Hi guys,
I'm thinking of knocking together a preamp to put in my guitar in order to lower the output impedance. I had a look at the available audio opamps (TL072, NE5532 et.al) but was dissatisfied with the quiescent current requirements. A web search only turned up simple single transistor FET buffers, with pretty average distortion specs. I figured I'd roll my own.
After playing with LTspice, I came up with the following:
It's a simple enough matched JFET diffamp, with a follower on the output. I've used constant current sources as loads to improve open loop gain. It's unity gain stable, with about 70 degrees phase margin.
Any comments or ideas for improvement would be appreciated. is 1V RMS adequate for guitar use? Any hints for further reducing power consumption...
Cheers,
Suzy
I'm thinking of knocking together a preamp to put in my guitar in order to lower the output impedance. I had a look at the available audio opamps (TL072, NE5532 et.al) but was dissatisfied with the quiescent current requirements. A web search only turned up simple single transistor FET buffers, with pretty average distortion specs. I figured I'd roll my own.
After playing with LTspice, I came up with the following:
- Iq = 200uA
- 0dB gain (buffer)
- >1Meg input impedance
- <10K Output impedance
- low distortion (<0.001% THD for 100mV RMS input)
- >1V RMS maximum signal swing (0.02% THD)
- 20Hz to 20KHz bandwidth
- low noise
An externally hosted image should be here but it was not working when we last tested it.
It's a simple enough matched JFET diffamp, with a follower on the output. I've used constant current sources as loads to improve open loop gain. It's unity gain stable, with about 70 degrees phase margin.
Any comments or ideas for improvement would be appreciated. is 1V RMS adequate for guitar use? Any hints for further reducing power consumption...
Cheers,
Suzy
Hi Suzy
It may be a matter of taste, but personally I'd go for something simpler. Your circuit is undoubtedly beautiful and state of the art, but is it really needed considering the application?
A differential input is useful when you need DC accuracy, but this is clearly not the case here; and I'm not sure striving for the highest possible open loop gain is the best of the ideas: the real load will certainly be far from a nice 10K resistance and will include a capacitive component from the cable, amplifier input, etc
Will it still be stable under those conditions?
Anyway, if you persist in your choices, you could probably add degeneration resistors in the emitters of Q2 & Q4. It would improve the gain, linearity and would minimize the mismatches between the transistors (you're not designing an IC).
You could probably reach similar results using a follower made of a composite N-Fet/P-Bjt, with a resistor between the base and emitter of the Bjt to set the Fet current at some nice, stable value.
The load could be a simple resistance or a current sink depending on the dynamic/current consumption compromise. And I'd add a 100ohm resistance in the output to avoid any HF instability driving the cable. Such a circuit could easily be powered from the other end of the cable.
Regarding the signal handling capability, 1V may or may not be sufficient; the output of transducers may vary widely from a model to the other, and will also strongly depend on the adjustment/clearance of each individual device. The attack waveform can reach huge levels; I recommend you make some real measurements.
LV
It may be a matter of taste, but personally I'd go for something simpler. Your circuit is undoubtedly beautiful and state of the art, but is it really needed considering the application?
A differential input is useful when you need DC accuracy, but this is clearly not the case here; and I'm not sure striving for the highest possible open loop gain is the best of the ideas: the real load will certainly be far from a nice 10K resistance and will include a capacitive component from the cable, amplifier input, etc
Will it still be stable under those conditions?
Anyway, if you persist in your choices, you could probably add degeneration resistors in the emitters of Q2 & Q4. It would improve the gain, linearity and would minimize the mismatches between the transistors (you're not designing an IC).
You could probably reach similar results using a follower made of a composite N-Fet/P-Bjt, with a resistor between the base and emitter of the Bjt to set the Fet current at some nice, stable value.
The load could be a simple resistance or a current sink depending on the dynamic/current consumption compromise. And I'd add a 100ohm resistance in the output to avoid any HF instability driving the cable. Such a circuit could easily be powered from the other end of the cable.
Regarding the signal handling capability, 1V may or may not be sufficient; the output of transducers may vary widely from a model to the other, and will also strongly depend on the adjustment/clearance of each individual device. The attack waveform can reach huge levels; I recommend you make some real measurements.
LV
Oh, by the way, it's so obvious I had nearly forgotten: you need some DC feedback too, otherwise the operating point will be completely unpredictable; probably the simplest method is to put R10 in parallel with C3. I figure you too know it, but it's better being reminded before the copper is etched...
LV
LV
suzyj said:
Nice www.littlefishbicycles.com guitar pre-amplifier!
