BC-610 (and other) Spark-Gap Help Needed

[ashart]

My restored transmitter (see www.w8vr.org) is rapidly depleting my bank account and sending me to the proverbial poor house.   Filter choke after filter choke suffers from insulation break down.  Why it does so now, and didn't have the problem in 1945 baffles me.  Perhaps insulation was better then.

However, it seems that some versions of the BC-610 used spark gaps across the filter choke to deal with such problems.

1.  Did the BC-610 use an off-the-shelf component or something fabricated by Hallicrafters?
2.  If an off-the-shelf part, does anybody know part numbers or specifications?
3.  Is there a commercially-built unit available today?
4.  Assuming all the parameters of the power supply circuit are unavailable and thus preclude calculation, is there a rule-of-thumb for specifying a spark gap to protect a filter choke?

All constructive and on-point help appreciated.

Thanks very much.

al hart
al@w8vr.org
www.w8vr.org

  

[w4bfs]

could you please elaborate .... is this a repeat failure of one choke or is this a group of chokes or all chokes

switching off dc current in a choke will produce a hv transient that can easily become damaging ...  spark gaps are just one way to deal with this

[ashart]

Hello to BFS, and thanks for your reply!

To attempt elaboration:
1.   I've attached a schematic.
2.   Xmtr built in 1945 had no problem over 16 years of use.
3.   Laid in a basement for about 45 years.
4.   Rebuilt and put back on the air.  Original chokes used.
5.   So far, about 3 filter choke failures.
6.   My 80-yr-old mind can't remember for sure, but  I think always the input choke.  Cannot guarantee this.
7.    1935 Hammond catalog rates the 10-300 as having "4000 volt Insulation Test."
8.   Replacement chokes have been with new cores, same  size as original cores. 
9.   Given that you, Mr. Faraday, and Mr. Lenz all recognize effects of sudden current interruption, why have the failures only been since the rebuild?
10.   Also, that v = -L di/dt and v = -N d(phi)/dt, was certainly well known in 1945.  Why did so few, practically none, of the power-supply designs of the era use spark gaps?   I know only of the BC-610.  The books of the day showing power-supply designs, don't seem to use them. An occasional textual reference to spark gaps, but they just didn't find their way into handbook schematics.

I hope somewhere here, you might find a clue.  It sure evades me.

73.

al hart

[w4bfs]

ok ...thanks for including schizmatic diagram .... nice drawing, I need the same here for my projects

possible rabbit trails:

     1.  those particular chokes were not made well ....a long shot .... you just never know

     2.  insulation degradation .... hard to justify if they are sealed in metal cans and in oil or pitch
 
     3.  original design maybe marginal on insulation .... this is more likely .... 4kv insulation rating is iffy at 2kv b+

so what can be done .... need to reduce turn off spike amplitude ... C1 and C2 will take care of L2 but as you see there is nothing to take care of L1 ... so lets add something...

     1.  add a resistor across L1 .... since all it is needed to do is snub, a high value resistor should be adequate say 10k ... this will have a small effect on the filtering action .... check with scope before and after .... use a series string of resistors

     2.  add a small value capacitor (say .1 mfd) either from T1 end of L1 to ground or across L1 .... be wary of resonance effects ... check with scope as before

be safe ....some of these tests require the use of hv probes and clip leads .... watch working voltage rating of parts ...73

[ashart]

Hi again -

I agree on the snubbing possibility.  I was, though, thinking of a diode snubber because I think they're more effective and because:

1.   Those old timers didn't have them in 1945, so they used spark gaps!
2.    I can get 'em!
3.    With power xfrmr secondary = approx. 2400 Vrms, then Vpeak = approx. 3.5 kv.  (Just running approximate numbers from memory of addled mind).   Need 10K PIV diode.
4.     Operating current through choke abt 0.3 amps.
5.     Energy in choke = 0.5 LI^2  =  0.5 (10) (0.09)  =  approx. 1/2 joule.
6.     I can probably find a 10K piv diode that will sweat out 0.5 joule.
7.     If not, I can use 2(or more) diodes.
7.     Will have to calculate peak discharge current, but haven't done that yet.

