Embedded Files
Having used DC step-up converters on several projects, I appreciate their size, weight, and utility. A common commercial unit (intended for capacitor charging) is widely used to power tube circuits. Its low cost makes it attractive, but it has some stability issues that are described in the GAD book, and it’s become hard to find from local US sources. Shipment from Asian sources is taking 2-3 months. I recently waited over 100 days for an order. To ensure access to step-up converters in the future, I decided to design a converter suited for my needs as an alternative, should supply problems persist.
Having used DC step-up converters on several projects, I appreciate their size, weight, and utility. A common commercial unit (intended for capacitor charging) is widely used to power tube circuits. Its low cost makes it attractive, but it has some stability issues that are described in the GAD book, and it’s become hard to find from local US sources. Shipment from Asian sources is taking 2-3 months. I recently waited over 100 days for an order. To ensure access to step-up converters in the future, I decided to design a converter suited for my needs as an alternative, should supply problems persist.
The requirements are simple. Input voltage is 24-36 V. 12 V operation would be a bonus. Output voltage should be about 360 V (maximum) at 20-25 mA - or about 7-9 watts of output power. Voltage adjustment (down to ~300 V) should be feasible in case less voltage is needed for an application. These requirements are sufficient to power a variety of tubes and solid-state circuits. The Solo 100, for example only requires about 18 mA at 360 V.
The requirements are simple. Input voltage is 24-36 V. 12 V operation would be a bonus. Output voltage should be about 360 V (maximum) at 20-25 mA - or about 7-9 watts of output power. Voltage adjustment (down to ~300 V) should be feasible in case less voltage is needed for an application. These requirements are sufficient to power a variety of tubes and solid-state circuits. The Solo 100, for example only requires about 18 mA at 360 V.
There are many switching regulator ICs out there and I spent some time surveying the options. I’m relatively new to switching regulator circuits so it also took some time to get comfortable with the issues and details of these systems. I can’t claim extensive expertise in these circuits. I’m sure there are more-seasoned designers out there who may take issue with my decisions, but the following describes my process and the results.
There are many switching regulator ICs out there and I spent some time surveying the options. I’m relatively new to switching regulator circuits so it also took some time to get comfortable with the issues and details of these systems. I can’t claim extensive expertise in these circuits. I’m sure there are more-seasoned designers out there who may take issue with my decisions, but the following describes my process and the results.
My survey of regulators was simply reading spec sheets and app notes looking for useful info and examples. After a while you start to see common patterns emerge about circuit types and device features. During the survey one device emerged and stood out. The LT1170 series offered several attractive features.
My survey of regulators was simply reading spec sheets and app notes looking for useful info and examples. After a while you start to see common patterns emerge about circuit types and device features. During the survey one device emerged and stood out. The LT1170 series offered several attractive features.
1) It is highly integrated and lends itself to complete systems that only require a few additional components.
1) It is highly integrated and lends itself to complete systems that only require a few additional components.
2) There is a LTSpice model available so I could try different design options and learn about operation and performance issues.
2) There is a LTSpice model available so I could try different design options and learn about operation and performance issues.
3) The device is available in a TO-220 package (5 leads) that incorporates the switching transistor and is easy to heat-sink. Assuming the final circuit is ~80 % efficient, the waste heat is about 1-2 watts. That should not stress the TO-220 package nor create excessive dissipation for a simple heat sink.
3) The device is available in a TO-220 package (5 leads) that incorporates the switching transistor and is easy to heat-sink. Assuming the final circuit is ~80 % efficient, the waste heat is about 1-2 watts. That should not stress the TO-220 package nor create excessive dissipation for a simple heat sink.
4) Lastly, the device comes in a high-voltage version that allows safe operation at 36 V, which is my preferred input voltage source.
4) Lastly, the device comes in a high-voltage version that allows safe operation at 36 V, which is my preferred input voltage source.
This combination of features made the choice for me. As I followed through with an actual design and build, I found no reason to revisit the device choice.
This combination of features made the choice for me. As I followed through with an actual design and build, I found no reason to revisit the device choice.
The spec sheet shows example circuits. In particular, the flyback converter (page 11 of the spec sheet) drew my attention due to its simplicity and familiarity. I’ve seen similar circuits before and I believe the commercial converter I’ve been using is also a flyback system. The flyback circuit has the crucial feature that a step-up transformer isolates the switching device from the high output voltage. The circuit is further simplified since my application allows the input and output voltages to share a common ground.
The spec sheet shows example circuits. In particular, the flyback converter (page 11 of the spec sheet) drew my attention due to its simplicity and familiarity. I’ve seen similar circuits before and I believe the commercial converter I’ve been using is also a flyback system. The flyback circuit has the crucial feature that a step-up transformer isolates the switching device from the high output voltage. The circuit is further simplified since my application allows the input and output voltages to share a common ground.
The next design issue was selecting a transformer among stock commercially-available options. My goal is a 360 V output from a 12-36 V input. I’m mainly interested in the 24-36 V range, but if the converter works down to 12 V, that’s a bonus. The LT1170 itself doesn’t have a problem with 12 V, it’s just a question of whether the whole system will operate properly.
The next design issue was selecting a transformer among stock commercially-available options. My goal is a 360 V output from a 12-36 V input. I’m mainly interested in the 24-36 V range, but if the converter works down to 12 V, that’s a bonus. The LT1170 itself doesn’t have a problem with 12 V, it’s just a question of whether the whole system will operate properly.
