John Larkin wrote:> On 14 Jun 2014 04:11:16 GMT, Jasen Betts <jasen@xnet.co.nz> wrote: > >> On 2014-06-13, John Larkin <jjlarkin@highNOTlandTHIStechnologyPART.com> wrote: >>> >>> We've used transmission-line transformers to drive mosfet gates. We've wound >>> them from micro-coax on ferrite toroids, with the shield being the primary and >>> the inner conductor the secondary. Sub-ns speed and low leakage inductance. But >>> it's labor intensive. >>> >>> So I was thinking about doing it on a multilayer PCB, like a 6-layer. Layers >>> 1/3/5 could be one to three layers of spiral trace, primary, and 2/4/6 ditto, >>> secondary. I'm not sure how to think about the impedances, but it ought to have >>> wide traces and thin dielectrics, I guess. >>> >>> Maybe one layer, one turn, per winding? That takes no vias. >> one layer will be better, with 6 interleaved layers you have the L2-L3 and the >> L4-L5 parasitic capacitance working against your goal, >> >>> It needs a ferrite core. Who makes the sorts of cores that work for PCB >>> inductors? >> with the coax transformer how much coax does it take before you don't >> need a core? > > The coax would have to be really long to sustain wide pulses, and the > the prop delay becomes a problem. I don't entirely understand this, > but it seems to me that the output impedance becomes the Zo of the > coax when the coax is long relative to the pulse rise time. > > Here's a version we did, but only for 200 ns 5-volt pulses. > > https://dl.dropboxusercontent.com/u/53724080/Parts/Inductors/Xfmrs.JPG >Caerful when soldering. When the solder wicks up that far and into the coax, especially in a bend like this, it can partially melt the insulation. Even PTFE can be nicked a bit. [...] -- Regards, Joerg http://www.analogconsultants.com/
PCB transmission line transformer
Started by ●June 13, 2014
Reply by ●June 15, 20142014-06-15
Reply by ●June 15, 20142014-06-15
John Larkin wrote:> On Fri, 13 Jun 2014 09:02:54 -0700 (PDT), Lasse Langwadt Christensen > <langwadt@fonz.dk> wrote: > >> Den fredag den 13. juni 2014 17.25.49 UTC+2 skrev John Larkin: >>> On Fri, 13 Jun 2014 00:25:47 -0500, "Tim Williams" <tmoranwms@charter.net> >>> >>> wrote: >>> >>> >>> >>>> "John Larkin" <jjlarkin@highNOTlandTHIStechnologyPART.com> wrote in >>>> message news:s4vkp95o83ps2l4hfpp9nn75v6unrhrp6d@4ax.com... >>>>> So I was thinking about doing it on a multilayer PCB, like a 6-layer. >>>>> Layers >>>>> 1/3/5 could be one to three layers of spiral trace, primary, and 2/4/6 >>>>> ditto, >>>>> secondary. I'm not sure how to think about the impedances, but it ought >>>>> to have >>>>> wide traces and thin dielectrics, I guess. >>>>> Maybe one layer, one turn, per winding? That takes no vias. >>>>> It needs a ferrite core. Who makes the sorts of cores that work for PCB >>>>> inductors? >>>>> Anybody done this? We might be up for some consulting, if the project >>>>> gets >>>>> serious. >>>> Should be pretty straightforward. Two identical windings, on layer 1 and >>>> layer 2, have a given width, spacing and dielectric the whole length. >>>> Interleaved windings of different construction, or layers connected in >>>> series, will have different results, but that's true of normal transformer >>>> windings (dispersive helical waveguide modes and such). You'd only want >>>> "wide" traces if you want a low impedance. You can also parallel up >>>> layers for that. >>>> I've used planar transformers before. Rather lossy and space-inefficient >>>> with few layers. No PCB manufacturer can produce a board with better than >>>> about 0.2 winding factor, so it can't compare to the density typical of >>>> wound parts. On the other hand, it's quite low profile, and with enough >>>> layers, the losses are good enough. Performance is great, since the >>>> coupling is good, the winding compact with few turns, and high frequency >>>> core materials are available. Frequency response will of course matter >>>> more to your purposes, unless you want to do superfast kW pulses with >>>> selectable matching or something. >>>> Ready made transformers, either made of PCB stacks, or just ridiculously >>>> thick PCBs of many layers, are available into the 5kW range, intended to >>>> operate at a sizable fraction of a MHz. >>> >>> >>> This will be a mosfet gate drive transformer, up to a few hundred ns, and I need >>> >>> very fast edges, namely low leakage L. 4KV isolation. >>> >> is a transformer wound with micro coax really rated for 4KV ? > > Probably. We'd test them, of course. >Just keep in mind that the longer your test the more compromised your barrier becomes. Hipot tests are partially sacrificial, they eat away at the insulation properties.