Current sensing overload shutdown for tube transmitters - Hall effect

[K1JJ]

Continuing on with the Fabio II project, I am trying to design up an overload shutdown system.

I'd like to have both the RF deck screens and modulator screen supplies shut down if a certain screen current is exceeded in either. (300 ma DC)   Also if the main HV 240 primary exceeds a certain current, (say 20A  ac) then have the 240VAC contactors open up. (Along with the screen supplies)

Sampling the screen current directly is not that easy because it's 500 volts. There are other issues I will not get into.  Instead of overload relays or resistive samplers, I want to try some Hall effect devices that put out a voltage from magnetic sampling. I'd like to feed these three Hall effect devices into some 339 comparators with latching relays that in turn would open up the supplies and the PTT line.

Here are two Hall effect transducer devices I would like to try.  One is for the HV 240VAC primary and the other is for the screen 500 volt line. I'd like the screens to trip out at anything above 300 ma or so.  

I realize sampling the HV primary is slow due to the capacitors being able to deliver current even with the primary off. Maybe there is a Hall effect device that will let me wrap a number of HV insulated wire turns around it and sample the 3KV line. That would be quick to respond.  I don't know if a Hall effect can work at less than an amp, however.  The low current screen board below requires a direct connection and is rated at only 2.1 KV.

Maybe I'll end up using the negative return line for both the screen and HV supplies and avoid the HV problems. Better idea.

I want to run this idea by some of the guys who know this stuff well before I order the parts and possibly make a vely, bery dumb mistake...  



1)  240VAC Hall transducer:  The 240VAC primary wire would just go thru the hole, no direct connection)

http://www.ebay.com/itm/Non-invasive-AC-current-sensor-SCT-013-000-100A-Split-Core-Current-Transformer/121031821946?_trksid=p2045573.m2042&_trkparms=aid%3D111000%26algo%3DREC.CURRENT%26ao%3D1%26asc%3D27%26meid%3D7896132986285299432%26pid%3D100033%26prg%3D1011%26rk%3D2%26sd%3D321058374836%26


(2)   Low current Hall sensor circuit board: This is interesting, since low current sensors do not seem very common. The 500V lead would actually connect to the board, but it gets sampled magnetically and is insulated for 2KV.

http://www.karlssonrobotics.com/cart/acs712-low-current-sensor-breakout/?ecn

DataSheet for the chip:
http://www.karlssonrobotics.com/datasheets/BreakoutBoards/0712.pdf


Any ideas?

T

[KA2DZT]

First thought,  magnetic coupling isn't going to work with DC.  Those ebay sensor are for AC.

Second thought,  why do you think you need any type of current protection??  Fabio II isn't a class E rig.

With the dissipation of two 4-1000s in parallel,  probably simple fuses might be all you really need.

Fred

[K1JJ]

Fred,

The second board will measure both AC or DC, so OK for the screens.  

The first device listed was intended to measure the 240VAC primary, thus AC is OK.  But I may change my mind and look to sample the fil CT's instead, which will require DC.

This new system was to mainly protect the screens. If I were to lose the HV and have full drive, it could damage them despite being rugged tubes.

I guess at this point I need to know if I can wind a number of turns thru the most sensitive DC Hall effect device and put thru a 1 Amp or less DC current - and get a reasonable signal level to put into the 339 comparator.  These Hall devices seem to be mostly high current devices, not made for tube-type use.  Winding more turns thru will increase their output.

T

[steve_qix]

Hall effect devices work fine, but it seems overly complex for shutting down a tube transmitter.  I use them in the class E rigs because a fast shutdown is required.  Not the case here.  I have used sensitive DC relays to great advantage in sensing current in tube transmitters.  Be *sure* to bypass the coil with a suitably large capacitor to bypass the audio, else the coil looks inductive.

Other methods:  If you can spare 5 volts or so voltage drop (suitably bypassed, etc.) you could use an opto-isolator.  I use these in the class E  pulse width modulators.  They work well up to a couple thousand volts of isolation.  There are very high voltage devices available as well.

The circuit could be self-powered using a bypassed zener diode or a resistor feeding another resistor and a zener.  You only need enough voltage to power up the sending side of the opto isolator.

The relays are the easy solution if all other conditions permit their use.

[steve_qix]

Another idea:  Use a vactrol (make sure you have some method of protecting the light emitting diode from too much current - I have ideas about this if you go further with the idea).

