I have been messing around with IC circuits, whcih I admittedly know almost nothing about. I was looking at the IC buffer schematic and a simple non inverting boost and realized they were basically the same and could be combined into one device. So I came up with the booster/buffer below:
(http://moosetone.com/images/schematics/boofer02.jpg)
It works and sounds good, and the buffer is great. There is a small issue: I have a little too much gain in the boost setting, even at full off it is above unity.
The biggest issue though is a wierd interaction between the cable connected to the boost and the gain. I put the 2cnd 47k resistor next to the pot to keep the pot from "turning off" the IC when turned too far. The signal would pop and get fuzzy and crackly, or cut out all together. Turning the gain back up would get rid if it. It took a lot of fiddling with the two gain resistor values to find a sweet spot where the pot will turn the full length and not screw up the IC. BUT, it turns out that this issue is dependent on the length of cable connected to the OUTPUT of the device. If I use a standard 20 foot cable I get full rotation with no issues. Connect another 20' and now the pot can't turn all the way down without the static and fuzzyness from the IC. I have used several different cables and connection methods, they all do the same thing.
What on earth can be going on here??
Just curious,why did you use 2M resistors for yout voltage divider ?
Instead of 100K or so ? I am still learning so I am curious not being a smart alec.
Also I had have issues with cables causing static cause the 1/4" plugs were slightly loose and intermitently shorting.Oh So slightly,causing static and crackle at times.
Lots of trouble shooting just to unscrew the plugs and re-solder the coax :icon_redface:
May not be your problem,but IDK how the output cable capacitance could effect your pedals gain.
Except for your lack of R out maybe,the cable is loading your circuit ?
Perhaps a 1-10K resistor before your 10uf cap would solve this ? IDK
Also maybe a 100uf/10n filtering would help stabilize your circuit ?
Thanks for the thoughts. This is issue seems to be with the cable loading the circuit, not a cable issue itself.
I used 2M as that was how the two original circuits were both designed. I just combine the two circuits into one. I don't make any claim to understand things well myself. I assumed this would create a desirably high input impdeance.
I put a 10K resistor in there at one point, before this issue was obvious, but I took it back out. I am under the impression that will make the output impedance of my circuit larger than I want for a buffer??
I put a 100u cap from +9v to ground in the circuit, it didn't make any obvious difference. Do I need a cap/resistor combo to effectively create a power filter?
Quote from: oliphaunt on April 08, 2010, 10:16:46 PM
Thanks for the thoughts. This is issue seems to be with the cable loading the circuit, not a cable issue itself.
I used 2M as that was how the two original circuits were both designed. I just combine the two circuits into one. I don't make any claim to understand things well myself. I assumed this would create a desirably high input impdeance.Quote
Ah yes that makes sense a 1M input impedance.
Quote
I put a 10K resistor in there at one point, before this issue was obvious, but I took it back out. I am under the impression that will make the output impedance of my circuit larger than I want for a buffer??Quote
IDK still learning too,a 1k might be a good fix though.
Quote
I put a 100u cap from +9v to ground in the circuit, it didn't make any obvious difference. Do I need a cap/resistor combo to effectively create a power filter?Quote
Again, I am still learning too,but adding a 100R resistor before the 100uf cap will help,I just cant tell you why.
Same for the 10n cap its good to have but I would have to search to tell you why,some circuits will sqeal or oscillate w/o the 10n in there.
Do you think a series output resistor would need to be inside the output cap or outside? Would it matter?
I took this off the breadboard and boxed it up already, I should have been more patient...
Quote from: Brymus on April 08, 2010, 10:34:42 PM
Again, I am still learning too,but adding a 100R resistor before the 100uf cap will help,I just cant tell you why.
Same for the 10n cap its good to have but I would have to search to tell you why,some circuits will sqeal or oscillate w/o the 10n in there.
It's because capacitive loading can cause oscillation with high-feedback circuits, especially if they are fed back to low gains (that is, all the open loop gain goes into feedback). It's a common flaw. Even some emitter followers can oscillate with capacitive loads. The 100R isolates the output and feedback from the capacitance somewhat.
