I'm about to start chassis drilling and fabrication on yet another BF Princeton Reverb build, but with a twist. Or rather a couple of twists and one delicious donut. Or Toroid rather.
After surveying the pile of parts ahead of my impending JM Sig Reverb and Sorado clone builds I decided I needed something simpler to tide me over, get some experience with Toroids, see if the Cap Coupled bias circuits work and also try out KOC's split push pull reverb driver. I also decided to upgrade from pencil and paper to KiCAD and DIYLC. I'm not sure KiCAD is that much of an upgrade, but OMG DIYLC where have you been on the last builds?
Purpose:
-> Pretty good small clean amp for noodling, church gigging and late night playing with great bass. (I play on the neck p/u. 80% of the time).
-> See if a bunch of these nutsy ideas work: Donut PT, DC + LM338 regulated heaters, AC coupled bias supply, split push/pull reverb driver ala KOC (since I had a hammond 125A sitting around).
-> See if my layout ideas work (preamp tubes accessed midships to the preamp board).
I figure a BFPR is a nice reference platform since we know what they should sound like.
I'm still working on the layout in DIYLC with about 80% done. Ok so the schematic is basically a BFPR with a lot of mods in the reverb driver, preamp heater supply and bias supplies and (as you'll see soon) a weird layout. I've also used big Solen fast caps + HEXfreds in both of the previous two builds and really like the sound.
MOST CORRECT FILES HERE. USE THESE FILES FOR AN ACTUAL BUILD.
The Donut Princess Schematic:
Updated 1.4 Sep 25, 2020):
- Corrected reverb driver values
- Adjusted regulator control resistors to give 5.8VDC heaters in real world.
- Added note on Hammond 125A taps to use.
- Had input jack wired tip/sleeve reversed.
Updated 1.3 (Aug 5, 2020):
- Added optional MOSFET buffer
- Final trem values
Updated 1.4 Layout (Sep 25, 2020)
- Error in a couple resistor values
- REverb driver grid input moved to correct node
- added proper screen resistor values.
Updated 1.3 Layout (Aug 5, 2020)
- Began process of matching R/C numbering between schematic and layout.
- Updated/resized power/bias/heater board to reflect actual dimensions.
- Corrected tremoscillator layout, added Mosfet buffer section.
- Moved a couple star grounding points.
This layout is designed to go in a mojotone "custom" 6.5" x 17" chassis which just barely will take a 47uF solen fast cap laid down sideways with a 1/4" standoff for the g4 board.
The Donut Princess Build
Moderators: pompeiisneaks, Colossal
-
mirage_indigo
- Posts: 84
- Joined: Thu Oct 02, 2014 3:34 pm
- Location: Longmont, CO
The Donut Princess Build
You do not have the required permissions to view the files attached to this post.
Last edited by mirage_indigo on Fri Sep 25, 2020 7:56 pm, edited 5 times in total.
We build because we must.
Re: The Donut Princess Build
Thanks for posting your schematic. The reverb driver circuit is interesting. Looks like a cool build!
-
mirage_indigo
- Posts: 84
- Joined: Thu Oct 02, 2014 3:34 pm
- Location: Longmont, CO
Re: The Donut Princess Build
I'll correct some minor errors in the schematic above later (notably cap polarity in the bias supply). Here is the semi-final layout. Note that I added a 1uF/ 630 V cap across the Reverb board power bus since the
ground is pulled from the preamp ground but the HV is pulled from V+C. This should help buffer things a bit. I'll add it to the schematic later.
It's a super tight layout even without the tube rectifier. I put the bias pot on the back panel, mostly because there wasn't room on the bias board. I've tried to keep thelayout colors consistent, but didn't quiote get it all.
I started fabrication last night. I forgot just how much drilling is involved when using a blank chassis. I've identified a couple mechanical fit issues that I need to address, but so far so good.
ground is pulled from the preamp ground but the HV is pulled from V+C. This should help buffer things a bit. I'll add it to the schematic later.
