Here is an overview of my project:
I am designing a curve tracer that can in theory work with all tubes that were found in audio circuits, radios, televisions, etc. The only tubes it is not designed to work with are high voltage rectifiers, CRT's, Thyratrons, or any kind of "specialty tube."
The unit consists of four programmable power supplies:
- The B+ supply uses SCR's, activated by an embedded system via optotriacs, to select one six of "ranges" on a custom wound multitap transformer. These transformer taps directly apply rectified B+ to the plate of the tube. The humps of the rectified mains are thus responsible for varying the current on the plate. This is how the old Tektronix tracers worked, only they used a clunky rotary switch to select the range taps.
- The screen supply consists of a PAD195 discrete opamp module, current limited to 20mA. A DAC sets the voltage output level via a non-inverting amplifier. The V+ rail is 600V, and the V- rail is -100V.
http://www.powerampdesign.net/images/PAD195_Datasheet_Rev_C.pdf
- The grid bias (C-) supply supplies very low current, as the grid in theory does not draw current unless it is positive relative the cathode. This supply uses an LT6090 opamp in an inverting configuration, programmed by a DAC. The V- rail is -100V unregulated. The positive is +5V.
http://www.linear.com/docs/42316
- Finally, we have the various filament supplies. I still am in the process of designing these. With linear regulators, these could dissipate up to 40 watts of heat, specifically with tubes like a 2A3. Therefore, I am using switchers. There will be 3 ranges. One will be a 0-6V 5A, for the high current, low voltage tubes. The next will be a 6-30V supply 3A, which will be most commonly used range. The third range will be for the high voltage heaters like the 35WC and 50C5. There are even some 117 volt tubes.
To connect tubes to the device, there will be a patch panel to connect each power supply to the appropriate pins. I thought of a relay based system to perform this operation automatically, but it is too expensive.
My questions:
To short circuit and idiot proof the device, what type of circuit do you recommend I implement to protect each supply? Uses can connect the wrong wires to the wrong elements, or a tube can be shorted, outgassed, or damaged in some other way. Even worse, users can short the supplies, say the B+ and C- together on the patch panel. I need a circuit breaker to kill power quickly to the unit before any damage occurs. A light is to be lit denoting a fault as well. High voltage shorts are deadly weight
I am designing a curve tracer that can in theory work with all tubes that were found in audio circuits, radios, televisions, etc. The only tubes it is not designed to work with are high voltage rectifiers, CRT's, Thyratrons, or any kind of "specialty tube."
The unit consists of four programmable power supplies:
- The B+ supply uses SCR's, activated by an embedded system via optotriacs, to select one six of "ranges" on a custom wound multitap transformer. These transformer taps directly apply rectified B+ to the plate of the tube. The humps of the rectified mains are thus responsible for varying the current on the plate. This is how the old Tektronix tracers worked, only they used a clunky rotary switch to select the range taps.
- The screen supply consists of a PAD195 discrete opamp module, current limited to 20mA. A DAC sets the voltage output level via a non-inverting amplifier. The V+ rail is 600V, and the V- rail is -100V.
http://www.powerampdesign.net/images/PAD195_Datasheet_Rev_C.pdf
- The grid bias (C-) supply supplies very low current, as the grid in theory does not draw current unless it is positive relative the cathode. This supply uses an LT6090 opamp in an inverting configuration, programmed by a DAC. The V- rail is -100V unregulated. The positive is +5V.
http://www.linear.com/docs/42316
- Finally, we have the various filament supplies. I still am in the process of designing these. With linear regulators, these could dissipate up to 40 watts of heat, specifically with tubes like a 2A3. Therefore, I am using switchers. There will be 3 ranges. One will be a 0-6V 5A, for the high current, low voltage tubes. The next will be a 6-30V supply 3A, which will be most commonly used range. The third range will be for the high voltage heaters like the 35WC and 50C5. There are even some 117 volt tubes.
To connect tubes to the device, there will be a patch panel to connect each power supply to the appropriate pins. I thought of a relay based system to perform this operation automatically, but it is too expensive.