The Vactrols are light dependent resistors.  Almost an open with no diode current - very low resistance with enough current.  Could easily trigger another circuit.  And, no power is required to make them work!  Look at the VTL5C3 as a good example.

[KA2DZT]

Tom,

Was reading the specs on the AC DC low current board,  it speaks about some sort of noise issue, take another look.

I see you're trying to protect the screens in case the HV fails.

You can use a simple relay.  110VDC small relay, pull some current off the HV supply through big resistors, no HV the relay opens which can open the screen circuit.  Just one way of doing it.  There are other ways.

Fred

[KA2DZT]

Tom,

Another thought,  If you use a current sensing device,  what about the changing screen current with modulation.  Seems that may be an issue to consider.

Fred

[K1JJ]

"You can use a simple relay.  110VDC small relay, pull some current off the HV supply through big resistors, no HV the relay opens which can open the screen circuit.  Just one way of doing it.  There are other ways."

Bingo.  Fred, that's an interesting idea about sampling the HV with a resistor/relay and killing the screen voltage if the HV disappears.  Very clever.

That may be the simplest solution since loss of HV is the biggest threat to the screen. I could also have that same relay kill the PTT at the same time.  Use a sensitive DC relay (like 5V) off the bottom of the HV metering divider.  Set it to come on when HV is at least 1500V.  It doesn't need to be a latching relay either - and no reset required..  It would work silently in the background. Cool.

Thinking more, when the sequencer activates both the HV and screen, the screen will not key up unless the HV comes up first.

Gotta give that some more thought.


Steve, OK on the suggestions.  I did go over the overload relay idea, though it requires a latching relay. They seem too expensive mechanically and to do the latching electronically seems as complex as doing a faster Hall effect.

I also see Hall effect switches and related stuff on the web. I find it hard to believe there is not an off-the-shelf device that can go in series with a lead, set the pot to a sensitivity / trigger current threshold and have an on-board relay to the outside whirl. I've been looking on eBay and can't seem to zero in on it, if it exists for 300 ma, etc.

I may still add in a 240VAC Hall device in the HV primary just as a breaker back-up. I could tailor it to exactly where HV current shutdown needs to be for Fabio II.    The linears have different requirements and could have their own presets.

T

[MikeKE0ZUinkcmo]

Why not a resistor in the negative screen and plate returns, and sense for over or no current with a comparator.   You'd need a negative supply but you can easily get that from the filaments.   

you could also just use a sens resistor and opto coupler in the HV lines to sens over current as well.   

[KA2DZT]

Tom,

The relay needs to be a higher voltage DC type like at least 110VDC.  Reason, it will need less current to work.  Lower voltage relays need much more current that requires much bigger resistors to drop 2-3KV to get down to the relay voltage.

Only bad feature of using this method is it will only protect the screens if the HV fails up to the point you connect the dropping resistors.  Meaning, if you connect them at the B+ terminal of the mod reactor, anything after that point that fails won't open the relay.  Things like the reactor, connections, plate RFC or you forget to put the plate caps on the tubes.

Fred

[W2VW]

Tom, I like sensing the screen current itself. Loss of plate is the most likely problem but since when did your failures adhere to most likely descriptions 

There's still parasitics, loss of protective bias, loss of proper plate circuit loading, operator error wrong band/antenna, and more.

The screen chassis return via relay bypassed for audio as mentioned looks FB. 

[K1JJ]

Tom, I like sensing the screen current itself. Loss of plate is the most likely problem but since when did your failures adhere to most likely descriptions  

There's still parasitics, loss of protective bias, loss of proper plate circuit loading, operator error wrong band/antenna, and more.

The screen chassis return via relay bypassed for audio as mentioned looks FB.  

Thanks for the suggestions guys.  Yep, looks like there is still a lot of risk using just the HV sensing.  Maybe I should sense the screen itself and get it over with - cover all bases.

Well the over-current screen relay might do.   I'm a little confused at which relay to pick for the job.  Whatever relay is used needs to drop the rated voltage across the coil, right?  So, will a 120V DC relay subtract 120V DC  from the screen supply?  At 280 ma, that is 33 watts being burned up by the relay and shunt resistor across the relay coil, no?     Now if it were a 5 VDC relay, then 280ma at 5V would be only 1.4 watts total wasted, and only 5V of supply drop.

What am I missing?   What would be the optimum coil to use for a negative lead screen relay?    

Then comes the latching relay problem. The bandits want $50 for most latchers.   I could use a more complex latching circuit with a simple relay. But at that point we are almost back to the complexity of a Hall effect and comparator which is probably faster when used with a vacuum relay.  This is why I've been going in circles with this design.