It's a super tight layout even without the tube rectifier. I put the bias pot on the back panel, mostly because there wasn't room on the bias board. I've tried to keep thelayout colors consistent, but didn't quiote get it all.
I started fabrication last night. I forgot just how much drilling is involved when using a blank chassis. I've identified a couple mechanical fit issues that I need to address, but so far so good.
Last edited by mirage_indigo on Fri Sep 25, 2020 7:55 pm, edited 1 time in total.
We build because we must.
-
mirage_indigo
- Posts: 84
- Joined: Thu Oct 02, 2014 3:34 pm
- Location: Longmont, CO
Re: The Donut Princess Build
A couple notes about the mechanical layout:
Preamp Tubes:
The single biggest thing is that the preamp tubes are laid out in the middle of the respective boards. In other words, the 3/4" hole for the 9 pin socket is in the middle of a 1" hole in the board above so the leads between board and socket are basically as short as practical. There's also a hole for the RCA jack Reverb Return that comes in right next to the second preamp tube. The huge advantage of this is it lets you run very very short leads and keeps the chassis a lot tidier (assuming it works). I couldn't figure out how to specify a 1" hole in the board layer in DIYLC, so if you play with the schematic, they show up as chassis cutouts, but they're actually board cutouts. It ends up being a very compact layout (if it works). I used a 3.25" x 9" main board and a 6" x 4" main power board. The filter cap for the preamp stage is actually smack in the middle of the preamp. That's the big black square in the preamp, a 27uF 630V Solen cap.
The huge disadvantage of this approach is you end up with the heater wires running right underneath all your lovely preamp components. This is why I'm using DC heaters everywhere, mostly to make the layout more forgiving wrt hum/buzz. I was going to run the power tube heaters with just the 100 ohm center taps, and not DC, but with an LM338T, you can run up to 5A of output, so you might as well use DC for everything, I not even need to heatsink the package (but I will).
Wiring across the tube:
Since I have pretty short leads, I also employ a couple tricks to avoid some routing issues on the board, namely wiring across the socket to the same node but on the other side. You can see this in the tone stack where the 100k resistor to V+D is on one side and the take off to the top of the tone stack is on the other side. This is another nice advantage of having the tube right in the mix.
Mechanical fit:
My two previous scratch builds, I found by far my biggest problems were mechanical, getting things to fit became the project, so hopefully this will help solve some of those issues. I will likely need to nudge the turret boards back ~1/4" into the amp a bit to get better clearance under the row of front panel pots.
The other thing to note is that the main 47uF solen cap has almost no height clearance in a 2.5" chassis if you use a 1/4" standoffs and 1/8" g4 (which I am). I might need to use a shorter standoffs, but then you hit problems of needing really hard bends coming out of the transformers under the boards and trying to clear screw heads, etc underneath the boards. I also need to be able to clear the tremelo enable switch which sticks out from the front panel right into where the power board wants to sit.
Reverb: I'm using the KOC self split reverb driver from TUT1, the reverb transformer is actually mounted inside the chassis right to the reverb board. I added a 1uF 630V cap right across the power mains on this board since the leads to gnd/power are pretty long.
Heat: So what about heat? This is for a head, so hopefully the heat will mostly be headed the other way.
Photos of progress:
https://photos.app.goo.gl/rukQggeK5MzYSnLe7
Preamp Tubes:
The single biggest thing is that the preamp tubes are laid out in the middle of the respective boards. In other words, the 3/4" hole for the 9 pin socket is in the middle of a 1" hole in the board above so the leads between board and socket are basically as short as practical. There's also a hole for the RCA jack Reverb Return that comes in right next to the second preamp tube. The huge advantage of this is it lets you run very very short leads and keeps the chassis a lot tidier (assuming it works). I couldn't figure out how to specify a 1" hole in the board layer in DIYLC, so if you play with the schematic, they show up as chassis cutouts, but they're actually board cutouts. It ends up being a very compact layout (if it works). I used a 3.25" x 9" main board and a 6" x 4" main power board. The filter cap for the preamp stage is actually smack in the middle of the preamp. That's the big black square in the preamp, a 27uF 630V Solen cap.
