Using AC relays for step start HV supply

[W1IA]

I’m looking through the tech posts. The old broadcast transmitters using AC relays to control in-rush current on start up.
Thanks
Brent

[W1ITT]

This is common in the broadcast business, big honking multiple pole contactors with shirt button sized contacts that pull in with a manly clunk after a delay controlled by an RC network, or sometimes a small time delay relay to activate the coil,  They short out fat resistors, then just carry current for the rest of the day.  The modern fashion is to use SSR solid state relays, that would probably be more easily available and not terribly expensive.  But style and fashion isn't everything.   My 4-1000 was mechanical but the upcoming GS35B has SSRs ready to be bolted in place.  Either way works, whatever pleases ya.
73 de Norm W1ITT

[WQ9E]

It is also a good idea to put a fuse in series with the resistor so you don't fill your power supply with exploded resistor if the relay doesn't close. 

I use a step start system with a couple of my linear amps and I use an extra pole on it in series with the control line so the exciter can't key the amp if the step start relay doesn't close.

Rodger WQ9E

[W1IA]

Thanks Norm,
Helping Greg K1YW with his build on a pair of GS35B’s.
Chris W2JBL and Tm HLR used a similar AC relay pair to handle in-rush and provided a second AC relay as part of the keying circuit on the secondary. Just can’t find my notes.
I used a triode controller board for keying control and a typical shorting relay across a low value resistor etc. for my pair of GS35’s.
Brent W1IA

This is common in the broadcast business, big honking multiple pole contactors with shirt button sized contacts that pull in with a manly clunk after a delay controlled by an RC network, or sometimes a small time delay relay to activate the coil,  They short out fat resistors, then just carry current for the rest of the day.  The modern fashion is to use SSR solid state relays, that would probably be more easily available and not terribly expensive.  But style and fashion isn't everything.   My 4-1000 was mechanical but the upcoming GS35B has SSRs ready to be bolted in place.  Either way works, whatever pleases ya.
73 de Norm W1ITT

[W1IA]

It is also a good idea to put a fuse in series with the resistor so you don't fill your power supply with exploded resistor if the relay doesn't close. 

I use a step start system with a couple of my linear amps and I use an extra pole on it in series with the control line so the exciter can't key the amp if the step start relay doesn't close.

Rodger WQ9E

Good tip Rodger. I don’t think I did that, but did use a large ceramic cylinder style across the relay.
Brent W1IA

[WQ9E]

Brent,

I can't remember which model Ten Tec amp I repaired for a friend about 15 years ago but the step start didn't close and the burning resistor took out a board.

I built my triple 4CX800 amp a couple of years later and used the extra relay poles to ensure it doesn't blow up a resistor.  

I used the G3SEK tetrode boards which taught me a valuable lesson about solid state relays.  The tetrode boards were designed to control a regular mechanical relay and with the very low load of a solid state relay, the control voltage tailed off instead of dropping sharply.  If the control voltage doesn't shift sharply between the on and off control levels, during part of the transition it turns the relay into a rectifier which tripped the 20 amp breakers in the amp power supply.  I added a load resistor across the control output and cured that issue.

I also replaced the plate contactor in my Johnson Desk with a pair of solid state relays and I had to add RF chokes and bypass capacitors on the control lines to keep RF out of the relays.  During my initial tests, the relays were trying to modulate the AC into the plate transformer which doesn't work well  I like solid state relays but they do have their own special needs/requirements.

Rodger WQ9E

[W1IA]

Brent,

I can't remember which model Ten Tec amp I repaired for a friend about 15 years ago but the step start didn't close and the burning resistor took out a board.

I built my triple 4CX800 amp a couple of years later and used the extra relay poles to ensure it doesn't blow up a resistor.  

I used the G3SEK tetrode boards which taught me a valuable lesson about solid state relays.  The tetrode boards were designed to control a regular mechanical relay and with the very low load of a solid state relay, the control voltage tailed off instead of dropping sharply.  If the control voltage doesn't shift sharply between the on and off control levels, during part of the transition it turns the relay into a rectifier which tripped the 20 amp breakers in the amp power supply.  I added a load resistor across the control output and cured that issue.

I also replaced the plate contactor in my Johnson Desk with a pair of solid state relays and I had to add RF chokes and bypass capacitors on the control lines to keep RF out of the relays.  During my initial tests, the relays were trying to modulate the AC into the plate transformer which doesn't work well  I like solid state relays but they do have their own special needs/requirements.

