Crowbars

You probably need a fast response, so a thyristor crowbar may be too slow. Maybe a TL431 and a beefy PNP or PMOS transistor to augment the current handling. John

søndag den 23. juli 2023 kl. 11.39.55 UTC+2 skrev Cursitor Doom:

how much current do they need and how much over voltage can they handle?

I recently saw a Diodes Inc chip that is a combination overcurrent fuse and overvoltage limiter.

But fix the bad supply! Or add yet another regulator after it.

A giant SCR crowbar would be easy. A zener and an SCR. And a fuse maybe.

What RF transistor is it?

If the RF transistors are feeding a nearby antenna, make sure the crowbar triggering circuit does not pick up some of the radiated RF. Such random RF pickup could fire the SCR, shutting down the power supply :-).

After a few blown fuses, I replaced the fuse with some soldering tin. As an additional bonus I also got a smell indication that the crowbar had operated. Moving the antenna outside solved the nuisance triggering problem.

MRF 485. Motorola don't make 'em any more. There was a company called Eleflow that was re-manufacturing some obsolete semis including this one, but they seem to have gone out of business unfortunately.

12A. Vceo max is 35V.

I would put in some heavy duty zeners and fuses. OV almost always kills zeners and shorting them and taking out the fuses. Having bunch of them in parallel can make sure at least one of them is shorted.

I am using plenty of 12V 1A zeners and 3A/5A fuses in my project.

søndag den 23. juli 2023 kl. 15.21.16 UTC+2 skrev Cursitor Doom:

not overly accurate but you have +5V to work with might want a zener for the short time it takes the relay to open

Sorry, I wasn't particularly clear at all in my initial message. I won't be using a relay. If an over-voltage condition arises, I plan to short the supply out with a MOSFET and blow a F12.5A fuse to protect the downstream circuitry. I think a relay would be too slow in this app.

That's all very fine, Fred, but as I said, I'm looking for a simple expedient here. It's only a temporary measure until I manage to source a better PSU (Kenwood shame on you!)

Transorb ???

I vote for that all so for a quick fix. Simple and no adjustment.

Soft triggering an SCR, from a zener, is a potential failure mode, but the SCR would fail shorted, which is better than blowing out the RF stuff.

------------------------------------

** The Vceo rating applies when the base and emitter terminals are left open circuit. However, the maker's published " Vcer " rating is 60V and applies when a low resistance exists between base and emitter ( < 100ohms) and is usually the same as the Vcb rating. This does not change things much because in RF power output operation the max Vce on the transistor is likely to approach or even exceed 2 times the DC supply.

...... Phil

Well, by a strange coincidence I only recently bought a load of diacs which breakover at 32V. I could pick the lowest breakover device out of the bunch and maybe shave a volt or so off of that and that sorts out the soft-trigger problem as well.

By the same strange coincidence, i brought a load of 13V zeners. They are too high to protect Lithium batteries, since I really want 12.5V. Two in series would clamp it at 26V. Zener is the same as Diacs without the negative cycle that you don't really need. There are many voltages to choose from as well. 12V and 13V are very common, but not 12.5V, unfortunately.

That's perfct if you can tolerate 32v. Get an SCR with an inherent gate-cathode resistor, or just add one so the diac sees the full rail voltage before it fires.

Diacs are wonderful gadgets. I've tried to replicate the diac function with discretes and it ain't easy.

An RC-Diac-LED or better yet DepletionFet-Diac-LED can make a bright warning flasher on a high-voltage power supply that draws very little current.

People who can design electronic circuits know how to set up an circuit that will start drawing current when the voltage at its input hits 12.5V or any other arbitrary voltage much over about 2V. (1.2V is about the lowest voltage reference you can buy easily) and its a matter of comparing a divided down version of the input voltage with your chosen reference voltage, and turning on a current sink when the input is higher than you chosen voltage.

The LM10 does offer a 200mV reference voltage output, but it isn't cheap.

Actually working up the circuit isn't worth the effort until you know how which much current you need to sink, and how fast the current sink has to turn on, which usually means telling the audience what you are actually trying to do. Keeping a Nissan Leaf running after it batteries have hit the end of their service life isn't detailed enough to be helpful.

I know i can sink more current with another FET or something. But if one part fail, that 50% chance of failing. However, i can put 10 sets of zeners with each sinking 1A. One failure is only 10%. 10 zeners aren't much more expensive than a FET.

Actually, around 3A to 5A for a 3KWhr batteries.

12V zener turns on at 80% SOC. It doesn't need to be fast. It can be overnight. I seldom charge it to 90% anyway. Even if I have to, it's OK to lose some energy to heat slowly.

I am adding new portable batteries (2 to 3KWhr). It's roughly 2A per KWhr MAX.

I have told the audience many times, but not everybody, everywhere and all at once.