> I've run RG58 at 20 KV, no problems. Teflon is rated somewhere between 1000 and > 4500 volts/mil. >However, there is a huge difference between a continuous duty working voltage on the isolation versus a single-even safety function. Easily 10:1.>> you can't find some small smd transformer that would work? > > I've never seen a commercial coax-type transmission-line transformer. The > MiniCircuits balun types wouldn't work. >In the ham radio world there are some. http://www.dj0ip.de/app/download/5794262625/Spiderbeam+Balun+Construction_Eng.pdf -- Regards, Joerg http://www.analogconsultants.com/
Reply by ●June 15, 20142014-06-15
John Larkin" <jjlarkin@highNOTlandTHIStechnologyPART.com> wrote in message news:99tpp9h3ivho86avm4ddgge98acl3ds1n6@4ax.com...>>Or if you can find SiC diodes in that range, too. The junction (not >>schottky) types are just as awful as silicon, for instance, the >>performance >>of SiC BJTs, or MOSFET body diodes. AFAIK, they are constructed exactly >>the >>same as their Si counterparts, just in different material. > > Schottky diodes don't do the step-recovery thing; diffused silicon PNs > seem to have the best doping profiles.I said "not schottky". Though the lower voltage ones have so much nonlinear capacitance, they are sometimes worse than true recovery. I once built a circuit to try and demonstrate step recovery; I got the best result from a schottky that couldn't possibly be giving me true step recovery. It was only a ~10ns blip. More like parametric sharpening. Broadly the same idea, in any case. Suppose it's worth wondering if a chain of SiC schottky junctions could do the same. Maybe a row of SMC or DPAK devices would have low enough stray inductance to do the job.> What I may wind up doing is talking the customer down to something > more reasonable, 2-4 ns. The sine-squared transfer function of a > Pockels cell speeds up the optical risetime a bit. > > I guess we may as well get some SiC transistors and bang the gates > until we blow them up. Then we'll know. The Cree, Rohm, and Ixys parts > seem to be remarkably similar.And there's always toobs. I recall reading research papers from the 80s-90s where they were using planar transmitter triodes in transmission line structures for driving nonlinear sharpening networks (magnetic saturation in that case, I think). Plenty of voltage, you'll have to see if the Vg, Qg, etc. is reasonable enough to work with. Tim -- Seven Transistor Labs Electrical Engineering Consultation Website: http://seventransistorlabs.com
Reply by ●June 15, 20142014-06-15
"Jeff Liebermann" <jeffl@cruzio.com> wrote in message news:l90qp9t860iq49skck0jmftqoris0jet7g@4ax.com...> Only works at low frequencies. A large diameter open wire "ladder" > transmission line would probably work equally well, but would require > very big baluns at both ends.Well, it would presumably work up to the cutoff frequency, limited by competing coaxial waveguide modes. And possibly degenerate modes involving wire-to-wire coupling, which shouldn't couple strongly to differential mode energy, but are nevertheless present. Guessing from the picture, it might be good up to... 10 to 100MHz? I think the downside of ladder line would be the prohibitive wire diameter, even for a high impedance (widely spaced wires) line. And too wide a spacing and you start having to worry about unwanted radiation. (They could go one further and twist it every half wave, so the far field is nulled.) But yeah, the less balun you need at those power levels, the better. Interesting to note, the voltage on the outer cage (with respect to ground) is probably nonzero. Or if it is, it's because they did indeed add a common mode choke, or something along those lines. Thought being: the ground impedance is nonzero, so the "ground" at the antenna feedpoint will have nonzero voltage.