About 7 years ago I built supply with 0-300V range. TL431, CA3130, IRF710, 2SC945 and IXFH50N50. I arranged IRF710 as common gate configuration with degeneration input resistor. Easier to fine tune and more stable. I chose 3 secondary winding-sets transformer which provide 15V for TL431 and CA3130, 10V for biasing IXTH50N50's gate (float) and 350V for output. Thank you
@@stevenbliss989 At that time, I drew a circuit schematic from a blank piece of paper, based on my dumb imagination. I don't know if there has ever been a similar circuit schematic. And I don't care about that. My friend needs it and I can help him. I just need friends.
Just some advice if you do proceed with this project. When you say "transistor" rated at 1.2kV at 2A, I suppose you mean MOSFET. In general MOSFETs are built for switching and not linear operation (there are a very few exceptions). Look at the SOA curves for a range of so called high power MOSFETs. At 800V you'll be looking at roughly 50W of continuous dissipation - even then you'll be close to the SOA limit. You could design a linear power supply using MOSFETs to give 800V at 1 amp, but you'd need multiple parallel MOSFETs ie 16 of them. You need to consider thermal run-away as well which is dangerous because MOSFETs blow short circuit. Have a look at some DC load designs, since they share similarity with this type of power supply design.
You probably misunderstood but the "1.2kV at 2A" comment I made in the video is for the MOSFET that is interfacing the op-amp to the pass transistors not the main pass transistors. As can be seen in the video, I was using the IRF710 for this test and when I built my 800V, 100mA supply, I used IRFBG30 and it only dissipates about 1.6W at 800V. This is because it is only used as common source amplifier to interface the op-amp to the main pass transistors (see below). Indeed I am aware that most MOSFETs are optimized for switching applications and SOA needs to be strictly overserved for linear application. That is why I built this supply in 2019 with two IXTX20N150 pass transistors :). The supply is rated at 0-800V, 0-100mA with CC and CV function. Stay tuned for an update video.
Very well executed and thoroughly tested with great effort, I am very impressed! I am just wondering why the op-amps are functioning correctly, as they are controlled by two potentials: 15V from the power supply and at the sensor input from the output potential 0-150V. I did not know that this was possible at all. Or is it a special function of this type of op-amp? I assume that both potentials are referenced to ground, is that correct?
No over temp (high risk in this), no over current or short circuit protection. You might find what useful what a TH-cam channel "FesZ Electronics" did. :)
Great content. Funny I have just designed a 200V variable PS with CV/CC function, and the circuit is remarkably similar. Although I'm curious as to why you have the UF4007 diode at the base of the output transistor?
Thank you and great question! The UF4007 is there to protect the output transistors's BE junction when the IRF710 switches on putting the base of the Darlington transistors at ground potential while the output filter capacitors are charged to more than 10V. As you might be aware, BJT's maximum BE junction reverse voltage rating is around 5 to 9V. Therefore the UF4007 diode will block any reverse voltage up to 1000V across the BE junction of the output transistors. I should have explained the purpose of that diode in the video.
The circuit is quite amazing. Where I have a question is the the part of TL431, the reference terminal seem to be excluded from the schematic. Is there any reason for that i would like to know. Thanks
Hello, nice circuit! Have a question - those capacitors at the output of the rectifier have voltage rating of 250V. But if you are using transformer with output of 240V that would make something like 340V DC at the output of the rectifier if you connect that circuit directly at mains. So if one wants to build this circuit those caps should be rated at least 350V, am I right?
Thank you! And great question. Yes you are absolutely correct that if the output of the 240V transformer is to be used from 120V mains directly, those filter capacitors need to have the proper voltage rating which should be greater than 340VDC. But if you notice when I am testing the circuit at 18:45, I am using a variac (autotransformer) to supply the 120V side of that transformer so that the rectified and filtered voltage sits around 200V. If you do the math that is about 70V on the 120V rated primary winding. That is why I said at the end of the video that all the components need to be scaled up properly for high voltage operation. In fact those capacitors on the supply that I have built (0-800VDC) are made of two 450V rated capacitors in series to give me the required voltage rating. Stay tuned for an updated video that shows the implementation of 0-800VDC and 0-100mA power supply soon. Cheers!
Why did you decide to use BJT over FET for pass transistor? Couldn't find one with sufficient SOA at HV of 800V? I read in the comments that you're limiting output current to 100mA. It should be easy to find a FET capable of that. Just curious, thanks for video and stay safe.