My questions:
To short circuit and idiot proof the device, what type of circuit do you recommend I implement to protect each supply? Uses can connect the wrong wires to the wrong elements, or a tube can be shorted, outgassed, or damaged in some other way. Even worse, users can short the supplies, say the B+ and C- together on the patch panel. I need a circuit breaker to kill power quickly to the unit before any damage occurs. A light is to be lit denoting a fault as well. High voltage shorts are deadly weight
Have a "sqizz" at what others have done.
The Link list:
The uTracer, a miniature Tube Curve Tracer / Tester.
On protections - don't forget that you maybe plugging in a tube that has internal shorts. Your various supplies must survive that.
Cheers,
Ian
Cheers,
Ian
The Link list:
The uTracer, a miniature Tube Curve Tracer / Tester.
On protections - don't forget that you maybe plugging in a tube that has internal shorts. Your various supplies must survive that.
Cheers,
Ian
Cheers,
Ian
Can't you avoid this? You can bring the supplies to panel sockets. Then have the plugs permenantly wired to the valve socket pins. That way it's harder to short the supplies together (OK, not impossible, as someone could accidentally connect say, one heater pin to the anode feed and one to the cathode feed).
A diode in series with the screen grid will protect the screen supply against anode-g2 shorts. A diode between grid and cathode will protect the grid supply from anode-g1 shorts. The anode could have a fuse (with a parallel resistor+LED for indication), and maybe one for the cathode too. Or perhaps polyfuses?
Is it going to test DHTs too?
Last edited:
Merlinb,
Yes, this device will test directly heated tubes. It will even do compactrons.
jazbo8,
For starters, this unit is just a DIY job. If I were to make it commercial, I would need to revise the unit to comply with the various agency standards.
mrneedle,
This unit will connect to the PC with a USB to UART. All the digital logic is isolated from the high-voltage using isolating amplifiers and optocouplers. Ground loops are avoided by similar techniques.
Here is the status of the project:
Three of the six systems in this unit are part of my EE degree senior design project. The PCBs are on order and are scheduled to ship Friday. They are on 2 day delivery so they should arrive Monday or Tuesday. After that, time for debugging.
The three systems are the Main Logic Board, the B+ sweep board, and the Step Amplifiers board. Each board is a system that is controlled from the Main Logic Board via SPI. I will post schematics as soon as the the three systems are working.
I am in the process of designing the three remaining systems. These are the Filament Supply, Circuit Protection, and the Socket Interface.
Further Idea Questions:
For safety, instead of a patch panel, it probably would be wiser to have an automated switching system to connect the various power supplies to the various tube elements.
My plan here is to have on the tester chassis itself a 9 pin socket, 8 pin socket, 7 pin socket, and a 4 pin for tubes like the 2A3. Rarer tubes like compactrons and loctal will connect via molex to external modules. To minimize the cost of relays, what combinations of what elements attach to what pins for all tubes made? For each of the sockets (never mind the loctal and 12 pin as these are for a later date)?
Yes, this device will test directly heated tubes. It will even do compactrons.
jazbo8,
For starters, this unit is just a DIY job. If I were to make it commercial, I would need to revise the unit to comply with the various agency standards.
mrneedle,
This unit will connect to the PC with a USB to UART. All the digital logic is isolated from the high-voltage using isolating amplifiers and optocouplers. Ground loops are avoided by similar techniques.
Here is the status of the project:
Three of the six systems in this unit are part of my EE degree senior design project. The PCBs are on order and are scheduled to ship Friday. They are on 2 day delivery so they should arrive Monday or Tuesday. After that, time for debugging.
The three systems are the Main Logic Board, the B+ sweep board, and the Step Amplifiers board. Each board is a system that is controlled from the Main Logic Board via SPI. I will post schematics as soon as the the three systems are working.
I am in the process of designing the three remaining systems. These are the Filament Supply, Circuit Protection, and the Socket Interface.
Further Idea Questions:
For safety, instead of a patch panel, it probably would be wiser to have an automated switching system to connect the various power supplies to the various tube elements.
