Understanding the Music Man Stingray 2-band preamp

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
The schematic here is of the original Music Man Stingray bass guitar preamp. It runs off 9VDC and I have left off the voltage divider part of the preamp schematic that provides +4.5V, rail split ground and -4.5V from the battery, just to keep it simple.

I redrew this schematic from the one on freestompboxes.org which put it up around 2004.
An externally hosted image should be here but it was not working when we last tested it.

The reputation of this design (c.f. talkbass.com) is that it sounds really good. It is also said about it that the treble is boost/cut, that the bass is theoretically boost only, but that in effect the bass has a slight cut designed into the bottom section of the pot, which I can't see in the schematic...can you? Another aspect of this design in that (it is said) the LM4250C op amp has such a low slew rate that it rolls off highs at something like 8-11K, and thus that op amp choice is actually a part of its sound, besides its having very low quiescent current. I'd like to know if this is really true.

The EQ technology has some unconventional features (to me at least) and I'm hoping one of you guys can explain to me how it works and most importantly, how to determine what the effective frequencies are for bass and treble sections.

I drew it with colored lines for the various feedback paths so it would be easier to talk about.

Feedback paths:

Path A in red through C4
This is a high cut, pretty standard. I'd like to know how to calculate the corner frequency for this cutoff effect. Doesn't seem possible to treat it as an RC filter with a HUGE R (the input impedance of the - pin of the op amp. Is there an easy way to calculate the f(c) of the 100p or 120p cap?

Path B in gold through C5
This would pass mids and highs for unity gain thru R2, the 220k feedback resistor, except for the effect of P2, R3, C6, which is the hardest part of the circuit for me to grasp.

Path B+C in turquoise/aqua through R3, C6
This is the big mystery. I checked the corner frequency to see what frequencies would be blocked by C6 to ground. It looks to me like applying 1/(2*pi*R*C) [in Excel this would be =1/(2*3.14*10000*0.000010)] gives a corner freq of 1.6 hZ. So my big questions are...what does this section do, and what frequencies are affected by it?

Path D in blue and path E in purple
For treble freqs passing thru C3. P2 provides treble cut or gain. I think I understand how it works, but how do you calculate the frequencies that are affected? And is it odd that the two caps (C2 and C3) don't match?

If someone can unravel these mysteries for me, it might help a bunch of us. On talkbass.com there was recently a big group buy of PCBs for this preamp. It's very popular.

John
 
A little mini-bump--here's how I think the treble EQ section works:

If the input to the op amp is positive, then the output is negative (and vice versa of course). For higher frequencies, P1 joins the input to the output. So, there must always be a spot in P1 where the voltage is at 0VAC. And if the wiper is set at that point, the treble will be neither boosted nor cut. If the op amp were operating at unity gain, this point (seems to me) should be at 500K ohms, i.e. half of one megohm.

If the wiper applies some input signal to the wiper the op amp will have to increase its output, and if the wiper applies some output signal to the wiper the op amp will have to decrease its output. In both cases C3 assures this only applies to the treble range.

If I'm wrong I'd appreciate being set straight...

The fact that P1 is a log pot suggests there's some boost in the circuit, but I am still hoping for someone to explain how R3 and C6 affect gain for bass, mids and treble, because I don't see the gain anywhere else.
 
>100 views, but no replies (except my bump)...

I certainly have no hard feelings about the lack of replies, but I am going to post this on freestompboxes.org instead, since no one has jumped in to explain the circuit (to me and to you).

So I'm asking the moderator to deal with this however is appropriate. I really don't like it when guys place the same post on multiple forums. Delete this one? Mark it inactive? as you wish...

But users who might feel moved to offer some clues about how the schematic works, please reply over there so we're not splitting the info over 2 forums.