I take it back. FET at DC 800V@100mA is a little hard and probably expensive. For someone looking for 0-500V range, here are my notes. AOT10N60 300V@800mA, 600V@400mA to-220 AOT12N65 300V@800mA, 650V@400mA D2PAK AOT27S60 300V@1000mA, 600V@600mA to-220 $3 STP33N65M2 300V@2000mA, 650V@1000mA to-220 $2.78 190W, no dc on SOA AOT42S60 300V@1500mA, 600V@700mA to-220 FCP104N60 300V@1250mA, 600V@600ma to-220 I haven't explored much in the past 2 years though. Have fun.
@@CraigHollabaugh Hey sorry for the late replay! Yes you are indeed correct regarding the availability of reasonably priced high voltage MOSFETS. As a matter of fact I am using two very expensive HV MOSFETs (IXTX20N150) from Littlefuse with excellent SOA for this project. They are $20 a pop and definitely not cheap to use in hobbyist projects. Normally I like to stick with BJTs for linear applications as they are easy to stabilize in closed loop application like this project. I had a hard time stabilizing these MOSFETs with their large parasitic capacitance. But the control loop is sable now with decent transient response. Stay tuned for an update! And thank you for providing part numbers of decent MOSFETs with good SOA for low voltage applications ;)
@@SmithKerona I looked at those IXTX mosfets but didn't need (or want) a 1000V circuit. The 20N150 is quite a part 300mA@1000V DC SOA, that's a killer part, both dangerous and awesome. I don't see my tube amp work going much above 500V. Look forward to your upcoming work. Thanks for the reply.
Not as much as you might think. Most of the items I acquired them in disrepair/broken state. Likely I manged to repair most of them and some still left for spare.
Hi ! Congratulations for your project !! I just smoked six TL783 ICs trying to assemble a 1,5V -120V high power supply as depicted on their datasheet .Even a MJE11015 got shorted . I gave up to use this " wonderful "IC and backing to use discrete components like you did. Snag is that I could not find MC34072 and IRF710 . Any suggestion ?
Hello, have a look at the schematic that I came up with using low voltage single supply opamp (MCP601) and HV BJT (MJE13005). drive.google.com/file/d/1AaTodBA69dSlvdN47gFvb6UYdPmTPd4l/view?usp=sharing Make sure to observe SOA of MJE13005 transistors carefully since they are switching type transistors and are not well suited for HV linear application. If you can cool them sufficiently you might get away with 70-80mA of current at 120V. You can parallel a few more with emitter degeneration for more load current. Also if you don't have access to MCP601 opamp you can replace it LM358. But at 5V supply voltage you will need to recalculate the feedback resistor Rf to give you the required output voltage since LM358 can't output more than 3.5V. Let me know if you build this circuit.
An inverting loop made so by the mosfet inverting to positive input, your are BEGGING for oscillation, no matter the mosfet C!!!!!!!! ....SERIOSULY!!!!!!!!!!!!!!
Sorry for the late replay. To answer your question, there are two reasons. First it is very difficult to find high voltage BJTs with good Hfe. Most high voltage BJTs have very low Hfe (~5 to 30). What that means is, you have to provide substantial amount of base current to control the collector current. In this instance the op-amp would need to provide quite a bit of base current to control the collector current of a high voltage BJT for example like the main pass transistors (BUT11). Therefore I wanted to avoid to source that base current from the op-amp (MOSFETs don't require static basic current like BJTs). Of course there is a dynamic base current being sourced from the Op-amp due to the junction capacitances of the MOSFET which will leads me to the second reason. Since that transistor (IRF710) is being used as a common source (emitter if using BJT) amplifier, it has very high voltage gain. This high voltage gain varies with frequency and can cause instability in the main control loop. Typically if using a BJT in this arrangement (common emitter configuration) a small base to collector capacitor is added to introduce what is called a pole that will stabilize the control loop. BJTs have relatively low base to collector capacitance (typically called Miller capacitance). But MOSFETs have quite a bit of Miller capacitance and that capacitance actually helps to stabilize the control loop in this application. Hope this has answered your question.
