I think I saw this item on eBay, I could not afford the price though at the time. I want a high voltage power supply for working on vacuum tube circuits. I am planning to build one using similar architecture with toroidal transformer and high voltage N channel MOSFET's. Shahriar I love your style, admire your proficiency, and respect your vast electronics knowledge. Great video again , I look forward to each and every new video.
I have one. Very precise and useful for my experiments with Nixie tubes. It survived some unintentional abuse, such as a sudden 300V dead short while I was testing some IN-28 lamps. Whoopsie. My only complaint is the user interface. To set a voltage, for example, you have to press "VSET", type the number and then press "Enter". On My Rigol DP821, for comparison, you just type the number and press "V" (for "Volts"), which is enough to let the PSU understand that you want to set a voltage. One less keystroke. It is even more annoying with current limits, since this PSU handles low currents but only accepts values in Amps (or, at least, I've not found a way to specify them in mA), so you have to type cumbersome values like .005. On the Rigol it is as simple as pressing "5" and then "mA". Done. The lack of a rotary knob is also annoying when you want to slowly increase a voltage. You can do it with arrow keys, but it feels more awkward. Just my 0.02€
That's a very useful power supply for what I like to work on, old valve (tube) gear. Reforming electrolytic capacitors is something it could find an instant use for here. Thanks for showing it.
excellent job. My test power supply is an old Lambda LQD-422. Dual output with dual volts/amps display. All linear of course. About 35 yrs old and works perfectly.
Fortuitous purchase and repair. I wonder if the power socket was removed to prevent accidental use? A little bit annoying, but your replacement is probably much better than the original socket and I suspect that the wires you used were a touch of overkill as the input power is only in the region of 150 Watts.
Thanks for sharing the video. How about a video on performing curve traces with a PSU and computer via USB. Maybe something like PyPSUTracer on diyaudio.
It might be interesting to do some characterisation of high voltage capacitors. Their characteristic are quite different to low voltages ones due to different compromises in materials at higher voltages.
Measuring capacitor leakages at / near their voltage limits would be nice with this. Very swish piece of equipment! Finding out what its output ripple would be under some normal loads would've been nice.
Nice unit, but it seems rather higher voltage than you need for your typical experiments with solid-state and optical communications hardware. Testing of possible protection circuits for front-ends? Exotic microwave electron tubes?
Sorry to sound dumb , but a fuse on the line and neutral is bad is it not ? Fused neutrals went out in the uk many many decades ago, at the very least before 1955. I mean if the neutral fuse blows and the line does not you could get a sock if the device is earthed .
So, if we cut out the acronyms (I'm not from US so have no idea about your electrician code) you wanted to say that you use two fuses for devices that are powered by phase-to-phase (which is for example 220V in domestic US centre-tapped thingie or typical European 400V from L1 to L2 or whatever from what I understand)?
Is there any reason to add the tvs diode? Because your replacement mosfet has already a built in back to back gate-source zener diode. I am asking because I used the same STW15NK90Z on an omron smps where the external 2Sk3875 gate-source diode shorted. I just replaced with stw15nk90z coming with the integrated diode.
external diode can be slightly lower voltage than the internal, and have a lot better pulse capability, so the energy driven into the internal diode is a lot less, which, given that the internal diodes are tiny and with limited pulse handling, can mean the difference between it clamping and surviving, or simply blowing up and killing the drive side along with the power switch.
10:19 Bottom left a MELF diode bodge with a SMD resistor tacked onto that both flapping in the breeze with no mechanical support to speak of. If they had put a glob of silastic over the components at least it would show they gave a little bit of a crap. It's disappointing to see that from BK more of what I'd expect on a no name supply from China.