John
 
1) please post the actual schematic, this redrawn version is incomplete and has errors, so it really does not tempt anybody to "explain" (what's perceived as a flawed design).
2) that said, and to start with something:
a) the treble control looks like a simplified version of the treble side in a Baxandall Tone Control
The original Baxandall is symmetrical, uses a linear pot and equal capacitors on both ends, so boost and cut frequencies are the same, and on "5" it's flat.
This modded one, uses a reverse log pot ( so knob "5" is approximately electrical "8" or more) and Boost cap 1800 pF is much larger than the 500pF cut one.
So on "5" it's already boosting (a lot) of highs and high mids; I'm sure that on 0 it kills highest treble but not high mids.
Why would they choose such a "weird" settings?
I'm quite certain that they wanted to provide a quite lively and snappy Bass, and didn't want Musicians to mess with it.
b) the Bass control provides boost only.
On "0" it shorts C5, low frequency gain is R2/R1=1 = Flat.
R3 C6 "do nothing", they are simply an extra load, easily driven (unaffected) by the Op Amp output.
On "10" it forms a 10X voltage divider with R3, so gain gets boosted 10X, but P2 is bypassed by C5 , so only low frequencies are boosted, in the band from 32 to 320 Hz.(the 3 lowest octaves)
Hope this is enough, but anyway post the original full schematic, in case somebody feels curiosity.

Side note; "100 views" is *nothing*, here threads sometimes reach 1500 posts or more, with , say, over 20000 views ;)
 
Last edited:
Here's the schematic Bajaman originally posted.

An externally hosted image should be here but it was not working when we last tested it.


I rearranged the drawing of the 47nF cap, but if you check the pathways the schematics are functionally the same there. I corrected the 1uF DC-blocking cap on the output with a 10uF, which is supported in a lot of discussion earlier in this thread. I showed 100pF and 120pF caps because in the discussion here or on pickup maker's forum, some ops have seen this variation. Ditto the 100K vs 220K (for more gain) in the feedback loop above. And of course 1n8 and 1800pF are the same. But my cap supplier(s) use only uF and pF and skip nF completely, so I just changed the notation, not the value. If there ARE errors (I don't know of any) please let me know what they are.

JM Fahey--thanks for your analysis. These non-matching caps on the treble side mean that the treble cut corner frequency is higher (500pF) than the boost frequency (1800pF)? Interesting--I'll try to measure that when I get it built.

One more question (for all electronics gurus): what's the formula to use to calculate the corner frequency for the 500pF, 1800pF and 120 or 100pF caps?
 
Fine.
If there ARE errors (I don't know of any) please let me know what they are.
In a nutshell, the errors are what you modified :(, go by the original one and you won't have problems :)
Just as an example:
1) why talk about +/- 4.5V supplies while you have an excellent, single supply, 9V circuit?
Which by the way is easier to turn on-off .
2) you didn't provide for a DC path for the Op Amp - input. <-this one is gross .
3) you disregarded the LM4250 current setting pin 8 and resistor 1M5 .
These non-matching caps on the treble side mean that the treble cut corner frequency is higher (500pF) than the boost frequency (1800pF)?
Yes.

One more question (for all electronics gurus): what's the formula to use to calculate the corner frequency for the 500pF, 1800pF and 120 or 100pF caps?
The guru hat doesn't fit my head, if anything because I'm *still* learning, but I can answer these doubts:
4) at max boost 1800pF is in parallel with input 220K, boosting highs at 6dB/octave above a turnover frequency of around 400 Hz.
As I said, it also boosts mids and *a lot* of mid-highs.
Remember it's a Bass guitar, a "difficult" instrument to amplify, mainly because the basic design is flawed from the beginning. Oh well.
5) on max cut, 500 pF is in parallel with 100K, cutting 6 dB/oct, above around 3200 Hz.
Meaning it will cut little music, if any, mainly hiss or some external RF or buzz.
As I said, they don't want to "muddy" the sound.
6) 100pF is also (always) in parallel with 100K, cutting above nominal 16 KHz.
Doubt the LM4250 has much gain there, if any at all, but it's a good anti-RF measure, never hurts.
I have been called to help in a rush to 48000 people filled Football Stadiums when Musicians about to play had to delay it because of unwanted RF intrusion, so it *is* a real, everyday problem.
7) get the "official" PCB design (I've seen it somewhere) and build it "by the book", do not Mod things before listening the original first.
If you can't find the LM4250, you *may* use a TL061, with no current setting capability of course.
Battery will last a lot anyway, over a year.
Don't fall for the temptation of using an "exotic" modern Op Amp here, you'll only increase out-of-band noise or make it susceptible to interference ... not forgetting the 10 to 40X shorter battery life, but it won't sound "better" ;)

EDIT: you didn't ask for it, but the bass boost comes from the 47nF in series with 10K, boosting below 320Hz, down to 32Hz (.047/100K).
 
