New 4-1000A modulated by 4-1000As AM Project - Fabio II Reborn

[KA2DZT]

Reading your comment again, you shut off the HV supplies before unkeying and there was no arcing, that's a good sign.  Seems that using the sequencer, it could be the drive is dropping out too soon, wereas there is still enough HV still present when the drive cutoff stops the current too suddenly.

[KA2DZT]

If it blew while transmitting, then the problem has nothing to do with the back emf that occurs at keyoff.  The thing to remember is that under normal transmitting there is still very high voltages that occur at the reactor on voice peaks.

I think the relay you used is just not insulated enough for such high voltage.

The one thing you did prove out is that just shutting off the supplies greatly reduced the voltage spike to where you had no arcs.  Just what I thought would happened.

I think you should increase time between HV shutoff and drive shutoff.  That 100uf cap may be holding the charge a bit longer than you figured.  Another thing, maybe a smaller filter cap may be helpful.  Maybe only 30-50ufd.

I use 30ufd in my HV supply and a 50hy Heising reactor, I never had any arcing problems.

Fred

[K1JJ]

Fred,

The back EMF may have started the carbon trail before - then when I ramped up to 3KV, it blew when transmitting. Hard to say. 

Yeah, looking at the relay that blew, the coil has a long metal cover that extends to near the HV contact arms. It is not visible inside, but appears not rated for the job, even though it says 15KV.  I had that feeling when I installed it - it is only 1" tall, tiny, really.  The same relay in the screen damper seems OK.

I ordered a used Jennings 20KV vac unit that is 3" tall and obviously better suited. I think the guy will give me a credit on the bad one.

As for the sequencing... my sequencer  has fixed intervals and the order is the same in unkey and key but reversed.  If I slow down the overall sequence, it is too slow.   I might need something that will let me put a long delay in the HV position and also turn the HV on as a 1st or second step and drop out as a first step.  That wud give it more time do discharge it.

I'm going to keep 100 uf in the supply. I use it for the linear amps and want the regulation.  I shud be able to get it working at 100uf.

Despite whether the new relay will be able to hold up with the HV damper, it still concerns me that there was a huge arc when I unkeyed without the HV damper in place. I am not keying the CTs and have a damper on the screens. So, it comes down to the heisting reactor dumping power and HV supply timing.


T

[KA2DZT]

You mentioned that the HV snubbing resistor was getting warm after many cycles.  It would cause it's damping the HV spike from the reactor.

You also found that letting the HV shutoff before unkeying you have no arcing.  Better to design the rig to have little or no back emf so you don't need to use the snubbing resistor.  At least that would be a target to try to hit.

We also know that dropping the drive causes a sudden drop in plate current.  The fact that you were getting some heat in the snubbing resistor tells you that the drive was causing the problem while there was still enough HV left to still have current flowing when the drive drops out.

You proved that shutting the HV before unkeying created no arcs.  Only thing left after that is to drop the drive and antenna.

Seems to me, you need to increase the time between the HV shutoff and the drive shutoff.  This way the HV has more time to decay down to nothing or a very low value before the drive drops.

Fred

[KA2DZT]

OK FB on the timing not being adjustable between steps.  No sure what to do, not sure what exactly drops the drive, a separate relay controlling the driver or what?  If it's a DC relay that is keyed by the sequencer and the relay controls the driver, maybe try adding a cap on the relay coil to delay it opening.

Having said all of these great ideas, I think you should go with the better relay you ordered and use the snubber to cover all bases.

Trying to fine tune the timing may be too difficult to achieve perfectly and you know that the snubber was working great until the relay failed.

Fred

[K1JJ]

Some good ideas, Fred.

Well, here's what I've come up with...

I need a sequencer that will let me go forward or backward in a different order. Also the ability to change the time between steps wud be nice, but not mandatory.