T

[W2PFY]

Are clamp tubes unsuitable for 4-1000 circuits? The Johnson 500 uses such a circuit as well as many others. Maybe they only protect against loss of drive?

[W2VW]

Tom, I like sensing the screen current itself. Loss of plate is the most likely problem but since when did your failures adhere to most likely descriptions  

There's still parasitics, loss of protective bias, loss of proper plate circuit loading, operator error wrong band/antenna, and more.

The screen chassis return via relay bypassed for audio as mentioned looks FB.  

Thanks for the suggestions guys.  Yep, looks like there is still a lot of risk using just the HV sensing.  Maybe I should sense the screen itself and get it over with - cover all bases.

Well the over-current screen relay might do.   I'm a little confused at which relay to pick for the job.  Whatever relay is used needs to drop the rated voltage across the coil, right?  So, will a 120V DC relay subtract 120V DC  from the screen supply?  At 280 ma, that is 33 watts being burned up by the relay and shunt resistor across the relay coil, no?     Now if it were a 5 VDC relay, then 280ma at 5V would be only 1.4 watts total wasted, and only 5V of supply drop.

What am I missing?   What would be the optimum coil to use for a negative lead screen relay?    

Then comes the latching relay problem. The bandits want $50 for most latchers.   I could use a more complex latching circuit with a simple relay. But at that point we are almost back to the complexity of a Hall effect and comparator which is probably faster when used with a vacuum relay.  This is why I've been going in circles with this design.

T

If yer screen supply uses the typical industrial primary power circuit (relay powered by one time application of voltage, voltage kept alive by second set of relay contacts) the screen relay can simply interrupt the screen primary relay's coil circuit. There's a proper name for this but I don't know it. Latching circus as you say maybe. To me that isn't complex compared to the comparator.

Steve said it well. You don't need the speed of the silicon control.

Same thing with all these threads. There are a dozen OK ways to make it work and many more not so OK ways. Great to see how others have set theirs up.

[KA2DZT]

Tom,

If you want to use the current sense method it's not that difficult.  Estimate the max screen current you'll allow.  Using that current,  use a low voltage DC relay 5-12 volts.  Measure the DC resistance of the coil. Calculate what resistance you need to drop the coil voltage using your max screen current.

Example, using 200ma, 12VDC relay.  You need 60ohms,  If the coil resistance is near 60ohms, just put the coil in line with the screen current circuit.  Since you're running 500 volts, probably put the relay in the neg return line, but most any relay coil and frame will handle 500 volts.  You may have to also figure in any other currents (like bleeders).

If the coil resistance is higher than 60 ohms, then just bridge the coil with the correct parallel resistance.  If the coil resistance is lower that 60 ohms, figure using a different voltage or allowable current level.

I forgot exactly the resistance of most common 12V relays, I'll check it later.

You'll right this method will need some sort of latching, otherwise the screen circuit will just cycle on and off.  Probably there is a way to do it with two relays, but I'll have think about it.

Fred

[Steve - K4HX]

If solid state is what you wish to use, a resistor and a differential amp will do the trick. Hall devices are overkill and probably overly complicated.

[KL7OF]

    I'm not saying that the circuit protection devices are not needed....I'm soliciting comments from those that have damaged unprotected transmitters by losing the HV with drive applied...I'm just wondering how common HV failure is....  I use a clamp tube in my 813 rig....

[K1JJ]

If solid state is what you wish to use, a resistor and a differential amp will do the trick. Hall devices are overkill and probably overly complicated.

Yep, that's what I have decided to do.   Using small value sampling resistors in the two negative screen leads and the two fil CT's, I will monitor  screen currents for both decks and plate currents for both.

I'm building up a master box with four separate inputs to a quad LM 339 comparator, driving a 2n2222A which latches a standard relay.    If one current exceeds a certain preset threshold, then the PTT will toggle off (latch off) and the sequencer will go thru its normal unkey sequence.    This will shut all the supplies off including the RF drive and not key up until the overload is reset.

This will give full control using four pots for these overload presets.  

I figger anything that could hurt the rig will probably show up with either a screen or plate current excess.   HV arcing is another thang, but this is protected by tiny wire fuses in the 3KV lead along with 240VAC primary breakers.  

The linears all have similar overload shutdowns and work FB, so might as well do it with Fabio II too.