@@SmithKerona thanks for the answer! I definitely plan on building one such supply but i don't think i have any high voltage bjts, maybe there are some in old crt monitor will have to look at it again
@@ljaeracer4152 Sorry for the delay making the video. Just don't seem to have enough spare time to do videos anymore. But I can tell you that 350V, 5A rated supply would be difficult to implement on the principles shown in the video since you are looking to have around 1.7kW rated power transformer (VA rating will be even worse), serious cooling for pass transistors and also expensive pass transistors that are designed for linear application. For that kind of output power, If I were you I would look into full bridge forward type switching converter.
That is incorrect! The DC bus that the 100k resistor is connected to can at best go to 200V. At that voltage, if the output is set to 0V, the 100k resistor will only dissipate 0.4 watts. And that is the worst case scenario. Now if you have watched the whole video I do mention that when I build the 800V supply, all the components need to be selected appropriately.
@@SmithKerona Ooops, my maths is a little off :) ...I get into such things because I have a pet peeve about wasted power, so my emotions got the better of me! Btw, I design electronics stuff , sometimes power supply, and it is almost always hybrid tracking, because I like clean power, but hate power waste. :)
Using the +ve input of the MC34072 as feedback (made -ve by the mosfet) is begging for oscillation. I see the 150pf in the feedback, I hope it is enough, BUT it is part of a 150pf/1,000pf divider at higher freq.
About 7 years ago I built supply with 0-300V range. TL431, CA3130, IRF710, 2SC945 and IXFH50N50.
I arranged IRF710 as common gate configuration with degeneration input resistor. Easier to fine tune and more stable.
I chose 3 secondary winding-sets transformer which provide 15V for TL431 and CA3130, 10V for biasing IXTH50N50's gate (float) and 350V for output.
Thank you
Hello sir. Can you share us your experiment ? thanks.
@@braveheart9275 Unfortunately I can't.
Hello ... O.K. and thanks for the reply.@@iblesbosuok
@@iblesbosuok Not your design?
@@stevenbliss989 At that time, I drew a circuit schematic from a blank piece of paper, based on my dumb imagination. I don't know if there has ever been a similar circuit schematic. And I don't care about that. My friend needs it and I can help him. I just need friends.
Very interesting! Any update on this? Any artefacts to share or specific things we should know about the final design?
Thanks mate your circuit solved my circuits ripple problems 👍, nice video
GOOD WORK !!! I'M CURRENTLY USING THIS CIRCUIT FOR A DESIGN !!!
Hi, greate video and explanation!. Any progress with the video and schematic for the 800VDC regulated power supply?
Nice video, you have a very nice lab!
Nice design I might build one. Thanks SK!
Just some advice if you do proceed with this project. When you say "transistor" rated at 1.2kV at 2A, I suppose you mean MOSFET. In general MOSFETs are built for switching and not linear operation (there are a very few exceptions). Look at the SOA curves for a range of so called high power MOSFETs. At 800V you'll be looking at roughly 50W of continuous dissipation - even then you'll be close to the SOA limit. You could design a linear power supply using MOSFETs to give 800V at 1 amp, but you'd need multiple parallel MOSFETs ie 16 of them. You need to consider thermal run-away as well which is dangerous because MOSFETs blow short circuit. Have a look at some DC load designs, since they share similarity with this type of power supply design.
You probably misunderstood but the "1.2kV at 2A" comment I made in the video is for the MOSFET that is interfacing the op-amp to the pass transistors not the main pass transistors. As can be seen in the video, I was using the IRF710 for this test and when I built my 800V, 100mA supply, I used IRFBG30 and it only dissipates about 1.6W at 800V. This is because it is only used as common source amplifier to interface the op-amp to the main pass transistors (see below).
Indeed I am aware that most MOSFETs are optimized for switching applications and SOA needs to be strictly overserved for linear application. That is why I built this supply in 2019 with two IXTX20N150 pass transistors :). The supply is rated at 0-800V, 0-100mA with CC and CV function. Stay tuned for an update video.
Very well executed and thoroughly tested with great effort, I am very impressed! I am just wondering why the op-amps are functioning correctly, as they are controlled by two potentials: 15V from the power supply and at the sensor input from the output potential 0-150V. I did not know that this was possible at all. Or is it a special function of this type of op-amp? I assume that both potentials are referenced to ground, is that correct?
No over temp (high risk in this), no over current or short circuit protection.
