Orange Squeezer...what exactly is it supposed to do?

[graylensman]

Just built an Orange Squeezer - the Tonepad version 2. I didn't lay it out on breadboard but went straight to PCB. It works, I guess - I get signal through the circuit, and can increase and decrease volume. But I'm not sure I'm hearing any change in my guitar tone. I turned the trim pot to extremes and still...maybe...there's a subtle difference? Before I launch into major troubleshooting, I just want to know - how will this circuit color my sound?

[MrStab]

i haven't used the Orange Squeezer specifically, but i do know that some of the more subtle compressors can be hard to recognise, especially with psychoacoustics messing with you. You probably know the theory behind dynamic range compression, but here's what i'd do to check if it works:

One possibility would be to repeatedly pick a string, starting out really gently and gradually picking harder and harder with the circuit OFF. With the circuit ON, you should hear less variation in level as you increase your picking intensity - ie. it should be instantly louder (once you're past a threshold). less dynamics. The reverse is also true: pick really loud and it'll be squashed back to roughly the same level.

If you record on computer a lot, another possibility would be to record a riff with the circuit off, then the same riff with it on, and look at the waveform. If the waveform looks flatter/less-varied in the second one, then it works.

If you decide you like your compressors, i highly recommend the Engineer's Thumb. It's one of my best friends. lol

[induction]

To get any compression from the Orange Squeezer, you need to adjust the bias trimpot correctly. The point of the trimpot is to set the jfets until they are just barely turned off. Then large signals will open up the jfet and attenuate signal to ground - voila, compression. Jfets have notoriously wide tolerances, so the trimpot is there to mitigate that. It's not really a compression control or anything, but if you have it set wrong, the circuit will not behave like a compressor.

At one extreme you will get a volume boost (jfets always off, no compression), at the other extreme you get very low volume (jfets wide open, all signal attenuated). The transition between these two states is a very small range on the trimpot. Park the trimpot somewhere in that range and you should hear the compression. It's more noticeable (to me) with fairly clean signals, but once you get an ear for it, you'll hear the effect on distorted signals as well. You can adjust the trimpot within this range to get slightly different amounts of compression, but the OS is a one-trick-pony (it's a really good trick, though), and it really only gives one sound. You don't have a wide range of compression to choose from, so find the sweet spot and keep it there.

There are tricks for making the trimpot easier to dial in (find where the usable range starts and ends, then replace the trimpot with an inline resistor and a smaller value trimpot to fill that range), but it's a set and forget thing, so if you can get it dialed in once you can just leave it alone. Then set the volume trim for unity and you're done.

The result will be a bit more evenness of volume. Volume spikes from variations in playing technique should be noticeably reduced. You will also hear a bit of a click at the beginning of each picked note. The attack time on the stock OS is long enough that the initial pick strike gets through the circuit before the compression kicks on. The click is also more noticeable on clean signals (dirt is naturally compressive, so it tends to remove the pick strike anyway), and can be fairly useful for giving definition to squeaky clean single notes that might otherwise get lost in the mix. 

[graylensman]

Thanks for the responses!

Is there a good starting point with the trimpot - i.e. put a meter on it, and dial in to a particular reading, and adjust from there? Or is it simply go by hearing?

MrStab, I'll give your approach a try. Maybe with a friend to pick while I twist.

[induction]

Is there a good starting point with the trimpot - i.e. put a meter on it, and dial in to a particular reading, and adjust from there? Or is it simply go by hearing?

It's probably best to go by hearing. There is no particular voltage or resistance reading that will work for all instances of this circuit. Remember the trimpot is only there to make up for variances in jfets. If all jfets had the same characteristics, you could just use a fixed resistor, no trimpot required.

Fortunately, you don't need golden ears to dial this in. The difference between clean boost on one extreme of the trimpot and little or no volume on the other extreme is very obvious, and the tiny band of trimpot settings between the two extremes is very easy to find.

There are other, more clever ways to set the bias with a DMM. I haven't tried this, but considering the source, it probably works pretty well. Nonetheless, once you know what you're looking for, bias-wise, it's very easy to set by ear.

Also, there are lots and lots of threads on biasing the OS on this forum. You might browse through them if you get stuck.

[karbomusic]

Thanks for the responses!

Is there a good starting point with the trimpot - i.e. put a meter on it, and dial in to a particular reading, and adjust from there? Or is it simply go by hearing?

MrStab, I'll give your approach a try. Maybe with a friend to pick while I twist.

- IMHO it is much, better to use a 10 or 20 turn trim pot. With a 1 turn pot, there isn't much wiggle room at all.

- Start in the middle of the pot's range

- Use the low E string or open E chord, you should unmistakably be able to hear the lull in the audio and it's recovery (aka compression) once you get into the correct range.

- Once you find that range, back it off some and/or adjust to taste. 

Or you could try...

- Adjusting the trim so that when the volume is 50% you have appx. unity gain which works for some (I prefer the above method by ear).

- Run a square wave through it and into a oscilloscope so you can see the compression action (again I prefer the first method by ear).

[peterg]

Try setting the trim pot to 1.5 to 1.7 vdc. Once set you should hear the compression. Try an auto wah after the OS and you'll really hear it work.

[midwayfair]

I set it by twisting the trim until the sound gets very quiet, and then backing it off a little until it loosens up a little. There will be popcorn noises in the louder settings, you're looking to minimize the noise. If you don't have a spot on your trim where the volume drops a ton, there's something wrong.