Rodger WQ9E

I’m familiar with the G3SEK issue you are referring to. Mark KA2QFX and I modified the control board. There was a condition where B+ could be applied before the bias was applied to the tubes? The board has been ok in my amp. Sadly, no longer produced. Some nice features of the circuit. Greg can use a conventional design for controlling the keying and start up.
Nice part was setting Ip and Ig limits locking out the Ptt and resetting automatically. My plate supply is a bit small for the pair, so we set the Ip point to 1.75 amps. Running 4200 volts for the pair. The plate supply is a hypersil @ 1.25 amps using choke input filtering.
Brent W1IA

[w9jsw]

Here is what i did on mine that has the heart of a Gates BC in it.

Key features -
Stepstart with resistor and fuze
HV Enable relay
Variac on one leg of the 220 feed
3PDT configured for power on that will drop out if the mains glitch.

[W1IA]

Here is what i did on mine that has the heart of a Gates BC in it.

Key features -
Stepstart with resistor and fuze
HV Enable relay
Variac on one leg of the 220 feed
3PDT configured for power on that will drop out if the mains glitch.

John can you provide a full schematic? I like the design. If its not to much trouble, send to my email. W1IA@comcast.net. Nice job on the rig by the way. I have been wanting to build a pair of 4-250's or 4-400's plate modulated. Keith WA1HZK has an old plate modulated rig he picked up with 4-250's and the modulator. It needs some tlc, but it would be nice to have glow in the dark bottles in the station. I'm getting bored with the 3 legged fuses. Time for old time radio.
Cheers
Brent

[w9jsw]

I need to make a couple of updates to the schematic to reflect reality and I will then send you a copy. Give me a few days.

[W1IA]

I need to make a couple of updates to the schematic to reflect reality and I will then send you a copy. Give me a few days.

Thanks!

[W3SLK]

Being that it is my nature, (and part of my job) to be analytical but wouldn't you benefit from using some kind of Eagle timer, albeit analog or programmable? Or if you really want to go full bore, use a micro PLC to make sure all conditions are made prior to operation. Just a view from afar.

[KB2WIG]

Sadly, the aformentioned Tetrode Board is no longer available. I wanted to get one for my project, but no mo.

http://www.ifwtech.co.uk/g3sek/

The site does have the manual available, so the design is not lost. Maybe the Chinese will make the boards available........

klc

[K9MB]

That is a nice site.
I copied this circuit from the site and it is exactly what I have used in the past for a soft start circuit. Very reliable if you use a big resistor and the circuit keeps that resistor in the circuit until the voltage drop across the transformer primary rises enough to pull the 24 volt relay in and short out the resistor.
It relies on the actual state of the capacitor being charged rather than some selected time delay, though the pull in threshold can be adjusted by the small resistor in series with the 24 volt contactor coil.

[W1IA]

That is a nice site.
I copied this circuit from the site and it is exactly what I have used in the past for a soft start circuit. Very reliable if you use a big resistor and the circuit keeps that resistor in the circuit until the voltage drop across the transformer primary rises enough to pull the 24 volt relay in and short out the resistor.
It relies on the actual state of the capacitor being charged rather than some selected time delay, though the pull in threshold can be adjusted by the small resistor in series with the 24 volt contactor coil.

Thanks Mike,
That schematic is what I recommended to Greg K1YW. KISS mode is the best.
Brent W1IA

[K9MB]

That is a nice site.
I copied this circuit from the site and it is exactly what I have used in the past for a soft start circuit. Very reliable if you use a big resistor and the circuit keeps that resistor in the circuit until the voltage drop across the transformer primary rises enough to pull the 24 volt relay in and short out the resistor.
It relies on the actual state of the capacitor being charged rather than some selected time delay, though the pull in threshold can be adjusted by the small resistor in series with the 24 volt contactor coil.

Thanks Mike,
That schematic is what I recommended to Greg K1YW. KISS mode is the best.
Brent W1IA

Totally agree. Nothing easier than to compete with Rube Goldberg and the more parts, the greater likelihood of failure.
One additional benefit might be, that a short in the transformer secondary circuit will result in the relay dropping out and the current limited by the resistor.
Not-perhaps a solution for a short circuit, but it makes a bit less smoke and gives time to hit the kill switch….😉

[KD6VXI]

A step start requires 2 parts.  A 120 volt contactor with the contacts rated at the current level needed (I prefer at least a 50 percent margin) and a resistor.  You can add a fuse as wanted, to ensure not blowing the step start resistor.