I couldn't have done it better myself ( EWB, MultiSim9 )
Only one thing:
Power Supply?
Voltage requirements for good operation?
I am sure it is battery device, 9 Volt maybe.
If you think this one or any other battery circuit
needs a low consuming matching precision! Regulator for 9V Alkaline Battery
I have a little such regulator, with only 100uA drawn
7.20 Volt output,
for 7.50 -50.0 Volt input!
This is a circuit I have built.
It is extremely precise, can be made small ( 20x20 mm, 8/10 of an inch) and good working
even though my practical build is using only 101.8 uA,
for circuit bias and own consumption.
Will provide 50 mA with no effect on output voltage: Still 7.200 Volt DC
My Circuit, it is my own design from beginning.
Uses 3 TO92 small signal low noise transistors
and some standard diodes and resistors.
No use of voltage references or other IC chips.
Of course could easily be set for other output voltage, if needed.
7.20 was chosen, because at 7.5V most 9V alkaline batteries are practically empty.
Regards
lineup
Lineup Audio Lab
Elvee said:
D'Oh! Thanks for that. I put R10 across C3, and made both R10 and R11 1Meg (to get the gate bias on J2 equal to J1 again).
lineup said:
Yes, it is. if I keep the quiescent current down to 200uA, that means I can get around 2500 hours (3 months) operation from a 9V PP3 alkaline battery (500mAh). I'll arrange it so that the battery is disconnected when I remove the lead from the guitar, so the battery lifetime should be quite long.
I don't think regulating the supply is all that important. I've set the operating point so that it works okay down to 7V or so. It still works well with two batteries in series (18V).
I'm actually tempted to use a pair of 3V CR123 lithium cells. My guitar is an Ibanez SA160, with a very thin body. Finding a spot for a 9V battery is rather difficult. I could squeeze in two CR123 cells, and 1500mAh would mean the battery would last for ages.
darkfenriz said:
The idea of this circuit is to buffer the pickups. In this application, I don't think that adding distortion is helpful. Certainly, really low distortion isn't necessary, but it isn't that hard to get, either
Cheers,
Suzy
Hi, that's interesting!
I'm sure there are some! op-amps that fit your requirements...
You want high input impedance, low quiescent current, low distortion (yes... is that really required?)
And possibly 9V battery block operation?
...wait a minute, I'll be back!
Edit: Now you posted just before me, in the meantime the power supply question is solved.
I'm sure there are some! op-amps that fit your requirements...
You want high input impedance, low quiescent current, low distortion (yes... is that really required?)
And possibly 9V battery block operation?
...wait a minute, I'll be back!
Edit: Now you posted just before me, in the meantime the power supply question is solved.
Well, the datasheet is a bit confusing, because it covers single and dual plus unity gain-stable and higher gain versions of the otherwise same op-amp...
...but the specs seem right: very high input impedance, low noise, distortion, a quiescent current of 0,45mA. A low 3MHz of gain-banwidth, but that will be enough
Max4249-4257 (you'll have to look for the right version and case, btw. they're all SMD) (Maxim-IC.com)
The only thing is I don't know how much voltage the pickup of a guitar produces! (To stay within the voltage rails)
That op-amp above only likes a 6V supply at max!
Then there's the OP1177, input impedance isn't given in the datasheet, but I guess it's rather high although it has a bjt input.
0,36mA quiescent current, no distortion data given. Supply from about 3 to 36V. (Analog.com)
LT6013 look s fine too with only about 0,15mA quiescent current. It's got only a slew rate of 0,2V/us, that's not much, but should be enough (still don't know how much signal voltage we are talking about). Distortion isn't given, but at gain=1, which op-amp will really produce too much distortion for an electric guitar? Supply from 2,7 to 40V
(Linear.com)
There are certainly way more op-amps for you out there, but I quickly found those because I already picked the data before. I just additionally had the requirement of low noise and low offset, that you don't have (to that extent) in your application.
Well, maybe you already decided to build your deiscrete version, but because you mentioned in your first statement that you initially searched for op-amps I wanted to show you that there's more than the common op-amps.
If you know the specs you need in an op-amp, you can nicely use the search functions at the OP-AMP manufacturer's sites. I didn't mention National yet...
Cheers,
Dominique
...but the specs seem right: very high input impedance, low noise, distortion, a quiescent current of 0,45mA. A low 3MHz of gain-banwidth, but that will be enough
Max4249-4257 (you'll have to look for the right version and case, btw. they're all SMD) (Maxim-IC.com)
The only thing is I don't know how much voltage the pickup of a guitar produces! (To stay within the voltage rails)
That op-amp above only likes a 6V supply at max!
