I have been attempting to design a 12B4A linestage which I plan to build in the next three months or so but I ran into some snags. I was going to build a conventional resistor loaded version with a B+ of 200v but when I made the load resistor calulations the plate resistor was too small to be effective as a load for the plate. I now am attempting to design one using a ccs as a plate load. The problem is I don't quite get how to adjust it for a current of 35mA. I am looking for 100v on the plate at 35mA. Is there anyone on the forum who has any experience with this? I plan on buying the C4S boards from Bottlehead but I would like to have some idea of what value cathode resistors and PSU components I need to order from Mr. Percy.
Attachments
I'm beginning to think that this part may be the way to go.:
http://pmillett.addr.com/current_source.htm
The C4S is really not designed for the amount of current that I'm asking it to handle. The problem with the IXYS part is that the tab is at the same potential as the + pin. I was going to simply bolt the regulator to the bottom of the chassis plate but I'm not sure how to do that without causing a fireworks show now. I'll post any further info I come up with.
http://pmillett.addr.com/current_source.htm
The C4S is really not designed for the amount of current that I'm asking it to handle. The problem with the IXYS part is that the tab is at the same potential as the + pin. I was going to simply bolt the regulator to the bottom of the chassis plate but I'm not sure how to do that without causing a fireworks show now. I'll post any further info I come up with.
The constant current sink sets the current through the valve at 35mA, but the anode voltage is set by the cathode bias. Referring to the 12B4 curves, you need 10V of cathode bias to set the anode to 100V with 35mA, so you need a 10V/35mA = 286R resistor. 270R will probably do just fine.
Don't screw the MJE350 to the case, fit a small (6C/W) heatsink to it, and don't let it touch anything (especially you).
Don't screw the MJE350 to the case, fit a small (6C/W) heatsink to it, and don't let it touch anything (especially you).
Gary Pimm has a site with lots of info. here on current sources
http://home.pacifier.com/~gpimm/
I have built the bottlehead C4S and run it at those current levels. It works well, and is easy to play with for a project in current sources. The key is to watch the power through the MJE350, which would be the current times the voltage drop across it at idle.
You may want to consider dropping the voltage from 200 to 175 or so, and running the 12B4 with 125V across the tube.
Here's a suggested design to start. Run the 12B4 at 20ma, with 140V on the plate, and -15V bias on the grid (140V -15V = 125V across the tube).
This would mean a 750 ohm cathode resistor, and R1 in the C4S = 50 ohms.
You would also need a 1.25k resistor in front of the C4S to drop the supply voltage from 200 to 175V.
You may need to tweak this value a bit, it should be rated 1-2W min. as it is dissipating 0.5W. I would order inexpensive parts from Mouser not Percy to get it going first, and order resistor values + and -, say 20% from values given here to play around.
They also carry the transistors needed.
The other C4S resistor, R2, used to set the 2ma for the LED, would be 140V/2ma = 70k ohms, dissipating .28W so 1W rated resistor. Many people would consider using two 1/2 W 35k resistors in series here, they may be easier to find.
This would give roughly +/- 30V output swing, (about 20Vrms) which is plenty for a line stage.
The MJE350 is dissipating 35V times 20ma = .7W, so it needs a heat sink. I have used ones as simple as aluminum angle colored black with a permanent marker. As another mentioned, unless you isolate the heat sink, which requires a proper mounting kit, it will be at high voltage, but usually this is not a big deal.
-Bob
http://home.pacifier.com/~gpimm/
I have built the bottlehead C4S and run it at those current levels. It works well, and is easy to play with for a project in current sources. The key is to watch the power through the MJE350, which would be the current times the voltage drop across it at idle.
You may want to consider dropping the voltage from 200 to 175 or so, and running the 12B4 with 125V across the tube.
Here's a suggested design to start. Run the 12B4 at 20ma, with 140V on the plate, and -15V bias on the grid (140V -15V = 125V across the tube).
This would mean a 750 ohm cathode resistor, and R1 in the C4S = 50 ohms.
You would also need a 1.25k resistor in front of the C4S to drop the supply voltage from 200 to 175V.
You may need to tweak this value a bit, it should be rated 1-2W min. as it is dissipating 0.5W. I would order inexpensive parts from Mouser not Percy to get it going first, and order resistor values + and -, say 20% from values given here to play around.
They also carry the transistors needed.
The other C4S resistor, R2, used to set the 2ma for the LED, would be 140V/2ma = 70k ohms, dissipating .28W so 1W rated resistor. Many people would consider using two 1/2 W 35k resistors in series here, they may be easier to find.
This would give roughly +/- 30V output swing, (about 20Vrms) which is plenty for a line stage.
The MJE350 is dissipating 35V times 20ma = .7W, so it needs a heat sink. I have used ones as simple as aluminum angle colored black with a permanent marker. As another mentioned, unless you isolate the heat sink, which requires a proper mounting kit, it will be at high voltage, but usually this is not a big deal.
-Bob
Thanks for the info Bob. When I stated I wanted to use a B+ of 200 volts I was showing my ignorance of CCSs. I thought I was going to need to drop 100 volts across the CCS and 100 volts across the tube like you would in a conventional design. I now see my mistake. I only planned on having 100 volts on the plate but I can use 140. I was told that a 12B4A sounds best at a 100 volts on the plate and 35mA current flow. I have never heard one so who am I to argue with someone who is nice enough to give me a hand. I would like to be able to get away with not using a heatsink but if I must I must. What do you think about using 100 volts on the plate and biasing it for 20mA? I can design the power supply to give me any B+ needed. Thanks again.
"What do you think about using 100 volts on the plate and biasing it for 20mA? "
Usually you would look at the plates curves to understand these choices. On this tube, most people are running it at a higher cathode bias voltage, which tranlates to a higher plate voltage with a C4S. In your example, you only have about 8V on the cathode, which does not give nearly the headroom that 15-18V would.
With a C4S, once you set the current and "fix it", the idle point follows a horizontal line at that current level on the plate curve. So more cathode voltage directly translates to higher plate voltage.
-Bob
Usually you would look at the plates curves to understand these choices. On this tube, most people are running it at a higher cathode bias voltage, which tranlates to a higher plate voltage with a C4S. In your example, you only have about 8V on the cathode, which does not give nearly the headroom that 15-18V would.
With a C4S, once you set the current and "fix it", the idle point follows a horizontal line at that current level on the plate curve. So more cathode voltage directly translates to higher plate voltage.
-Bob
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.