Please evaluate this brick for DRSSTC suitability /anything to note please?
astralhighway, Wed Oct 22 2014, 10:43PMI have obtained one of these[http://ixdev.ixys.com/DataSheet/l152.pdf][/ link] brick devices, NOS. I'm not intending to use it anything close to its capacity but it's now mine, so I'm curious.
Can this candidate work? I have wound a secondary and fabricated a topload, identical to EVR's microbrute design. Sec is approx 10 in winding length, pri 3-4 ins dia, 3 turns. MMC is 75nF.
Switching topology exactly follows EVRs half-bridge design, including external PWM. I also read good feedback somewhere about Steve Ward's recommendation of adding of a 7K-ish resistor in parallel with the primary.
(I crudely measured a fundamental of just 40Khz fres with this coil and small, 15 pF ish topload - seems very low, but that's what I got. There were so many harmonics , way up...
Any comments very welcome, thank you.
Re: Please evaluate this brick for DRSSTC suitability /anything to note please?
GrantX, Thu Oct 23 2014, 03:37AM
A secondary of similar dimensions to EVR's Microbrute normally wouldn't have a self resonant frequency that low. I have a 295mm x 50mm secondary wound with ~1400 turns and without a toroid (only a small brass terminal) that was somewhere around 630kHz. The addition of an 200mm aluminium toroid dropped the frequency to below 500kHz.
What is your method for measuring the secondary's resonant frequency? Remember that any object near the secondary ("near" as in within a metre or so) will add parasitic capacitance and will drop the frequency considerably. And watch out for harmonics, both above and below the actual resonant frequency. This is a decent guide for measuring a coil with an oscilloscope and function generator: http://www.hvtesla.com/tuning.html. You first need a rough idea of the resonant frequency, using JavaTC or similar.
As for the IGBT, in the datasheet's features list it says "switching frequency up to 30 kHz," but I think that is for hard switching at full power. It's switching delays are: Td(on) 100ns, Tr 50ns, Td(off) 650ns, Tf 50ns, for a total delay of 850 nanoseconds, which is a little bit faster than the venerable CM300's used in much larger coils. You might be able to coerce it into switching beyond 100kHz if you were careful and didn't abuse it with huge power levels.
You could make a test rig to drive the IGBT with a variable frequency source and very low DC bus power, and scope the output over a dummy load to see what the switching waveforms look like, if you really want to see what it's limits are.
GrantX, Thu Oct 23 2014, 03:37AM
A secondary of similar dimensions to EVR's Microbrute normally wouldn't have a self resonant frequency that low. I have a 295mm x 50mm secondary wound with ~1400 turns and without a toroid (only a small brass terminal) that was somewhere around 630kHz. The addition of an 200mm aluminium toroid dropped the frequency to below 500kHz.
What is your method for measuring the secondary's resonant frequency? Remember that any object near the secondary ("near" as in within a metre or so) will add parasitic capacitance and will drop the frequency considerably. And watch out for harmonics, both above and below the actual resonant frequency. This is a decent guide for measuring a coil with an oscilloscope and function generator: http://www.hvtesla.com/tuning.html. You first need a rough idea of the resonant frequency, using JavaTC or similar.
As for the IGBT, in the datasheet's features list it says "switching frequency up to 30 kHz," but I think that is for hard switching at full power. It's switching delays are: Td(on) 100ns, Tr 50ns, Td(off) 650ns, Tf 50ns, for a total delay of 850 nanoseconds, which is a little bit faster than the venerable CM300's used in much larger coils. You might be able to coerce it into switching beyond 100kHz if you were careful and didn't abuse it with huge power levels.
You could make a test rig to drive the IGBT with a variable frequency source and very low DC bus power, and scope the output over a dummy load to see what the switching waveforms look like, if you really want to see what it's limits are.
Re: Please evaluate this brick for DRSSTC suitability /anything to note please?
loneoceans, Thu Oct 23 2014, 05:19AM
The IGBTs look like they'll do fine for a medium-small coil :)
loneoceans, Thu Oct 23 2014, 05:19AM
astralhighway wrote ...
I have obtained one of these[http://ixdev.ixys.com/DataSheet/l152.pdf][/ link] brick devices, NOS. I'm not intending to use it anything close to its capacity but it's now mine, so I'm curious.
Can this candidate work? I have wound a secondary and fabricated a topload, identical to EVR's microbrute design. Sec is approx 10 in winding length, pri 3-4 ins dia, 3 turns. MMC is 75nF.