The huge disadvantage of this approach is you end up with the heater wires running right underneath all your lovely preamp components. This is why I'm using DC heaters everywhere, mostly to make the layout more forgiving wrt hum/buzz. I was going to run the power tube heaters with just the 100 ohm center taps, and not DC, but with an LM338T, you can run up to 5A of output, so you might as well use DC for everything, I not even need to heatsink the package (but I will).
Wiring across the tube:
Since I have pretty short leads, I also employ a couple tricks to avoid some routing issues on the board, namely wiring across the socket to the same node but on the other side. You can see this in the tone stack where the 100k resistor to V+D is on one side and the take off to the top of the tone stack is on the other side. This is another nice advantage of having the tube right in the mix.
Mechanical fit:
My two previous scratch builds, I found by far my biggest problems were mechanical, getting things to fit became the project, so hopefully this will help solve some of those issues. I will likely need to nudge the turret boards back ~1/4" into the amp a bit to get better clearance under the row of front panel pots.
The other thing to note is that the main 47uF solen cap has almost no height clearance in a 2.5" chassis if you use a 1/4" standoffs and 1/8" g4 (which I am). I might need to use a shorter standoffs, but then you hit problems of needing really hard bends coming out of the transformers under the boards and trying to clear screw heads, etc underneath the boards. I also need to be able to clear the tremelo enable switch which sticks out from the front panel right into where the power board wants to sit.
Reverb: I'm using the KOC self split reverb driver from TUT1, the reverb transformer is actually mounted inside the chassis right to the reverb board. I added a 1uF 630V cap right across the power mains on this board since the leads to gnd/power are pretty long.
Heat: So what about heat? This is for a head, so hopefully the heat will mostly be headed the other way.
Photos of progress:
https://photos.app.goo.gl/rukQggeK5MzYSnLe7
We build because we must.
-
mirage_indigo
- Posts: 84
- Joined: Thu Oct 02, 2014 3:34 pm
- Location: Longmont, CO
Re: The Donut Princess Build
I've gotten most of the holes drilled in the chassis with the exception of the bias pot and maybe a ground attachment or two. The layout will be tight inside. I will post a preliminary gut shot later today when I get the rough mechanicals all figure out. Here is the outside top view (ignore the plastic wrap still on the chassis).
You do not have the required permissions to view the files attached to this post.
We build because we must.
Re: The Donut Princess Build
I am wondering IF the one triode bias tremolo will be able to work with 6L6's???
IF not, look at Sluckey's Trem-O-nator mod on the Hoffman forum which will work.
https://el34world.com/Forum/index.php?topic=25762.0
https://el34world.com/Forum/index.php?topic=26246.0
With respect, 10thtx
IF not, look at Sluckey's Trem-O-nator mod on the Hoffman forum which will work.
https://el34world.com/Forum/index.php?topic=25762.0
https://el34world.com/Forum/index.php?topic=26246.0
With respect, 10thtx
-
mirage_indigo
- Posts: 84
- Joined: Thu Oct 02, 2014 3:34 pm
- Location: Longmont, CO
Re: The Donut Princess Build
Thanks for you help Steve, I could find values that were changed but not any topological issues. I did draw it out differently than the stock schematic, but it looks correct. I won't be using a footswitch, it'll be a front panel as I find I never really switch the trem on or off in the middle of a song. That's what the univibe is for. It didn't help that KiCAD seemed to have stripped the actual values off some of the parts. The only real difference over stock on the oscillator side is it's powered off the same power node as the PI, mostly since the power point is right there next to the other half of the tube so it made the layout easier.
I'll very likely run 6v6s, although the heater tap will handle 6l6GCs. The main reason to use 6L6 tubes was that was what I had the symbol for and the symbol editor keeps crashing KiCAD on OSX. Either should be fine.
I'm wondering if I should add a 1M safety resistor across the vbias pot wiper and supply in case the wiper lifts. On a stock adjustable bias fender, there's a resistive path back to V- in case the wiper lifts, but the marshal bias I copied doesn't seem to have a pure resistive path to V- that isn't the bias pot wiper. I can add it right at the pot, so it's not a big deal to add later.
