1000uF?!? I don't think I've ever seen a tube amp with that much input capacitance. Isn't that going to put a lot of strain on the power transformer and the diodes? Seems like there will be very very large inrush current. I can't imagine you need that much. Really, unless you construct a fully regulated supply for the output tubes, you're better off keeping the power supply "light on it's feet," IMO. The amp will sound more musical and responsive. Yes, you'll get some sag at peaks, but with a stout power tranny and a choke in each leg, it should be plenty stiff.
I'm stupid I putted the wrong secondary impedance in the "PSUD2 Duncan" and now to get stable around 350 450uF is ok, its way better, I was also very surprised by those 1000uF !!! So JJ's will be
You may want to adjust this down somewhat. Such big caps mean very large peak currents through transformer and rectifiers, larger radiated noise, etc. Might need to compromise or adopt some other scheme, maybe choke input. Just a caution about balancing conflicting requirements.
All good fortune,
Chris
All good fortune,
Chris
Hello Chris, this afternoon I'll run more simulations to see if I can do better with those caps, for the rectifiers its not an issue they can handle a lot lot lot of current I'll post some graphics
Regards
Alexis
Regards
Alexis
You are not using a tube rectifier, so you have more leeway. With a 5AR4, the first cap would need to be 30uF or less…or inrush would cause fireworks.
Can a moderator delete the last message please #68 ? I didn't know about 30 min edit and I had a big mistake in the PCB 😆
Hello Everyone
So I made two version, they are not finalized they need few more things adjusted but the main things are here 🤓
The first version is inspired by Mircea with all the GND in the top copper but as the PCB is gonna be cut, is a big bridge in the middle of the pcb to the another one is gonna be ok the a correct current flowing ?
And a second version inspired by my guitar amps PCB, every triode GND is link directly to its own main capacitor to allow the current flow to be focused correctly:
Regards
Alexis
Hello Everyone
So I made two version, they are not finalized they need few more things adjusted but the main things are here 🤓
The first version is inspired by Mircea with all the GND in the top copper but as the PCB is gonna be cut, is a big bridge in the middle of the pcb to the another one is gonna be ok the a correct current flowing ?
And a second version inspired by my guitar amps PCB, every triode GND is link directly to its own main capacitor to allow the current flow to be focused correctly:
Regards
Alexis
Maybe my understanding is bad, but why don’t you make 2 different boards? One will be the amp, and the other one will do the PSU.
The left card is gonna go for the Right chan the center one the PSU for both R+L, and I'll cut one card to use it as Right Chan. The minimal production of JLCPCB is 5 card and I'm gonna build only one amp.
I can make a full card but the PCB is gonna cost way more especially with 2mm 2OZ and ENGI 😵
I can make a full card but the PCB is gonna cost way more especially with 2mm 2OZ and ENGI 😵
I have a quad of Mullard KT88. Will follow your build, maybe I’ll build something similar. Output transformers are from Piemme also?
Looks good! But you really don't need regulated DC on the 6SN7's, just wasted space and heat. A Williamson with properly balanced output tubes will be dead silent.
Nice one Alexis!
I do really like the shape of the copper traces - I'm talking about the radius on the corners. Reminds me of vintage one-side copper traces.
By looking at the PSU board, may I recommend something. You can add some more GND connections, just in case you need them later. It is not a functional matter. Is just that in case you need an extra GND connection when assembling the amp, you have provisions for that.
Cheers,
Mircea
I do really like the shape of the copper traces - I'm talking about the radius on the corners. Reminds me of vintage one-side copper traces.
By looking at the PSU board, may I recommend something. You can add some more GND connections, just in case you need them later. It is not a functional matter. Is just that in case you need an extra GND connection when assembling the amp, you have provisions for that.
Cheers,
Mircea
Did you mean to neck-down the separation of the PT HT incoming traces before they reached the diode pads, and to use that pad separation to gnd? Many software products have a keep clear mechanism for restraining minimum separation of circuit nodes, so as to comply with creepage and clearance requirements. Somewhat similar concern between KT88 pin 3 to anything nearby it, as that pin has largest signal voltage on it, and is not uncommonly seen arcing to pin 2 in guitar amps. The external connections to anode/screen and heater on one side of V4 KT88 may be better to swap over to minimise high voltage signal from coupling to other circuitry.
Did you also mean to increase parasitic capacitance of C3-C6 to the gnd flood ?
There is a concern with pin 1 of the KT88's. You have chosen to make them NC. Some KT88 connect pin 1 to the metal shell - and it is plausible that could be a screening and/or a safety function, where pin 1 was originally meant to be connected to gnd in the equipment. No issue for KT88 where pin 1 is NC in the valve, but can become an issue if other well-known valves are used in lieu of KT88 at some stage. Perhaps worth adding a hazard note to use KT88 only.
The input stage and PI stage grid pins need minimal capacitance coupling to gnd and other signals. There is a lot of added gnd flood aligned with those nodes. Also driver stage pin 5 has close coupling (capacitance) to pin 2 and pin 1 - perhaps don't try and squeeze traces between pads, but route them around.
Did you also mean to increase parasitic capacitance of C3-C6 to the gnd flood ?
There is a concern with pin 1 of the KT88's. You have chosen to make them NC. Some KT88 connect pin 1 to the metal shell - and it is plausible that could be a screening and/or a safety function, where pin 1 was originally meant to be connected to gnd in the equipment. No issue for KT88 where pin 1 is NC in the valve, but can become an issue if other well-known valves are used in lieu of KT88 at some stage. Perhaps worth adding a hazard note to use KT88 only.
The input stage and PI stage grid pins need minimal capacitance coupling to gnd and other signals. There is a lot of added gnd flood aligned with those nodes. Also driver stage pin 5 has close coupling (capacitance) to pin 2 and pin 1 - perhaps don't try and squeeze traces between pads, but route them around.