Switching topology exactly follows EVRs half-bridge design, including external PWM. I also read good feedback somewhere about Steve Ward's recommendation of adding of a 7K-ish resistor in parallel with the primary.
(I crudely measured a fundamental of just 40Khz fres with this coil and small, 15 pF ish topload - seems very low, but that's what I got. There were so many harmonics , way up...
Any comments very welcome, thank you.
I have obtained one of these[http://ixdev.ixys.com/DataSheet/l152.pdf][/ link] brick devices, NOS. I'm not intending to use it anything close to its capacity but it's now mine, so I'm curious.
Can this candidate work? I have wound a secondary and fabricated a topload, identical to EVR's microbrute design. Sec is approx 10 in winding length, pri 3-4 ins dia, 3 turns. MMC is 75nF.
Switching topology exactly follows EVRs half-bridge design, including external PWM. I also read good feedback somewhere about Steve Ward's recommendation of adding of a 7K-ish resistor in parallel with the primary.
(I crudely measured a fundamental of just 40Khz fres with this coil and small, 15 pF ish topload - seems very low, but that's what I got. There were so many harmonics , way up...
Any comments very welcome, thank you.
The IGBTs look like they'll do fine for a medium-small coil :)
Re: Please evaluate this brick for DRSSTC suitability /anything to note please?
astralhighway, Thu Oct 23 2014, 07:14AM
GrantX, thank you. I used a signal generator with anti-parallel LEDs in the output and swept up and down until I hit what appeared to be the brightest output. But I hadn't thought in terms of harmonics below the fundamental. I'll now follow your suggested method instead. And I'll watch out this time for my impact on the measurement (i was about 30 cm away)
I'll also think about making some kind of test rig for the IGBTs. I was curious as to why they are tagged only 30Khz when in fact their total rise time, switch off time and dead time is still very short in comparison.
I'll check out Java TC first. Thanks again
loneoceans, thanks for the reassurance!
Well, from your total delay of 850 ns (let's round it up to 1uS for good measure) we should be able to switch at 1MHz, right????
astralhighway, Thu Oct 23 2014, 07:14AM
GrantX, thank you. I used a signal generator with anti-parallel LEDs in the output and swept up and down until I hit what appeared to be the brightest output. But I hadn't thought in terms of harmonics below the fundamental. I'll now follow your suggested method instead. And I'll watch out this time for my impact on the measurement (i was about 30 cm away)
I'll also think about making some kind of test rig for the IGBTs. I was curious as to why they are tagged only 30Khz when in fact their total rise time, switch off time and dead time is still very short in comparison.
I'll check out Java TC first. Thanks again
loneoceans, thanks for the reassurance!
GrantX wrote ...
...as for the IGBT, in the datasheet's features list it says "switching frequency up to 30 kHz," but I think that is for hard switching at full power. It's switching delays are: Td(on) 100ns, Tr 50ns, Td(off) 650ns, Tf 50ns, for a total delay of 850 nanoseconds, which is a little bit faster than the venerable CM300's used in much larger coils. You might be able to coerce it into switching beyond 100kHz if you were careful and didn't abuse it with huge power levels.
You could make a test rig to drive the IGBT with a variable frequency source and very low DC bus power, and scope the output over a dummy load to see what the switching waveforms look like, if you really want to see what it's limits are.
...as for the IGBT, in the datasheet's features list it says "switching frequency up to 30 kHz," but I think that is for hard switching at full power. It's switching delays are: Td(on) 100ns, Tr 50ns, Td(off) 650ns, Tf 50ns, for a total delay of 850 nanoseconds, which is a little bit faster than the venerable CM300's used in much larger coils. You might be able to coerce it into switching beyond 100kHz if you were careful and didn't abuse it with huge power levels.
You could make a test rig to drive the IGBT with a variable frequency source and very low DC bus power, and scope the output over a dummy load to see what the switching waveforms look like, if you really want to see what it's limits are.
Well, from your total delay of 850 ns (let's round it up to 1uS for good measure) we should be able to switch at 1MHz, right????
