Each SSR channel will get 0.85 square inches on the board, and that allows two DPAKS, the gate drive transformer, and the wirewound resistor on the top side.
We'll put the small stuff on the bottom. It's a VME module, so one has to be careful about putting parts on the bottom side, specifically nothing tall or easy to shear off.
I don't think I could fit many joules of thickfilm resistors in the available board area. I'm spiking an axial wirewound hot each shot, and it has a lot of mass. I don't know what a, say, 1206 could absorb, but I'd guess millijoules. Does anybody know?
--
John Larkin Highland Technology, Inc
picosecond timing precision measurement
jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com
There's a nichrome-to-other-metal weld, and an other-metal-to-copper wire bond, so there's dissimilar metals involved; a little humidity can be a problem, eventually.
Better battery packs are assembled with nickel straps, 'cuz the nickel-to-nickel spot welds aren't subject to corrosion due to dissimilar metals.
I have some open-frame wirewound (almost rod-wound) resistors with corrosion. Green spots. Like you sometimes see on ruddy old chrome bathroom fixtures. (It's the color of chrome oxide, of course.)
Although, to be fair, those resistors did survive the numerous chemical experiments I did in high school. HCl fumes are bad for a lot of metals, in fact; who'd have thought? :-)
How fast will you switch off the fets? If too fast then any external series inductance could overvolt the fets, or are you placing transzorbs/movs across the SSR?
We can probably shut down the fets in under 500 us, given a big overload. We'll measure the voltage drop across the SSR, and PCB surface temp, and run a realtime simulation of resistor internal temperature, and limit that. Gate drive rise/fall times might be something like 100 us, maybe less.
The fets that we will probably use are avalanche rated to about 80 mJ, so we'll tell our customers to provide external clamping above that.
--
John Larkin Highland Technology, Inc
lunatic fringe electronics
I'm banging a Vishay AC05W axial wirewound at 20 joules, 0.1 Hz. It has a green, sandy sort of coating. Every shot, I can see in my Mantis a brief red flash in the spiral pattern of the nichrome wire. But the surface temp only peaks around 100C. And if I switch off the illumination and look in the dark, nothing is visible. So it's not getting red hot, something weird is happening to the chemistry of the coating. Fluoresence? If I illuminate it with a UV flashlight, the effect is more visible.
It smokes a bit for the first ten shots or so; I think it's burning off fingerprints.
My weekend test died at some unknown point. The resistor is fine, but my cute little HP6212 power supply quit. I fatigued the line fuse with all that pulsing.
--
John Larkin Highland Technology, Inc
picosecond timing precision measurement
jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com
Why not use constant current loads using MOSFETs or bipolar. Spread the power over a longer time and reduce the peak values. MOSFETs are rated for this kind of work. They have good thermal conduction to heatsinks.
Dumping the load into a resistor seems like applying power to a device that is not designed for the application. First task is to find a device that is designed for this work. Then design the rest of the control circuitry.
Headaches from mold spores. I have been fighting this problem for 20 years. Constant day and night. Cannot sleep. Headach pills have no effect. No solution so far. Kind of gives you a shitty outlook on life. I thought I was doing a bit better. You are very sensitive. Thanks.
No, nichrome has to be welded. I was thinking about actually measuring the wire peak temperature and its decay time constant. I'd have to make my own resistor, using copper instead of nichrome, and use the copper tempco to measure its temperature after a big pulse. But that's too much work, so we'll just blow up resistors to ballpark their thermal dynamics. There are papers and stuff online about thermal modeling of wirewound resistors, but they are mostly a lot of equations and few numbers.
I had to upgrade my test gear:
formatting link
There are now three 22mF 80 volt caps (210 joules) and an IXYS mosfet (IXFH-FT400N075T2) rated for 1000 watts dissipation and 1000 amps peak current.
This experiment is slow: occasional setup and days-long experiments. Meanwhile the product PCB layout is in progress.
--
John Larkin Highland Technology, Inc
lunatic fringe electronics
You are very knowledgeable about mold. It will grow on anything, including silicon shower sealants and ordinary concrete. They look for the sulfur and bury the hyphae throughout the concrete until it disintegrates.
Boulder, Colorado, was a very nice place. Dry, good building codes to eliminate basement humidity and kill mold. I am very sorry I ever left.
The only problem was the buildings tended to tilt a couple of degrees due to the gravity well from the nearby Flagstaff mountains. You can barely notice. I can live with that.
Moving is not a good option. No matter where I go there will be mold. I have to conquer the problem at the source.
Conventional filters, including HEPA, do not work. They suck in the stray lint particles which cling to the input. The mold loves these particles as a food source. Within 3 days, the filter is now a generator of mold spores.
Conventional elecrostatic cleaners do not work. The spores float past the collection electrodes as it they were not there.
I have invented a new type of electrostatic air cleaners. Most ionic generators require high voltage to ionize the air. The voltage is applied to a thin electrode that must be thick enough to support cleaning. The conductor radius determines the applied voltage that is needed. This determines the minimum spacing you can use between elcrodes. It also generates ozone, which is harmful.
Instead, I use the aluminum foil from party balloons. The metal thickness is around 40 nm, which creates an extremely sharp radius. The electrode separates the need for strength which is needed for cleaning, and the need for sharp radius which is needed for ionization. This means you can reduce the electrode voltage and reduce the electrode spacing without problems with arcover.
Eventually, I will solve these problems. I have a lot of new ideas for products for industry. But it is a race against time. I am old, and growing older faster. I don't have the strength I used to have. I need to find someone younger and stronger who can absorb what I have learned though the years, and apply it to a new company.
I have made millions in the past, and I'm sure I can do it again. But I'm starting to recognize my weaknesses, and I have to figure a way around them.
I'm quantifying when the resistor will fail. Isn't that good to know?
A slo-blo fuse fixed the HP supply.
I'm designing an SSR; the customer supplies the current. My job is to set the specifications, and do my best to protect the SSR if the customer exceeds those specs, specifically applies a hard voltage across the SSR contacts.
A 5-watt wirewound resistor can absorb ballpark 100x as many joules as a DPAK mosfet, and mosfets have better SOARs than bipolars.
But power resistors are exactly designed to have power dumped into them. But most are poorly characterized for joules-vs-time, so I'm blowing some up to see their limits. Looks like, with the right resistor, I can have a safety factor at least 4:1.
We were just noting that we're grossly overkilling this product from a design standpoint, but it's fun.
--
John Larkin Highland Technology, Inc
lunatic fringe electronics
I was just wondering if one of the stainless steel fluxes would work.
I've tinned SS and then soldered that to some other hunk of copper.
I was thinking about actually measuring
I don't know how many resistors you'll need, but I wonder if it's worthwhile talking to some pulse power resistor person/company?
There is most likely all sorts of weird stuff going on.
It sucks to use some part in a way that the manufacturer didn't intend, and then have them change the process. (I've had that happen with optical components.)
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.