain', so I'm stumped by what you meant there as well ;)
 
Last edited:
Hi Schotky,

Got your PM - good thing you sent it - sorry I didn't see this project sooner. Figure it's better to answer here where others can 'check my work'.

Others have mentioned, so this isn't news: This circuit has a crap-ton of gain designed into it. One or both Source bypass caps will need removal. Is there any chance that you copied two instances of a circuit originally designed as a single-stage amp?

That's ok Rick, thanks for weighing in :)

Remove the 50uF caps?

The circuit is as I found it and was built by others, for some works fine for is as mine. 2 zener diodes are omitted on the output (out to ground) which were the for output protection.

The main thing is this circuit is basically Aguilar OBP-1 which is everywhere on the market, and I have one too. Of course the Aguilar is adjusted not to distort and has gain of 1 with EQ set to flat. This one has a lot of gain with flat EQ, and when boosted it creates this "heavy compression like" even more distorted sound. Both get more dB boost on 18V, just yeah the Aguilar is clean and normal this one distorted.

I doubt if it has anything to do with different pin-out or fake parts -- JFETs often vary over 4:1 or 6:1 range (and are still within specs per the PDF!) in the current they conduct at Vgs=0! That alone likely explains the widely varying results. Your PM mentioned 'matching at 2.2mA'; don't understand that, since a 10k resistor (the Drain loads) conducting that current would have 22 Volts across it.

When matching I measured the current draw and the voltage as suggested by the "matcher circuit" so I got 2.2mA reading whatever that means. Sorry if it's confusing I don't understand that bit that we'll to explain it better :/

Misters Tillotson and Fahey have already hit this nail on the head -- you're getting rail-slamming clipping.

Then post the Drain and Source voltages and the one from the divider, when operating on 9V, since that was the 'original design'. Then try lifting one leg of C7 and extracting the signal from the wiper of the treble pot. It'll be much lower amplitude (gain will only be about 6 dB) but maybe less-clipped; should give us a starting point.

Not sure what rail-slamming clipping is or how to fix it, so please bear with me till I get everything mentioned here.. I may need simpler explanation for some bits.. I'll try that thing with C7 then I get back home and provide feedback.


I may be kicking myself for weeks for doing anything to discourage your high ambitions, but you've just got to buy a 'scope if you're going to continue at this level ..

Haha well these are the only two that I plan on building now, so I'll try to do some mods you guys suggest here and see if that will work, if we're stuck then I guess I'll have to source an oscilloscope........ I can do V readings if that'll help now? Current on the whole circuit on 18V is just below 1mA though where the Aguilar is as well.

edit: Forgot from a couple posts back -- this circuit cannot be 'adjusted to unity gain', so I'm stumped by what you meant there as well ;)

Sorry, I meant gain of 1 (or acting like a buffer) with EQ set flat [emoji16]
 
Start by removing 50uF caps C7 and C9 , only then can preamp gain be near 1 with tone pots set halfway.

Rail to rail clipping: signal at Drains will start at some idle voltage (which I asked you to read and post) which in principle will be roughly about half battery voltage, so around 4.5V or 9V respectively if you use 9 or 18V batteries.

When driven, FET will amplify signal which will swing up and down.

Now to the problem: it can not rise higher than battery level (9 or 18V) and not lower than 0V .

Battery voltage and ground are usually called "rails".

IF input signal or gain are too high, output signal will try yo go above battery voltage or below 0V, failing miserably. (you can´t beat Physics) , signal peaks will be harshly chopped, that´s ugly clipping distortion.

So start by lowering excessive gain by removing those capacitors.

If you have a notebook with a microphone input available, you can use it as a very useful hobby scope, by installing Scope freeware and building a *simple* attenuator (2 resistors, 2 diodes and 1 capacitor)

I published it in The Gear Page.
Sadly for some unknown reason my ISP was banned there so I can´t get in, but if you wish feel free to read and follow it:

Search Results
Web results

Scope Died. Anyone Use the PC-Based Ones? | The Gear Page

Search for: "fahey simple software scope attenuator the gear page"

For the signal levels you are using I suggest 47k>1k attenuator values and add a .1 or .047uF capacitor in series to keep DC out.

sample search images:
 

Attachments

  • PC Scope.gif
    PC Scope.gif
    119.5 KB · Views: 354
Couldn´t edit because of time limit, so here it goes:

EDIT:I think (can´t get in to check) that the attenuator schematic was lost , so just in case here it goes again, with pedal/preamp friendly values:

You make 2 probes: red and black crocodiles (and cables) to connect to preamp ground and visualized points, and attenuated output goes to a 3.5mm microphone/earphone jack into notebook Mic input.
 

Attachments

  • PC Scope Attenuator basic.gif
    PC Scope Attenuator basic.gif
    5.4 KB · Views: 397
Couldn´t edit because of time limit, so here it goes:



EDIT:I think (can´t get in to check) that the attenuator schematic was lost , so just in case here it goes again, with pedal/preamp friendly values:



You make 2 probes: red and black crocodiles (and cables) to connect to preamp ground and visualized points, and attenuated output goes to a 3.5mm microphone/earphone jack into notebook Mic input.
This seems save worthy, thanks JMFahey. If this works nicely it will do the job for once in while testing
 
Ok @Rick PA Stadel and @JMFahey, no matter how simple this was you two saved me a lot of time wasting and frustration, hats down.

I've removed only the C7 and wanted to test just without that one. The circuit is back to normal life! Everything works as it should, drain control, bias for boost, tone control, no audible issues.

Current draw has also halved, it's down to 0.41mA on 18V now and RF is completely gone.

Should I remove C9 too now or should I leave the circuit as is at this point? I'm asking if I should remove this one too before actually removing and testing cause to remove C9 I have to take other components out to access it.
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.