Dale type resistors are garbage for this application.  They really don't like surge currents.  I tried paralleling multiples, tried series a couple, always ended up having to replace with a decent ceramic wire wound.

See the attached pic.  It really is that easy.  If you don't want to have a neutral line for 120 volt operation, substitute a 220 vac contactor coil.

As the capacitor in the power supply charges the resistor drops less and less voltage. 

At the point the contactor sees it's rated "pull in voltage", it will trip and apply full operating voltage to the power supply.

It's self regulating, self timing, doesn't require any extra parts and just..... works.

Some people report a buzzing with the contactor.  For me, that was contactor dependant.  Any newer contactor that was off-the-shelf available from electric supply houses didn't have this problem.

I've used this circuit from a .750 kva power supply to power supplies weighing hundreds of lbs.

--Shane
WP2ASS / ex KD6VXI

[K9MB]

I used this exact circuit on a 4x811a amp. Very similar to the 24vdc version but simpler, as you say with lower parts count than a DC relay.
I use 30 or 40 amp contactors from hvac or industrial applications.

[DMOD]

Late to the party but here is the Inrush Protection I use for up to 2kW:

Phil - AC0OB

[Opcom]

Sorry to the late reply.

Been using Mercury contactors or HVAC contactors, and heater-type inrush resistances for about 10-15 years now.

Been getting the Hg units for a song when they show up. Most are 20-30A per pole. Those seem to be burn-proof on the contacts and are silent. HVAC contactors withstand very high inrush intended for condensing unit compressor-starting and occasional locked rotor occurrences, and come in 25-50A sizes. They are usually 24VAC coils though coil voltage is a trivial matter.

Inrush resistors are the 500-600W radiant 'heating cone' type elements or, presuming the cabinet has room, large enclosed heating elements designed to bolt to a flat surface like a tank, but it is mounted to cabinet one way or another. Those type components will take it all day long. I expect the operator would notice the poor regulation or low B+.. But no burnouts or fuses to replace.

These things probably matter most when the plate supply is what us being keyed rather than just leaving it on for the whole session and cutting off the tubes with bias as some designs do.

[VE7RF]

I used a 0-15 second adjustable timer module, plugged into an 8 pin flush mount octal socket.  I wanted every last bit of B+ on the filter caps before the resistor is shunted, so I set the timer to 15 seconds. With just 1-2 seconds, once resistor shunted, I would get a massive secondary surge. The higher the value of resistor used, the bigger the secondary surge.  Reason is, the magnetizing current on the primary of the plate xfmr is flowing through the resistor, creating a V drop.   Hence the B+ will never get to max normal level. 

That jump from  6-7 kv is was causes the big secondary surge. I had a 0-50 amp, ac ammeter in one leg of the 240 line, and could see both the mag current, and also the sec surge.  The mag current   was 1.9 amps on my 120 lb  hypersil / 4800 vac pole pig.  The fix was to reduce the value of the resistor down to just 10 ohms.

When  T=0, the entire  240 is dropped across the 10 ohm resistor..then the caps charge up.... but not quite to max.  The jump, with resistor shunted, is then minimized quite a bit.

The 0-15 sec timer module contained a DPDT relay.   I used one set of contacts to activate the coil in the massive spst contactor used at the time.  The other set of contacts in the same timer module were routed to a pair of  RCA jacks, and the key line went in .....and out of those connectors.   That way, the amp could not be driven, until the timer had finished it's timing sequence.

A huge  2 pole contactor applies the 240 vac,,and also feeds  120 vac into the timer module input.  The DPDT relay in the timer module eats up 6 of the 8 pins, the 120 vac input eats up the remaining 2 pins.  I used  a pair of  wire wound 20 ohm @  50 watt  resistors in parallel for the step start R.  IF one ever opened up, the 2nd one is still in the circuit.

On start up, the initial current is just 240 vac / 10 ohms = 24 amps.... no big deal with big contactors.   I had  plenty of agastat / P+B plug in timer modules, and the flush mount sockets, so used em.   That 6700 vdc supply had a 150 uf filter cap in it.  With that much C, and the big xfmr, the step start had to be carefully thought out. 