You might find what useful what a TH-cam channel "FesZ Electronics" did. :)
Great content. Funny I have just designed a 200V variable PS with CV/CC function, and the circuit is remarkably similar.
Although I'm curious as to why you have the UF4007 diode at the base of the output transistor?
Thank you and great question! The UF4007 is there to protect the output transistors's BE junction when the IRF710 switches on putting the base of the Darlington transistors at ground potential while the output filter capacitors are charged to more than 10V. As you might be aware, BJT's maximum BE junction reverse voltage rating is around 5 to 9V. Therefore the UF4007 diode will block any reverse voltage up to 1000V across the BE junction of the output transistors. I should have explained the purpose of that diode in the video.
Regarding CC operation, I will include fold-back type over-current protection set to a maximum of 100mA.
@@SmithKerona That makes total sense. I didn't think of that. Now I have to add that diode to my design. Thank you!!
Hello The LED artist, Do you mind to share you project? I am looking for something similar.
Good tutorial
How do you calculate the value of the input filter capacitor?
The circuit is quite amazing. Where I have a question is the the part of TL431, the reference terminal seem to be excluded from the schematic. Is there any reason for that i would like to know. Thanks
Hello, nice circuit! Have a question - those capacitors at the output of the rectifier have voltage rating of 250V. But if you are using transformer with output of 240V that would make something like 340V DC at the output of the rectifier if you connect that circuit directly at mains. So if one wants to build this circuit those caps should be rated at least 350V, am I right?
Thank you! And great question. Yes you are absolutely correct that if the output of the 240V transformer is to be used from 120V mains directly, those filter capacitors need to have the proper voltage rating which should be greater than 340VDC. But if you notice when I am testing the circuit at 18:45, I am using a variac (autotransformer) to supply the 120V side of that transformer so that the rectified and filtered voltage sits around 200V. If you do the math that is about 70V on the 120V rated primary winding. That is why I said at the end of the video that all the components need to be scaled up properly for high voltage operation. In fact those capacitors on the supply that I have built (0-800VDC) are made of two 450V rated capacitors in series to give me the required voltage rating. Stay tuned for an updated video that shows the implementation of 0-800VDC and 0-100mA power supply soon. Cheers!
Why did you decide to use BJT over FET for pass transistor? Couldn't find one with sufficient SOA at HV of 800V? I read in the comments that you're limiting output current to 100mA. It should be easy to find a FET capable of that. Just curious, thanks for video and stay safe.
I take it back. FET at DC 800V@100mA is a little hard and probably expensive. For someone looking for 0-500V range, here are my notes.
AOT10N60 300V@800mA, 600V@400mA to-220
AOT12N65 300V@800mA, 650V@400mA D2PAK
AOT27S60 300V@1000mA, 600V@600mA to-220 $3
STP33N65M2 300V@2000mA, 650V@1000mA to-220 $2.78 190W, no dc on SOA
AOT42S60 300V@1500mA, 600V@700mA to-220
FCP104N60 300V@1250mA, 600V@600ma to-220
I haven't explored much in the past 2 years though. Have fun.
@@CraigHollabaugh Hey sorry for the late replay! Yes you are indeed correct regarding the availability of reasonably priced high voltage MOSFETS. As a matter of fact I am using two very expensive HV MOSFETs (IXTX20N150) from Littlefuse with excellent SOA for this project. They are $20 a pop and definitely not cheap to use in hobbyist projects. Normally I like to stick with BJTs for linear applications as they are easy to stabilize in closed loop application like this project. I had a hard time stabilizing these MOSFETs with their large parasitic capacitance. But the control loop is sable now with decent transient response. Stay tuned for an update!
And thank you for providing part numbers of decent MOSFETs with good SOA for low voltage applications ;)
@@SmithKerona I looked at those IXTX mosfets but didn't need (or want) a 1000V circuit. The 20N150 is quite a part 300mA@1000V DC SOA, that's a killer part, both dangerous and awesome. I don't see my tube amp work going much above 500V. Look forward to your upcoming work. Thanks for the reply.
WoW How much all those equipment costed? You have like two or three backup of each.
Not as much as you might think. Most of the items I acquired them in disrepair/broken state. Likely I manged to repair most of them and some still left for spare.
Hi ! Congratulations for your project !! I just smoked six TL783 ICs trying to assemble a 1,5V -120V high power supply as depicted on their datasheet .Even a MJE11015 got shorted . I gave up to use this " wonderful "IC and backing to use discrete components like you did. Snag is that I could not find MC34072 and IRF710 . Any suggestion ?