Last edited:
I have bought a few LM4250s from Digikey, but for my first iteration on the preamp, which is almost ready to test, I'm using an OPA2244, which also has a very low gain-bandwidth product (thus slew rate) and very low quiescent current. I hope this counts as not a 'modern' op amp (which I take to mean a very fast and very current-hungry design). I left out pin 8 from the schematic--the current setting pin--since it doesn't apply to my case and also because it is not directly a part of the audio. I'm using a TLE2426 virtual rail ground splitter, so I won't implement the part with the 2.2M resistors, but really, that's not a part of the audio path either, so I felt it was just a distraction to the readers.

Where you said "you didn't provide for a DC path for the Op amp - input" I am confused about that (a) because I thought that was necessary only for non-inverting op amp circuits, (b) because it seems in inverting op amp circuits to be provided by the 'real' ground on the non-inverting pin (pin 3) and (c) because I don't see that in the original schematic either. Are you sure about this? Can you clarify?

I get your point about not modifying the design before trying it out, so I'll try to cut back on the urge to innovate. Thanks for explaining about the frequency calculations.

John
 
One more thing about the bass boost. Am I right that when you first start to dial in boost and the pot is just coming off 0 ohms that the boost corner frequency is around 320 hz, then as you gradually roll on more boost the bass gain increases but the corner frequency drops so that at max gain (100K ohms on the pot) the corner frequency is down to 32hz?

If so then on a 4 string bass, whose lowest string (open E) is 41 hz, you may actually hear more bottom end with the pot half way up than with it all the way up, boost peaking at 32hZ which is roughly the fundamental of a low B. And most bass cabs don't really provide any 32hZ output anyway, just the harmonics. Interesting?

Finally, if you have the time and patience to give a little discourse on what you see as the inherent flaws in Leo's design of the bass guitar, I'd be interested in hearing that too.
 
Eh, decided to just go ahead and do the sim. I used 180K in the feedback loop where 100K/220K is spec'ed, since that's what I have in the one in my bass (Travis Bean fretless). 1uF output cap, going to 10 doesn't do much at all and I wanted to use nonpolar coupling caps. 330K on the voltage divider resistors since I'm using LT1351 which is more current hungry.

So here are some curves that LTspice threw out. Scale for bass and treble controls is 0-10, I use 01 and 99 in file names to reflect min and max rotation because that's how LTspice specs them. Freq sweep is 20Hz-20KHz.

T=0, B=0

An externally hosted image should be here but it was not working when we last tested it.


T=10, B=10

An externally hosted image should be here but it was not working when we last tested it.


T=4, B=5

An externally hosted image should be here but it was not working when we last tested it.


T=4, B=3

An externally hosted image should be here but it was not working when we last tested it.


T=4, B=7

An externally hosted image should be here but it was not working when we last tested it.


T=10, B=0

An externally hosted image should be here but it was not working when we last tested it.


T=0, B=10

An externally hosted image should be here but it was not working when we last tested it.


Hope this helps?
 
No, using the Uncle Fluffy board would be the easy way and I usually do things the hard way. Plus I want to do a lot of customizing and experimenting, which is why I was so eager to find out how the EQ sections really work.

Here's the work in progress:

An externally hosted image should be here but it was not working when we last tested it.


The huge resistor is the 220K, which I ordered from Digikey and somehow it sneaked through my filter for 1/4 watt parts--it's a 3-watt. I'll replace it when I go shopping for more parts locally next week. I'm in India until March and there are somethings you can't get here, like exotic op amps. The white ready mate connectors are there so I can modularize the parts that won't change--pickup and blend control (3 pins), +4.5, 0, -4/5v supply with battery (3 pins), treble and bass pots (the 6-pin connector) and output (3 pins, only need 2). That way I don't need a full set of pots for each preamp I solder up.

Sometimes the hard way is more fun. I take JM Fahey's point about starting off plain before making changes, but I am being faithful to the parts values of the original, except for the rail splitter, which is a no-brainer, and using a dual op amp because of the input buffer.

And to make things really challenging I am winding my own low impedance split coil pickups...
 
Sometimes the hard way is more fun.

Yes, I can totally appreciate that. For me, learning to do Spice models is a lot harder than just huffing a lot of solder smoke and physically swapping out components. I've done plenty of that though, and still will. :p

You might be able to get one of the big opamp manufacturers to send you some engineering samples, even in India.
 
That sim is fantastic! I was planning to try to do those measurements with a laptop freeware signal generator feeding the circuit, then a desktop (nearly) freeware real time analyzer to plot the results. It looks to me like I really need to learn how to do simulation.