On buddy's  3x6 amp, it used a 253 lb dahl + 163 uf filter and was almost 8 kv on  RX. We installed a BUSS  HVU-3 fuse in one leg of the plate xfmr secondary.... then fed to  FWB  assy.  With a 20 ohm resistor for the step start, and a plug in timer, the sec surge was so great, it would blow open the HVU-3 HV fuse abt 50% of the time.  Fix was to reduce it's value to just 10 ohms, problem solved.   That amp used surplus, nib 135 amp, 3 pole contactors. We strapped all 3 x poles to make one big 400 amp single pole contactor, then ditto with a 2nd identical contactor.  IE: one contactor in each leg of the 240 line.
 Strapped with 1" x 1"  angle aluminum,  1/8" thick..on both input + output sides.   A 3rd  identical 3 pole contactor was also strapped in parallel..and used to shunt the 10 ohm resistor assy.   Those 135 amp, 3 pole contactors were things of beauty,  $50.00  each, nib, from some surplus place in the se part of the usa.....and I'll be damned if I can remember the name of the surplus electrical place.

Each contactor also came with  3 x aux contacts.   We used 120 vac neons + 120 vac to neutral, to light em up, used as status lamps, to indicate the contactor had actually closed.

[VE7RF]

On my latest  B+  supply,  conventional step start  was a non starter..... not with strings of 10,000 uf  surplus  lytics used. The only way the step start would work, is to use real low values, like 3.3 ohms.  ( 6 x 20 ohm, 300 watt, metal finned arcols in parallel).   That results in an initial  73 amp surge, then a  big secondary surge.

Plan B used a small 0-280 vac @  4 amp  variac to bring up the B+ slowly, till it's 5%  higher than normal B+ levels.   Output of variac is fed to a dpdt relay... whose contacts apply the  vac  to the primary of the 253 lb dahl xfmr.   Once dialed up to  5% higher, a DPDT- center off toggle is thrown from one extreme to the other.  DPDT relay drops out..and main 300 amp 2 pole contactor slams down...applying  normal 238 vac to the primary.  B+ sinks  from 5%  above normal..down to normal.   NO surges anywhere, at any time.   I wired a 0-150 amp AC ammeter remote current xfmr around one leg of the incoming 240 vac.  It barely moves as the supply is slowly brought up.

With the dpdt- center off toggle in off position, everything  is OFF.  Down is the variac (+ it's small dpdt relay)..and up is the main 300 amp contactor.

On sage advice, I also installed a simple cam on the variac shaft, that when fully  CCW, activates a micro switch, which operates yet another relay..that latches to itself then applies 120 vac to the  dpdt relay coil... used to feed variac output to plate xfmr primary. 

Done this way.... in case I do something stupid, like forget to turn the variac down to zero / full ccw  when done for the night. OR if the commercial AC power goes off for whatever reason.  The simple circuit is known as 'fail safe'.  Nothing starts up until the variac is fully  CCW initially.   Input + output of variac is fused / breaker anyway, so a non issue really, but I wanted it safe, no surprises.

Sri for the diatribe.

Jim   VE7RF

[W1IA]

I want to thank everyone for all the tech advice. I was researching the topic for K1YW Greg. He’s building a pair of GS-35B’s similar to mine.
Greg went shopping at Keith’s WA1HZK’s for parts. His trunk was loaded with 300 lbs of plate transformer, chokes and assorted parts.
Back to the bench…Putting together a Clapp design 160/80 vfo for the E rigs. Took Jays W1VD’s design he posted on his webpage. http://www.w1vd.com/16080VFO.html
I’m almost finished. Check out my QRZ page.
Cheers,
Brent W1IA

[K8DI]

I had wanted to add to this another way to trigger a relay to short the inline step-start resistor. 

https://www.icmcontrols.com/product/icm102/

Simply put it in series with the relay and set the time.  Works with a 24v control system or directly on 120 or 240v. Time from 1.5 seconds to ten minutes. These are intended for HVAC systems and thus are both cheap and reliable — $10.00 or so.  I like them much better than the charge a capacitor until the pull-in voltage is reached method because it does not chatter, it just switches. I also use these for plate switching delay, providing filament warmup time. Once you set the time, it just works.

Ed

[WQ9E]

Ed, I also really like these electronic delay devices.  My homebrew contesting amp has one that keeps the blower energized for 3 minutes after the power switch is turned off so that it can conveniently cool without the filaments on and it doesn't require me to babysit it while waiting for post-operation cool down.

Rodger WQ9E