I also want to experiment with using a Hall effect device directly on the negative HV supply output lead driving a vacuum relay to open up the HV hot lead in milliseconds. This wud feed all amps in the shack.


Thanks for the ideas, guys.

T

[K1JJ]

I finally found a set of boards that will work together to sample both the screens and fil center taps - and then latch a relay.

The first board is the Hall effect pickup that will work up to 2.1 KV.  I will put one board in each of the leads - four boards.  Mod screen current -  RF screen current  -  Mod fil CT plate current  -  RF fil CT plate current.

Each Hall effect board directly drives a second relay board. It is an eight channel relay unit that takes a 2.5V DC signal to close a relay.  I will use only four channels.  The four relays will drive ONE latching relay that will have a reset.   This one latching relay will control the PTT of the sequencer. After being reset, unless all Hall effect boards are below their preset currents when keyed up, the PTT will drop out.  If ANY of the Hall effect boards exceed a pre-set current, the sequencer will unkey in an orderly manner.

I plan to put all boards into an RF shielded box with the bypassed screen and CT leads coming into the box for sampling.  Another lead will go to the sequencer for overload disable. Needs +5  and +12V to function.


Combined with the sequencer, this appears to be an effective and inexpensive way to build a commercial shutdown overload system. I ordered all the boards, so it's a done deal...


Hall effect board:  (4 needed  - This board will work with low currents and is fully adjustable)

http://www.karlssonrobotics.com/cart/acs712-low-current-sensor-breakout/?ecn


Eight channel relay board with DPDT relays:  (1 needed)

http://www.ebay.com/itm/Eight-DPDT-Signal-Relay-Module-Board-12VDC-SKU177004-/400495273680?pt=LH_DefaultDomain_0&hash=item5d3f60e6d0


T

[KB2WIG]

T,

I dont trust this automatic stuff. Beware.


klc

[w4bfs]

just got in some Salon boards .... rreasonably good build quality .... how can they make this stuff so cheaply ?

[K1JJ]

just got in some Salon boards .... rreasonably good build quality .... how can they make this stuff so cheaply ?

I always wonder about this myself.  I look for the catch, but it's never there.  They sell all kinds of boards. Some cost only $3 and free shipping from Hong Konk. Impossible.   The boards are great quality.

Maybe they made a huge stockpile of all different kinds and got stuck.  These boards and parts just dominate eBay. I see some USA vendors buy qualities and mark them up X4 and charge shipping.  Or maybe these manufacturers have a long range plan of putting all competition out of biz and later raising prices.

At the present prices, gas would be at 50 cents a gallon again.  Maybe it's the cheap labor and government support, I dunno.

All I know is except for the 2 week wait for shipping, they always work FB.  The customer service is courteous and fast.


T

[W2PFY]

Are clamp tubes unsuitable for 4-1000 circuits? The Johnson 500 uses such a circuit as well as many others. Maybe they only protect against loss of drive?

May I address my question above again? Are clamp tubes impractical for the 4-1000 tubes. Would such a circuit be too slow to protect the tubes? Perhaps not enough protection for tetrode modulators? I would like someones take on this. Thanks!

[K1JJ]

Are clamp tubes unsuitable for 4-1000 circuits? The Johnson 500 uses such a circuit as well as many others. Maybe they only protect against loss of drive?

Hi Terry,

I've never used a screen clamp tube, nor spent the time thinking it out, but they appear to sense a no-RF drive condition and then limit screen current.  Unless a clamp tube circuit can be configured to work with a fixed supply, a screen dropping resistor is impractical for 4-1000A use due to power waste.

However, just adding a small amount of fixed grid bias would solve this problem and also limit plate current when no drive exists.

Probably the manufacturers who use them depend on fuses to handle the other tube elements.

Maybe some of the other guys have more info, like,  if plate voltage is removed with full RF drive, (and other risky situations) does a clamp tube help at all to limit screen current or is it usually just for no drive conditions?  I suppose the basic circuit could be modified to sense other conditions too.

T
 

[KA2DZT]

Terry

Clamp tube circuits generally are used to protect the finals if the drive drops out.  The clamp tube conducts when the neg grid voltage drops out.  The circuit only works if you modulate the screens through a screen resistor from the modulated plate voltage.  If you use a fixed screen supply and a choke the clamp circuit can't be used.  Also, if you are using some protective neg fixed grid bias the clamp circuit is not needed.  They work as fast as the grid voltage drops.

Other circuits and relays are used to protect screens if the HV plate voltage drops.  This is different from clamp tube circuits.

Fred