The sad thing is scraping off part numbers won't stop someone determined. I don't get why manufacturers even bother. The micro controller for example determine where the unique pins are VCC, GND, oscillator inputs, Reset, etc and cross reference that to the pin out from common micro controller manufacturers. Once you've determined the manufacturer and the family it belongs to from the pin out query the chip for it's unique ID number (with a PICKIT 3, AVRISP, etc) and you have the exact micro controller part number. All of the real work is in the code anyway who cares if they know the micro part number just enable code protection and leave it at that IMO.
Bushougoma That takes a lot more time than go in and read the part numbers. Nothing is 100% safe and this stupid grinding is just to delay cheap ripoffs or make them harder so it might be not worth it. It is all just an overall economic benefit if it works say for 2 month on a certain product that generates this and that revenue and that can be more than is needed for grinding part numbers away. Might however be to also stop people from seeing cheap crap inside there
The delay that grinding provides is in the order of hours, not month. If someone wants to rip off such a product, we are talking about people who know what they are doing (after all you need to understand the whole system to program a working software for it). It is 100% clear what each IC is supposed to do just by looking at the circuit. And with a little bit of experience you most likely already have your list of "suspects". Figuring out what a chip is, is a one time thing. Grinding on the other hand has to be done on every single device manufactured. So, yeah, the only thing left is that they want to hide cheap off the shelf components (which in no way means they are bad).
I am also curious about the addition of the diode. Was there one originally? Was it just a precaution or was it absolutely necessary? (Great video BTW)
Hi Shahriar, at 12:09 in the video the resistors, (current balancing, perhaps degenerative feedback?) you installed are what appear to be 101 or 100 ohms. All of the rest of the parallel devices are using 1000 ohm. Are you sure the replaced resistors are correct? BTW love the cat! All the best.
Yes, they are the same resistor value. This is because the resistors with 1000 marking on them are higher precision resistors and have 4 numbers. The last number is the multiplier. Therefore 1000 marking is 100 x 10^0 which is 100 ohms. The three number resistors are 10 x 10^1 which is also 100 Ohms.
One of my cats could be Pooch's twin, though without the white patch. He's also intensely interested in whatever I'm working on when I'm at the bench. And one of the other cats also seems strangely drawn by the Q--tips, too. Must be in the DNA in there, somewhere..
They look like little bones to them. My other cat stole a chicken bone the other day and ate it. They can have the shape of q-tips. It also had en entire magpie that crashed into the window and was already dying.
The most interesting I find about this supply, is that they are using MOSFETs that are NOT rated for DC linear operation. That is simply wrong and may have been the cause of the blowup in the first place.
Linear operation is the most basic form of operation of any transistor. They do not need to be specifically rated for it. As long as the device current and power dissipation limits are not exceeded, there is no reason why those MOSFETs cannot be used for this job.
Of course, there IS a reason. The result of designers not getting grasp of it you see in the video: transistor blown. Vertical mosfets are NOT designed to be used in a DC application. The internal mosfet structure does not allow the current to be shared equally across the die, when it is operated in linear mode - leading to a local meltdown of the die. Modern switching optimized mosfet use a multiple hundreds of minitransistors in parallel on the die.
I think you should take a look at modern day high end ATX PSU... Because that 'switching mode Power Supplys have high noise' thought could be broken with that. Modern things are at under 20mV at full load (=650W or so) on the +12V rail and on the +5V and 3,3V rails the best are at around 15mV or under...
Is Pooch a Korat ? (he looks just like my cat!). (I have to put the screws away when I open a device or I latter find I'm short as Eccles has carried one or more off 8-)
In production shops where they do testing, there is an expansion card to give it Ethernet. With that, you can have a device under test submitted to various voltage/current profiles to see what happens. Ethernet is for control
Your soldering joints don't look well at all. The surface looks like a cold junction, and has an irregular shape, not sphere-like. Either your soldering iron was not strong enough for the FETs mounted already to the heat sink, or have you used lead free solder? Anyhow, I recommend to redo these. Otherwise, a fine teardown. root cause analysis, and repair.
