Relay Sequencer glitch

[K6JEK]

I'm using the Advanced Receiver relay sequencer in one of my transmitters.  I think some of you might be doing the same, maybe some of you Class E guys.

Mine energizes all the relays for just a fraction of a second when it first sees power, i.e, when I first power up the 12V supply.  Clackity clack. Wakes a fellow up.    Does this happen for you guys too?  Is there a workaround?

It doesn't matter for me.   The HV isn't on.   But I want to recommend it to someone sequencing a big RCA broadcast transmitter.  He would not be amused by a quick pop of the B+.

I also pointed him to Don's two-step sequencer in ER which I've adapted for various uses, but my buddy (for reasons I don't remember) is enamored with more than two steps.

Here's the link to the little beast.

http://www.advancedreceiver.com/page20.html

[K6JEK]

I contacted the guys at AR.  They said they consider it a feature.  I think I'll look for a simple mod to their board to stop it.

[k4kyv]

They said they consider it a feature. 

Sounds like a bug to me, and they're trying to put lipstick on a pig.

I can think of absolutely no useful purpose for turning on all the relays at once when you first power up a T/R system, but it certainly could result in major damage.

[W1EUJ]

It would let you know the board and the relays are working; a power-up self-test.

[W1VD]

The relays don't turn on all at once during power up...it completes one normal ramp up and ramp down as a self test to show that everything in the system is working.

It was designed with this in mind.   

[k4kyv]

That would be a good feature if you could turn it on or off at will.  If there is no way to turn it off, that's a classic example of feature creep.

[W4EWH]

It was designed with this in mind.   

I heard of a company that made control circuits for traffic signals: the timer mechanism they used depended on a neon bulb (I don't know why), and it turned out to be sensitive to temperature. In other words, the signals would take longer to change in cold weather because the timer circuit wasn't operating at its intended frequency.

The company patented the design and sold it as a automatically-adjusting, weather-sensitive "smart" controller that allows for longer start and stop times when roads are icy and motorists are being cautious.

Not quite sure why, but I just thought of it for some reason ...

Bill W1AC

[WA1GFZ]

I've worked on sequencers and interlocks for years and the more complicated they are the bigger hunk of junk they become.  I look at some out there and get a good laugh.

[Mike/W8BAC]

Yeah, but think about the stand alone nature of that sequencer product. I built mine instead of buying somebody's product. When starting the station up I have to remember to cycle certain things first and second because of the sequencer closures that happen as tubes in the mixing board and other things warm up.
I turn on the exciter and transmitter after all of that is finished and I'm on the air. This is all analog and it's no inconvenience. I think buying a nice sequencing board like that would have saved me many hours of R&D and I don't think I would mind a few seconds of start up system check. As a matter of fact I could design a TX interlock to wait until the self test is complete.

[WA1GFZ]

Every someone else's product I have seen is full of failure modes and way too complicated. A good sequencer is dirt simple and has built in interlocks so you can't advance until you have a valid step.

[steve_qix]

Is the schematic available?  With that, it may be possible to come up with a SIMPLE fix (or workaround).  If you can post it or email it, I bet some solution can be found 

Regards,

Steve

[K6JEK]

Is the schematic available?  With that, it may be possible to come up with a SIMPLE fix (or workaround).  If you can post it or email it, I bet some solution can be found 

Regards,

Steve

Yes, I have the sequencer schematic somewhere.  I'll find it.

Jon

[WA1GFZ]

If the sequencer is supposed to make one pass up then down, why not just sequence the way you power up your rig? I

[w1vtp]

If the sequencer is supposed to make one pass up then down, why not just sequence the way you power up your rig? I

Right on!!  I think it's great that the product is available.

Al VTP

[steve_qix]

Hi Jon,

I took a look at the circuit - looks very straight-forward and fairly foolproof.  Here is one possible workaround to the issue you are having.  I made things as absolutely simple as possible.  The circuit uses a single NPN transistor, that will hold the output of U1 low (preventing any relays from firing).  The output of U1 is short circuit protected.  The 2n3904, as configured will sink well over 100mA, about 10x what U1 will deliver, effectively holding the output low for a time.

The hold down time is determined by the time constant of C1 - R1.  When power is first applied, C1 is discharged, allowing current to flow to the base of Q1, turning it on hard.  When C1 charges completely (through the base-emitter junction of Q1 and R2), no current will flow, and Q1 will turn off completely, allowing the output of U1 to rise - ultimately triggering the relays when the sequencer is set to the transmit mode.

