Hi guys,
I hope you all had a merry Xmas.
I have been building a guitar pre-amp to put in some of my home made guitar builds. I managed to get my hand on a schematic from: https://robkiddhomepage.wixsite.com/robkiddmusic/project-berocca
This is the schematic (which I re-designed to have a bypass switch):
The schematic is great and works ok. However, when I implemented it with some of the recommended op-amps (lm358 for example - not the lm4562, I changed that later on), I found the sound to be muffled and not as detailed as I would have liked.
So in my haste I bought some lm4564/lme49720 which seemed to be a great idea at the time. However swapping the lm358 for the lm4564/lme49720 yields some strange behaviours (to me):
1- When I first plug the system the sound is quite thin, lower in volume and quite distorted compared to the original signal. The filters seemed to work but since the original signal is bad the output sounds bad.
2- If I pull SW1A out (it is setup with a 3 pin connector) and touch pin 3 with a short resistor's pin (I can hear a loud buzz), then plug SW1A back the sound becomes normal (very detailed and exactly what I was looking for).
3- If I now use SW2 or remove the 9V battery to open the power supply circuit (setting off V+ to 0V), when I close it back to provide power again the circuit is back to the initial thin distorted sound.
I have tried a couple of things (including adding a 0.1uF capacitor to connect V+ to the ground) but nothing will do. I am starting to lose it and I hoped some of you guys could help me out.
Any help would be much appreciated!
Cheers
Rejerh
I hope you all had a merry Xmas.
I have been building a guitar pre-amp to put in some of my home made guitar builds. I managed to get my hand on a schematic from: https://robkiddhomepage.wixsite.com/robkiddmusic/project-berocca
This is the schematic (which I re-designed to have a bypass switch):
The schematic is great and works ok. However, when I implemented it with some of the recommended op-amps (lm358 for example - not the lm4562, I changed that later on), I found the sound to be muffled and not as detailed as I would have liked.
So in my haste I bought some lm4564/lme49720 which seemed to be a great idea at the time. However swapping the lm358 for the lm4564/lme49720 yields some strange behaviours (to me):
1- When I first plug the system the sound is quite thin, lower in volume and quite distorted compared to the original signal. The filters seemed to work but since the original signal is bad the output sounds bad.
2- If I pull SW1A out (it is setup with a 3 pin connector) and touch pin 3 with a short resistor's pin (I can hear a loud buzz), then plug SW1A back the sound becomes normal (very detailed and exactly what I was looking for).
3- If I now use SW2 or remove the 9V battery to open the power supply circuit (setting off V+ to 0V), when I close it back to provide power again the circuit is back to the initial thin distorted sound.
I have tried a couple of things (including adding a 0.1uF capacitor to connect V+ to the ground) but nothing will do. I am starting to lose it and I hoped some of you guys could help me out.
Any help would be much appreciated!
Cheers
Rejerh
I have quickly soldered the circuit without the switches (just the active path and no power on/off) to try to eliminate the switches as potential culprits. I also tried to change C9 to 1uF (that's all I got on hand at the moment) and it has not solved it.
It is such a strange behaviour. If I put the IC in its socket very slowly (not all pins connected at the same time and starting from 1-8, then 2-7, etc), I get the loud buzz. Jusst like when in step 2. After that the behaviour becomes as expected. But then if I remove the battery and put it back in place the sound is back to be thin and distorted.
I am so puzzled.
Hi Rejerh
If you do a search on "LM4562" "latch" - there are quite a few hits.
One of the first on my list was describing the same issue as you seem to have.
In short: the high DC-impedance of your Vref circuit may allow the internals in the opamp to latch.
I would try following: reproduce the problem and then measure Vref - I suspect it will be close to 0V and not 4.5V as expected.
Note: not all opamps exhibits this behaviour and that may be the reason why the LM358 did not have the problem.
Cheers, Martin
If you do a search on "LM4562" "latch" - there are quite a few hits.
One of the first on my list was describing the same issue as you seem to have.
In short: the high DC-impedance of your Vref circuit may allow the internals in the opamp to latch.
I would try following: reproduce the problem and then measure Vref - I suspect it will be close to 0V and not 4.5V as expected.
Note: not all opamps exhibits this behaviour and that may be the reason why the LM358 did not have the problem.
Cheers, Martin
Rejerh, I can sympathize with your frustration. Been there, done that, decades ago, when dealing with my first high-gain audio integrated circuits. It drove me nuts for days, until I finally fixed my problems.
The symptoms you've listed read a lot like a circuit that is unstable, and erratically bursting into (high frequency) oscillation.
Firstly, using a very fast op-amp (55 MHz unity-gain bandwidth, 20V/uS slew rates) in a DIY preamp is basically asking for trouble. Unless your PCB layout, shielding, and wiring are all impeccable, you are more likely to end up with a low-power RF oscillator than an audio amplifier.
A guitar produces no usable signals whatsoever above 10 kHz (and nothing good above 5 kHz), so there is no benefit at all to using a very fast op-amp. All it does is cause trouble.