The desired KEY-UP:

(Dampers are already resting ON)

1) HV ON   Screen ON   (HV Step-start on early will keep clacking noise out of key-up audio)

(no plate or screen current yet due to fixed grid bias)

2) Ant On

3) RF drive ON

4)  Both Dampers OFF

5) Audio ON



Un-key

1) Audio off   - Both Dampers ON

2) HV off   -  screens off

3)  Long Space or no assignment

4)  RF drive off

5)  Antenna relay off


Notice the HV and screen needs to key-on early -  and then un-key early.

It all makes sense to me doing it this way. However, I need a different sequencer to do it.

T

[KA2DZT]

Those two sequences look good.  Having the HV turn-on ahead of the antenna shouldn't be a problem as without drive there wouldn't be any output.

The only thing I don't understand is why the plate current stops dead when the drive drops.  I would think the plate current would rise without drive.  If I shut off the drive on my rig the plate current increases.  To prevent that I use a 6Y6 clamp tube on the screen voltage.  The 6Y6 is controlled by the neg. grid leak voltage.

Not sure why your rig does the opposite.

F

[WD5JKO]

This thread has been most interesting. I'm curious though, was the class E rigs flame out more/less scary than FabioII's? For sure, silicon does not recover easily after a flameout.

A good read on the use of a smaller reactor:

http://www.vias.org/eltransformers/lee_electronic_transformers_07_09.html

If I follow the article linked above, using Heising cap and reactor as parts of a high pass filter, can be done using much smaller cap (more Xc), and smaller reactor (less Xl), and yet still get 30hz to 10Khz response. The voltage transient issue would also be less extreme.

Jim
WD5JKO

[W3RSW]

Good read.
So if this was a seminal and industry changing article in the late thirties , would not most of the designs and BC transmitters constructed since then have the PI net, lighter weight components already installed?  Are not all the RCA's , Gates, etc. and our copies thereof already incorporating the PI AF filter method.?

Tom, sorry about your zorch.  Man, you are buying some high class replacement components.  Speaking of designs;
How are the high power short wave xmitter boys sequencing their stuff?  Harris n'gang.

[w4bfs]

Some good ideas, Fred.

Well, here's what I've come up with...

I need a sequencer that will let me go forward or backward in a different order. Also the ability to change the time between steps wud be nice, but not mandatory.


The desired KEY-UP:

(Dampers are already resting ON)

1) HV ON   Screen ON   (HV Step-start on early will keep clacking noise out of key-up audio)

(no plate or screen current yet due to fixed grid bias)

2) Ant On

3) RF drive ON

4)  Both Dampers OFF

5) Audio ON



Un-key

1) Audio off   - Both Dampers ON

2) HV off   -  screens off

3)  Long Space or no assignment

4)  RF drive off

5)  Antenna relay off


Notice the HV and screen needs to key-on early -  and then un-key early.

It all makes sense to me doing it this way. However, I need a different sequencer to do it.

T

let me think about this .....

[KA2DZT]

Tom,

Thought about my last question.  You're using -90V fixed bias, which probably is enough to bring the plate current down to a very low amount, I remember class B level.  What you need is for the finals to pull more current with no drive.

If you can, reduce the fixed bias to a less neg amount (use more grid leak bias if needed) so as to allow the finals to draw current (with no drive) which is closer to the normal operating current.  I think you're running about 400ma during operation.  This way with no drive the finals will bring down the supply at a constant rate just as if you still had drive.  You already did this somewhat when you manually shut off the HV before you unkeyed and you had no arc.

This test may be worth a try.  You can do this with the same sequencing arrangement you're using now.

The thing to remember is that even with 500 volts left in the supply you'll still get an arc when the drive drops out because of the sudden drop in current to the class B level.  It's that sudden change in current that's causing the problem.

The screens only run 500 volts and you were getting arcing in that circuit before you made the improvements.  With that big Heising reactor you wud probably get some arcing even with lower voltages.

Fred

[AMLOVER]

Tom,

If the HV damper became hot means that the reactor's relay often blocked in the "off position" and part of the modulation energy passed to the ground through the damper and the HV bleeders. From the kind of failure seems that your relay was very week and couldn't follow the on-off orders.
My HV is 5500v under load and I use a big Jennings for 15KV, 26.5v coil, never had a problem with hot resistor or relay coil failure.
When you'll feel secure with on-off by using dampers you will realize that keeping all voltages on during receiving is not at all dangerous as soon you use 2 basic protections.

