I have developed a regulator for directly heated tubes that is capable of 1.5A continuous output current. The topology is that of a voltage controlled current source that is high impedance for all but the lowest frequency's of the audio band.
It has a tracking pre-regulator in the negative return consisting of a LM337, this is connected to an adjustable shunt reference and keeps the voltage over the output transistor of the constant current source at a constant voltage.
See the attached images of my prototype, it works flawlessly. The heatsink is a
RAD-A5723/50 STONECOLD AVAILABLE FROM TME.EU
Specs are as follows:
Dropout voltage : 4V worst case scenario
Output voltage : Adjustable between 2.5 and 7.5V higher is possible
output current : 1.5A continuous, about 2A short circuit
Output noise : hard to measure, its below the noise floor of my instruments. There is some popcorn noise on my scope but its low amplitude.
Input voltage : 6.5-30V with no troubles, if the input electrolytic is rated for 35V
Features: Slow start of filaments, thermal overload protection and short circuit proof on a sizeable heat sink up to 7.5V output voltage. The Current source is hard limited to 2A of output current, and the 337 will also limit the current at about the same voltage.
Drawbacks: Higher input-output voltage than some other regulator designs due to the addition of a tracking pre-regulator, you lose about 2V extra due to this addition. But this also means that the LM337 soaks up most of the dissipation at higher input voltages. And gives the benefit of the LM337 internal protection circuitry.
Parts kit contend.
the parts kit will consist of a green ENIG-ROHS board and all the parts pictured on the prototype except the heatsink and m3 stand-offs and to220 thermal silpad. The schematic will be included. And you just have to solder everything yourself. Its all through hole with generous pads so it shouldnt be a problem for anyone who has ever held a soldering iron. One note: the resistors will all be in the same bag so you need a ohms meter or be able to read the color codes for assembly.
price
the cost of one parts kit for one module will be €15. Shipping at current postage rates through POSTNL or DHL whichever is cheapest.
Currently taking firm orders with no down payment, for any questions please ask them here and i will answer them when i have the time.
It has a tracking pre-regulator in the negative return consisting of a LM337, this is connected to an adjustable shunt reference and keeps the voltage over the output transistor of the constant current source at a constant voltage.
See the attached images of my prototype, it works flawlessly. The heatsink is a
RAD-A5723/50 STONECOLD AVAILABLE FROM TME.EU
Specs are as follows:
Dropout voltage : 4V worst case scenario
Output voltage : Adjustable between 2.5 and 7.5V higher is possible
output current : 1.5A continuous, about 2A short circuit
Output noise : hard to measure, its below the noise floor of my instruments. There is some popcorn noise on my scope but its low amplitude.
Input voltage : 6.5-30V with no troubles, if the input electrolytic is rated for 35V
Features: Slow start of filaments, thermal overload protection and short circuit proof on a sizeable heat sink up to 7.5V output voltage. The Current source is hard limited to 2A of output current, and the 337 will also limit the current at about the same voltage.
Drawbacks: Higher input-output voltage than some other regulator designs due to the addition of a tracking pre-regulator, you lose about 2V extra due to this addition. But this also means that the LM337 soaks up most of the dissipation at higher input voltages. And gives the benefit of the LM337 internal protection circuitry.
Parts kit contend.
the parts kit will consist of a green ENIG-ROHS board and all the parts pictured on the prototype except the heatsink and m3 stand-offs and to220 thermal silpad. The schematic will be included. And you just have to solder everything yourself. Its all through hole with generous pads so it shouldnt be a problem for anyone who has ever held a soldering iron. One note: the resistors will all be in the same bag so you need a ohms meter or be able to read the color codes for assembly.
price
the cost of one parts kit for one module will be €15. Shipping at current postage rates through POSTNL or DHL whichever is cheapest.
Currently taking firm orders with no down payment, for any questions please ask them here and i will answer them when i have the time.
Attachments
Boards arrived today.
50 boards arrived in the mail today, Green ENIG-ROHS boards.
I will try to work on a Bill of materials this week, so it is clear what parts are included in a kit, and what parts you have to source yourself.
Currently i feel that it is too much work to include all the 0207 resistors in the package, and most people have a assortment of them anyway.
50 boards arrived in the mail today, Green ENIG-ROHS boards.
I will try to work on a Bill of materials this week, so it is clear what parts are included in a kit, and what parts you have to source yourself.
Currently i feel that it is too much work to include all the 0207 resistors in the package, and most people have a assortment of them anyway.
Attachments
Power supply boards to go with these regulators gerbers attached
Here is a picture of some completed power supply boards to go with these regulators.