Re: Please evaluate this brick for DRSSTC suitability /anything to note please?
dexter, Thu Oct 23 2014, 11:17AM
if you switch at 1MHz then there is almost no on time...
usually alter summing and inverting all the times from the datasheet to get the frequency that number is divided by 10 or more and that will be the max operating frequency
dexter, Thu Oct 23 2014, 11:17AM
if you switch at 1MHz then there is almost no on time...
usually alter summing and inverting all the times from the datasheet to get the frequency that number is divided by 10 or more and that will be the max operating frequency
Re: Please evaluate this brick for DRSSTC suitability /anything to note please?
astralhighway, Fri Oct 24 2014, 02:24AM
Right, so retesed Fres. Secondary + topload is 398Khz, call it 400Khz/ See what you mean about the effect of close-by objects! I had to stand 2 metres away to not affect this.
Primary + series MMC had a couple of bumps 40Khz either side of 400Khz.
Hmm. Do I now have to acquire a new IGBT brick? Has to be 2 devices in one brick package as I've started the DC Bus round this design.
Thanks!
astralhighway, Fri Oct 24 2014, 02:24AM
Right, so retesed Fres. Secondary + topload is 398Khz, call it 400Khz/ See what you mean about the effect of close-by objects! I had to stand 2 metres away to not affect this.
Primary + series MMC had a couple of bumps 40Khz either side of 400Khz.
Hmm. Do I now have to acquire a new IGBT brick? Has to be 2 devices in one brick package as I've started the DC Bus round this design.
Thanks!
Re: Please evaluate this brick for DRSSTC suitability /anything to note please?
loneoceans, Fri Oct 24 2014, 09:51PM
The IGBTs will not be happy at 400khz, but will do great if it were at your original 40kHz. I would either swap out the transistors to something faster like TO-247 IGBTs, or made a much lower frequency coil no higher than around 100kHz, lower if possible.
loneoceans, Fri Oct 24 2014, 09:51PM
The IGBTs will not be happy at 400khz, but will do great if it were at your original 40kHz. I would either swap out the transistors to something faster like TO-247 IGBTs, or made a much lower frequency coil no higher than around 100kHz, lower if possible.
Re: Please evaluate this brick for DRSSTC suitability /anything to note please?
astralhighway, Fri Oct 24 2014, 10:07PM
Ah, shame. I've already blown up 40N60 minibricks. What T0-247's would you recommend and why does the package make a difference? Just curious
I am operating from 120V doubler bus, if that helps.
Many devices I've looked at don't have any faster recommended switching frequency from the mfr. I do want to migrate to a larger coil later on, so my brick won't be wasted, but I want to get this build right and master the magnetics side at the same time.
Thank you
astralhighway, Fri Oct 24 2014, 10:07PM
loneoceans wrote ...
I would ... swap out the transistors to something faster like TO-247 IGBTs...
I would ... swap out the transistors to something faster like TO-247 IGBTs...
Ah, shame. I've already blown up 40N60 minibricks. What T0-247's would you recommend and why does the package make a difference? Just curious
I am operating from 120V doubler bus, if that helps. Many devices I've looked at don't have any faster recommended switching frequency from the mfr. I do want to migrate to a larger coil later on, so my brick won't be wasted, but I want to get this build right and master the magnetics side at the same time.
Thank you
Re: Please evaluate this brick for DRSSTC suitability /anything to note please?
RateReducer, Sat Oct 25 2014, 07:57AM
minibricks are much more rugged and nearly as fast as TO247 IGBTs so i guess it is a better idea to improve your gate driver respectively your whole setup before u use/kill other IGBTs ^^
Smaller packages are faster and have a lower gate charge than IGBT bricks but they can not switch such a high amount of power or current than the big IGBT bricks. Those bricks are not that fast and have a higher gate charge than the small TO247 but they can handle much more power dissipation. So for a small coil TO247 are better because they are faster and for a big coil IGBT bricks are the better choice because the can handle higher power levels.
RateReducer, Sat Oct 25 2014, 07:57AM
Ah, shame. I've already blown up 40N60 minibricks. What T0-247's would you recommend and why does the package make a difference? Just curious I am operating from 120V doubler bus, if that helps.
minibricks are much more rugged and nearly as fast as TO247 IGBTs so i guess it is a better idea to improve your gate driver respectively your whole setup before u use/kill other IGBTs ^^
Smaller packages are faster and have a lower gate charge than IGBT bricks but they can not switch such a high amount of power or current than the big IGBT bricks. Those bricks are not that fast and have a higher gate charge than the small TO247 but they can handle much more power dissipation. So for a small coil TO247 are better because they are faster and for a big coil IGBT bricks are the better choice because the can handle higher power levels.