> I'm also wondering about the losses. At low RF frequencies, skin > depth and resistive losses predominate. I don't think there's enough > copper in the wires to be really low loss. I've built broadband wire > cage dipoles where such losses almost became a problem. At megawatt > power levels, any transmission line losses can be a serious problem.Hmm, not obvious if they used litz -- suppose that's a possibility. Probably not worth it though. I would guess the impedance is high enough not to be a big problem (obviously, they would've thought of that; question is, we want to know why). Seems like it would be around 50 ohms if it were solid coax, but since it's not solid, it's probably higher. But by how much? 100 ohms? 300 ohms? Still quite a few amperes at that power level, either way. If the load is a standard 1/4 wave tower, it would match near 50 ohms, which would require a matching network or transformer, which is again undesirable. Given the potential for ground currents, they may well have a transformer out there anyway, in which case it doesn't matter much what the line or load are; they just tap them off as needed. And lightning isolation, might be a plus.> I wouldn't worry (much) about the wind. The wires should all swing in > unison. For wide gaps, spreaders can be added. However, I suspect > the real danger is some bird landing inside the cage, resulting in an > RF arc barbequed bird.>:-DTim -- Seven Transistor Labs Electrical Engineering Consultation Website: http://seventransistorlabs.com
Reply by ●June 15, 20142014-06-15
On Friday, June 13, 2014 6:33:20 AM UTC+2, John Larkin wrote:> We've used transmission-line transformers to drive mosfet gates. We've wound > > them from micro-coax on ferrite toroids, with the shield being the primary and > > the inner conductor the secondary. Sub-ns speed and low leakage inductance. But > > it's labor intensive. > > > > So I was thinking about doing it on a multilayer PCB, like a 6-layer. Layers > > 1/3/5 could be one to three layers of spiral trace, primary, and 2/4/6 ditto, > > secondary. I'm not sure how to think about the impedances, but it ought to have > > wide traces and thin dielectrics, I guess. > > > > Maybe one layer, one turn, per winding? That takes no vias. > > > > It needs a ferrite core. Who makes the sorts of cores that work for PCB > > inductors? > > > > Anybody done this? We might be up for some consulting, if the project gets > > serious. >I have done a safety approved transformer this way: Layer 1 and 2, primary winding, about 12 turns. Blind via from Layer 1 to 2 Layer 3 and 4, secondary winding, same turns, same stuff with a via It ran resonant at about 8MHz with added capacitance to trim it. (you could do a SW controlled sweep to make the trimming) I got an overall input to output efficiency of 70%, which I thougth was good for a air core transformer and 0.5mm distance from layer 2 to 3 Cheers Klaus
Reply by ●June 15, 20142014-06-15
On Sun, 15 Jun 2014 14:17:00 -0500, "Tim Williams" <tmoranwms@charter.net> wrote:>"Jeff Liebermann" <jeffl@cruzio.com> wrote in message >news:l90qp9t860iq49skck0jmftqoris0jet7g@4ax.com... >> Only works at low frequencies. A large diameter open wire "ladder" >> transmission line would probably work equally well, but would require >> very big baluns at both ends.>Well, it would presumably work up to the cutoff frequency, limited by >competing coaxial waveguide modes. And possibly degenerate modes involving >wire-to-wire coupling, which shouldn't couple strongly to differential mode >energy, but are nevertheless present. Guessing from the picture, it might >be good up to... 10 to 100MHz?I like order of magnitude guesses. I don't think I can do better. Let's see if I can still function after eating my own cooking. Looking at the photo at: <http://upload.wikimedia.org/wikipedia/commons/6/61/Solec_Kujawski_longwave_antenna_feeder.jpg> there's not much in the way of known references from which to extract dimensions. The best I can do is the roadway, which I would guess(tm) is about 2.5 meters wide. Translating the track width to the square metal frame and adjust for parallax, perspective, and indigestion, I get 2.0 meters for the shield diameter and 0.4 meters for the center conductor. (5:1 diameter ratio extracted from my monitor with a ruler, which I think is fairly accurate). For air dielectric: Z = 138*log(D/d) = 138*log(5) = 96 ohms I'll call it 100 ohms. The original photo