First protection is the RF amp's screen overcurrent relay that cares if you loose RF hv, in this case this relay disactivates the RF amp's screen voltage and the modulator's screen voltage, too (cutting primaries).

Second protection is the RF amp's safety bias setting which must be so that in case of loosing excitation the RF amp stays idle at 10-50ma which in my case was many times my good angel and as Fred mentioned earlier will keep the reactor not to dive suddenly to the no current/huge voltage situation.

2 protections, one bias setting and one current relay let me keep voltages always on and on-offing as fast as the linears do...

However step start and voltages on-off make things simple but considerably slow,
2-5 secs is a safe time for using this way.
In this case you don't even need the dampers and the sequence is simle like this :

                            
                                 ON and oposite OFF

1)ant relay to tr
2)exciter on
....1 sec delay....or 3 empty clicks in the rotary switch
3)RF hv and screen on (step start ac)
....1 sec delay...or 3 empty clicks in the rotary switch
4)Modulator hv and screen on (step start ac)
5)audio on

A 12position 1pole rotary switch with the 1st or 12th position connected to a huge "on air" is the key for all traditional amers here in Greece....

Stefano

[K1JJ]

Great info, thanks.

I checked the screen damper relay that had its coil in PARALLEL with the blown HV damper - it was OK. So was the sequencer relay controlling them. I'm amazed that the HV relay melted down its coil but the screen relay was untouched.

Fred, yes, I have -90 fixed safety bias and about -200V grid leak.  I am trying to find a smaller 50w zener diode for say, 30 to 50 volts so I get SOME plate and screen current idle. Right now -90V fixed just cuts it off. Having that small idle would stabilize things more during bleed off.  Seems DigiKey and Mouser do not carry them anymore, or maybe I missed it. I want the stud mount type.

Stefano - OK on the sequencing and rotary switch.  We use fast break-in here, so need the PTT.  I'm finding about 3/4 second for the HV step-start is about right to bring it up to 3KV. If I activate the HV step-start first, it will be ready by the time the other functions key up.  I plan on screen overcurrent relays soon. I have them in all my tetrode linears and they have saved me many times.

Jim, I've had my share of class E rig shutdowns during initial testing, but the shutdown board worked well. I usually like to shake down a rig by putting it thru its paces and ramping it up, expecting to break parts. It then becomes a reliable rig once shaken down. I just feel more comfortable breaking parts in a tube rig rather than a solid state rig. Troubleshooting is easier for me there. There probably isn't a big difference in flashover reliability once they both are working as designed.

John / BFS:  I think your experimental sequencer might be my next try if it is not too complex.   Is this something that you made a board for or is it a kit?

I did look into Bruce / XR's suggestion to use the NCC timer relays. I am not sure if they can do a reverse order different than forward like I need.

Look at the "Timer Delay Relays"
http://www.nationalcontrols.com/products/productcategory.cfm?ProductCategory_Id=4307&Product_ID=1323#Prod1323

Maybe a string of these wud work.


I think Fabio II is very close to a working rig now - just need to get over the hump of these timing issues.  The audio peaks certainly looked impressive on the scope running at 3KV before it puked.


T

[KA2DZT]

OK FB, I thought you were using a bias PS.  Since you're using a zener I guess it's in the cathode circuit.

I don't have a plate current vs grid voltage graph for the 4-1000 so I don't know what grid voltage to use.

Any way you can build up a neg bias supply and use it instead of the zener.  You should be able to add it in right at the grid meter.  Although, I'm not exactly sure what your circuits look like.

You can do this just for testing purposes and then go back to the zener if you can find them.  As for zeners, you can use a string of them, lower voltage and wattage using more than just one big zener.

[K1JJ]

Fred,

I'm using a standard bias supply applied to the grid, but regulating it by a 50 watt  -90V zener.  I just found some zeners at Newark and ordered a 62V, 33V, 24V.  Some combination should give the RF amplifier some idle when the drive is off.