Then there's the OP1177, input impedance isn't given in the datasheet, but I guess it's rather high although it has a bjt input.
0,36mA quiescent current, no distortion data given. Supply from about 3 to 36V. (Analog.com)
LT6013 look s fine too with only about 0,15mA quiescent current. It's got only a slew rate of 0,2V/us, that's not much, but should be enough (still don't know how much signal voltage we are talking about). Distortion isn't given, but at gain=1, which op-amp will really produce too much distortion for an electric guitar? Supply from 2,7 to 40V
(Linear.com)
There are certainly way more op-amps for you out there, but I quickly found those because I already picked the data before. I just additionally had the requirement of low noise and low offset, that you don't have (to that extent) in your application.
Well, maybe you already decided to build your deiscrete version, but because you mentioned in your first statement that you initially searched for op-amps I wanted to show you that there's more than the common op-amps.
If you know the specs you need in an op-amp, you can nicely use the search functions at the OP-AMP manufacturer's sites. I didn't mention National yet...
Cheers,
Dominique
Dominique said:Well, the datasheet is a bit confusing, because it covers single and dual plus unity gain-stable and higher gain versions of the otherwise same op-amp...
...but the specs seem right: very high input impedance, low noise, distortion, a quiescent current of 0,45mA. A low 3MHz of gain-banwidth, but that will be enough
Max4249-4257 (you'll have to look for the right version and case, btw. they're all SMD) (Maxim-IC.com)
The only thing is I don't know how much voltage the pickup of a guitar produces! (To stay within the voltage rails)
That op-amp above only likes a 6V supply at max!
Then there's the OP1177, input impedance isn't given in the datasheet, but I guess it's rather high although it has a bjt input.
0,36mA quiescent current, no distortion data given. Supply from about 3 to 36V. (Analog.com)
LT6013 look s fine too with only about 0,15mA quiescent current. It's got only a slew rate of 0,2V/us, that's not much, but should be enough (still don't know how much signal voltage we are talking about). Distortion isn't given, but at gain=1, which op-amp will really produce too much distortion for an electric guitar? Supply from 2,7 to 40V
(Linear.com)
There are certainly way more op-amps for you out there, but I quickly found those because I already picked the data before. I just additionally had the requirement of low noise and low offset, that you don't have (to that extent) in your application.
Well, maybe you already decided to build your deiscrete version, but because you mentioned in your first statement that you initially searched for op-amps I wanted to show you that there's more than the common op-amps.
If you know the specs you need in an op-amp, you can nicely use the search functions at the OP-AMP manufacturer's sites. I didn't mention National yet...
Cheers,
Dominique
There are plenty of "micro power" op-amps out there that would fit the bill, but for hobbyist stuff, who want’s to design with IC op-amps? BORING!
Cheers,
Glen
The LT6013 looks pretty cool, but I note it's not unity gain stable, which would make it harder to play with.
I think the main advantage of a discrete implementation is that I can optimise the circuit to match a specific application. For example, if I use a pair of 3V lithium cells rather than a 9V alkaline, I can increase Iq to get back more dynamic range...
I've been reading about pickups over the last couple of days, and thought it might be fun to get some inexpensive pickups, unwind them, and then use the cores to make active ones (by rewinding them with fewer turns and adding some gain). Having a preamp with a maximum of tweakability would really help for this.
Cheers,
Suzy
I think the main advantage of a discrete implementation is that I can optimise the circuit to match a specific application. For example, if I use a pair of 3V lithium cells rather than a 9V alkaline, I can increase Iq to get back more dynamic range...
I've been reading about pickups over the last couple of days, and thought it might be fun to get some inexpensive pickups, unwind them, and then use the cores to make active ones (by rewinding them with fewer turns and adding some gain). Having a preamp with a maximum of tweakability would really help for this.
Cheers,
Suzy
Suzyj,
I've been wanting to make my own pickups too.
Been thinking about making them with neodymium magnets and winding them for low impedence, then using low noise opamps to add the necessary gain, while buffering the output.
Been wanting to make at least 6 individual pickups, one for each individual guitar string. Then I can mix / process the outputs externally.
Then I want to try using standard RJ-45 jacks and Cat-5 cables for multichannel output, which should also provide enough leads for an external power supply. With a built-in buffer, and its low impedance output, noise should not be much of a problem, and hopefully, sheilded cable will not be necessary. I hate 1/4 phono jacks for guitar!!!
Let me know how well your circuit works.
You have my attention, and I wish you great success!
