Build Your Own Clone Message Board

The para compressor comes with some extra resistors and 10uf caps that are not mentioned in the instructions. However, the high resolution schematic from the schematic download page shows the extra parts as adding bandpass EQ signals to both the LF and HF compression sections. Looking at the parts layout picture posted in the compressor parts layout thread, it looks like only one of the 10uf caps is to be installed. I am referring to C102, C101, R102, and R101.

So, does anybody know the story on this? It looks like the state variable bandpass output signal was missing from the original PCB, and perhaps there is an option to route it to one or both of the LF/HF compressor sections?

Looking at the circuit some more and testing with a signal generator and scope, a couple of interesting things turned up: (1) It tends to ring up around 400 kHz, which I fixed with some power supply bypass caps on each of the opamps, and with a few little tiny 10pf feedback caps on a few of the opamp stages (also added extra tantalum and ceramic caps to clean up 4.5v reference voltage, since the 33uf electrolytic it came with wouldn't do much good up at 400kHz where the former instability was at -- due to parasitic inductance of that type 33uf cap), and (2) the band-pass filter output is inverted, so adding it back in (as shown on the updated schematic with C101 and C102) to the LP and HP signals causes a scooped tone effect, rather than filling in the previously missing mid-band frequencies. Possibly, an inverter stage could put the band-pass output back in phase with the LP and HP signals before adding them all together, but don't know if it would sound any better, or maybe it might not sound not as good as the way it is now.

Also, the way the signals are summed using IC3a and IC3b together is kind of strange. The same thing could be done with one single opamp stage. That might leave over a spare opamp that could be used to invert the band-pass signal, if desired.

Will think about that some more later. The main reason for cleaning up the power supply and HF stability was in order to allow trying some newer, less noisy opamps in the circuit. The compressor is a bit noisy at some settings, so now ready to give it a try with some quieter opamps to see if it makes any improvement.

Don't know if its worth the bother of trying to put the mid-band frequencies in phase or not. Maybe something for another day.

After completing the build and trying it out for awhile on bass guitar, I find there is a problem. The inverted phase bandpass signal mixed in with the HP and LP signals causes bass notes around the cross-over frequency to sound very thin. Not good for bass, where it is common to play only one note at a time. Not good to have some notes sound thin and higher and lower notes to sound full. Maybe the BP signal was added for acoustic guitar use where the scooped frequency response is more complementary to the instrument?

Not knowing anything about the history of how the circuit came to be the way it is, is not helpful. Now I'm not sure what to do to make the compressor more usable for bass guitar. Maybe disable the BP signals? It would sure help if I didn't have to start from scratch to redesign this thing. Some history on it would be very much appreciated. Makes for a much better starting point to make it better for bass.

Also, with respect to the BP signal getting mixed into the HP and LP compressor circuits, please recall that the schematic in the assembly instructions is wrong. There is a link in the instructions to another version of the schematic which is the correct one: http://www.byocelectronics.com/pmbcschematic.pdf This version of the schematic correctly shows the BP signal as it is now used, and as it is now causing thin sounding bass notes if their frequency is near the cross-over frequency.

Thanks for any info on why the circuit is the way it is, or how it came to be that way.

Mark.

Since no reply to my questions, I decided to disconnect the band bandpass signals and see what the original version of the multi-band compressor sounded like. I see it is much more sensitive to various control settings, and it also seems like the 12 db/octave roll off of a state variable filter may not be enough to separate the HF and LF bands very well. Adding the bandpass notch back in seems to make the HF and LF sound more distinct from each other and smooth out overall control operation, at a cost, as previously noted, of making notes around the crossover frequency rather thin sounding. Oh well, maybe that's why some bass multi-band compressors use a fixed crossover frequency -- makes is more practical to use sharper filters, maybe 18 or 24 db/octave?

Anyway, still trying to sort out if there is some simple, yet better way to make the work more to my liking for bass.

After some more experimentation, I decided that for bass I will run the multi-band compressor at 18 volts, with the bandpass signals disconnected. If they are connected, there is very little compression of mid-band notes, since there is a dip in the frequency response there making the volume too low to activate the compressor circuitry more or less to the same degree as very high or very low notes.

With the bandbass disconnected, control sensitivity is high. Bass seems to sound best with all controls pretty near the 12 o'clock position. Small adjustments one way or the other, or reducing bass volume pot a little can fine tune the sound.

To see if this pedal actually makes bass sound better, it is necessary to adjust the output level controls so that the apparent loudness is the same with the pedal switched in or bypassed. Otherwise, louder always sounds better. I think with the apparent volume dialed in to be equal, and with the compression and frequency controls pretty near 12 o'clock, the pedal does make bass sound better. So, I consider the purchase and build a success.

R202 was overlooked when drawing the working schematic. The CAD was done from that schematic. It wasn't till after the production PCBs had arrived that we realized there was a lot of midrange missing from the signal.

_________________
*patience is a virtue*

Please do not PM me. email is prefered. keith@buildyourownclone.com