drew.spriggs
Well-known member
Howdy,
Thought I'd document some of a build I'm doing here in case anybody is interested in how you approach making custom boards/etc for an amp.
I've got a stack of submini tubes, and a guy has been harassing me about making him a mini Trainwreck-inspired head for a while. Decided to see what I could do using as many existing parts as I've got at home.
This is the schematic I'm working from - using a Wreck preamp layout, but instead of ignoring that spare triode I'm using it as a follower. Unsure why anybody wouldn't as it's free gain! I was tossing up an idea for something like switchable mids instead of that mids pot but settled on a 'deep' switch which gets rid of the scoop entirely. It might be a little too much body but I can tinker once it's built.
The cathode/plate values are very different from what people are probably used to seeing from full size tubes, but somebody else had done some pretty extensive testing and discovered 220k/3k9 gives the 6112's a very similar response to AX7's with 100k/1k5, and these can be scaled directly to emulate any other cathode resistor values - if you see most peoples submini builds they're just using AX7 values and calling it a day!
Rest of the schematic is fairly self-explanatory, but C7's value should be 10n and not 2n2. R14 is there as it made the layout simpler, but it could be replaced with a wire link if it needed to be.
For the power tube I was originally going to use a PSEP but couldn't get a load line that worked well with the transformers I have. Going back to the single, I ended up with something that looks pretty reasonable:
Power was the thing that was confusing for a bit. I do have those cheap inverter modules I've previously posted about, and napkin math puts them at around a 200v output with a 6-7w draw on a 9v supply - but the preamp wants to see about 170v and the power amp wants to see about 120v. Decided to go with a split-rail approach, and with only 3 tubes I could stick with a 317 for heaters (using a thermistor to manage inrush).
Went from this to a basic board layout, and then Kicad shit the bed - it refused to export half the tracks or pads, and after much messing around I discovered that I'd found 3 separate .step export bugs! The eventual solution was to export the board outlines as a .pdf then manually make them in CAD. Even then there was a stack of issues (ie. the amp board had 40k segments!), but I finally got something that worked. Boards were from from standard 1.6mm copper clad FR4 and routed with a 1mm endmill using 1mm of isolation which was an issue - doing it again I'd increase isolation to at least 1.2mm as there's a stack of issues that pop up, but I only had to manually finish a handful of incomplete routes. Didn't get many photos of the boards at this stage as I was keen to start populating, then discovered I couldn't find a stack of parts I had ordered.
A quirk of these inverter modules is without any load at all, the output swings way above what you'd expect - with a 12v supply, you're getting close to 400v out! On the last lot of modules I've done a couple of mods (removing the AC output and adding onboard bleeder resistors), however I must have grabbed a module without the bleeders. Even without all the droppers populated, the first lot of caps were a spicy 370v which I discovered after testing the heaters. Good thing it was only with one hand!
I'm just waiting on the last few components to turn up, and will start working on the head cabinet when I've got 5 minutes spare.
Thought I'd document some of a build I'm doing here in case anybody is interested in how you approach making custom boards/etc for an amp.
I've got a stack of submini tubes, and a guy has been harassing me about making him a mini Trainwreck-inspired head for a while. Decided to see what I could do using as many existing parts as I've got at home.
This is the schematic I'm working from - using a Wreck preamp layout, but instead of ignoring that spare triode I'm using it as a follower. Unsure why anybody wouldn't as it's free gain! I was tossing up an idea for something like switchable mids instead of that mids pot but settled on a 'deep' switch which gets rid of the scoop entirely. It might be a little too much body but I can tinker once it's built.
The cathode/plate values are very different from what people are probably used to seeing from full size tubes, but somebody else had done some pretty extensive testing and discovered 220k/3k9 gives the 6112's a very similar response to AX7's with 100k/1k5, and these can be scaled directly to emulate any other cathode resistor values - if you see most peoples submini builds they're just using AX7 values and calling it a day!
Rest of the schematic is fairly self-explanatory, but C7's value should be 10n and not 2n2. R14 is there as it made the layout simpler, but it could be replaced with a wire link if it needed to be.
For the power tube I was originally going to use a PSEP but couldn't get a load line that worked well with the transformers I have. Going back to the single, I ended up with something that looks pretty reasonable:
Power was the thing that was confusing for a bit. I do have those cheap inverter modules I've previously posted about, and napkin math puts them at around a 200v output with a 6-7w draw on a 9v supply - but the preamp wants to see about 170v and the power amp wants to see about 120v. Decided to go with a split-rail approach, and with only 3 tubes I could stick with a 317 for heaters (using a thermistor to manage inrush).
Went from this to a basic board layout, and then Kicad shit the bed - it refused to export half the tracks or pads, and after much messing around I discovered that I'd found 3 separate .step export bugs! The eventual solution was to export the board outlines as a .pdf then manually make them in CAD. Even then there was a stack of issues (ie. the amp board had 40k segments!), but I finally got something that worked. Boards were from from standard 1.6mm copper clad FR4 and routed with a 1mm endmill using 1mm of isolation which was an issue - doing it again I'd increase isolation to at least 1.2mm as there's a stack of issues that pop up, but I only had to manually finish a handful of incomplete routes. Didn't get many photos of the boards at this stage as I was keen to start populating, then discovered I couldn't find a stack of parts I had ordered.
A quirk of these inverter modules is without any load at all, the output swings way above what you'd expect - with a 12v supply, you're getting close to 400v out! On the last lot of modules I've done a couple of mods (removing the AC output and adding onboard bleeder resistors), however I must have grabbed a module without the bleeders. Even without all the droppers populated, the first lot of caps were a spicy 370v which I discovered after testing the heaters. Good thing it was only with one hand!
I'm just waiting on the last few components to turn up, and will start working on the head cabinet when I've got 5 minutes spare.