My plan here is to have on the tester chassis itself a 9 pin socket, 8 pin socket, 7 pin socket, and a 4 pin for tubes like the 2A3. Rarer tubes like compactrons and loctal will connect via molex to external modules. To minimize the cost of relays, what combinations of what elements attach to what pins for all tubes made? For each of the sockets (never mind the loctal and 12 pin as these are for a later date)?
Interesting.
One of the downsides of the aforementioned kit from Europe is that it still doesn't handle power tubes (not even 6L6 size) although the latest version might, I haven't kept up.
Automating the pin assignments would be neat, but probably expensive to implement, and require a uproc to handle - a project within a project?
No elements attach to the same pin for every tube. You will need to use one of the many methods for assigning the pins... choose ur poison.
One of the downsides of the aforementioned kit from Europe is that it still doesn't handle power tubes (not even 6L6 size) although the latest version might, I haven't kept up.
Automating the pin assignments would be neat, but probably expensive to implement, and require a uproc to handle - a project within a project?
No elements attach to the same pin for every tube. You will need to use one of the many methods for assigning the pins... choose ur poison.
I calculated an automated switching solution will cost in the low hundreds of dollars for relays alone.
Here is an idea to safety proof the patch panel. An aluminum chassis box on hinges will close over top of the sockets (and tube under test) and patch panel. A perforated plexiglass window will fit inside a cutout on the top of the chassis. A bolt in the lip on the bottom of the chassis will push a switch in the main chassis through a hole drilled in the panel. This switch kills power to the unit when the cover is lifted.
I am very concerned about safety because up to 600 volts at 320 mA will be present on that panel. Exposed banana plugs with that voltage can kill someone who does not know what they are doing or are not careful. There are times I have not been careful and have been shocked by a 400 volt rail while working on a preamp. I also have been shocked by even higher voltages in charged caps. The shock is so powerful your finger feels the current about 1/4 inch away from the conductor. I am a big guy, weighing in at 213 pounds. A young woman I know only weighs 90. I hate to see what that charge would do to her.
Here is an idea to safety proof the patch panel. An aluminum chassis box on hinges will close over top of the sockets (and tube under test) and patch panel. A perforated plexiglass window will fit inside a cutout on the top of the chassis. A bolt in the lip on the bottom of the chassis will push a switch in the main chassis through a hole drilled in the panel. This switch kills power to the unit when the cover is lifted.
I am very concerned about safety because up to 600 volts at 320 mA will be present on that panel. Exposed banana plugs with that voltage can kill someone who does not know what they are doing or are not careful. There are times I have not been careful and have been shocked by a 400 volt rail while working on a preamp. I also have been shocked by even higher voltages in charged caps. The shock is so powerful your finger feels the current about 1/4 inch away from the conductor. I am a big guy, weighing in at 213 pounds. A young woman I know only weighs 90. I hate to see what that charge would do to her.
You might want to consider isolating the device from the computer with an isolated USB adapter -- it's not difficult to fry a laptop or motherboard -- no kiddin. If it's a student project you can probably get a sample of the development kit from Analog Devices.
If you take a look at "Art of Electronics" by Horowitz and Hill, there's an easy to implement HV amplifier using a pair of MOSFETs -- will be a lot less expensive than employing an OEM device for a one-off project.
If you take a look at "Art of Electronics" by Horowitz and Hill, there's an easy to implement HV amplifier using a pair of MOSFETs -- will be a lot less expensive than employing an OEM device for a one-off project.
Update:
For a while this project has basically been dormant. I finally found some time to start working on it again. In the process, I found a better way to engineer the B+ supply. Applying rectified pulses requires continuous operation and some control problems were experienced.
- Instead of relying on the antequated Tektronix approach, I am now using a different configuration (read more below).
http://www.diyaudio.com/forums/tubes-valves/270602-transformer-sizing-etc.html
- The screen supply only switches on while the ramp is being injected. The circuitry is the same.
- There are no changes to the grid bias supply.
- I have not fully designed the heater supply, but it will be some kind of switcher most likely.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.