V 1.1
- Schematic edited for clarity in spots by dragging around labels.
- filed in Trem part values from stock schematic. Might need to add a mosfet emitter follower buffer.
- Added 1uF 630V filter cap across V+B to GND on the reverb daughter board.
- Changed Intensity back to 250kB.
- Updated Tone stack to more correctly reflect how the stock fender tone stack is actually wired. (ie: bass pot is just lugs 2 and 3, not really 1 tied to 2 and 3).
I'll very likely run 6v6s, although the heater tap will handle 6l6GCs. The main reason to use 6L6 tubes was that was what I had the symbol for and the symbol editor keeps crashing KiCAD on OSX. Either should be fine.
I'm wondering if I should add a 1M safety resistor across the vbias pot wiper and supply in case the wiper lifts. On a stock adjustable bias fender, there's a resistive path back to V- in case the wiper lifts, but the marshal bias I copied doesn't seem to have a pure resistive path to V- that isn't the bias pot wiper. I can add it right at the pot, so it's not a big deal to add later.
V 1.1
- Schematic edited for clarity in spots by dragging around labels.
- filed in Trem part values from stock schematic. Might need to add a mosfet emitter follower buffer.
- Added 1uF 630V filter cap across V+B to GND on the reverb daughter board.
- Changed Intensity back to 250kB.
- Updated Tone stack to more correctly reflect how the stock fender tone stack is actually wired. (ie: bass pot is just lugs 2 and 3, not really 1 tied to 2 and 3).
Last edited by mirage_indigo on Fri Sep 25, 2020 7:55 pm, edited 1 time in total.
We build because we must.
-
mirage_indigo
- Posts: 84
- Joined: Thu Oct 02, 2014 3:34 pm
- Location: Longmont, CO
Re: The Donut Princess Build
Updated Layout:
- Moved Reverb daughter board power to screen node, moved ground to main power amp ground.
- Updated some rulers for actual as built dimensions.
- Updated colors of transformer leads to be correct w.r.t. actual transformers.
- Moved some of the power board nodes around to reflect what got actually laid out in chassis.
- Moved Reverb daughter board power to screen node, moved ground to main power amp ground.
- Updated some rulers for actual as built dimensions.
- Updated colors of transformer leads to be correct w.r.t. actual transformers.
- Moved some of the power board nodes around to reflect what got actually laid out in chassis.
Last edited by mirage_indigo on Fri Sep 25, 2020 7:55 pm, edited 1 time in total.
We build because we must.
Re: The Donut Princess Build
I think you'll like it. The Solens sound much cleaner and no regular replacement. I use them in all builds. The push pull drive on reverb is the only circuit I've used for reverb in the last 10 years. It sounds so clear and 3D. Of course some of that may be a result of the return being sent to the P.I. non-inverting input, instead of being run with the main signal through the other circuitry. The 125 is serious overkill, I used one as that's what I had. I later fitted a more reasonably sized xformer and immediately re-installed the 125 as it sounded better.
Tube junkie that aspires to become a tri-state bidirectional buss driver.
Re: The Donut Princess Build
Look againmirage_indigo wrote: ↑Tue Jul 28, 2020 4:00 pm Thanks for you help Steve, I could find values that were changed but not any topological issues. I did draw it out differently than the stock schematic, but it looks correct.
You do not have the required permissions to view the files attached to this post.
-
mirage_indigo
- Posts: 84
- Joined: Thu Oct 02, 2014 3:34 pm
- Location: Longmont, CO
Re: The Donut Princess Build
Wow that's emberassing. The dreaded off-by-one error. In retrospect I should have realized the grid of the trem triode was floating. PEBKAC
Schematic v 1.2:
- Fixed trem circuit to not be grossly incorrect
- Updated some more annotations and dragged around some more labels. I did the rough mechanical fit for the chassis last night. Funnily on this build, I've spent way more time than previous builds thinking about the mechanicals than my last two builds. So far I think the main issues are:
- I have some clearance issues from the main filter cap (you can see in the photo it's sticking out above the chassis top). I can fix them with some lower standoffs.
