Seems to me the old Halli warning was related to opening the switch, not closing! How much voltage spiking and or arcing takes place in the switch when it is opened under load as opposed to under standby?

In the event that this is the problem, the solution might be to open the switch only near zero current---

?

--
don??? va3drl

?

FYI, in case my posts don't seem to be in logical order, it's because they aren't.? My messages appear on the forum in a sequence different from that in which I submitted them for posting.? Maybe I'm posting too often.
Halden VE7UTS ?

?

Hi Don,
Thanks for bringing us back to the original question and for mentioning one of the options I hadn't listed!? Turning the transformer OFF at a particular time relative to the AC voltage cycle might cause a voltage spike as the flux collapses faster than during a regular cycle!? If that's what causes this failure, how would a pause in STBY mode mitigate it?
Your point about low magnetic remanance interests me.? If low remanance means the amount of magnetization still present after removing voltage at the primary is low but persists indefinitely, then the timing for opening the switch might not affect inrush current next turn-on.? If low remanance means that the magnetization remains for a short time after power removal, then simply waiting a little while before turning the switch back on could avoid a heavy inrush current but probably not affect a turn-on HV spike at the secondary.
Here's another idea about the possibility of a turn-on timing causing a voltage spike at the secondary.? If the remanent magnetic field has one polarization and the voltage at the moment of turn-on moves it towards the other polarity, then there's even more room for a fast change in flux, and the accompanying voltage spike.
I think it's time to hook this transformer up to the storage 'scope and try a few things.
Contemplating that setup (how shall I simulate loads on the secondaries?) raises another idea.? When filament secondaries are connected to cold filaments, the heavy current draw on those windings might load down the rapidly rising flux at turn-on, reducing the turn-on voltage spike magnitude.? Turn-on immediately after turn-off means no cold filaments would be available to do this.
Cheers,
Halden VE7UTS

?

?

Thanks, all, for the comments and lessons!?
It seems we're converging on using zero-crossing transformers that are capable of switching whatever current the transformer would draw if only the winding's DC resistance limits the current.? I measured the HT-37 primary to be 0.616 ohms (4-wire, Keithley 2100 DMM).? A #18 AWG power cord 1.5 meters long would add 63 mohms to this and a 4A SB fuse about 31 mohms more.? The wiring from the AC socket to the breaker panel might add another 100 m ohms.? This means the maximum surge current could be as high as 150A if there is really no back emf.? The estimate from the Opto22 document suggests 10 times nominal which is much less than this.
The HT-37 transformer is probably close to 250 VA, so the [M]P120D4 is the OPTO22 recommendation.? 4A continuous and 85A for a single cycle, 66A for 3 cycles.? This seems ample capacity if using Opto22's "10 times" estimate.? If a turn-on drives the transformer into saturation, then the additional current that flows doesn't actually cause any more flux and thus no voltage surge at the HV secondary.? OTOH, a relay that turns on immediately could cause the flux to rise more quickly, to saturation or not, causing the voltage spike.? Did I get this right?
In many cases, I the zero-crossing SSR in use actually does have the capacity to handle the surge current even though the person installing it wasn't paying attention to this issue.? Thus, many people using ZC-SSRs haven't had problems.?
The part I don't understand is:
"Large inrush currents can occur during the first half cycle of line voltage if a zero-voltage SSR
happens to turn on during the positive half cycle of voltage when the core is already in positive saturation." (Opto22 document cited earlier in this thread)
How could a transformer already be in positive saturation if it has been unenergized for a while prior to turn-on?
Halden VE7UTS

Jeff, as a guess, your likely insulation breakdown was enough to continue building a carbon track between windings even though the 5v was isolated; and must have eventually destroyed the T-T [turn to turn] insulation on a winding and eventually became a real short. Any T-T short in a transformer is fatal . don



From: [email protected] [mailto:[email protected]] On Behalf Of Jeff AC0C
Sent: Monday, July 1, 2024 5:55 PM
To: [email protected]
Subject: Re: [HallicraftersRadios] HT-37 transformer failure due to rapid STBY-->OFF-->STBY



Two the fellow's earlier comment on not using the 5V line, when
reworking my 32A, that was the first thing I did - move to SS
specifically to idle the 5V line and avoid the risk of the well reported
short involving the 5V line. Toward the end of the process, I was
getting ready to finish up alignment and had let the transmitter run
overnight to ensure it had reached stable temp across the chassis. When
I came down the next morning, the rig was dead. Painfully confirmed
later that some amount of windings on the primary or secondary had
shorted (there were no detectable winding-to-winding shorts at idle -
and I did not have a spare HVPS available at that time to test higher).
However, as the variac was ramped up, the current draw of the
disconnected transformer soared. So the 5V to HV short is apparently
not the only transformer failure mode with this model, unfortunately.

I solved the problem by cannibalizing another HT32. But assuming Murphy
may like to come see me soon, I bought one of those shiny Dahl
transformers and it's here on the bench ready to run to the rescue when
the current transformer eventually dies. Of course, with my luck,
simply having bought the Dahl transformer may keep Muphy away allowing
the current transformer to run the rest of my days. ha ha.

73/jeff/ac0c
alpha-charlie-zero-charlie
www.ac0c.com