I use a zener for the fixed grid bias cuz without it the RF drive will sometimes rectify and charge up this bias line and cause the bias to get higher and higher until it's too hard to drive.


I took the burned relay apart. Strange, even though it had been arcing and didn't click anymore, inside there was no noticeable signs of damage or arcing.  The coil was still good as well as the main contacts.  But the bottom at the 24V pins was all black.  It's like the HV contacts were arcing to the negative lead of the 24V.  The polarity EMF diode was shorted across the 24V pins.

I guess the new relay will work FB, being three times the size and all glass.

Still wondering what to do about the sequencer. Maybe John's design will fit the bill.  Making it go backwards in a different order seems to be the challenge.

T

[KB2WIG]

"  I took the burned relay apart. Strange, even though it had been arcing and didn't click anymore, inside there was no noticeable signs of damage or arcing.  The coil was still good as well as the main contacts.  But the bottom at the 24V pins was all black.  It's like the HV contacts were arcing to the negative lead of the 24V.  The polarity EMF diode was shorted across the 24V pins.  "

Could it be Satan?


klc

[KA2DZT]

Tom,

OK FB on everything.

I think you're getting close to having the rig work correctly even though you may not think so.  I hoping that an increase in idle current may be the final answer.  The sequencer you're using should work with the way you have it set up.

I'm trying to guesstimate what idle current you should try to use. Figure what your operating current is, I'm thinking 300-400ma (at 3KV 300ma = 900W input). I would guess that the plate current, at the exact instant the drive drops, is maybe half the operating current (just a estimate).

Meaning, from the time the HV is shut off to the time the drive drops, the plate current has maybe decayed to half.  So, if you set the idle current to about the same, when the drive drops out the plate current will still be about the same.  The HV PS and the Heising reactor will never know the difference (provided you don't spill the beans and tell them).

F

[w4bfs]

I am going to be busy at church for next couple of days ... we cook a meal for our shut-ins

I think Stephano and you Tom are right about the relay .... probably a partial loss of vacuum .... when you get the new one see iffin you can hi-pot test it to at least 8kv peak or so

I think I can get a sequencer to do what you need .... these only exist as a design ....I don't have time to lay out boards .... I was going to use a prototype board for my gpt-750 redo

[W2NBC]

Dear Father of the Fabio's,

The ultimate sequencing (using the time delays mentioned by Bruce)-

Take 4 time sequenced 'delay on make' relays for transmit 'on' sequence, through a 12 or 24 volt latching relay(s).. Set sequence to "Electron Wrangler" standards..

Take 4 parallel 'delay on break' relays  (now in circuit by the latching relay(s) being thrown), and set 'Un-Key' sequence  to "Electron King" standards..

On "un-key" the latching relay(s) comes home to start the day all over again..

It should work.. you throw the details together..  At LEAST use it in Fabio III

It's nice to have Sequence Freedom  

[K1JJ]

Interesting, Jessie.

Look at the "Time Delay Relays"

http://www.nationalcontrols.com/products/productcategory.cfm?ProductCategory_Id=4307&Product_ID=1323#Prod1323

Which relays would you suggest for the job?  These are the NCC ones I think Bruce uses.


So we would be using four relays for make and four different relays for break.

I'll have to draw this out and see if I can make it work on paper.  Thinking more, I definitely need a sequencer that will reverse in different order to get the HV on early and turn it off early. This will give the HV a longer time to decay and also clack it's brains out before the audio goes on.  


TNX, OM.


John,  I'll look to see what you come up with. If I go with the Electronic Wrangler idea, I will let ya know.

Thangs happen very fast in OZ.


T

[W2NBC]

Bruce specifically mentioned the ones with the dip switch (numerical) set control as opposed to the pot.. I would contact him directly (sure you will anyway).. The cool part is 120 volts for make and break rails, through a latch.. Simple Wrangler Techno.. 


 at work, so I will give it more thought probably while caught in traffic later..