Quote from: oliphaunt on April 08, 2010, 10:40:52 PM
Do you think a series output resistor would need to be inside the output cap or outside? Would it matter?
Yes, it matters. It should go right on the output pin of the opamp.
IDK ,I would guess before the coupling cap to provide a load for the IC.
But we are speculating on the premise that the cable fixes it by loading the circuit.
I would breadboard another one and try a few different ways to see what works best.
Otherwise someone who isnt a newb like me will need to chime in.
Ah RG chimed in while I was typing so there you have it. :icon_cool:
Thanks guys!
Would this fix it? Is 1K a good value to keep the output impedance as low as possible?
(http://www.moosetone.com/images/schematics/boofer03.jpg)
Based on other circuits I have seen I would say at least a 1K if not higher,but I am just guessing ,breadboard and see.
I am headed out of town and can't get to a breadboard until at least Tuesday, but my curiousity is killing me! Any thoughts on my last schematic?
My thoughts:
-The input bias can be done with the "noiseless biasing" for improved noise performance (see GEOFEX)
-With 1M input resistance, you can use a smaller cap without losing bass... smaller caps typically have lower leakage
-If your minimum gain in "booster" mode is too high, reduce the value of the limiting resistor in the feedback loop.
-1k is on the high side for a buffer's output impedance, but still better than a lot of well-loved guitar effects. More resistance helps isolate the opamp from the cable, but also forms a lowpass filter with the cable capacitance. Another approach is to increase output current, so it can drive those reactive loads without oscillation. Yet another technique is to incorporate a snubber filter for those frequencies at which the circuit is prone to oscillate.
-Electrolytics in the audio path can cause trouble. The one on the output is probably going to be reverse biased at least some of the time, which is unhealthy for it.
-Do the calculations to determine just how big of capacitors you need, in general. Depending on what this effect is plugged in to, you might be able to get away with an output cap 100x smaller than what's in there now.
-There is one good thing about the electrolytic output cap; electros have higher ESR compared to other cap types, which may help in isolating the op-amp from the cable capacitance. But this is not a big enough benefit to outweigh the disadvantages of using a polarized AC coupling cap.
Quote from: earthtonesaudio on April 10, 2010, 12:00:55 AM
-Electrolytics in the audio path can cause trouble. The one on the output is probably going to be reverse biased at least some of the time, which is unhealthy for it.
I'm probably not understanding the terminology or something, but how will it be reversed biased? In terms of DC the end connected to the opamp will be 4.5 volts. The "Out" side will presumably be referenced to ground, unless you connect it to something with positive DC on its input, so it will be 0v DC.
For AC the opamp side can only go between 0 and 9 volts at the very most, which will give +-4.5 volts on the "out" side, so the opamp side will always be a higher voltage than the "out" side, or have I just made a complete fool of myself :)
I probably should have said "possibly" instead of "probably." :icon_confused:
You're right that whatever you plug into is "presumably referenced to ground," but that doesn't cover every possible scenario. In practice it's really not likely to reverse bias the output cap, but it is possible. For example if the thing you plug into uses a higher bias voltage, and a leaky electro cap, then you might have +4.5V on the (+) side of the cap, but maybe +7V on the (-) side. With some series output resistance and an aluminum electrolyte, it might still work problem free, but if you used a tantalum output cap... they really don't like being reverse biased.
I am back in town and finally had a chance to get this project back on the breadboard. It turns out that the ouput resistor is the key to stabilizing the circuit. I ran it through the "coil of tonal doom" again and the boost gain stays solid with a 100R resistor, so I am hoping this will give the circuit a low enough impedance to do a good job as a buffer. In either buffer or boost mode it does a great job of cleaning up the signal loss from the coil's 70+ feet of cable.
I removed the 47K resistor on the boost pot, it was no longer needed now that the circuit is stable, and allowed me to get unity gain at the lowest boost setting. Lowering the remaing 47K resistor will add more gain to the ciruict, but this was plenty for my purposes.
I also found that the 100R/100u filter on the power does clean up some low end hum when in boost mode, but it does not not seem to contribute to the circuit stablility.
