flexmod P3.3 catastrophic failure

[dkumpula]

I'm glad you took me up on my suggestion all the way back in Post #3.

I've got a couple dozen of Vladim's (BBE/X-Wossee) drivers now and am really pleased with their performance in my lower power builds. They are feature rich, dirt cheap and shipping cost is very reasonable. Shipping times have been all over the map sometimes taking a full month, but I had a new batch land on my doorstep yesterday just 10 days after ordering.

-David

[catalanjo]

All of this Flexmod info is NOT GOOD !
I will have to bin a few hundred $ of Flexmods, most already installed and running!
Thanks for all the info everyone ... sad that we haven't discovered the CAUSE of all this grief, but at least I'll probly save myself a lot of future hassle!

Still bright side is Vlad gets a new client !
Cheers
PS. Have a few unused flexmods for sale cheap .... if anyone is interested drop me a PM.

[VDX]

Hi Draco,

I forgot to ansure Kecked's replay, yeah you can add more fets in parallel for higher output current and with out the emitter resistors that a transistor needs, i think just small value resistors on the gates will do the job, you just want to make sure that all of the fet's share the load and not have one turn on a litter sooner then the others, that would cause that fet to take more current then the others and possibly fail, if i am wrong some one please correct me.
I have done that with transistors as well and it worked out quite well

... look into my schematics - my old drivers are built with more LM's in parallel and works without problems with pretty high currents and pulsing frequenzies

First I've added some 'averaging' resistors (the 0.22 Ohm types) while fearing some cross-driving, but they can be dropped - the current regulating is only sensible to the shunt resistor, so won't react with the other branches.

In my newer drivers I'm driving single mosfets, each capable of up to +50 Amps pulsewise (rated for 140Amps, but overheat+melt pretty fast when >80Amps), but in parallel strands with double- and quad-OpAmps in analog set current regulator modes, and can combine the current outputs ... so no problem to setup a 200 Amps driver with 1MHz switching capabilities with 4 to 6 MOSFETS (2 or 3 double-OpAmps) and a pretty big cooler

Viktor

[Draco]

VDX
I am going to build some of the drivers you supplied a schematic for so you info is still very helpful, i finally found my bulk order of the lm317 and am going though my stash of heat sinks to get a driver up and running, the bar diode i have will be the load for them, until i saw your schematic i had not known they could be driven in parallel so easily so fear not your info is very helpful and i will be building at-lest one of them , the other plan i have for them is for some high power white leds i some times use for back up lighting and have a few 20 watters waiting to be used, since i blew an array of 395nm leds i have been wary of the Chinese switch mode constant current devices, while they may be more efficient they are a bit touchy on set up so i think they are better for battery chargers then led drivers. i sometimes rebuild battery packs and have been working on a way to weld tabs on to the cells and your desighn may work great for the interment load, i had a low voltage 80 amp power supply that will do constant current as well but it was not a cheap supply, i have 20 lm317's and wonder if that would work for that application to protect the glassman made supply http://www.glassmaneurope.co.uk/downloads/LVP1kWdc.PDF mine is the 80 amp version but works quite well, little off topic but i have an older high voltage power supply from them that experienced damage and had two broken meters, the 30kvdc meter and the current meter and even though the units almost 30 years old now they custom made me two new meters. allthough it cost some cash but these type of supplies are expensive and not easy to come by so 250 bucks was reasonable, that though i can not let my toad army near it, dont want any cooked toads, lol

[VDX]

Hi Draco,

the LM317 are rated for 1.5A max. amperage - I'm driving them with 1 Amp max. to reduce the generated heat.

For higher currents you can use the pin-compatible LM338, which are rated for up to 5 Amps (7 Amps peak!) ... I'm driving them with 2.5A continuous for the heat issue too.

When driving them with PWM you can go for max. or peak current, if the pauses are long enough (1:1 or more) in relation to the pulse-time ...

Viktor

[JohnYayas]

Is it common to use LM317s in parallel? I don't really know much about it but I wonder how well it works to equally distribute the current across them. Do you have to tune them or do they take care of it themselves. Are you worried about one possibly opening and overloading/killing the remaining ones? Not sure if it is a practical worry but wondered about it.

[VDX]

... I have an old 10Amps-voltage regulator with 8x LM317 parallel were 0.22 Ohm resistors were inserted after each LM317+shunt-Combo to avoid this 'cross-driving', so I've copied this in my first schematics too.

But, when configured for constant current regulator, this 'averaging' resistors aren't needed anymore, as the LM317 in combination with the shunt will 'open' only to the specified current, with which the voltage drop across the shunt resistor is equal to 1.25 Volts relative to the input.

As they are 'floating' at the plus voltage level without reference to ground (this is the main difference from the more common configuration as voltage regulators), they should be 'selfcontrolling' without interfering with the others ...

Viktor

[Bradfo69]

Attachment 49780 Jouleus Squeezer !
Cheers

LOL This is why I sometimes wish PL had a "Like" button.
.
This thread is pretty much why as much as I'd like to, I don't mess around with diodes that much. I have all the crap to do it but, I fear I'd be just another one of those that blows money up in poofs. I've killed 3 diodes in my limited experience and that's even been following peoples directions.

[VDX]

... I've killed some too with overdriving them - but this was either with wrong settings (I should know better), or for testing their 'absolute max' and 'safe' limits.

But I'm mostly doing this sort of "finding the limits" with spare diodes ... have some +30 pump-diodes out from wrecked fiber-lasers with 5Wats or 9Watts optical output, so enough for testing and developing the next ten years

Viktor

[mixedgas]

I tried VDX's schematic with a 317 clone from ST Microdevices.. No joy, that particular brand could not current share..., even with diodes and sharing resistors..

Sorry VDX... Not all 317 are created equal... ST parts got hot...


Steve