Where might I be going wrong with that analysis?

Thank you again.  I really do appreciate the conversation.

-al h

[w4bfs]

ok, diode snubber .... interesting .... are you going to use regular diodes or Zeners ... another possibility are movs and transorbs

all of those present possibilities BUT you will need to evaluate effects on L1 and the energy that it stores and releases ... a good way to test this is to use a dual trace scope set up to do a differential measurement (reverse polarity on one channel and add)  
(you will need X10 or X100 probes)   ..  this will let you see what is going on in L1 as it does it's thing

[KD6VXI]

What about a diode in series with the hv?

Seems I've seen that used on lower powered rigs before.....

--Shane
KD6VXI

[flintstone mop]

A dangerous suggestion would be to lift the choke from the chassis. Insulate it from ground,,,,,and YOU WILL have always remember not to touch it while the rig is energized. And always discharge the HV before any surgery to the unit.
Any way that you can scrounge around for the original chokes?
How are you exciting the rig? Externally or using the on board crystal osc/RF driver?
There seems to be something serious kicking back to the power supply.
Does this happen when the rig is keyed or when un-keying? Or when you simply turn it on and the power supply energizes after the rectifiers warm up?

Fred

[KA2DZT]

Just looked at your schematic.  Get rid of all the junk in the plate xfmr CT, you don't need any of it.  The first choke in the supply needs to have a very high RMS Test Voltage rating.  Most likely the over load circuit in opening for some unknown reason which can cause high back EMF in that input choke.

You need to understand the different ratings you see on filter chokes.  A choke with a 5KV RMS Test Voltage rating is only good for 2KV RMS.  The rating is reduced by first subtracting 1KV and then dividing by 2.  Chokes that have a RMS or DC "WORKING" voltage rating can be used at that voltage (the key word is working).  Some chokes will have some sort of a insulation rating which you can never be sure what the xfmr maker was thinking.  If it says 4KV insulation TEST then most likely you would not run more than half that voltage at best.

The BC 610 plate xfmr runs 2500 volts DC out of the filters.  This means the RMS voltage out of the rectifiers is probably over 2800 volts.  You need a choke rated for at least 7-8 KV RMS Test.  After the common 5KV chokes there is only the 9KV RMS Test Voltage rated ones.  Chicago made some, the R-1050, 10hys at 500ma or the R-67 which is 6hys at 700ma, both rated at 9KV test.  Collins chokes from some of their BC xmtrs, are Chicago R-67s.

I've been building power supplies for over 50 years, never lost a choke.  Get rid of the overload circuit in the CT lead.

What chokes have you so far blown up??  Let us know.

Fred

Took a second look at the schematic.  The over load relay opens what circuit??  I thought it was opening the CT tap lead, that's not the case, so what does it open??

[KA2DZT]

OK, I guess your supply is not the stock 610 supply.  Disregard my comments about the 610 supply. (Should have studied your entire posting before commenting).


Anyways,  looks like you have about a 1800-2000 volt supply.  Those 3KV Hammond chokes are most likely 3KV RMS Test Voltage rated,  good for about 1000 volts, maybe a little more, but not 2KV, especially on the input.

Fred

[VE3AJM]

Al, the information that I have on the Hammond 10-300X, rate this choke for max. 800 vdc peak operating volts, which is the DC operating voltage plus the peak of any transients or ripple.This choke wouldn't work well or survive use in a 1800 volt PS.

This is from a 1953 Hammond catalogue. It doesn't give an insulation test voltage though.

Your original choke seems like it was an older vintage Hammond. Does your 1935 catalogue have a maximum peak operating voltage rating for the 10-300 or 10-300X choke?

Al VE3AJM

[KA0HCP]

Why would 70 year old chokes/transformers die?

-Paper insulation breakdown/disintegration
-Wire enamel coating breakdown/flaking
-Wire corrosion causing mechanical shorts or high resistance/heating open/shorts
-Internal wire vibration causing shorts