Hello, have a look at the schematic that I came up with using low voltage single supply opamp (MCP601) and HV BJT (MJE13005). drive.google.com/file/d/1AaTodBA69dSlvdN47gFvb6UYdPmTPd4l/view?usp=sharing
Make sure to observe SOA of MJE13005 transistors carefully since they are switching type transistors and are not well suited for HV linear application. If you can cool them sufficiently you might get away with 70-80mA of current at 120V. You can parallel a few more with emitter degeneration for more load current. Also if you don't have access to MCP601 opamp you can replace it LM358. But at 5V supply voltage you will need to recalculate the feedback resistor Rf to give you the required output voltage since LM358 can't output more than 3.5V. Let me know if you build this circuit.
An inverting loop made so by the mosfet inverting to positive input, your are BEGGING for oscillation, no matter the mosfet C!!!!!!!! ....SERIOSULY!!!!!!!!!!!!!!
I know this is quite a while ago but why are you using the IRF710 ? is there any particular reason you're using a mosfet there instead of a bjt?
Sorry for the late replay. To answer your question, there are two reasons. First it is very difficult to find high voltage BJTs with good Hfe. Most high voltage BJTs have very low Hfe (~5 to 30). What that means is, you have to provide substantial amount of base current to control the collector current. In this instance the op-amp would need to provide quite a bit of base current to control the collector current of a high voltage BJT for example like the main pass transistors (BUT11). Therefore I wanted to avoid to source that base current from the op-amp (MOSFETs don't require static basic current like BJTs). Of course there is a dynamic base current being sourced from the Op-amp due to the junction capacitances of the MOSFET which will leads me to the second reason. Since that transistor (IRF710) is being used as a common source (emitter if using BJT) amplifier, it has very high voltage gain. This high voltage gain varies with frequency and can cause instability in the main control loop. Typically if using a BJT in this arrangement (common emitter configuration) a small base to collector capacitor is added to introduce what is called a pole that will stabilize the control loop. BJTs have relatively low base to collector capacitance (typically called Miller capacitance). But MOSFETs have quite a bit of Miller capacitance and that capacitance actually helps to stabilize the control loop in this application. Hope this has answered your question.
@@SmithKerona thanks for the answer!
I definitely plan on building one such supply but i don't think i have any high voltage bjts, maybe there are some in old crt monitor will have to look at it again
can i use 3A output current in this this design?
Can you provide the schema of this?
Did you finish the tutorials for component calculations?
Yes I did. I just didn't have time to make a video about it. I will try to link the schematic here when I have a chance.
@@SmithKerona please do so. much appreciated. the tutorial for component calculation would be beneficial for us hobbyist.
@@SmithKerona any updates on this project? I have some interest in seeing what's needed to scale this up to 0-350v 5a cc/cv adjustable power supply.
@@ljaeracer4152 Sorry for the delay making the video. Just don't seem to have enough spare time to do videos anymore. But I can tell you that 350V, 5A rated supply would be difficult to implement on the principles shown in the video since you are looking to have around 1.7kW rated power transformer (VA rating will be even worse), serious cooling for pass transistors and also expensive pass transistors that are designed for linear application. For that kind of output power, If I were you I would look into full bridge forward type switching converter.
Wow, the 100k will be dissipating 6.4Watts when the voltage is set to 0V output!
That is incorrect! The DC bus that the 100k resistor is connected to can at best go to 200V. At that voltage, if the output is set to 0V, the 100k resistor will only dissipate 0.4 watts. And that is the worst case scenario. Now if you have watched the whole video I do mention that when I build the 800V supply, all the components need to be selected appropriately.
But for the purpose of this video and this particular circuit, the 100K resistor is more than adequate.
@@SmithKerona Ooops, my maths is a little off :) ...I get into such things because I have a pet peeve about wasted power, so my emotions got the better of me! Btw, I design electronics stuff , sometimes power supply, and it is almost always hybrid tracking, because I like clean power, but hate power waste. :)
@@SmithKerona Yep! :)
Using the +ve input of the MC34072 as feedback (made -ve by the mosfet) is begging for oscillation.
I see the 150pf in the feedback, I hope it is enough, BUT it is part of a 150pf/1,000pf divider at higher freq.