Nice repair, and good to see a proper cat scan of the burned resistors. My dog (Sophie the "lab" dog) like to steal gloves ;-)
A video with pooch helping is always a great video👍
Always good to start with a cat scan...
I think I saw this item on eBay, I could not afford the price though at the time. I want a high voltage power supply for working on vacuum tube circuits. I am planning to build one using similar architecture with toroidal transformer and high voltage N channel MOSFET's. Shahriar I love your style, admire your proficiency, and respect your vast electronics knowledge. Great video again , I look forward to each and every new video.
Being an electronics technician I really appreciate the detail you put into each one of your videos. thank you
Total side note: You have a beautiful cat! (interesting video btw)
Come clean. Pooch is the real brains of this channel.
I have one. Very precise and useful for my experiments with Nixie tubes. It survived some unintentional abuse, such as a sudden 300V dead short while I was testing some IN-28 lamps. Whoopsie.
My only complaint is the user interface. To set a voltage, for example, you have to press "VSET", type the number and then press "Enter". On My Rigol DP821, for comparison, you just type the number and press "V" (for "Volts"), which is enough to let the PSU understand that you want to set a voltage. One less keystroke.
It is even more annoying with current limits, since this PSU handles low currents but only accepts values in Amps (or, at least, I've not found a way to specify them in mA), so you have to type cumbersome values like .005. On the Rigol it is as simple as pressing "5" and then "mA". Done. The lack of a rotary knob is also annoying when you want to slowly increase a voltage. You can do it with arrow keys, but it feels more awkward. Just my 0.02€
Should definitely test it against an Switched Mode Power Supply and check for the noise in both.
That's a very useful power supply for what I like to work on, old valve (tube) gear. Reforming electrolytic capacitors is something it could find an instant use for here. Thanks for showing it.
excellent job. My test power supply is an old Lambda LQD-422. Dual output with dual volts/amps display. All linear of course. About 35 yrs old and works perfectly.
Very nice repair job. I love your workmanship.
That’s an awesome supply, I might have to see if I can pick one up too!
Fortuitous purchase and repair. I wonder if the power socket was removed to prevent accidental use? A little bit annoying, but your replacement is probably much better than the original socket and I suspect that the wires you used were a touch of overkill as the input power is only in the region of 150 Watts.
He tried to help you to keep desk clean and organized! ;)
Very nice repair video, thanks for sharing.
You should try a new Keysight handheld DMM and replace the old Fluke.
It is so slow !
Thanks for sharing the video. How about a video on performing curve traces with a PSU and computer via USB. Maybe something like PyPSUTracer on diyaudio.
Awwwww look at that cute kitty tail twitching curiously!
Are the rectifiers around the 2:50 mark FFUUUULLLLL BRIIIDDGGGEEEE?
And here we have a..
FULL BRIDGE RECTIFIER!!!
No. They're full bridge rectumfriers...
Please stress test some X5R or X7R ceramic capacitors, 25V or 50V. I couldn't get high enough for any sort of leakage here.
It might be interesting to do some characterisation of high voltage capacitors. Their characteristic are quite different to low voltages ones due to different compromises in materials at higher voltages.
Measuring capacitor leakages at / near their voltage limits would be nice with this. Very swish piece of equipment! Finding out what its output ripple would be under some normal loads would've been nice.
6:06 33ohm purr transistor... :D
Pooch reminds me of my dog, is too funny. If I drop any paper towels, cotton swabs, etc. the dog snatches it up runs and eats it.
With high voltage like that you could have some fun with RF modulation and demodulation using heterodyne tubes, like pentagrid converters or hexodes.
Nice unit, but it seems rather higher voltage than you need for your typical experiments with solid-state and optical communications hardware. Testing of possible protection circuits for front-ends? Exotic microwave electron tubes?
Great video.
Any particular reason "TSP#" designators were dropped from video titles?
amazing as always!
it would be of interest what you paid for the broken equipment
I think I paid $300.
