mounting a pot core

That's the sort of scheme you get with people who don't know what they are doing.

If you don't want two pairs of leads to interact, twist each pair of leads and take them out through different slots.

Twisting each pair of leads separately is a much more potent protection against interaction than just keeping them physically separate.

That's the sort of reply you get from someone who hasn't understood the previous post.

I know it doesn't fit within your present constraints, but maybe for a future upgrade, how about something like the arrangement below? Coax T1 and T2 sit in the same magnetic core. You could use a single 25 Ohm coax. The actual lengths do not matter, as long as T1=T2 and T3=T4. T3 needs its own core, but T4 does not.

It produces pulses with the full supply amplitude and aligns the positive and negative outputs. It's also a heavier load on the switch, of course.

Jeroen Belleman

=============================== Version 4 SHEET 1 1476 680 WIRE 240 0 64 0 WIRE 864 0 240 0 WIRE 416 128 272 128 WIRE 576 128 512 128 WIRE 624 128 576 128 WIRE 736 128 704 128 WIRE 1040 128 896 128 WIRE 1168 128 1168 96 WIRE 1168 128 1136 128 WIRE 240 144 240 0 WIRE 272 144 272 128 WIRE 864 144 864 0 WIRE 896 144 896 128 WIRE 416 160 352 160 WIRE 576 160 512 160 WIRE 1040 160 976 160 WIRE 1232 160 1136 160 WIRE 1296 160 1232 160 WIRE 1456 160 1376 160 WIRE 64 176 64 0 WIRE 352 176 352 160 WIRE 576 176 576 160 WIRE 976 176 976 160 WIRE 272 256 272 240 WIRE 320 256 272 256 WIRE 896 256 896 240 WIRE 944 256 896 256 WIRE 64 288 64 256 WIRE 240 304 240 240 WIRE 864 304 864 240 WIRE 864 304 240 304 WIRE 240 336 240 304 WIRE 240 368 240 336 WIRE 192 384 96 384 WIRE 192 432 96 432 WIRE 96 448 96 432 WIRE 240 480 240 448 WIRE 96 544 96 528 FLAG 64 288 0 FLAG 240 480 0 FLAG 736 128 0 FLAG 96 384 0 FLAG 96 544 0 FLAG 576 128 outp FLAG 240 336 D FLAG 320 256 0 FLAG 944 256 0 FLAG 1456 160 0 FLAG 976 176 0 FLAG 1168 96 0 FLAG 1232 160 outm FLAG 576 176 0 FLAG 352 176 0 SYMBOL voltage 64 160 R0 SYMATTR InstName V1 SYMATTR Value 100 SYMBOL tline 256 192 R90 SYMATTR InstName T1 SYMATTR Value Td=5n Z0=50 SYMBOL sw 240 352 R0 SYMATTR InstName S1 SYMATTR Value myswitch SYMBOL res 608 144 R270 WINDOW 0 32 56 VTop 2 WINDOW 3 0 56 VBottom 2 SYMATTR InstName R1 SYMATTR Value 50 SYMBOL voltage 96 432 R0 WINDOW 123 0 0 Left 2 WINDOW 39 0 0 Left 2 SYMATTR InstName V2 SYMATTR Value PULSE(0 1 1n 1n 1n 15n) SYMBOL tline 880 192 R90 SYMATTR InstName T2 SYMATTR Value Td=5n Z0=50 SYMBOL res 1392 144 R90 WINDOW 0 0 56 VBottom 2 WINDOW 3 32 56 VTop 2 SYMATTR InstName R2 SYMATTR Value 50 SYMBOL tline 1088 144 R0 SYMATTR InstName T3 SYMATTR Value Td=5n Z0=50 SYMBOL tline 464 144 R0 SYMATTR InstName T4 SYMATTR Value Td=5n Z0=50 TEXT 296 408 Left 2 !.model myswitch SW(Ron=10m roff=1G Vt=0.5) TEXT 96 -16 Left 2 !.tran 50n

In applications where power levels are important, pot core orientation will affect 'N' in the flux density concentration, and produce early saturation at the location where minimum x-sectional area occurs.

RL

It throws a "Multiple Instances of Flag" error

John Larkin Highland Tech Glen Canyon Design Center Lunatic Fringe Electronics

Dream on.

That should have a small effect on my pulser. Ill try it.