(http://www.moosetone.com/images/schematics/boofer04.gif)
I am a little concerned that in boost mode it may add a little too much lower end for my taste. It may be that a linear boost (which I assume this is) is nnot exactly what my ear wants to hear and as it gets louder I percieve more bass, and whant to get rid of some of it. Can anyone suggest a good way to do this without changing the input/output caps? I will try putting a high pass filter on the gain loop and see what that does...
I added 100R to 200R resistors to ground after the 10u outout cap and that rolls off the sub lows that can make the boost seem too rumbly. Making this a trimmer or a pot could be a really nice way to dial in just the right amount of "umph" for a specific setup.
I thought I would add a verified vero layout for a 1590A to finish this thing up:
(http://moosetone.com/images/schematics/BooferVero02.gif)
It is based off this final schematic:
(http://moosetone.com/images/schematics/Boofer04.gif)
I dont have a small value trimmer right now so I made the low cut a fixed resistor (R5) and socketed the part on the board so I can experiment with it.
I appreciate the comments above, and wish I had taken more time to think about the output cap. At some point I experimented with different values and I liked a film 1u for the output cap. I worried though that I might lose some low end I that I wouldn't notice until I got it loud on stage (it's happend before). So I left it, but I will have to keep that in mind if work on this any further. I will also look into the noiseless biasing.
I've been using this booster/buffer for a while and it causes a popping problem. It pops occasionally and it causes effects after it to pop. I assume there is voltage leaking from the electrolytic output cap. I replaced the cap wth another one of the same kind and that helped but did not solve the problem entirely. A few thoughts I had to fix this are:
- replace the electrolytic output cap with a film cap
- add a pulldown resistor after the output cap
Will the pulldown resistor change the output resistance? How much voltage is needed at the output to cause a problem?
just make the 10uF 1uF (film) and the pot a 2k and the 100R to ground 1k. for the same frequency response. Or change it to 100n cap, 20k pot (or 25k if that's what you have) and the 100R to ground into 10k to keep everything similar or the same. If it's really that 10uF cap leaking that causing the pop, than that will do it.
http://www.muzique.com/schem/filter.htm (http://www.muzique.com/schem/filter.htm)
you can put a small cap: 27p, 47p, etc on the opamp loop, since it helps with some some weird oscillations that may happen.
Your "low cut" performs the same function as an output pull-down. A really quick test would be to see if it still pops with the "low cut" engaged. If you hear any pop at all, a leaky cap is not the culprit. Well, maybe a "short circuited" cap, but not a normally leaky one. It would be good to check the DC voltage on the output jack in any case.
The way you've implemented the low-cut filter is not ideal for a TL071 op-amp. They don't like to drive much lower than 2k (I think) of resistive load, let alone 300 ohms through a huge cap.
A "nicer" way to do the low cut would be in the negative feedback portion of the circuit. There, your components could be high impedance and let the output focus on driving the long cable and nothing else.
Quote from: earthtonesaudio on June 15, 2010, 08:58:26 AM
Your "low cut" performs the same function as an output pull-down. A really quick test would be to see if it still pops with the "low cut" engaged. If you hear any pop at all, a leaky cap is not the culprit. Well, maybe a "short circuited" cap, but not a normally leaky one. It would be good to check the DC voltage on the output jack in any case.
Earthtones, thanks for the suggestions. My schematic is still not exactly correct. The low cut is on a switch, and is only engaged when the buffer is switched to boost mode. So it doesn't act as a pulldown when in buffer mode, which is where the voltage builds up.
Quote from: earthtonesaudio on June 15, 2010, 08:58:26 AM
The way you've implemented the low-cut filter is not ideal for a TL071 op-amp. They don't like to drive much lower than 2k (I think) of resistive load, let alone 300 ohms through a huge cap.
I am not following you here. If I remove the low cut resistance will that raise the resistive load to acceptible levels somehow? Sorry, I'm not familiar with what you are seeing happening here at all.
Quote from: earthtonesaudio on June 15, 2010, 08:58:26 AM
A "nicer" way to do the low cut would be in the negative feedback portion of the circuit. There, your components could be high impedance and let the output focus on driving the long cable and nothing else.