at: <http://en.wikipedia.org/wiki/Transmission_line#Coaxial_cable> indicates that it's 225 KHz at 1200 kW. Power = E^2 /R E = sqrt(Power*R) = sqrt(1.2*10^6 * 100) = 11 kV volts RMS. Peak voltage (which is what makes the arc) will be 1.4*11kV = 15kV. At 3x10^6 V/meter, a 0.8 meter gap will flash over at 2.4 kV. So either my guesswork and math suck (a likely possibility), or this thing isn't going to work at rated power. (I really hate it when reality doesn't agree with my reverse engineering). Here's the station and tower: <http://en.wikipedia.org/wiki/Longwave-transmitter_Solec_Kujawski> More photos including better pictures of the open wire coax arrangement: <http://radiopolska.pl/wykaz/pokaz_lokalizacja.php?pid=610> Google street view works, but doesn't get close enough to the towers to show details: <http://maps.google.com/?t=h&om=0&ll=53.022778,18.261111&spn=0.023387,0.058537>>I think the downside of ladder line would be the prohibitive wire diameter, >even for a high impedance (widely spaced wires) line. And too wide a >spacing and you start having to worry about unwanted radiation.Agreed. If we accept my 100 ohm impedance guess(tm), the current in the coax would be: P = I^2 * R I = sqrt(P/R) = sqrt(1.2*10^6 / 100) = 110 amps However, that's distributed equally among the individual wires. I count 12 wires each in the shield and center conductor or about 9 amps per wire. No problem there.>(They could >go one further and twist it every half wave, so the far field is nulled.)That's what the old telegraph wires did to minimize crosstalk.>But yeah, the less balun you need at those power levels, the better.Yep, that's what's nice about using coax cable. If the antenna and transmitter impedances were fairly close to my speculated 100 ohms, the baluns are not needed.>Interesting to note, the voltage on the outer cage (with respect to ground) >is probably nonzero. Or if it is, it's because they did indeed add a common >mode choke, or something along those lines. Thought being: the ground >impedance is nonzero, so the "ground" at the antenna feedpoint will have >nonzero voltage.I beg to differ. If you look at the photo: <http://upload.wikimedia.org/wikipedia/commons/6/61/Solec_Kujawski_longwave_antenna_feeder.jpg> the coax shield wires appear to be directly connected to the metal frame. If they were isolated, insulators would have been needed. Methinks the outer cage is at ground potential.>> I'm also wondering about the losses. At low RF frequencies, skin >> depth and resistive losses predominate. I don't think there's enough >> copper in the wires to be really low loss. I've built broadband wire >> cage dipoles where such losses almost became a problem. At megawatt >> power levels, any transmission line losses can be a serious problem.>Hmm, not obvious if they used litz -- suppose that's a possibility. >Probably not worth it though.Litz wire would certainly have helped a little, but I suspect the cost and maintenance problems might have been prohibitive. Besides, if you've got 1.2 Megawatts to play with, what's a few extra watts of loss?>I would guess the impedance is high enough not to be a big problem >(obviously, they would've thought of that; question is, we want to know >why). Seems like it would be around 50 ohms if it were solid coax, but >since it's not solid, it's probably higher. But by how much? 100 ohms? >300 ohms?See my previous guesswork. I would say 100 ohms.>Still quite a few amperes at that power level, either way.110 A RMS total or 9 A per wire. VOA made use of open wire transmission lines at high power levels: <http://cryptome.org/eyeball/voa/voa-antennas.htm> <http://www.lbagroup.com/blog/wp-content/uploads/2010/08/open-wire-antenna-transmission-lines.jpg>>If the load is a standard 1/4 wave tower, it would match near 50 ohms, which >would require a matching network or transformer, which is again undesirable.1/4 wave at 225 KHz is 333 meters (1100 ft). If the tower had a capacitive hat, it could be made considerably shorter. The web site says 289 and 330 meters. <http://en.wikipedia.org/wiki/Longwave-transmitter_Solec_Kujawski> A 1/4 wave monopole over a decent ground works out about 35 ohms impedance. That's not going to match very well to a 100 ohm feedline. I couldn't find much techy detail on the station, so I have no clue how their doing the matching. A clue is that there are small brick building at the base of both towers, which suggest a matching network.