I'm sure MM worked hard on the component values and crossover points. I like the even transition from bass to treble, like in the T=4, B=3 and the T=4, B=7 curves. You don't get a notchy or scooped result except in the T=10, B=10, which is what you'd be trying for with those settings.

And looking at the T=0, B=0 curve at the top of your post, I see the evidence for the claim that although the bass pot is supposed to be boost only, it actually has a cut (very small, like -2dB) when set to zero.

And the T=10, B=0 is interesting because this is a shelving EQ and would rise indefinitely, except that the high end is limited by the 120pF cap, and maybe also by the slowness of the op amp. I'm thinking I'd like to try a bigger cap, maybe 150pF to bring this peak a little lower. On the other hand unless you have a tweeter, you're not going to reproduce that 10-12K stuff anyway. And unless you slap and pop a lot I don't think you'd miss it.

Thanks again--this is really helpful.
 
That sim is fantastic! I was planning to try to do those measurements with a laptop freeware signal generator feeding the circuit, then a desktop (nearly) freeware real time analyzer to plot the results. It looks to me like I really need to learn how to do simulation.

I'm sure MM worked hard on the component values and crossover points. I like the even transition from bass to treble, like in the T=4, B=3 and the T=4, B=7 curves. You don't get a notchy or scooped result except in the T=10, B=10, which is what you'd be trying for with those settings.

And looking at the T=0, B=0 curve at the top of your post, I see the evidence for the claim that although the bass pot is supposed to be boost only, it actually has a cut (very small, like -2dB) when set to zero.

And the T=10, B=0 is interesting because this is a shelving EQ and would rise indefinitely, except that the high end is limited by the 120pF cap, and maybe also by the slowness of the op amp. I'm thinking I'd like to try a bigger cap, maybe 150pF to bring this peak a little lower. On the other hand unless you have a tweeter, you're not going to reproduce that 10-12K stuff anyway. And unless you slap and pop a lot I don't think you'd miss it.

Thanks again--this is really helpful.

Jay, keep in mind that this sim more or less includes the effect of your buffer section too, as I just used a low impedance sig generator to feed the model. I haven't really fully sussed modeling a pickup as the signal source yet, just one of many things I hope to get to eventually. PM me if you want the LTspice files, this is a pretty simple circuit and could be a good way to get your feet wet. LTspice itself is freeware and there is wide support online.

I do use RTA software (True RTA, Holm Impulse, Room EQ Wizard) quite a lot to test DIY builds, and RMAA as well. On Talkbass there's a DIY tube bass preamp thread where I compared modeled and measured frequency response curves, the correlation was excellent in that particular case.
 
Last edited:
Hi,

(Bass) guitar pickups have resistance and inductance, values are not hard
to find. Split coil P type bass pickups can have a series / parallel switch
as can the less common humbucker types, e.g. Gibson's / Musicman's.

(Jazz type pickups can be split to allow the same arrangements.)

The S/P switch is rare on actives, but works fine on passives. Series
is deeper, with more output, parallel brighter with less output, they
interact with the usual passive volume and tone controls.

rgds, sreten.

Higher the output, usually means the higher impedance and inductance.
 
Hi,

(Bass) guitar pickups have resistance and inductance, values are not hard
to find. Split coil P type bass pickups can have a series / parallel switch
as can the less common humbucker types, e.g. Gibson's / Musicman's.

(Jazz type pickups can be split to allow the same arrangements.)

The S/P switch is rare on actives, but works fine on passives. Series
is deeper, with more output, parallel brighter with less output, they
interact with the usual passive volume and tone controls.

rgds, sreten.

Higher the output, usually means the higher impedance and inductance.

Hi Sreten,

There's rather a bit more to it, actually. Nice discussion here.
 
I'm sure MM worked hard on the component values and crossover points. I like the even transition from bass to treble, like in the T=4, B=3 and the T=4, B=7 curves. You don't get a notchy or scooped result except in the T=10, B=10, which is what you'd be trying for with those settings.
As you see, the "original" in any commercial product which became a Classic is usually very good.
H*ck! There must be a reason behind that! :eek:
It can't be just good luck!!
Which it never is, of course, but *lots* of hard work.
That's why I always suggest everybody to start with the original one, whatever it is, learn to use it and what it does, and only after that, you should start tweaking.
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.