These are dual CRC topology boards, i usually use SB540 skottkeys on these boards for low voltage work. And four 7.5mm pitch 18mm dia capacitors.
The 0R47 resistors normally fitted dont filter too much 100Hz but the harmonics of 100Hz are better attenuated than the 100Hz itself.
These are available as ROHS-ENIG for 5 euro each once the GB picks up.
Here is a picture of some completed power supply boards to go with these regulators.
These are dual CRC topology boards, i usually use SB540 skottkeys on these boards for low voltage work. And four 7.5mm pitch 18mm dia capacitors.
The 0R47 resistors normally fitted dont filter too much 100Hz but the harmonics of 100Hz are better attenuated than the 100Hz itself.
These are available as ROHS-ENIG for 5 euro each once the GB picks up.
Attachments
completed and tested modules now €25 each.
I have a dozen of modules that i built in expectation of this picking up,
Letting these go completed and with all parts/mounting hardware for €25 each.
See the attached pic for boards nearing completion. These will come tested and trimmed to 5.0V. Trim range on these is 2.5-7.5V with dropout voltage of 4V max and maximum output current of 1.5A continuous/2A surge during startup.
I will include the m3 spacers, and with each set you get an additional board to serve as a drill guide on your heat sink.
I have a dozen of modules that i built in expectation of this picking up,
Letting these go completed and with all parts/mounting hardware for €25 each.
See the attached pic for boards nearing completion. These will come tested and trimmed to 5.0V. Trim range on these is 2.5-7.5V with dropout voltage of 4V max and maximum output current of 1.5A continuous/2A surge during startup.
I will include the m3 spacers, and with each set you get an additional board to serve as a drill guide on your heat sink.
Attachments
You could try filament bias with those tubes, that is safer for the fragile heaters as well because i have only tested these regulators with higher power tubes to be fair and that is where these absolutely shine.
Price on a kit is reduced to €10 for a kit with all the parts, except m3 spacers and transistor mounting hardware.
See the attached schematic, and Gerbers, you are free to order the boards yourself.
You can try a mosfet instead of the 2SC6144 but you may need to tweak the frequency compensation. i still want to try this on a prototype.
The part i had in mind for this is a RFP12N10L
Price on a kit is reduced to €10 for a kit with all the parts, except m3 spacers and transistor mounting hardware.
See the attached schematic, and Gerbers, you are free to order the boards yourself.
You can try a mosfet instead of the 2SC6144 but you may need to tweak the frequency compensation. i still want to try this on a prototype.
The part i had in mind for this is a RFP12N10L
Attachments
All out of boards, GB didnt pick up
I'm a bit disappointed in the forum. At €10 for a parts kit, that leaves €2.50 for my troubles and time. Yet i found no takers, apparently it takes being talked up by one of the big names on a forum to even get so much as a response.
I ended up selling most of the boards domestically. Improved version on the way. Not posting designs for free any more.
This design Beats the Tentlabs regulators in noisefloor any day of the week. And the softstart is just an added bonus.
And the Colemans in that if you build them on the specified heat sink you can feed 25V in and 5V out, hook up a 300B and it wont blow up period. I had that running on my desk for 6 hours before i figured i'd seen enough.
V4lve out.
I'm a bit disappointed in the forum. At €10 for a parts kit, that leaves €2.50 for my troubles and time. Yet i found no takers, apparently it takes being talked up by one of the big names on a forum to even get so much as a response.
I ended up selling most of the boards domestically. Improved version on the way. Not posting designs for free any more.
This design Beats the Tentlabs regulators in noisefloor any day of the week. And the softstart is just an added bonus.
And the Colemans in that if you build them on the specified heat sink you can feed 25V in and 5V out, hook up a 300B and it wont blow up period. I had that running on my desk for 6 hours before i figured i'd seen enough.
V4lve out.
Yeah, i just found out D1 is reversed on the silk
If your gonna use the gerbers in this thread for the regulators, just a friendly heads-up the diode D1 is drawn the wrong way. It will work, however the benefit of the temperature protection inherent in the 337/1033 will be lost if you dont reverse it or cut it out.
I did my tests with a LT1033 and i wanted to measure noise, and figured the diode capacitance would mess with the readings too much, so i omitted it at some point during testing.
If your gonna use the gerbers in this thread for the regulators, just a friendly heads-up the diode D1 is drawn the wrong way. It will work, however the benefit of the temperature protection inherent in the 337/1033 will be lost if you dont reverse it or cut it out.
I did my tests with a LT1033 and i wanted to measure noise, and figured the diode capacitance would mess with the readings too much, so i omitted it at some point during testing.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.