The total useful range of the trimpot is probably no more than 700 ohms, with the actual optimal setting likely being within a 100Ohm range on the pot. That's why a 10-turn trim is good here.

As far as what it's doing, the decay is short for a guitar compressor (470mS), and unless your guitar has hot pickups, it can easily sound like not much is happening. One way to test without anything except your guitar in hand is to put your volume at 10, put it in a noisy pickup position, and hit a VERY loud staccato chord. You should hear the pickup noise swell back up over the course of a half second. You're also more likely to hear compression with high output pickups or a booster between the guitar and compressor.

[LightSoundGeometry]

I use mine, and thanks guys because its working just fine now, as an overdrive or stacked with other pedals. sometimes I just turn it down and leave it on , let it do its magic

[graylensman]

OK, well - I am getting no response at all out of the trimpot. I'm not sure if that means it isn't soldered in correctly (I had to finagle it a bit to seat it on the board) or if the component itself is crappy or incorrect. Fortunately, I purchased duplicates of all components, so I guess I'll toss it together on the breadboard - which is what I should have done in the first place. Oh well, lesson learned.

[R.G.]

The Orange Squeezer is an idiosyncratic compressor.

It works by amplifying the signal through that opamp, with a relatively fixed gain of about 23. The output of the opamp goes to the audio output, but is rectified by the germanium diode into a DC signal coupled to the gate of one of the JFETs.

The JFET is set up to attenuate the input signal. When the signal is big, the opamp amplifies it to bigger, and this causes a higher voltage on the gate of the JFET. The JFET's source is held at some bias voltage by the other JFET and resistors.

The higher the voltage on the JFET gate, the lower the resistance of the JFET between drain and source. This causes a voltage divider action that cuts the input signal back down. So the higher the input signal, the less of it actually gets through. The smaller the input signal, the less it's attenuated, and the more it's amplified.  Small signals get more gain, big signals get less - it's a compressor.

The two JFETs do NOT have to be matched. They just have to have characteristics like Vgsoff that let them work within the 9V of the pedal's power supply. The JFET connected directly to +9V is used only as a constant current source to let the trimpot set the bias for the other JFET.

It's worth noting that **some 2N5457s will not work in this circuit**. The tolerance range of Vgsoff for the 2N5457 is big enough that some of them will not be pulled into the "sweet spot" of compressing with any adjustment of the trimmer. It is possible that this is the OP's problem. But it's also a subtle effect, as noted, so the OP may not have noticed.

I really like the 2N5485 or 2SK30A for the orange squeezer. But that's just me.   

[graylensman]

R.G., thanks for that bit of info. I think I'll order some different JFETs to see what happens; I had the foresight to socket the trannies on the board.

FWIW, I breadboarded the circuit, and made a harsh-sounding fuzz.    - now I don't want to give up: I want to make this sucker work.

[rankot]

It's worth noting that **some 2N5457s will not work in this circuit**. The tolerance range of Vgsoff for the 2N5457 is big enough that some of them will not be pulled into the "sweet spot" of compressing with any adjustment of the trimmer. It is possible that this is the OP's problem. But it's also a subtle effect, as noted, so the OP may not have noticed.

RG, I built OS, but have only heavily distorted sound. I've no idea what's going on. Could it be bad/wrong 2N5457? What Idss/Vgs(off) values are needed for OS to work properly?

[PRR]

> have only heavily distorted sound

If it is only distorted sounds, then check your pin voltages at the opamp. It should pass small signals cleanly; many build-errors will foil this.

[Mark Hammer]

I find it a little ironic that one of the most desirable aspects of the OS - its' "transparency" - is precisely what can make it such  a source of exasperation and confusion for some builders.  When it is working well, you have to listen very closely to notice it's on.

More familiar compressors often have a much slower gain recovery time.  The gain is instantly reduced in response to pick attack, and comes back up slowly.  One can often hear this as a progressive increase in hiss, if you pick hard then stop playing for a bit.  The musical role of this slowed gain-recovery is to create an impression of sustain, where the same level is maintained as a note is held.  The string is dying out, but the slowly-increasing gain makes it seem like the string just keeps on going.

As RG so eloquently described, the OS uses a fixed gain, and varies the degree of attenuation of the signal feeding that fixed-gain stage.  The net effect is more or less the same as gain adjustment of a fixed-level input signal.  Where the OS differs from many other units (or at least differed, since others have long since caught up) is in the use of a fast time constant, that functions like an ultra-quick gain-recovery.  In this regard, players notice several things about it:

• it tends to feel more like a peak limiter than a compressor which aims for an overall average level
• you don't get any of the "breathing" (gradual increase in his that sounds like an inhale) that accompanies slow gain-recovery
• you keep more of the treble and pick attack; especially when picking fast

Thanks to Jeorge "Mr. Huge" Tripps, I modded a friend's Dynacomp the other day to have a choice between stock and fast gain-recovery, and the improvement in treble response with fast gain-recovery is very noticeable.

Slower gain-recovery IS very handy for facilitating sustain (moreso at louder volumes), so I don't want to badmouth it.  But what many clean studio pickers always loved about the OS was that the notes were clear, the noise was low, and the peak level was constant: three traits that made it hard to know if you had built and set up the circuit right!