>Given the potential for ground currents, they may well have a transformer >out there anyway, in which case it doesn't matter much what the line or load >are; they just tap them off as needed.There are going to be ground currents anyway. A monopole requires a buried counterpoise in the form of ground radials in order to form the other half of the dipole.>And lightning isolation, might be a plus.The antenna is in Poland (and my family is from Poland). I suspect that using cheap electrical wire and just replacing the wire when it gets hit by lightning, might be all that's required.>> I wouldn't worry (much) about the wind. The wires should all swing in >> unison. For wide gaps, spreaders can be added. However, I suspect >> the real danger is some bird landing inside the cage, resulting in an >> RF arc barbequed bird. > >>:-DI once designed an antenna that included weight loading measured in standard bird loads, which I vaguely recall was based on the number of Gooney Birds (albatross) that could be lined up on the antenna radials times 10Kg per bird. I took exception to the spec claiming that it was an unlikely possibility, and was presented with a photo showing a rather large number of birds neatly lined up on a seriously sagging antenna element. I didn't protest any further.>Tim-- Jeff Liebermann jeffl@cruzio.com 150 Felker St #D http://www.LearnByDestroying.com Santa Cruz CA 95060 http://802.11junk.com Skype: JeffLiebermann AE6KS 831-336-2558
Reply by ●June 15, 20142014-06-15
On Sun, 15 Jun 2014 14:16:27 -0700, Jeff Liebermann <jeffl@cruzio.com> wrote: Why do I only see my math errors AFTER I post them? Argh...>The original photo at: ><http://en.wikipedia.org/wiki/Transmission_line#Coaxial_cable> >indicates that it's 225 KHz at 1200 kW. > Power = E^2 /R > E = sqrt(Power*R) = sqrt(1.2*10^6 * 100) = 11 kV volts RMS. Peak >voltage (which is what makes the arc) will be 1.4*11kV = 15kV. >At 3x10^6 V/meter, a 0.8 meter gap will flash over at 2.4 kV.At 3x10^6 V/meter, a 0.8 meter gap will flash over at 2,400 kV so coax air gap is sufficient to prevent flash over. -- Jeff Liebermann jeffl@cruzio.com 150 Felker St #D http://www.LearnByDestroying.com Santa Cruz CA 95060 http://802.11junk.com Skype: JeffLiebermann AE6KS 831-336-2558
Reply by ●June 15, 20142014-06-15
On 15 Jun 2014 03:34:47 GMT, Jasen Betts <jasen@xnet.co.nz> wrote:>On 2014-06-14, John Larkin <jjlarkin@highNOTlandTHIStechnologyPART.com> wrote: >> On 14 Jun 2014 04:11:16 GMT, Jasen Betts <jasen@xnet.co.nz> wrote: >> >>>On 2014-06-13, John Larkin <jjlarkin@highNOTlandTHIStechnologyPART.com> wrote: >>>> >>>> >>>> We've used transmission-line transformers to drive mosfet gates. We've wound >>>> them from micro-coax on ferrite toroids, with the shield being the primary and >>>> the inner conductor the secondary. Sub-ns speed and low leakage inductance. But >>>> it's labor intensive. >>>> >>>> So I was thinking about doing it on a multilayer PCB, like a 6-layer. Layers >>>> 1/3/5 could be one to three layers of spiral trace, primary, and 2/4/6 ditto, >>>> secondary. I'm not sure how to think about the impedances, but it ought to have >>>> wide traces and thin dielectrics, I guess. >>>> >>>> Maybe one layer, one turn, per winding? That takes no vias. >>> >>>one layer will be better, with 6 interleaved layers you have the L2-L3 and the >>>L4-L5 parasitic capacitance working against your goal, >>> >>>> It needs a ferrite core. Who makes the sorts of cores that work for PCB >>>> inductors? >>> >>>with the coax transformer how much coax does it take before you don't >>>need a core? >> >> The coax would have to be really long to sustain wide pulses, and the >> the prop delay becomes a problem. > >ok. you said how fast in the original post, but not how slow.Gate drive rise time should be ~~1 ns. Max pulse width might be a few hundred us.> >as for delay: > > what if you do this > > > .------> gate > | > drive >-------------(|) > | | > | | > \/\ > coax longer than shown. > \/\ > | | > | | > --------(|) > __|_ | > //// `-----> souce > >That will make the gate drive 50 ohms, which would radically slow things down. That same thing, but short and with a ferrite, would be lower impedance.