- I was going to mount the reverb chassis xformer inside the chassis. I thought I would mount it on the board, but the mounting would cover up the reverb out RCA jack, so I think I'll mount it on the back wall.
- I'm still trying to figure out the way to mount the toroid. I had the wires coming out on top, but I think it was designed to mount the other way. Does anyone with Antek donut experience have any ideas hwo these are supposed to be mounted? Based on the docs on Antek's site, oh wait, there aren't any.
I think the tube socket amidship to the amp board is going to work just fine. I didn't quite center the holes correctly, but it's close enough to work.
Also huge shout out for the Digikey Assorted machine screw variety pack. Once more having 20 of every size and length #2 -> #10 machine screw is super handy.
You can see the filter cap too high for the chassis. I had to drop the filter cap board by about 1/8". Starting to look like an amplifier.
Schematic v 1.2:
- Fixed trem circuit to not be grossly incorrect
- Updated some more annotations and dragged around some more labels. I did the rough mechanical fit for the chassis last night. Funnily on this build, I've spent way more time than previous builds thinking about the mechanicals than my last two builds. So far I think the main issues are:
- I have some clearance issues from the main filter cap (you can see in the photo it's sticking out above the chassis top). I can fix them with some lower standoffs.
- I was going to mount the reverb chassis xformer inside the chassis. I thought I would mount it on the board, but the mounting would cover up the reverb out RCA jack, so I think I'll mount it on the back wall.
- I'm still trying to figure out the way to mount the toroid. I had the wires coming out on top, but I think it was designed to mount the other way. Does anyone with Antek donut experience have any ideas hwo these are supposed to be mounted? Based on the docs on Antek's site, oh wait, there aren't any.
I think the tube socket amidship to the amp board is going to work just fine. I didn't quite center the holes correctly, but it's close enough to work.
Also huge shout out for the Digikey Assorted machine screw variety pack. Once more having 20 of every size and length #2 -> #10 machine screw is super handy.
You can see the filter cap too high for the chassis. I had to drop the filter cap board by about 1/8". Starting to look like an amplifier.
You do not have the required permissions to view the files attached to this post.
We build because we must.
Re: The Donut Princess Build
Tremolo oscillator looks right now. However, I think you will not be satisfied with the effect. It's hard to wiggle the bias enough to get a nice effect on 6L6s. And using a single triode is gonna be even harder. The bias level has a big influence on the tremolo effect. You will probably spend a lot of time trying to compromise between your desired bias point and a decent tremolo effect. An easy way to greatly increase the tremolo signal is to replace the cathode resistor/cap with a red or yellow LED. Decreasing the value of R27 will also give an increase in tremolo signal, but nothing like the LED. As a side, the LED can be mounted on the front panel for eye candy.
-
mirage_indigo
- Posts: 84
- Joined: Thu Oct 02, 2014 3:34 pm
- Location: Longmont, CO
Re: The Donut Princess Build
I'd been studying another princeton clone I also found around here by Tubeswell (Princestoner Reverb) and I'm debating maybe also adding a mosfet buffer. I've already got at least one TO-220 part in there and I have the mosfets and a 12V zener in stock already. I updated the schematic.
Updated Layout:
- Fixed the Trem circuit to match stock.
- Added lugs for an optional mosfet buffer using IRF820 or similar MOSFET (I'm using the mighty 2nk100z).
Updated Layout:
- Fixed the Trem circuit to match stock.
- Added lugs for an optional mosfet buffer using IRF820 or similar MOSFET (I'm using the mighty 2nk100z).
You do not have the required permissions to view the files attached to this post.
We build because we must.
-
mirage_indigo
- Posts: 84
- Joined: Thu Oct 02, 2014 3:34 pm
- Location: Longmont, CO
Re: The Donut Princess Build
I've updated the initial post with the latest corrections. I started fabing the board last night. I'd forgotten the drilling takes all of like 5 minutes after weeks of planning. Swaging is another 5.
We build because we must.