I'm hardly an expert but I have designed and made a few PCBs and all worked as hoped so I'll throw in my $.02.
Basically I tried to follow the same rules I would have followed if wring point-to-point. Keep low signal stuff away from AC heaters and high signal stuff, try not to cross DC and AC etc. The schematic provides a basic starting point but of course you can change the physical layout.
Probably not all the info you hoped for but I'm sure someone with more experience will chime in.
Basically I tried to follow the same rules I would have followed if wring point-to-point. Keep low signal stuff away from AC heaters and high signal stuff, try not to cross DC and AC etc. The schematic provides a basic starting point but of course you can change the physical layout.
Probably not all the info you hoped for but I'm sure someone with more experience will chime in.
Don't put the heater supplies on the PCB - twist them and wire them point-to-point.
There seems to be a HUGE amount of prejudice (*) against PCBs for tube use, some of it may well be justified (I'm not sufficiently qualified to comment), some of it maybe subjective rather than quantitive - who knows?
If you follow the "use the least number of connections in the signal path" law, then PCBs are bad news as you have typically two solder joints per single point-to-point connection (actually, you often get two joint in PtP wiring).
Anyway, there is a calculator on my web site which may help in your current calculations & voltage isolation distances.
(*) the use of the word "prejudice" is, in itself, prejudicial. q.v.
There seems to be a HUGE amount of prejudice (*) against PCBs for tube use, some of it may well be justified (I'm not sufficiently qualified to comment), some of it maybe subjective rather than quantitive - who knows?
If you follow the "use the least number of connections in the signal path" law, then PCBs are bad news as you have typically two solder joints per single point-to-point connection (actually, you often get two joint in PtP wiring).
Anyway, there is a calculator on my web site which may help in your current calculations & voltage isolation distances.
(*) the use of the word "prejudice" is, in itself, prejudicial. q.v.
You have to use the real component's footprint...
and go from there with the following in mind; 1. Grounding (two sided -ground plane???) , 2. Power inputs, 3. Signal inputs 4. Signal outputs 5. Size limitations. You need to keep inputs away from outputs both on the whole board and on each tube. You also have to keep minimum lead length on grid inputs. Lead and lag feedback compensation needs to be minimum length. The schematic is only a guide. You have to use the real world part in an x-ray view from the top. The board is laid out on 1/5 inch grid line paper and numerous circuit paths and versions are tried before I ever sit down at the computer. I assemble all the parts and know their lead spacing before I begin a design. Tube sockets are bad about causing board fractures in the tracings by insertion and pulling out numerous times. If you're using high transconductance tubes then their grids need grid stoppers mounted as close to the pins as is possible. Some hi gm tubes have numerous grid pins and all of these have to be used- not just one. Some of these grids have to have an additional shield tracing around the grid pin. Plate resistors need to be oriented away from cathode resistors. Neither should they lay side by side. You have to be thinking what is an input and what is an output and where could crosstalk enter (because the parts especially metal film resistors can broadcast to one another). This circuit board can if multi planed can become a swell capacitor. I don't run the heaters circuit on the circuit boards far away from the tube socket. Each tube has its own twisted pair, shielded heater lead that goes to a power supply board. Another approach is to run the heater traces on top and the signal traces on the bottom at severe right angles to one another with a peripheral ground plane. Only run the ground plane on one side of the board. This works for me.
and go from there with the following in mind; 1. Grounding (two sided -ground plane???) , 2. Power inputs, 3. Signal inputs 4. Signal outputs 5. Size limitations. You need to keep inputs away from outputs both on the whole board and on each tube. You also have to keep minimum lead length on grid inputs. Lead and lag feedback compensation needs to be minimum length. The schematic is only a guide. You have to use the real world part in an x-ray view from the top. The board is laid out on 1/5 inch grid line paper and numerous circuit paths and versions are tried before I ever sit down at the computer. I assemble all the parts and know their lead spacing before I begin a design. Tube sockets are bad about causing board fractures in the tracings by insertion and pulling out numerous times. If you're using high transconductance tubes then their grids need grid stoppers mounted as close to the pins as is possible. Some hi gm tubes have numerous grid pins and all of these have to be used- not just one. Some of these grids have to have an additional shield tracing around the grid pin. Plate resistors need to be oriented away from cathode resistors. Neither should they lay side by side. You have to be thinking what is an input and what is an output and where could crosstalk enter (because the parts especially metal film resistors can broadcast to one another). This circuit board can if multi planed can become a swell capacitor. I don't run the heaters circuit on the circuit boards far away from the tube socket. Each tube has its own twisted pair, shielded heater lead that goes to a power supply board. Another approach is to run the heater traces on top and the signal traces on the bottom at severe right angles to one another with a peripheral ground plane. Only run the ground plane on one side of the board. This works for me.
!
Although I haven't built any tube amps yet ( moving there slowly!), I'm old enough to have serviced a fair amount of various tube gear, and early days SS too, for that matter.
Heat and PCBs are very far from being good friends. Although it might work OK for pre amps and small signal, I am reluctant to use PCBs for power amps. At the moment, I'm planning on a tube power and a guitar amp for my son, but I think I'll use tag boards and PtP wiring for both. In adddition to the heated tubes themselves, the dropper resistors generate quite a lot of heat. Rotten solder joints used to be the main source of failure for e.g tube TVs, and also for the older and hotter SS stuff.
The idea of larger holes in the PCB and short lead connections from a chassis mounted tube socket, is actually a very good one, and not really uncommon in production either.
Heat and PCBs are very far from being good friends. Although it might work OK for pre amps and small signal, I am reluctant to use PCBs for power amps. At the moment, I'm planning on a tube power and a guitar amp for my son, but I think I'll use tag boards and PtP wiring for both. In adddition to the heated tubes themselves, the dropper resistors generate quite a lot of heat. Rotten solder joints used to be the main source of failure for e.g tube TVs, and also for the older and hotter SS stuff.
The idea of larger holes in the PCB and short lead connections from a chassis mounted tube socket, is actually a very good one, and not really uncommon in production either.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.