Boss GE-7 HiFi mod for you :D

[Dai H.]

thx R.G.

I'll add another one:

never add extra power supply bypass caps to a 488 mkII Portastudio even though there is a space for them on the mixer PCB because if you do, you'll blow out your headphones from excessive low frequency response... 

which I guess goes back to...

There is no real substitute for knowing the circuit and doing the math.

[g3rmanium]

When doing mods like this, you need to consider at least 3 specs:  noise performance, current draw and input impedance.  Here are some comparisons for the TL022 vs. the 5532.

Noise Performance (at 1kHz)
NE5532 = 5 nV/√Hz
TL022 = 50 nV/√Hz

Current Use (per amplifier)
NE5532 = 8.000ma
TL022 = 0.130ma

Interesting. I have a DOD 201 that also uses the TL022. I bought a couple of AD820s and TL072s a while ago. I guess the AD820 would work better?

[MartyMart]

.  Even the TL072 would be a decent sub:  TL072 = 18 nV/?Hz and 1.4ma per amp.

Just about anything is better than the 50nV noise on the TL022 but most dual chips will draw more current since the 022 was designed for low current consumption...

regards, Jack

Thanks Jack, I realized this, though I use PSU's 99% of the time, so I went with TL072CP's in mine
and a couple that I've done.
For those who do use batteries, that NE5532 power consumption is a no-no !!
Quiet as heck though
Regards,

Marty.

[rocket]

Just a correction to Jack / amz - the TL022 is NOT a FET opamp - it's totally bipolar.

[amz-fx]

Yes, even though the TL022 has fets in it, the chip is really bipolar input...   TI claims "High Input Impedance" on the datasheet but does not specify a value so the comments about the 022 vs. 5532 input Z should still be relevant.

re: DOD 201   The AD820 draws slightly less current and has slightly less noise but the TL072 is cheaper and easier to find. Either would do the job well.

regards, Jack

[Dragonfly]

I'm not a GE-7 type of guy, but good work. I'm sure LOTS of people will find this useful.

AC

[MartyMart]

Hell yeah, this makes a HUGE difference to the unit, it's NOT a subtle "hi-fi"
bullshit type of mod  ! 
You hear it right off the bat ....

MM.

[g3rmanium]

Interesting. I have a DOD 201 that also uses the TL022. I bought a couple of AD820s and TL072s a while ago. I guess the AD820 would work better?

Of course I mean the AD823 since the AD820 is just a single opamp.

[GibsonGM]

I picked up one of these on the cheap, it should be coming UPS and I'll be checking for noise.  I'd go in and do just the opamps, do ya suppose that'll make a noticeable difference?  I dunno about changing the caps; might have to if the switch to '072's isn't enough, tho...

[thehallofshields]

Tone caps:
Remove C10 ( 1u5 tantalum ) and replace with a 1uf poly cap ( closest value but may effect the 100Hz band a little, pushing it up to 150hz ?)
Remove C9 ( 0.68uf tantalum) and replace with a 0.68uf poly cap
Remove C8 (0.33uf tantalum) and replace with a 0.33uf poly cap
Remove C7 (0.15uf tantalum) and replace with a 0.15uf poly cap
Remove C6 (0.1uf tantalum) and replace with a 0.1uf poly cap

Sorry to bring this back from the dead. But if someone were to use more 'conventional' cap sizes such as:

C10 1uf
C9  .68
C8  .47
C7  .22
C6  .1

To what extent and how would the frequency bands get f'd up?

[ayayay!]

It's only conventional if that's the only thing you have on your bench.   

More than one thing would happen.  For those two, they're larger than the stock values, so they'd start moving into the other bands territory.  Not fun.  The other thing is you'd need to compensate some resistors as well so you don't end up with a nasty bit of boost on those two bands.  I speak from experience here. 

Seems you have a .1 and .22.  So wire those two in parrallel to make a .32.  (Close enough for our purposes.)