I've been wanting to make my own pickups too.
Been thinking about making them with neodymium magnets and winding them for low impedence, then using low noise opamps to add the necessary gain, while buffering the output.
Been wanting to make at least 6 individual pickups, one for each individual guitar string. Then I can mix / process the outputs externally.
Then I want to try using standard RJ-45 jacks and Cat-5 cables for multichannel output, which should also provide enough leads for an external power supply. With a built-in buffer, and its low impedance output, noise should not be much of a problem, and hopefully, sheilded cable will not be necessary. I hate 1/4 phono jacks for guitar!!!
Let me know how well your circuit works.
You have my attention, and I wish you great success!
suzyj said:
suzyj
I agree with your discrete approach. It would be my choice, too.
Most flexible, regarding most any parameter:
- power consumption,
- ensure Class A operation in the load in question
- choice of input transistors, JFET or BJT and etc. etc.
Of course this is for persons with at least basic knowledge
how to use transistors. In this case a nice dual JFET with one current mirror.
------------------------
But also agree, that for anyone wanting to use Op-Amp,
after some search at good Op-Amp manufacturers website
it is no problem to find a VERY Low Power consumption device
with good suitable data and working perfectly
at lower voltage and single DC supply ( +2.70V to +9.60V ).
There are plenty of these Op-Amps these days.
This comes from the fact that
portable battery driven Freestyle and MP3-players has became so popular .
(we all can see many everywhere with their freestyle earplugs attached in ears)
These portable gears needs low supply power small ICs, op-amps,
working good at for example: 2-4 x 1.35V ( 2-4 AA, LR6 batteries ).
So anyway we chose to build our little preamp for guitar pickups
we have many good options for making such an amplifier
with very good quality.
No problems here!
As seen by
lineup
http://lineup.awardspace.com/
It's probably irrelevant now that you are considering an adaptable design with voltage gain, but if you're still interested in a unity gain buffer, have you considered a discrete low-power implementation of the "diamond buffer", like that shown here:
http://focus.ti.com/lit/an/sboa071/sboa071.pdf
Particularly the circuit in fig 11. For a discrete implementation you'll need to add emitter balast resistors to each output device and also directly at the emitters of the input transistors to set the output pair's bias current. Being a straight buffer without negative feedback, stability wouldn't be an issue and distortion could still be kept relatively small.
Cheers,
Glen
http://focus.ti.com/lit/an/sboa071/sboa071.pdf
Particularly the circuit in fig 11. For a discrete implementation you'll need to add emitter balast resistors to each output device and also directly at the emitters of the input transistors to set the output pair's bias current. Being a straight buffer without negative feedback, stability wouldn't be an issue and distortion could still be kept relatively small.
Cheers,
Glen
G.Kleinschmidt said:
Very good option, G.Kleinschmidt.
In fact, such a buffer
is seen more than often! in good Op-Amp internal schematic.
Used as the output buffer for very good performance (chip-)amplifiers.
It is a design element (part of amplifier) that is proven to work very well.
So that is why it is so commonly used.
lineup
There is a pretty uniform agreement that load resistors sound better than a current mirror. If you search you will find Nelson Pass, John Curl, and most other respected designers making this recommendation.....which I also follow in my designs.
Something worth trying when you solder up a prototype.
Something worth trying when you solder up a prototype.
G.Kleinschmidt said:
Hey, that's a pretty cool circuit. Looks like a great replacement for the follower on the back end of the diffamp. I just knocked one together with LTspice. It drives to within 0.7V of the rail, has a nice low output impedance and high input impedance, and 0.7% THD...
<grin> Eight bipolars and a pair of matched JFETs... Where will it end!
Cheers,
Suzy
Wavebourn said:
I wont comment on the supposed sonic merits of a pentode audio amplifier, but I will point out that fitting a hot running pentode along with it's necessary heater and plate power supplies into a guitar might be a little more difficult than slipping in a solid state amp running off a 9V battery
However, if you wanted to build something like that Yahoo Serious guy did in that 'Young Einstein' movie...................
G.Kleinschmidt said:
I wont comment on the supposed sonic merits of a pentode audio amplifier, but I will point out that fitting a hot running pentode along with it's necessary heater and plate power supplies into a guitar might be a little more difficult than slipping in a solid state amp running off a 9V battery
However, if you wanted to build something like that Yahoo Serious guy did in that 'Young Einstein' movie...................
Good idea, thank you!
I have already couple of pentodes fitted into pair of Marshall condencer mics. I'm afraid my sister in law may not like me to fit a tube in her $800+ guitar, but I can put it in mine.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.