If you are building this circuit on a solderless breadboard, by the way, it is virtually guaranteed to be unstable. There is a lot of stray capacitance between adjacent rows of holes in a solderless breadboard, and this turns fast, high-gain circuits into unstable high-frequency oscillators.
Secondly, a circuit that has rising treble response is more likely to become unstable. When the treble control is turned up, that is exactly what you have - higher gain at high frequencies.
Thirdly, running a wire from the output of a fast, high gain op-amp back to the vicinity of its input is almost a guaranteed recipe for creating a high frequency oscillator. But that is exactly what your bypass circuit does - it brings the output of the circuit back to the same tiny switch (S1A). There will be a few picofarads of capacitance between the various pins of the switch, and with a 55 MHz op-amp, that is more than enough to feed back at high frequencies and cause oscillation.
Fourthly, a fast op-amp is least stable when set up for unity gain - and that is exactly how U1A is configured. Again, this is practically begging for instability.
Using U1A as a unity-gain buffer is also not good practice from the noise point of view. It adds noise without amplifying the signal, which means it makes the signal-to-noise ratio worse.
The 4562 is a very low-noise device, so this particular issue isn't much of a concern here. But confguring U1A for even a small voltage gain - say 2x or 3x (+6 dB or +10 dB respectively) will make the op-amp a little less unstable, and also slightly improve S/N ratio.
My suggestion is to replace the LM4562 with a tried-and-true TL072 op-amp. The TL072 and family has FET inputs, and input bias current is a thousand times less than the LM4562's, which makes it better suited to working with very high-value bias resistors (like R2 in your circuit). The TL072 has been used in literally millions of guitar circuits over the years, and, if you have a halfway-decent layout, will provide all the detailed sound you want - without constantly bursting into RF oscillations at frequencies of tens of megahertz.
And yes, please do add a 0.1 uF ceramic or film cap from power to ground, as close as possible to the power pins of the op-amp.
Finally, do not, EVER, under any circumstances, unplug or plug any IC into a socket while DC power is on, unless you want to deliberately destroy your IC!
In summary, pop in a TL072, and if your layout is even halfway reasonable, the circuit should work. And do post a photo or two, in case there are any obvious layout issues that another pair of eyes might spot.
Good luck!
-Gnobuddy
Thanks a lot for your reply!
This helps a lot, and you have educated me heaps. I was thinking of trying out the TL072. It seems that was the right thing to do. I did manage to get my hand on an OPA2134 (couldnt get the TL072 without some long delivery online shipping) and tried it in the circuit. It yielded great results and I am thinking in sticking with it for now (although pricey).
For anyone looking for a 2 band preamp. This circuit works great with the OPA2134 and I would recommend it.
Thanks again!
Cheers
Rejerh
PS: the circuit is pcb layout and shielded!
This helps a lot, and you have educated me heaps. I was thinking of trying out the TL072. It seems that was the right thing to do. I did manage to get my hand on an OPA2134 (couldnt get the TL072 without some long delivery online shipping) and tried it in the circuit. It yielded great results and I am thinking in sticking with it for now (although pricey).
For anyone looking for a 2 band preamp. This circuit works great with the OPA2134 and I would recommend it.
Thanks again!
Cheers
Rejerh
PS: the circuit is pcb layout and shielded!
Glad you solved it, or at least got much closer to what you attempted to.
Gnobuddy is right.
Let me add that yes, agree with ultrafast modern Op Amps being more of a problem than an asset, while LM358 will "work" (in fact I keep a stick of them for some special/emergency uses) but let me suggest TL062, THE workhorse in active instruments: similar to TL072, slower (not a problem here), identical in all other aspects including high impedance Fet input but most important:
* optimized for 9V battery supplies (much better than TL072)
and
* 1/5 the consumption of TL072
now you know why it´s the de facto standard there
Other points.
* your switch is at the wrong end: let the output jack switch between input signal or preamp out.
A true bypass DPDT would be even better, but an SPDT is usable, as I mentioned.
* even better, leave the input buffer always in circuit, and switch output jack between tone control out and buffer out Pin 1
You will need to add an extra "C7 - R8" net at buffer out because you have 4.5V DC there but no big deal.
Check Boss pedals (and many others), they always leave input buffer in place.
Gnobuddy is right.
Let me add that yes, agree with ultrafast modern Op Amps being more of a problem than an asset, while LM358 will "work" (in fact I keep a stick of them for some special/emergency uses) but let me suggest TL062, THE workhorse in active instruments: similar to TL072, slower (not a problem here), identical in all other aspects including high impedance Fet input but most important:
* optimized for 9V battery supplies (much better than TL072)
and
* 1/5 the consumption of TL072
now you know why it´s the de facto standard there
Other points.
* your switch is at the wrong end: let the output jack switch between input signal or preamp out.
A true bypass DPDT would be even better, but an SPDT is usable, as I mentioned.
