|
Components and Circuits For discussions about component types, alternatives and availability, circuit configurations and modifications etc. Discussions here should be of a general nature and not about specific sets. |
|
Thread Tools |
31st May 2017, 11:49 am | #1 |
Dekatron
Join Date: Jul 2009
Location: Sleaford, Lincs. UK.
Posts: 7,637
|
Crowbar circuit?
I've made a small 12.6v PSU to power some valve heaters. Powered it up today and noticed the valves were very bright. They would be with 34v on them, the PSU LM317 had gone.
Any idea's for a simple protection circuit I can put on the OP of the PSU? I thought about a zener, maybe a 13v one or better still some sort of crowbar circuit. As room on the already made PCB is tight, even better if it can be done with one or two components. Andy.
__________________
Curiosity hasn't killed this cat...so far. |
31st May 2017, 11:57 am | #2 |
Moderator
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,801
|
Re: Crowbar circuit?
Zener into the gate of a large thyristor. A few hundred ohms from gate to cathode so it only triggers once there is a decent amount of zener current.
Popular in amateur radio 13.6v supplies where a fault in a cheap PSU can take out an £3000 transceiver. David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done |
31st May 2017, 12:08 pm | #3 |
Dekatron
Join Date: Jun 2016
Location: London, UK.
Posts: 3,687
|
Re: Crowbar circuit?
See Art of Electronics 3e chapter 9, conveniently available as the free sample chapter: http://artofelectronics.net/wp-conte...3_chapter9.pdf
Page 690. Highly recommended. Pass transistor failure was what killed my Marconi 2019A dead by dumping 16v onto the 5v rail. If it's AC, then two zeners back to back of 7.5v and a few watts each and a slow blow fuse should do the trick. |
31st May 2017, 12:20 pm | #4 |
Rest in Peace
Join Date: May 2003
Location: Harlaxton, Lincolnshire, UK.
Posts: 3,944
|
Re: Crowbar circuit?
Why did your LM317 fail? Did you perhaps exceed one or more of its maximum ratings?
Follow David's advice and test your circuit with a dummy load and fit a suitably rated fuse on the unregulated side of the LM317. |
31st May 2017, 12:57 pm | #5 |
Dekatron
Join Date: Jun 2016
Location: London, UK.
Posts: 3,687
|
Re: Crowbar circuit?
This might be the kickback from driving inductive circuits that blows out the pass transistor. Did the circuit have the relevant protection diodes on it? You need one across the output and one across the regulator itself.
|
31st May 2017, 2:49 pm | #6 |
Heptode
Join Date: May 2015
Location: Bradford, West Yorkshire, UK.
Posts: 510
|
Re: Crowbar circuit?
The SCR circuit described in the book is excellent, no need for more than 4 junkbox components and works well.
__________________
worried about the electrons entering the circuit and the smoke leaving Andrew |
31st May 2017, 3:57 pm | #7 |
Heptode
Join Date: Mar 2009
Location: Portsmouth, Hampshire, UK.
Posts: 674
|
Re: Crowbar circuit?
How about TL431 and a Triac, see fig 32 of the datasheet.
This has the advantages of an easily settable threshold and a harder 'knee' than a zener. (Turning a Triac or SCR on slowly with low gate current can lead to device failure caused by a uneven distribution of current across the die.) Jim |
31st May 2017, 5:58 pm | #8 |
Dekatron
Join Date: Jul 2009
Location: Sleaford, Lincs. UK.
Posts: 7,637
|
Re: Crowbar circuit?
Some nifty ideas there, I'll try those out tomorrow. I have the AOE, keep forgetting to look in it for ideas.
The LM317 failed due to the PCB slipping onto a metal chassis and shorting Colin, I had it temporarily wired up. I had tested it on a dummy load, but fitting no protection wasn't a good idea. And no, I didn't have diodes on it as per datasheet, another oversight, I will pop some on. Will 1N4148's do or better to use 1N4001's? Luckily no valves were harmed, I had my hand on the off switch. Thanks, Andy.
__________________
Curiosity hasn't killed this cat...so far. |
31st May 2017, 6:24 pm | #9 |
Dekatron
Join Date: Jun 2016
Location: London, UK.