Great video, thanks!
Your cat can smell the burned components. That explains the interest.
Think the cat has a secret ether addiction. Looking for used cuetips
Sorry to sound dumb , but a fuse on the line and neutral is bad is it not ? Fused neutrals went out in the uk many many decades ago, at the very least before 1955. I mean if the neutral fuse blows and the line does not you could get a sock if the device is earthed .
Can I get a sock? I'd like an argyle sock, please.
Argyle socks went out with braces and spats.
Boonedock Journeyman you already made the helpful reply.
So, if we cut out the acronyms (I'm not from US so have no idea about your electrician code) you wanted to say that you use two fuses for devices that are powered by phase-to-phase (which is for example 220V in domestic US centre-tapped thingie or typical European 400V from L1 to L2 or whatever from what I understand)?
I'm a bit disappointed that you didn't measure the output noise, or do a simple transient test...?
Is there any reason to add the tvs diode? Because your replacement mosfet has already a built in back to back gate-source zener diode.
I am asking because I used the same STW15NK90Z on an omron smps where the external 2Sk3875 gate-source diode shorted. I just replaced with stw15nk90z coming with the integrated diode.
external diode can be slightly lower voltage than the internal, and have a lot better pulse capability, so the energy driven into the internal diode is a lot less, which, given that the internal diodes are tiny and with limited pulse handling, can mean the difference between it clamping and surviving, or simply blowing up and killing the drive side along with the power switch.
The reaction time of a tvs is far better than a zener diode.
Power handling is way better too.
10:19 Bottom left a MELF diode bodge with a SMD resistor tacked onto that both flapping in the breeze with no mechanical support to speak of. If they had put a glob of silastic over the components at least it would show they gave a little bit of a crap.
It's disappointing to see that from BK more of what I'd expect on a no name supply from China.
That, and scratching markings from components. Both things that give a bad vibe.
superdau They just don't have the bucks to do it like HP/Agilent/Keysight with their own obscured part numbers to protect against cheap ripoffs
The sad thing is scraping off part numbers won't stop someone determined. I don't get why manufacturers even bother.
The micro controller for example determine where the unique pins are VCC, GND, oscillator inputs, Reset, etc and cross reference that to the pin out from common micro controller manufacturers. Once you've determined the manufacturer and the family it belongs to from the pin out query the chip for it's unique ID number (with a PICKIT 3, AVRISP, etc) and you have the exact micro controller part number.
All of the real work is in the code anyway who cares if they know the micro part number just enable code protection and leave it at that IMO.
Bushougoma That takes a lot more time than go in and read the part numbers. Nothing is 100% safe and this stupid grinding is just to delay cheap ripoffs or make them harder so it might be not worth it. It is all just an overall economic benefit if it works say for 2 month on a certain product that generates this and that revenue and that can be more than is needed for grinding part numbers away. Might however be to also stop people from seeing cheap crap inside there
The delay that grinding provides is in the order of hours, not month. If someone wants to rip off such a product, we are talking about people who know what they are doing (after all you need to understand the whole system to program a working software for it). It is 100% clear what each IC is supposed to do just by looking at the circuit. And with a little bit of experience you most likely already have your list of "suspects".
Figuring out what a chip is, is a one time thing. Grinding on the other hand has to be done on every single device manufactured.
So, yeah, the only thing left is that they want to hide cheap off the shelf components (which in no way means they are bad).
That Fluke model is slow, that isnt annoying for you?
On full auto-range it is slow and can be annoying for sure.
Excellent
I am also curious about the addition of the diode. Was there one originally? Was it just a precaution or was it absolutely necessary? (Great video BTW)
Hi Shahriar, at 12:09 in the video the resistors, (current balancing, perhaps degenerative feedback?) you installed are what appear to be 101 or 100 ohms. All of the rest of the parallel devices are using 1000 ohm. Are you sure the replaced resistors are correct? BTW love the cat! All the best.
