To all Ross/Dyna compressor builders

[shredgd]

Hi,

I have just finished to build a Ross/Dyna compressor (I used Tonepad's layout) and I have a question for you.

Being a compressor it obviously can be noisy (I used all metal film resistors to reduce it), but if I crank the sustain or the level pot in mine not only do I hear a hiss, but I also hear a "crackling" noise floor (which reminds me the slightly leaky germanium transistor in my treble booster): now, I must confess I accidentally inverted the power supply connections in mine for about 30 seconds. There is a protection diode in the circuit, which should have lowered the inverted power supply to about 1v (I did a test separately for this).
In the end my question is: do you think the noise I hear is the consequence of a damage to the IC and trannies because of my mistake, or yours make this particular kind of noise as well, when cranked?
I'm asking because I must decide if I have to order another set of trannies and IC or not.

Thank you,

Giulio

[Primus]

My Ross is mousey quiet. I suggest first looking fo cold solder joints. Also, did you socket your IC when you installed it?

[hairyandy]

I'm thinking that your compressor is fine.  My Ross comp clone is noisy as well when the sustain knob is cranked up pretty high.  Set to standard levels (9 o'clock to 2 o'clock or so) it's not bad at all.  If your transistors or IC were bad I think that it wouldn't work at all.  These things are pretty good at amplifying any noise so make sure you have good cables, put it near the front of your pedal chain and try to stay away from fluorescent lights!

[Primus]

All good advice. I also socketed the IC and tried a few different parts to find one that was very low noise. Are you using the reverse log pot for effect level?

[analogmike]

I think you have a noisy transistor. We test them for noise before installing in a pedal. You can put a socket in for Q1 and test them there. Good luck!

[Ed G.]

Before I got into DIY, I sent my dynacomp to Mike to be modded to Ross specs. The comp was also noisy, a crackling sound like you describe. After modding it, the noise was still there, so Mike went through it to see what was causing it. It turned out there was a capacitor making that noise, if I remember right. If I had to guess, an electrolytic. He replaced it, and the noise was gone.

[shredgd]

Thank you guys for the fast answers. Yes, I socketed all the semiconductors, as usual. I'll try swapping each transistor with a couple other transistors I have (gotta check the pinout) and see what happens, I only hope I didn't damage the OTA! Then I'll let you know.

[shredgd]

I swapped the transistor in Q1 and put a 2n2222 I have for another project and "ta-dan" most of the crackling background noise went away! Then I put the other 2n2222 I have in Q2 and the crackling was totally gone, leaving just that bit of physiological hiss which is inerent to the nature of a compressor.
In the end the temporarily inverted power supply, probably because of the protection diode, didn't cause any damage; instead the three 2n3904 I had before buying the other parts to build the pedal where ALL noisy, as also was another one I had in another pedal (I tried each one at Q1 as a test), bought at the same time as the other three. I already sent an email to tell the store I bought them from, to suggest them to change their supplier!

As I need the two 2n2222 for another project, I had to find a solution: so I looked at the schematic carefully and realized that only Q1 and Q2 are really in the signal path: so I put there the two new 2n3904 which previously were in Q3 and Q4, and now the noisy trannies can't let their noise get into my sound!
Obviously, as 99% of the people who write in this forum, I'm SICK about these things, so next time I'll make an order for parts I'll buy some new trannies to replace the noisy ones...

In conclusion, the Ross/Dyna compressor has revealed a very good device to test noisy transistors!!

Also, SORRY for starting this topic when I simply could think about swapping the trannies! I was just too worried I had messed up the OTA with my mistake!

Ciao!

Giulio

[Mark Hammer]

There are 5 discrete transistors in the Dynacomp/Ross.  Only two of them are actually in the audio path: the one immediately before the CA3080, and the one immediately after.  The other 3 are only involved in supplying the control signal that is used to adjust gain in the 3080.  Consequently, you need only be concerned about the noise characteristics of those two.  The other 3 matter, just not in the same way.

Although those other 3 do not directly impact on the audio quality, they do so indirectly.  How?  Take a look at the Tonepad layout ( http://www.tonepad.com/getFile.asp?id=9 ).  Q3 and Q4 form a full-wave rectifier that provides a signal coresponding to the present audio signal level coming out of the CA3080.  The smoothest envelope/rectified control signal from this pair will occur when they are reasonably well matched (along with their accompanying diode, resistor and cap) so that the two provide precisely complementary signals where they sum at the junction of the 10uf cap and the 150k resistor.

Think of it like an octave-up fuzz using a phase-splitter (Green Ringer, Tone Machine, etc.).  The more closely matched the two half signals are, the more you get obvious doubling because the doubled frequency is more apparent, relative to the fundamental.  In this instance, getting "perfect doubling" means the envelope ripple is less audible as the note decays.  More envelope ripple will yield a faint "trill" to the note as it decays.  The trill is really microadjustments of the gain of the 3080.  If you can feed it a more perfect FWR signal, you can keep that to a blessed minimum.  Again, note that the matching only pertains to those 2 transistors.  They don't have to be matched to anything else other than each other.  Insomuch as the achievement of perfectly complementary rectified signals depends on identical output levekls from Q3 and Q4, the voltage drops of the diodes should be similar, and the use of 1% 1M resistors may be smart too.