J

[K1JJ]

OK Jessie -

I drew up a schematic of a timer-based sequencer, but although it may work, I feel it may be a little Rube Goldberg-ish.  Could you draw up something of the basic design? I don't see how the latching relays are needed or how the relays work together and independently very clearly.

What is the difference between a make and break timer?  Can't we just use a standard make timer and use the other set of NC contacts?  I must be missing something overall.

I sent a PM to Bruce for his schematic, but haven't heard back yet.

This appears the best way to go with full control as you said.  Might as well build it custom for the job, caw mawn.


Tnx.


T

[K1JJ]

Hmmm...  Take a look at these delay timers for only $3.61 each.  I wonder if they could be used in a group of ten to give five steps of the desired sequencing forward and backward in any order?


http://www.ebay.com/itm/DC-12V-Delay-Timer-Switch-Adjustable-Module-0-to-10-Second-w-NE555-Oscillator-/230964055573?pt=LH_DefaultDomain_0&hash=item35c687ce15#shId


The NCC relays are very expensive and maybe these Hong Konks can do the trick instead.

Anyone have a circuit?

T

[W2NBC]

T,

Yes, Rube would be proud! (it's the thought that counts, right?) 
Thought about this some more..  All this is the raw brute force approach to sequencing..  It's the "Switch and Command Sequencer" caw mawn..

The point is you require a different sequence for key-up and un-key, simple right? **NOT**

*(These NCC timers are all available in specific time ranges and voltage  ..you'll find the range you want, and most are available on E-bay.. I used 120 Volts as a reference!)

Let's take 8 standard say DPDT 120 volt relays ..  4 for the Key up side (left)  , 4 for the un-key side (right) and lay them side by side .. Line up (4) Q1T series "delay on make" next to the left side and (4) Q3T series "delay on break" to the right.. Parallel all relay poles (left side to right side) except for one (see below)..

1: To sequence the key-up (left side) apply 120 volts to the RELAYS (the load) through the (4) Q1T-Series timers (delay on make) set to your requirements..   OK on key-up sequence.. simple stupid so far!

The trick is the LAST RELAY in your key-up sequence..
2: Use an unused pole of that relay to switch the un-key side (right) into the act by closing the "start" switches on the (4) Q3T series "delay on break" timers all connected to your right side (un-key) relays. Remember, the right side timers will have 120 volts applied full time (see PDF Q3T), and you will only be closing the start switches..  The right side relays are in parallel and they will be closed only after your last key-up sequence..  No latching required!  Then, the right side is in "control and command" because when you unkey (remove 120 volts from the left side and De-energize the relays and start switch) the right side stays energized for as long as you want ("delay on break") and release in whatever sequence needed..

I just got home (it's 3 in da morning -Tornado news coverage)  so I will make a drawing up in the next few days..

http://new.ametek.com/content-manager/files/NCC//Q1T-Series-Series-Load.pdf           (key -up delay on make)
http://new.ametek.com/content-manager/files/NCC//Q3T-Series.pdf                                (un-key delay on break)

PS. I use the Q3T with my "Johnson-Gates 500" through relays to delay on break the antenna relay and RF drive (transceiver used as a vfo) and it is VERY reliable..

Just saw your last post.. they would work.. do they make delay on break ones too? Anyway, the theory above should work. It really is the Land of Oz..  I'm off to dream about relays and sleep 

[K1JJ]

Hi Jessie,

I appreciate the explanation.  

I'll start looking for the timers and see if I can accumulate five make and five break.

Please post the schematic when you can.

After drawing out some other ideas while trying to use the Hong Konk timers, I gave up on it. The next step wud be a digital controller of some kind.  Your idea seems to be the only practical way to get the reverse order changed.  There are great advantages to being able to do this.  Lots of relays and timers, though...

Tnx, OM.

T

**UPDATE:  I found three NCC break-delay and four make-delay NCC   .05 - 1 second  120VAC timers at good used prices.

Also found a lot of ten  DPDT 120VAC standard relays at a great used price.

So maybe I can find the remaining timers soon.  I want five-five total.   Interesting that this whole system runs on 120V.   No big deal cuz it will interface to the outside whirl using floating relay contacts.