Yes, they are the same resistor value. This is because the resistors with 1000 marking on them are higher precision resistors and have 4 numbers. The last number is the multiplier. Therefore 1000 marking is 100 x 10^0 which is 100 ohms. The three number resistors are 10 x 10^1 which is also 100 Ohms.
HA! As soon as I entered that replay I had the uneasy feeling I was incorrect. Thanks for the explanation. Again all the best!
This was a fun episode. I have no idea what I would do with 600v (other than blow my fingers off)
One of my cats could be Pooch's twin, though without the white patch. He's also intensely interested in whatever I'm working on when I'm at the bench. And one of the other cats also seems strangely drawn by the Q--tips, too. Must be in the DNA in there, somewhere..
Nice job.
My cat loves qtips as well...
Oh, that's why you are so great at finding problems, you use cat-scan:-)
How about millimetre wave tubes like a klystron or magnetron video
My cat also steals Q-tips from my lab table. And bags with little shiny TO-220 heatsinks. No idea why....
They look like little bones to them. My other cat stole a chicken bone the other day and ate it. They can have the shape of q-tips. It also had en entire magpie that crashed into the window and was already dying.
Lol my cat used to, then one day he decided to play with an energized output... he doesn't seem to want to get on the bench anymore....
Does the USB interface work?
Yes, works fine.
So you could sweep it through a range of currents or voltages and read the corresponding voltage or current, over the USB?
If you mistake those two it could be your last one ahahah fantastic
What happened to pootch the repaircat? In "newer" videos is he missing!
The most interesting I find about this supply, is that they are using MOSFETs that are NOT rated for DC linear operation. That is simply wrong and may have been the cause of the blowup in the first place.
Linear operation is the most basic form of operation of any transistor. They do not need to be specifically rated for it. As long as the device current and power dissipation limits are not exceeded, there is no reason why those MOSFETs cannot be used for this job.
Of course, there IS a reason. The result of designers not getting grasp of it you see in the video: transistor blown. Vertical mosfets are NOT designed to be used in a DC application. The internal mosfet structure does not allow the current to be shared equally across the die, when it is operated in linear mode - leading to a local meltdown of the die. Modern switching optimized mosfet use a multiple hundreds of minitransistors in parallel on the die.
I think you should take a look at modern day high end ATX PSU...
Because that 'switching mode Power Supplys have high noise' thought could be broken with that. Modern things are at under 20mV at full load (=650W or so) on the +12V rail and on the +5V and 3,3V rails the best are at around 15mV or under...
Is Pooch a Korat ? (he looks just like my cat!). (I have to put the screws away when I open a device or I latter find I'm short as Eccles has carried one or more off 8-)
Toroid transformers are sexy sexy sexy 😍
FeedbackLoop manually rewound one when 1 tap shorted, boy was that a big video because of all the work.
👍👍
For a guy that can repair a Tek RSA 6114A this was no challenge.
do some microwave/ optical experiment
You had to order the resistors? tsk ;)
I always keep my cats away from electronics when I work on them. I am afraid of a static charged cat zapping components.
MichaelKingsfordGray
The cats might have nine lives but, sadly, the components don't! 😸⚡
Why does a power supply need an IP address.
In production shops where they do testing, there is an expansion card to give it Ethernet. With that, you can have a device under test submitted to various voltage/current profiles to see what happens. Ethernet is for control
because of the "Internet of things", so you can destroy the connected device with a control command over the Internet with 500V ;-)
Cat :)
Q-tip thief haha...
Your soldering joints don't look well at all. The surface looks like a cold junction, and has an irregular shape, not sphere-like.
Either your soldering iron was not strong enough for the FETs mounted already to the heat sink, or have you used lead free solder?
Anyhow, I recommend to redo these.
Otherwise, a fine teardown. root cause analysis, and repair.