Posts: 3,687
|
Re: Crowbar circuit?
1N4001 will be fine. 1N4148's are a little wimpy.
I've lost count of the things I've blown up doing such things |
31st May 2017, 6:53 pm | #10 |
Dekatron
Join Date: Jul 2007
Location: Evesham, Worcestershire, UK.
Posts: 4,241
|
Re: Crowbar circuit?
I think the clue to that is in the first post: 34V.
That's a hell of a lot to drop across an LM317 (T, I assume), that is supplying valve heaters (so 100s of mA). The poor thing will have been running rather warm. If possible, lowering the input voltage would be well worth doing. As the load is fairly constant, even a fixed resistor ahead of the 317 would do. |
31st May 2017, 7:08 pm | #11 |
Rest in Peace
Join Date: Feb 2004
Location: Hexham, Northumberland, UK.
Posts: 2,234
|
Re: Crowbar circuit?
I protect the shack 13.8 volt supply bus by using an overvolt sense circuit driving a very large automotive relay. When the 13.8 volts is 13.8 volts, the relay is held in and everything is happy. If the voltage climbs above about 14.5 volts the relay drops out and disconnects the load. Very agricultural but it works. This was designed after my trusty linear PSU lost a series pass transistor and started sending 15 volts to my radio equipment. Luckily they all survived after I noticed a display flickering and switched off quickly. The circuit is well decoupled to RF to prevent spurious tripping but I haven't experienced any problems so far.
Alan. |
31st May 2017, 9:37 pm | #12 |
Dekatron
Join Date: Jun 2016
Location: London, UK.
Posts: 3,687
|
Re: Crowbar circuit?
Agreed. I've got a 15v 1W zener and a 2A slow blow fuse on the input of the transceiver I'm building. Confirmed the fuse pops before the zener of course
|
1st Jun 2017, 10:52 am | #13 |
Dekatron
Join Date: Jul 2009
Location: Sleaford, Lincs. UK.
Posts: 7,637
|
Re: Crowbar circuit?
Firstly Mark makes a good point that 34v to 12.6v is a lot for the LM317 to drop even with a big HS. I screwed up when I wound the tfmr. Had a play this morning and found a big 20 ohm resistor drops the IP V to the regulator to around 19v, it is a lot happier at this, the HS remains comfortably warm to the touch. the dropper R is mounted on metalwork
An aside here, where is the best place to put the dropper R? On the AC off the tfmr or after the bridge rectifier that feeds the reservoir cap? Am I right in thinking it would be better to put it the AC? I read page 690 on of The Art of Electronics last night concerning crowbar protection. A simple zener and SCR crowbar has issues as you probably are aware. The main one is that it's imprecise. Jims suggestion of using a TL431 is also covered in the good book. It is more precise and it is adjustable so can be set more precisely. I've ordered a few. I can't remember exactly, but think valve heaters should be kept +/- 10% of their recommended value, but 5% would be better if we can get the crowbar to trigger at 13.2v. I looked in my zener box and didn't have suitable values really, to try out a circuit. The best I had was a 9v1 and 4v7 which gives us 13.6v. Am I right in thinking that zeners over 5v are true zeners and have a sharper knee? if this is the case then maybe this combination may well do. Depends on the tolerence though. Having wrote this have had a rethink so will knock up a circuit. Andy.
__________________
Curiosity hasn't killed this cat...so far. Last edited by Diabolical Artificer; 1st Jun 2017 at 11:20 am. |
1st Jun 2017, 11:10 am | #14 |
Dekatron
Join Date: Jun 2016
Location: London, UK.
Posts: 3,687
|
Re: Crowbar circuit?
LT1084 is a good replacement for LM317 if you want more PD.
Only zeners around 6.2/7.5v have the best dynamic impedance (sharp knee). The ones higher and lower voltage are pretty rubbish. For the sake of what you're doing though, I don't think absolute precision is required. TL431 is one of those ICs that finds numerous usages all over the place. Worth having a few in. |
5th Jun 2017, 6:58 am | #15 |
Heptode
Join Date: Oct 2005
Location: Olympia, Washington, USA.