After much searching among available transformers, I selected a few to try in Spice simulations. It soon became clear that a relatively low inductance primary was needed to get proper operation at the LT1170’s 100 KHz switching frequency. A significant step-up turns ratio would also help keep the primary flyback voltages low and safe for the regulator. In the end, the Coilcraft DA2033-AL became an obvious choice.
After much searching among available transformers, I selected a few to try in Spice simulations. It soon became clear that a relatively low inductance primary was needed to get proper operation at the LT1170’s 100 KHz switching frequency. A significant step-up turns ratio would also help keep the primary flyback voltages low and safe for the regulator. In the end, the Coilcraft DA2033-AL became an obvious choice.
More component surveys and simulations led to the selection of rectifier and other components. The final circuit is shown below. It is almost exactly the configuration shown in the device spec sheet. However, in bread boarding the circuit I discovered the need for C3. The regulator lacked stability at varied input voltages and loads without that cap. Its value is not all that critical and I recall any value between 2nf-10nf worked well. The Spice simulation works fine with or without the cap.
More component surveys and simulations led to the selection of rectifier and other components. The final circuit is shown below. It is almost exactly the configuration shown in the device spec sheet. However, in bread boarding the circuit I discovered the need for C3. The regulator lacked stability at varied input voltages and loads without that cap. Its value is not all that critical and I recall any value between 2nf-10nf worked well. The Spice simulation works fine with or without the cap.
After breadboard testing, I used this simple circuit as motivation to learn KiCAD and do a PCB layout using that tool. KiCAD took a while. I started with version 4 and later realized I had to relearn many things when switching to version 5.
After breadboard testing, I used this simple circuit as motivation to learn KiCAD and do a PCB layout using that tool. KiCAD took a while. I started with version 4 and later realized I had to relearn many things when switching to version 5.
As expected, creating and managing schematic part libraries and PCB part footprints are among the biggest challenges to overcome. Fortunately, the online help and tutorials are pretty good and after a while I managed to complete a board design.
As expected, creating and managing schematic part libraries and PCB part footprints are among the biggest challenges to overcome. Fortunately, the online help and tutorials are pretty good and after a while I managed to complete a board design.
I’ve since used KiCAD to make the Solo Verb 3 card and I’m fully converted and comfortable with the tool at this point. It’s really a very good piece of software and I’m grateful to those who make it available for free. The PCB for the convert circuit is 1.8 x 2.65 inches, including a heat sink for the regulator.
I’ve since used KiCAD to make the Solo Verb 3 card and I’m fully converted and comfortable with the tool at this point. It’s really a very good piece of software and I’m grateful to those who make it available for free. The PCB for the convert circuit is 1.8 x 2.65 inches, including a heat sink for the regulator.
The regulator is available in several versions. The LT1171HV is the version I’m using. It has a medium peak switch-current and a high voltage rating.
The regulator is available in several versions. The LT1171HV is the version I’m using. It has a medium peak switch-current and a high voltage rating.
Tests are ongoing. I’ve used the board in the Solo 100 amp for several hours without any problems. I don’t yet have calibrated temperature sensors, so I can’t measure accurate temperature rises. The regulator heat sink gets warm, although it can be safely touched so it is clearly operating within device specs (100 deg C). The transformer also gets warm, but again, it’s not beyond the “safe-finger” test. All the other devices remain cool to touch.
Tests are ongoing. I’ve used the board in the Solo 100 amp for several hours without any problems. I don’t yet have calibrated temperature sensors, so I can’t measure accurate temperature rises. The regulator heat sink gets warm, although it can be safely touched so it is clearly operating within device specs (100 deg C). The transformer also gets warm, but again, it’s not beyond the “safe-finger” test. All the other devices remain cool to touch.
This circuit is quite flexible and achieves a range of lower output voltages by changing R3. RV1 is a 1KL trimmer pot and roughly centered at a 350 V output.
This circuit is quite flexible and achieves a range of lower output voltages by changing R3. RV1 is a 1KL trimmer pot and roughly centered at a 350 V output.
The board does operate down to 12 V, but with some reduction of achievable output voltage and current. Ripple and switch noise vary with loads and voltages, but aggregate noise generally stays below 200 mV peak. That’s good (IMO) given that C4 (1uf 400 V) is the only onboard filter. Additional off-board filtering is needed to reduce noise levels. The Solo 100 active filter produces excellent results and is a filter method to consider.
The board does operate down to 12 V, but with some reduction of achievable output voltage and current. Ripple and switch noise vary with loads and voltages, but aggregate noise generally stays below 200 mV peak. That’s good (IMO) given that C4 (1uf 400 V) is the only onboard filter. Additional off-board filtering is needed to reduce noise levels. The Solo 100 active filter produces excellent results and is a filter method to consider.
The stuffed card occupies less space and is more efficient than the dual converter system described in the GAD book. Its only disadvantage is cost. It costs a bit more to build this card than to buy the commercial converters. Having an option is good, however, since it ensures availability.
The stuffed card occupies less space and is more efficient than the dual converter system described in the GAD book. Its only disadvantage is cost. It costs a bit more to build this card than to buy the commercial converters. Having an option is good, however, since it ensures availability.
Page updated
Google Sites
Report abuse