* even better, leave the input buffer always in circuit, and switch output jack between tone control out and buffer out Pin 1
You will need to add an extra "C7 - R8" net at buffer out because you have 4.5V DC there but no big deal.
Check Boss pedals (and many others), they always leave input buffer in place.
OPA2134 & OPA2604 are used by tech21 on their pedals (Sansamp GT2 for example). OPA1641 should be great too for audio (absolutely no noise). TL072 & NE5532 are 2 classical OPA for audio but they are "vintage" now. One way to test these différents part numbers in your circuit is to use a electrical/mechanical support for IC
That may be true, but it doesn't mean the venerable TL062 is a bad choice for this application.
My humbucker-equipped guitars have a 500 kilo-ohm volume pot inside them. Turned down to half-resistance (that's just a smidge on a log pot), the source impedance is one quarter of that, or 125 kilo ohms.
The thermal noise spectral density from a 125k resistor at room temperature (25 C) is 45 nV/root-Hz.
From the Texas Instruments datasheet, the noise spectral density of a TL061 is 42 nV/ root-Hz.
The two numbers are about equal, so a TL061 does not add any significant amount of hiss to a guitar. It doesn't matter if newer op-amp designs exhibit lower self-noise; the guitar's own self-noise will set the noise floor. Not even a mathematically perfect magic zero-noise op-amp can lower the noise floor by more than 3 dB compared to a TL061. And 3 dB is simply too small an amount to matter.
The OPA2180 seems very nice and has one-fourth the equivalent input noise - but with a guitar plugged in, it will be almost as noisy as the old workhorse TL062.
Also, from the TI datasheet, the OPA2180 comes only in tiny surface-mount packaging; a huge pain in the rump for most hobbyists to deal with.
For me, this is the deal-breaker. If the OPA2180 was also available in 8-pin DIP packaging with 0.1" lead spacing, sure, go ahead an use it; it won't improve anything, and it's overkill, but at least it won't hurt anything either. But it doesn't come in through-hole packaging.
As a matter of philosophy: part of being a good engineer is using appropriate parts and materials for the application. Wheels for your car are either stamped from cheap sheet steel, or made from aluminium; no car manufacturer makes them out of, say, chrome vanadium steel rather than cheap carbon steel, though the former is stronger. There's simply no need for the extra strength, and there are no benefits for the extra cost.
IMO, using an OPA2180 with an electric guitar plugged into it is in the same category. SMD packaging makes it more cumbersome to use, and most importantly, it gives you no worthwhile noise (or other) benefits.
The TL072 I suggested, has noise performance in between the TL062 and OPA 2180. And you can still buy them in 8-pin DIP packages - no need to deal with fiddly SMD parts. But, as Fahey pointed out, the TL072 has more power consumption than the TL062, which may be a factor if it's used in an onboard guitar preamp powered by a 9V flat battery.
I wonder if the OP (Rejerh) is still around, or have we lost him?
-Gnobuddy
Yo,
I am still around and reading it all. You guys are just talking technicalities way above my understanding.
But to summarise this what I got:
Thanks a lot for all your messages.
It turned out that I was getting some feedback into the active circuit by leaving the bypass connected on the output side. So i used a DPDT switch to cut off both the input side (as shown on the schematic) and the output side of the active circuit when turned into passive mode.
It works a treat now!!
I am still around and reading it all. You guys are just talking technicalities way above my understanding.
But to summarise this what I got:
- TL062 great on the battery even though a bit noisier (but should not be affecting much an onboard preamp)
- TL072 and NE5532 classic/vintage and should work treats
- OPA2180 better performer but overkill and only comes in soic form (which would be fine for me since my preamp is PCB-ed)
Thanks a lot for all your messages.
It turned out that I was getting some feedback into the active circuit by leaving the bypass connected on the output side. So i used a DPDT switch to cut off both the input side (as shown on the schematic) and the output side of the active circuit when turned into passive mode.
It works a treat now!!
See 3rd point in my post #5...and the 2nd point Kay Pirinha made in post #2.
The high bandwidth of the LM4564 makes the circuit much more prone to accidental positive feedback like this.
Glad you got to behave the way you want!
-Gnobuddy
Hi rejerh,
Welcome to diyAudio!
Glad you've gotten this working somewhat better. Lot's of excellent advice available here. If I can be forgiven for being so late with my two bits . .
Some of the wild results you've observed are evidence of the unsuitability of 'that particular op-amp' to 'this particular circuit', and little more.
Cheers
Welcome to diyAudio!
Glad you've gotten this working somewhat better. Lot's of excellent advice available here. If I can be forgiven for being so late with my two bits . .
- NE5532 is an excellent op-amp, but it is not suitable for 9V battery operation.
- Some of the op-amps mentioned will not function properly at these input impedances.
- There are a few op-amp circuits that will tolerate a wide range of parts; many that are optimized for a particular class of part will not.
Some of the wild results you've observed are evidence of the unsuitability of 'that particular op-amp' to 'this particular circuit', and little more.
Cheers