I got a 2% change in inductance when I rotated the core halves.

John Larkin Highland Tech Glen Canyon Design Center Lunatic Fringe Electronics

That's a nuisance. I downloaded it again from your answer and it works fine here. I tried googling for that error message, but it typically seems to occur when subcircuits are included, and I don't do that. Eyeballing the .asc text, looking for anything duplicated, didn't help either. I'm a bit at a loss here.

Could you try removing all the lines starting with 'FLAG'? The file will still open in the schematic editor, but now all the GND connections are gone, as are the node names. Insert GNDs again by hand. They are all at the ends of bits of wire.

Jeroen Belleman

Weird. If I edit it down to one FLAG, it complains about multiple SYMATTRs. I think multiple symattrys are legal.

Your other circuit worked fine.

What is a FLAG anyhow? The help doesn't say.

John Larkin Highland Tech Glen Canyon Design Center Lunatic Fringe Electronics

Your netlist is fine.

It denotes a graphical (such as ground symbol) or text (such as net name) item.

Go to tools/color preferences and you can give change the flag item colours to see where they are on the schematic.

I fixed it. I copied the text from Jeroen's post, into a text editor, and saved it but didn't include the .asc extension. LT Spice then threw the weird errors.

LT is great but kinda weird and definitely under-HELPed.

Looks like FLAG x y 0

is a ground.

John Larkin Highland Tech Glen Canyon Design Center Lunatic Fringe Electronics

Wurth seems to be giving their book away for free. It's probably the closest to a user guide.

It doesn't mention FLAG.

Wurth is pretty weird.

John Larkin Highland Tech Glen Canyon Design Center Lunatic Fringe Electronics

Mind you, you can get a similar change just by forcing out the fluff and detritus present at the contacting surfaces. That might be what you're actually seeing, even if the core is gapped.

In pot cores, minimum x-section usually occurs where the centre core meets the top and bottom plates. In parts shapes designed for power applications, this is usually corrected.

This core rotation can be used as a tolerance trim, where needed, but anything like that before impregnation is probably just biting fart bubbles.

RL

There is a way of doing trolerance trims on some gapped pot cores.

The manufactures put cylindrical hole down through the centre of the core, and sell an adjustor which plugs into the hole. There a ferrite slug in the adjustor which you can screw up and down to fully or completely bridge the the gap between two core halves.

I've used them, and they provide enough adjustment to let you trim out the residual tolerance on the inductance of a gapped core which can get up to perhaps +/-5%, and a bit more.

The one time when I designed one in to get a precise 15MHz source for TV style video, one of my colleague engineers copied the design for a much higher volume project, he left out the adjustor - we were also using a varactor diode to accomodate the frequency difference between production TV displays and he figured that that - one its own - could cope with both sources of variation. Varactors are very non-linear, and I was worried about getting out of the region where they were linear enough.

Jerry was a very good engineer - if not all that self-confident - and I'd been warned off expressing any doubts about his designs, because he'd take my doubts much too seriously.

Pot cores were developed for tube-amplitude cctry. I expect that adjustible varieties had added cost. How did it compare in cost to the addition of a varactor?

For semiconductors, you'd get away with slug tuned 'can' - still candidate for a varactor.

Lots of headaches with pot cores - mounting, termination etc. The smaller parts actually increased in price . . . .

RL

Impregnation?

John Larkin Highland Tech Glen Canyon Design Center Lunatic Fringe Electronics

The varactor was there anyway.

The hole down the middle of the pot core did mean the part was dearer than it's plain vanilla equivalent, and you had to pay extra for the adjustor, but only a few cents. The time for final test to set the adjustor to give the right frequency and fix it there with a blob of wax was probably the most expensive element.

The gapped pot cores were a more up-market option. There's nothing about pot cores that make them more appropriate for tube-amplitude applications - they tend to be very flexible if you know what you are doing. People who want to buy all their components off the shelf tend to be less aware of this.

There are lots of headaches with every component. Pot cores need windings, but you can buy a coil former with mounting pins designed to integrate with the pot core. RM cores are sold with pinned mounting clips that you can use to lock the RM-core halves together around the coil former, and you can plug the whole assembly into a set of tinned holes on your circuit board and wave-solder the lot.

There are surface mount equivalents, but I've not kept up with them.