Same with the .15.  Use your .1 and a .047.  Even the Rat Shack has those.  

[earthtonesaudio]

Multilayer ceramic makes a good space-saving alternative to poly film (often even smaller than tantalum!), at a moderate price premium.

[thehallofshields]

It's only conventional if that's the only thing you have on your bench.   

You  seem to have interpreted what I was said as: "I don't want to drop any more money on Metal-Film Caps." - You read me correctly.

I'll give the parallel caps a shot. I really appreciate the the quick advice; I'm glad I didn't settle for warping the Frequency bands.

[georgyyy]

I'm thinking of applying this mod to my behringer eq700. it's not very noisy but at lower frequencies it does a noise...
Do you know if the chip is the same, cause google search says that it might be the same, nothing solid...

[beedoola]

I did these mods but something isn't right. I did the mods from the OP. Would slider lube help this? I used Deoxit contact cleaner to clean the sliders but that didn't help. I've gone over all the solder joints and they all check out...I used TL072 chips.

[earthtonesaudio]

http://www.diystompboxes.com/wiki/index.php?title=Debugging

[jgroover]

Noise Performance (at 1kHz)
NE5532 = 5 nV/√Hz
TL022 = 50 nV/√Hz

the RC4559 is a better choice than the 5532 as a sub...  noise is not quite as low but input Z is higher and current use is only about 1.6ma per amplifier.  Even the TL072 would be a decent sub:  TL072 = 18 nV/√Hz and 1.4ma per amp.

regards, Jack

Hello, resurrecting this one again. My GE-7 is certainly noisy, so I'm thinking about swapping the ICs.

Question 1 - I looked up the RC 4559 data sheet and it lists the EI noise as "2uV/Hz max" while the figures above use square root of Hz, so how do I compare them? Is there an answer that doesn't involve trig? 8~)

Can I get the square root symbol on my standard keyboard ?

Question 2 - How would just changing the op-amps compare to changing both the op-amps and caps listed in terms of noise performance?

[blackieNYC]

by all means try the opamps by themselves.  take a very consistent noise listening (or reading) test of some kind so you can compare before and after.  I got noticeably good results and I only did the op amps.  I think I did 4559s?  or were they in there already?  Use what is recommended above somewhere.

[PRR]

Welcome.

> RC 4559 data ..."2uV/Hz max" while the figures above use square root of Hz, so how do I compare them?...

The '4559 data *also* specifies "20Hz-20KHz". It is directly the number over the audio band.

Say TL072 = 18 nV/√Hz. To figure this for 20Hz-20KHz, a 19,980Hz band, we find the square-root of 19,980, 141. 18nV/rtHz times 141 is 2,540nV or 2.54u, a hair higher than '4559.

(The '4559 has significant input *current* and above 1.4uV/25pA= ~~56K source impedance it will hiss more than a TL072.)

> Can I get the square root symbol on my standard keyboard ?

"rtHz" works with anybody who knows noise. If a nearby post has the sign, try copy/paste. There are keyboard shortcuts but they may not work for all readers. There is an "HTML entity" for square root √ but this forum software does not interpret it as a special character.

[Andrekp]

This is what cheeses me off about Boss.  (And I do like Boss)

They have a subset of their pedals, or which the GE-7 is a nice example, which are just plain noisy, and which have long-known and simple mods that make them quiet. Yet Boss won't take the trouble to fix them.  The CS-3 is another one of these.

I tried a GE-7 a year or so ago and immediately returned it it was so noisy. If I pay $99 or a new pedal from a reputable company, I should NOT have to immediately mod it to make it usable. I'd even be willing to pay an extra $20, if I knew that Boss made these mods and fixed the pedal after all these years.

Instead, Boss makes a big deal out of its Waza stuff and charges an extra $100 or those. 

There will probably be a new Waza version of this with a new Waza price eventually, incorporating these very mods.