Figure the value of C1 for however long you want the circuit to stay dormant.

Anyway, that's one quick-and-dirty fix, and there are probably others.

Maybe we can get Jay to comment about this, too!  :-)

[K6JEK]

Thank you Steve.  I will try this and let everyone know. 
 
GFZ suggested I sequence the way I power up the rig. I do  I have this sequencer in a class H modulator.  I fire that guy up with the HV switch in neutral.  Everything clicks.  Nothing happens.

The application I'm worried about is for an RCA BTA-1R.

Jon

[WA1HZK]

I'm pretty sure Brent W1IA uses a few of these with good results. I use a more complex one and have had RF issues with the 833 rig. I would use this one in a heart beat for the next transmitter.
Keith

http://www.w2drz.ramcoinc.com/Sequencers.htm

[WA1GFZ]

Hey Steve,
Q1 will sink nothing until you give it some base drive....

[Opcom]

Not all off-the-shelf sequencers are alike. The only reason to have relays clattering during a device turning on is to exercise them. A power-on-self-test should include a passive indication as an option so that the energizement can be avoided if necessary.

The only reason a relay needs exercise@power-up is if it is passing very sensitive signals and the exercising claps the microscopic crud off the gold-plated contacts (for example so the -110dB signals are passed faithfully). Appropriate for a telecom line length simulator. Not appropriate for a control sequencing device. For every software "feature", there is usually a hardware fix.

It's not quite fair to complain about the advancedreceiver sequencer's behavior, since it is made for a transverter, not really for sequencing a large TX rig. Most components of a transverter won't even be up and running before the POST is done. A time delay could also be added to the controlled equipment, although that partially defeats the purpose of the sequencer. Nonetheless, with the sequencer's little relays, you would want to add bigger ones, so the time delay relay could be inserted in the common returns of the big relays.

Without some way to tame the beast, I can see the MV rectifiers in a big rig getting a quick flash during this little P.O.S.T..

This example of a possible malfunction threatened by a design feature of O.T.S. acessories is one reason I prefer to make my own controls, thyratron-infested and retro as they might be. 2D21, 884, 2050.. ahh the thyratrons. At least they won't go on a conducting spree until a few filaments in the subject equipment have warmed.

The W2DRZ unit looks pretty good and with 10A relays, you'd only need maybe one contactor for the KW section.

As to Q1, it would get base drive through the capacitor, as the cap charges. The bigger the cap, the longer the base is driven. The basic circuit is well-used in many television sets of old, to make a voltage to start the power supply control IC. Once the supply was then started the IC got "run" power off one of the HV supply's low-voltage windings. This avoided an additional cost to the standby power supply.

[WA1GFZ]

It will take one large cap to make a long pulse.

[steve_qix]

Hey Steve,
Q1 will sink nothing until you give it some base drive....

Yes, that's right.  The instant the unit powers up, current will flow through the [discharged] capacitor, C1, through the 4.7k resistor, and into the base of Q1, turning it immediately on.  

Assuming that the +V supply for the transistor is connected to the same +V as the remainder of the sequencer, the transistor will be turned on as soon as there is power, and everything else in the circuit can do what it wants - nothing's going to happen until that capacitor [C1] charges up, and at that time, current will no longer flow to the base of the transistor, turning it off, and allowing the circuit to work.



The key to the circuit's operation is that the cap is essentially low resistance for some finite period of time, until it charges up.  It charges through the base-emitter junction of transistor Q1, and that current flow turns the transistor on.

Let me know if there is some flaw in my logic !

Thanks and Regards,

Steve

[steve_qix]

It will take one large cap to make a long pulse.

I was figuring on starting with somewhere between 47uF and 100uF (at 16V, so the part is very small).  100uF will give a .47 second time constant, but it will probably delay the circuit longer because the capacitor really has to charge all the way to stop the current flow, and that will take longer than the .47 seconds.  It may perform perfectly with as little as 25uF.  I didn't plot or calcuate the charging curve.  It's so non-critical, I figured it would be just as easy to wing it 

Regards,

Steve

[WA1GFZ]

A reset pulse will be generated

[steve_qix]

A reset pulse will be generated

Hi Frank,

I'm not sure I follow what you mean here.  Let me know !

Thanks !!!

Steve

[Opcom]

It will take one large cap to make a long pulse.

Would a darlington pair with a 2N2222 front side to feed the existing 2N3905 and a 47K or 100K resistor instead of the 4.7K one increase the time constant sufficiently while also allowing enough base drive?