Posts: 663
|
Re: Crowbar circuit?
Andy-
Be sure to fuse the PS. The crowbar will blow the fuse and protect things. I saw one where they didn't fuse it. Charcoal & smoke. BTW You can no longer use "Dummy Loads" here. To be politically correct, they must be called "Disadvantaged Loads" now. :P |
5th Jun 2017, 1:48 pm | #16 | |
Nonode
Join Date: Oct 2014
Location: West Midlands, UK.
Posts: 2,181
|
Re: Crowbar circuit?
Quote:
Having been brought up ( electronically speaking ) in the religion of always under run, ,PERHAPS something like a 7815 in series might be better, so that 317 is not dissipating so much power. As for a crowbar circuit - looking at my old Bible"Electronics circuits", idea would be to fit a zener & resistor in series( zener to +) across the final 12.6 and use this to control a thyristor across the pre regulated supply Book gives formula as Vth = Vz + V gt where Vth = crowbar volts Vz = zener volts V gt =thyristor gate voltage With of course a fuse lower than the max current of the SCR . |
|
6th Jun 2017, 2:55 pm | #17 |
Dekatron
Join Date: Jun 2009
Location: Staffordshire Moorlands, UK.
Posts: 5,263
|
Re: Crowbar circuit?
I'd use a TVS diode. There are tables in some of the datasheets for what voltage to use to protect what.
__________________
Kevin |
7th Jun 2017, 10:12 am | #18 |
Dekatron
Join Date: Jul 2009
Location: Sleaford, Lincs. UK.
Posts: 7,637
|
Re: Crowbar circuit?
Have a fuse on the PSU Frank.
Am using a LM317AHV which can handle up to 57v. Still it runs better with a 20ohm resistor to drop the voltage on the in pin to 19v. I'm waiting on a delivery of SCR's to try this circuit out. A TVS diode looked a neat solution, however it doesn't really tell you if anythings blown,wheras a crowbar will blow the fuse, alerting you to a fault condition, it will just clamp the voltage. Still I had a look at available TVS diodes at Farnell. Most are SM which eliminates them unfortunately (homemade PCB track separation too big. Am also a bit puzzled by some of their specs. EG reverse standoff voltage. Also on one example it says min breakdown voltage 8.33v, max BDV 9.21v, this for a 12.9v clamping V. So does this mean it starts clamping at 8.3v to 9.2v? Need to look at these further. Posty just been, later.... Andy.
__________________
Curiosity hasn't killed this cat...so far. |
7th Jun 2017, 11:16 am | #19 |
Dekatron
Join Date: Nov 2006
Location: London, UK.
Posts: 3,496
|
Re: Crowbar circuit?
A TVS has a completely different function - it's for fast transients, which may last nanoseconds but are often repetitive incidents. In this case its capacity to handle colossal currents and overvoltsges is incredible. You need something for a fault condition, which unless you take appropriate measures , will go on lasting until something blows up, melts or catches fire.
__________________
Al |
19th Jun 2017, 5:14 am | #20 |
Dekatron
Join Date: Jul 2009
Location: Sleaford, Lincs. UK.
Posts: 7,637
|
Re: Crowbar circuit?
A bit of an update on this one. After a bit of experimentation and making a few mistakes I got a circuit working. In the end I used a 5w 12v zener I had and two 1N4001's to give me a latch on V of 13v. The circuit works well breaking a 350mA fuse quickly. the C106D SCR I used doesn't get hot. It's mounted to the body of a chassis.
All good there but I had less luck with using the TL431C's. Tried as I might, I couldn't get the things to work, though I did have a working circuit as per TAOE p690 version C using a triac instead of an SCR briefly. However when I turned the circuit on again I couldn't replicate my results. I was using the triac to turn on a green LED for visual indication/load. The LED would light straight away or wouldn't turn on using the TL431C. So off with the triac etc to see if I could get it to work as a straightforward zener. Nope. Either I bought fakes (ebay not my usual Farnell ) or my understanding of these devices is in error. A.
__________________
Curiosity hasn't killed this cat...so far. |