Reply by ●June 15, 20142014-06-15
On Sunday, June 15, 2014 10:23:27 AM UTC-4, John Larkin wrote:> On Sun, 15 Jun 2014 06:44:58 -0700 (PDT), dagmargoo...@yahoo.com > wrote: > >On Saturday, June 14, 2014 12:50:47 PM UTC-4, John Larkin wrote: > > > >> I guess I could float the gate driver electronics on the mosfet source > >> (and power that somehow, kilovolts off ground) and run the transformer > >> at signal levels. That's less interesting (aka less risky.) > >> > >> Not to change the subject, but I've been exploring mosfets for fast > >> kilovolt switching. Most kilovolt-level power fets kill you from gate > >> charge requirements and/or source lead inductance. SiC looks like a > >> winner until you see the internal gate resistances, 5 ohms or so, too > >> much to allow you to drive the gate hard and fast. > >> > >> GaN is nice, but too low voltage. Maybe a GaN gate driver into one of > >> those planar Ixys mosfets is the best one can do. My customer wants > >> 4000 volts in 1 ns, at 100 KHz, but maybe he can't have it. > > > >GaN cascoded into a MOSFET = fast, HV MOSFET? > > Maybe. At the point that the peak gate current approaches the peak > drain current, somewhere south of 2 ns maybe, the GaN will have to > sink twice the drain current.Yes. It might be that it's a lot easier to parallel a few low-voltage GaN fets to simply muscle though that, it solves the MOSFET driving problem to boot, and placates Mr. Miller.> It might be better to just ground the > mosfet source and use the GaN as the gate driver. A 7 AM, not having > had any coffee yet, I could well be wrong.Could be. It would take actual parts & numbers to figure out what's better. Cheers, James
Reply by ●June 15, 20142014-06-15
"Jeff Liebermann" <jeffl@cruzio.com> wrote in message news:jptrp958ju1opod0ftdendvvan9p31bde2@4ax.com...> For air dielectric: > Z = 138*log(D/d) = 138*log(5) = 96 ohms > I'll call it 100 ohms.Sounds about right. Any guess how much the open vs. solid construction changes it, if at all? Without measurements of similar structures, I really don't know what to guess... more than double or triple seems rather unlikely, but I can't see it being equal to the theoretical (solid wall) coax figure either.> VOA made use of open wire transmission lines at high power levels: > <http://cryptome.org/eyeball/voa/voa-antennas.htm> > <http://www.lbagroup.com/blog/wp-content/uploads/2010/08/open-wire-antenna-transmission-lines.jpg> >I can't even tell what's going on here, what a mess! I don't remember if VOA was just one massive station or several frequencies; would that be for improved propagation in, ahem, certain target regions, or just for bandwidth?> A 1/4 wave monopole over a decent ground works out about 35 ohms > impedance. That's not going to match very well to a 100 ohm feedline. > I couldn't find much techy detail on the station, so I have no clue > how their doing the matching. A clue is that there are small brick > building at the base of both towers, which suggest a matching network.Seems unlikely that the line is anything below 100 ohms, so that sounds like a good bet.> I once designed an antenna that included weight loading measured in > standard bird loads, which I vaguely recall was based on the number of > Gooney Birds (albatross) that could be lined up on the antenna radials > times 10Kg per bird. I took exception to the spec claiming that it > was an unlikely possibility, and was presented with a photo showing a > rather large number of birds neatly lined up on a seriously sagging > antenna element. I didn't protest any further.Needs a periodic surge of high power to scare them away. 'Ey boys, meat's back on the menu!... They'll screw up your tuning, too... what's an albatross, 20pF? Wings down or outstretched? (African or European Swallow?...) Tim -- Seven Transistor Labs Electrical Engineering Consultation Website: http://seventransistorlabs.com