I would love to know more about this, can you be more specific about how I would implement the low cut in the feedback portion of the circuit?
It appears that you want a low output impedance (to drive long cables), and so you have a large value output cap (which has a low impedance at most signal frequencies). The trouble comes from the low-value resistance to ground. The op-amp's output can be thought of as a pretty decent voltage source, so it will put out any current necessary to make the requested output voltage. So if you ask it to make a fast +7V output swing, the cap will be basically a short circuit (for a fast signal) and then the voltage is mostly across the 300 ohm resistor. 7V/300ohms=23mA. Now if you look at the datasheet you'll notice that the output resistance of the TL071 is about 30-odd ohms. So trying to drive 300 ohms will automatically take your output voltage down by 10%.
Contrast this with the performance into 10k, and it's almost ideal.
More simply, the 300 ohm value on the output is too low. It will degrade the performance of the op-amp.
A low-cut in the feedback loop would be something like putting a small value (10nF maybe) cap in parallel with the 10uF to ground, then adding a 10k resistor in series with just the 10uF cap. Or take a look at the Timmy, I think it uses that sort of technique. Basically you're presenting the inverting input with either a small cap or a large cap, which means either a little bit of low end (small cap) or a lot of low end (large cap) is boosted.
The point of putting the tone controls at this location (feedback loop) is that they're relatively decoupled from the input and output impedances, and therefore don't mess with the buffering action.
You could alternatively implement a low-cut at the circuit's input, just like the one you have at the output, but this would mean the input impedance would vary based on the low-cut setting. Similarly, placing it on the output means the output impedance varies based on the setting. But in the feedback loop, neither the input or output impedance varies all that much as you make adjustments to the control. This leads to a more predictable behavior from the circuit, which is generally (but not always) desirable.
http://www.analogzone.com/acqt1211.pdf
http://www.analog.com/library/analogDialogue/archives/31-2/appleng.html
http://www.national.com/appbriefs/files/AppBrief108.pdf
http://www.analogzone.com/acqt0704.pdf
Now for the bass cut with gain. Look for the Dist + schematic note the RC leg and that the resistance goes down as the gain goes up this can be adjusted in a noninverting amp for the break from a gain of one frequency.
Look at this for another idea for a movable high pass. I am surprised it has not been used by the "newtweekers" but maybe because it is hand drawn and opens big, people miss it. It should open in windows picture and fax viewer
http://www.aronnelson.com/gallery/main.php/v/gus/plus.GIF.html?g2_imageViewsIndex=1
Stick you circuit in the middle.
IIRC GFR posted a TS mod years ago to have the frequency change as the gain was increased.
Do people look at my schematics? I post some good stuff and people miss it or others get credit.
:o Your poor TL071 is trying to drive only 400R resistance, or less if the cut control is turned down! That's way too much for a TL, even running at 9V. Get rid of the cut control and sub in a 100k pull-down resistor or similar.
Bass cut should always be performed at the input (or somewhere other than the output). If you do it at the output then you completely ruin the circuit's output impedance, which means you lose the ability to shunt any hum picked up on the cable.
EDIT: Oops, didn't see there was a 2nd page to this thread, looks like I'm preaching to the converted!
Quote from: oliphaunt on April 08, 2010, 10:53:55 PM
Thanks guys!
Would this fix it? Is 1K a good value to keep the output impedance as low as possible?
(http://www.moosetone.com/images/schematics/boofer03.jpg)
If you remove the switch won't the pot take you between unity (buffer) and a gain of 6? I'm working on something similar and it works on the breadboard.
Found the Plate to Plate article
http://web.archive.org/web/20091021025317/http://geocities.com/SunsetStrip/Studio/2987/fatts.html
Does anyone remember when that was first posted?
> won't the pot take you between unity (buffer)
The diagram you quote: He's back to two 47K resistors. Pot full-down gives gain of 2.
The added 47K (at the pot) was an early attempt to avoid oscillation when feeding loooong lines with huge capacitance. 100+ ohms at the output jack is a better fix for that. Now he can eliminate (short) the 47K and let the pot dial-down to unity.