Re: Please evaluate this brick for DRSSTC suitability /anything to note please?
loneoceans, Sat Oct 25 2014, 05:01PM
The actual package itself doesn't make much of a difference, since they are nothing more just a way to package the actual silicon die (however the bigger the package, the circuit inductance will increase slightly). Some of the old IGBTs are actually quite rugged, probably due to the large physical die size, and high manufacturing de-rating. as the technology improves, they've managed to shrink the die in a more efficient package but as a result, are 'less overengineered'. The original 40N60 MiniBlocs from half a decade ago used to be very rugged but I don't think they make them anymore, and now offer a similar one in a TO247 package. A similar good miniBloc is the 60N60C2D1. Regardless, they're usually quite expensive so I just go for TO247s. I've found FGA60N65SMD to be good value due to their low cost. IRGP50B60PD1 from IR is a great choice as well, along with the similar FGH60N60SMD from Fairchild. The 30N60 and 40N60 line from Fairchild are also ok for a small coil. I've also be meaning to try the IXGH60N60C3D1 from IXYS. They have some very nice PT and XPT IGBTs. Would be happy to hear from others what their favourite transistors are.
loneoceans, Sat Oct 25 2014, 05:01PM
astralhighway wrote ...
Ah, shame. I've already blown up 40N60 minibricks. What T0-247's would you recommend and why does the package make a difference? Just curious I am operating from 120V doubler bus, if that helps.
Many devices I've looked at don't have any faster recommended switching frequency from the mfr. I do want to migrate to a larger coil later on, so my brick won't be wasted, but I want to get this build right and master the magnetics side at the same time.
Thank you
loneoceans wrote ...
I would ... swap out the transistors to something faster like TO-247 IGBTs...
I would ... swap out the transistors to something faster like TO-247 IGBTs...
Ah, shame. I've already blown up 40N60 minibricks. What T0-247's would you recommend and why does the package make a difference? Just curious
I am operating from 120V doubler bus, if that helps. Many devices I've looked at don't have any faster recommended switching frequency from the mfr. I do want to migrate to a larger coil later on, so my brick won't be wasted, but I want to get this build right and master the magnetics side at the same time.
Thank you
The actual package itself doesn't make much of a difference, since they are nothing more just a way to package the actual silicon die (however the bigger the package, the circuit inductance will increase slightly). Some of the old IGBTs are actually quite rugged, probably due to the large physical die size, and high manufacturing de-rating. as the technology improves, they've managed to shrink the die in a more efficient package but as a result, are 'less overengineered'. The original 40N60 MiniBlocs from half a decade ago used to be very rugged but I don't think they make them anymore, and now offer a similar one in a TO247 package. A similar good miniBloc is the 60N60C2D1. Regardless, they're usually quite expensive so I just go for TO247s. I've found FGA60N65SMD to be good value due to their low cost. IRGP50B60PD1 from IR is a great choice as well, along with the similar FGH60N60SMD from Fairchild. The 30N60 and 40N60 line from Fairchild are also ok for a small coil. I've also be meaning to try the IXGH60N60C3D1 from IXYS. They have some very nice PT and XPT IGBTs. Would be happy to hear from others what their favourite transistors are.
Re: Please evaluate this brick for DRSSTC suitability /anything to note please?
CiscoOPA, Tue Nov 18 2014, 04:05PM
So many choices. :) I've a full bridge of 60N60 I've been tinkering with.
Thread crashing a bit too...I also came across a half dozen used Semikron skm400gb124d. Specs seem pretty decent even if they're on the older side. Thoughts? No idea how I'm getting 1200V for them to switch though ^.^ http://pdf1.alldatasheet.com/datasheet-pdf/view/126784/SEMIKRON/SKM400GB124D.html
Beyond being more durable, wiring up the large bricks is much easier IMHO. They nicely separate out the gate pins from the power so OCD-ish folks like myself can make things prettier.
CiscoOPA, Tue Nov 18 2014, 04:05PM
So many choices. :) I've a full bridge of 60N60 I've been tinkering with.
Thread crashing a bit too...I also came across a half dozen used Semikron skm400gb124d. Specs seem pretty decent even if they're on the older side. Thoughts? No idea how I'm getting 1200V for them to switch though ^.^ http://pdf1.alldatasheet.com/datasheet-pdf/view/126784/SEMIKRON/SKM400GB124D.html
Beyond being more durable, wiring up the large bricks is much easier IMHO. They nicely separate out the gate pins from the power so OCD-ish folks like myself can make things prettier.
Print this page