- Build Rating
Here is my "Dynamic Haircut", intended to be a clone of the Barber Gain Changer.
The PCB is my own design, created from the service schematic Barber has on the product page. I bought an actual Gain Changer a while ago and found I really like it. The only "problem" is that I didn't build it myself. Easily remedied!
I made a lot of errors along the way. When I did a build of the first PCB I made, my first mistake was drilling the holes for the switches in the wrong place. Easily remedied, and only an aesthetic problem. When I powered it up to test, I realized I had wired all three pots backwards. All three! I expected some silly mistakes like that, so wasn't too frustrated, easily remedied in the next board revision.
But the biggest problem was that it sounded like total garbage. Not even a bit off, or a "fun" type of distortion, just extremely harsh, unpleasant noise that screams "something is terribly wrong!" My first step was to compare opamp voltages against the real GC I have. Voltages were just about spot-on for the first gain (clipping) stage, but a bit off for the second half. Not crazy off, but enough that it didn't feel right. Then I used an audio probe, things sounded more or less reasonable all the way up to the input of the second opamp.
So, two pieces of evidence pointing at the second opamp stage. The voltages weren't that far off, I would expect the sound to be "mostly right", maybe be too loud/quite or maybe a bit too distorted... But anyway, I thought, I need to fix those voltages. So I went through and checked all my component values. I triple checked all the resistors, because the gain switch does play some games with VREF. All component values were good!
When all else fails... check the schematic! I compared the schematic I drew in KiCad to Barber's service schematic. Rest assured, I did that before having the boards fabricated, but obviously I didn't do a good enough job! It turns out, I had swapped the inverting and non-inverting inputs on the second opamp! That's a problem that's actually hard to cleanly fix. I was afraid to cut traces on the PCB, in case I did more harm than good. Since my opamp was socketed, what I did was bend the two flip-flopped pins up, then soldered wires to them. This allowed me to use the six correct pins as normal, and I could use the soldered-on wires to cross the bad pins, making the circuit correct. Once I did that, it sounded reasonably close to the real GC.
So I revised my PCB, fixing the pots and the opamp wiring. Build #2 went fairly smoothly. I forgot to solder the electrolytics before putting it in the enclosure, and didn't want to take everything out, so I just soldered "from the top", leaving the leads longer than I would if I had easy access to the back of the board. I re-used the botched case from build #1 (the remains of which are now in my "to be recycled" pile). As I was testing it, I thought the EQ switch didn't seem to match the actual GC too well. I realized another subtle mistake I had made: I swapped the silkscreen labels for two resistors of the switchable EQ sub-circuit! What's interesting though, when I discovered this, I was comparing to the actual GC, and the EQ resistors are different than what is shown on the schematic. The service schematic is dated 2012, and my actual GC is dated 2020. Since my build and the real one sound very similar, I'm guessing that the topology has remained mostly the same, but maybe individual component values have been tweaked over time.
I built up a #3 to put into a nice, properly drilled enclosure with a waterslide decal. I used Tayda's "winked silver" color, which I think looks pretty sharp. I also used a PedalPCB Basic Relay Bypass for a "premium" feel. As I mentioned in my Cattle Driver build report, I want to make another board revision that incorporates the relay bypass onto the actual effect PCB. And maybe a 9v/18v toggle switch, if I have room! (I haven't actually tried the GC at 18v, but some report success with that.)
Anyway, pics below are of prototype build #2 and final build #3. If anyone is interested in build #2, PM me, if you cover shipping, you can have it.
The PCB is my own design, created from the service schematic Barber has on the product page. I bought an actual Gain Changer a while ago and found I really like it. The only "problem" is that I didn't build it myself. Easily remedied!
I made a lot of errors along the way. When I did a build of the first PCB I made, my first mistake was drilling the holes for the switches in the wrong place. Easily remedied, and only an aesthetic problem. When I powered it up to test, I realized I had wired all three pots backwards. All three! I expected some silly mistakes like that, so wasn't too frustrated, easily remedied in the next board revision.
But the biggest problem was that it sounded like total garbage. Not even a bit off, or a "fun" type of distortion, just extremely harsh, unpleasant noise that screams "something is terribly wrong!" My first step was to compare opamp voltages against the real GC I have. Voltages were just about spot-on for the first gain (clipping) stage, but a bit off for the second half. Not crazy off, but enough that it didn't feel right. Then I used an audio probe, things sounded more or less reasonable all the way up to the input of the second opamp.
So, two pieces of evidence pointing at the second opamp stage. The voltages weren't that far off, I would expect the sound to be "mostly right", maybe be too loud/quite or maybe a bit too distorted... But anyway, I thought, I need to fix those voltages. So I went through and checked all my component values. I triple checked all the resistors, because the gain switch does play some games with VREF. All component values were good!
When all else fails... check the schematic! I compared the schematic I drew in KiCad to Barber's service schematic. Rest assured, I did that before having the boards fabricated, but obviously I didn't do a good enough job! It turns out, I had swapped the inverting and non-inverting inputs on the second opamp! That's a problem that's actually hard to cleanly fix. I was afraid to cut traces on the PCB, in case I did more harm than good. Since my opamp was socketed, what I did was bend the two flip-flopped pins up, then soldered wires to them. This allowed me to use the six correct pins as normal, and I could use the soldered-on wires to cross the bad pins, making the circuit correct. Once I did that, it sounded reasonably close to the real GC.
So I revised my PCB, fixing the pots and the opamp wiring. Build #2 went fairly smoothly. I forgot to solder the electrolytics before putting it in the enclosure, and didn't want to take everything out, so I just soldered "from the top", leaving the leads longer than I would if I had easy access to the back of the board. I re-used the botched case from build #1 (the remains of which are now in my "to be recycled" pile). As I was testing it, I thought the EQ switch didn't seem to match the actual GC too well. I realized another subtle mistake I had made: I swapped the silkscreen labels for two resistors of the switchable EQ sub-circuit! What's interesting though, when I discovered this, I was comparing to the actual GC, and the EQ resistors are different than what is shown on the schematic. The service schematic is dated 2012, and my actual GC is dated 2020. Since my build and the real one sound very similar, I'm guessing that the topology has remained mostly the same, but maybe individual component values have been tweaked over time.
I built up a #3 to put into a nice, properly drilled enclosure with a waterslide decal. I used Tayda's "winked silver" color, which I think looks pretty sharp. I also used a PedalPCB Basic Relay Bypass for a "premium" feel. As I mentioned in my Cattle Driver build report, I want to make another board revision that incorporates the relay bypass onto the actual effect PCB. And maybe a 9v/18v toggle switch, if I have room! (I haven't actually tried the GC at 18v, but some report success with that.)
Anyway, pics below are of prototype build #2 and final build #3. If anyone is interested in build #2, PM me, if you cover shipping, you can have it.
