I am working with some coaxial connectors. I would like to get a better handle on maximum voltage. I have consulted 3 sources from Microwaves101 to Pozar but cannot recreate or support lab results with theory. To start, we know the voltage is limited by the coax section where the inner and outer conductor come into closest contact. This is the rear of the connector, outer conductor at dia. .048" (0.0012192m) and inner at .0210" (0.0005334m) with an air dielectric for the section. It's also known the electrostatic field in a coax is Epeak(r)=Vpeak/(r*ln(b/a)) where a is inner and b is outer radius, respectively. The maximum voltage for this coax is the voltage just before dielectric breakdown, in this case air, which we will take at 3MV/m, which happens at radius a. Therefore solve for Vpeak:
3000000 = V/( (0.0005334/2) * ln (0.0012192/0.0005334))
Vpeak= 661 V = 467Vrms.

The problem is the connector is rated for 670 Vrms. How could this be? I would say it is an error, but we had four devices tested by an outside lab and indeed they achieved 670 Vrms at 5MHz according to MIL-PRF-39012. Basic calculations do not support this.

 

The maximum voltage for this coax is the voltage just before dielectric breakdown, in this case air

Why do you assume it's air?
It would be the dielectric of the connector.

 

Why do you assume it's air?
It would be the dielectric of the connector.

I'm not assuming. It is a 2.92mm connector, the rear dielectric is air. The only non-air dielectric is the capture bead near the center of the body, not in the rear section, which should be the voltage limiting section.

 

I would go by the manufacturer's specification, if there is one.

 

Interesting question
I've never looked in details at the connectors max voltage
only ever used the manufacturers specification,

Your saying the things test as per spec, but your calculations show it should be much lower, by factor of two ?

I dont know
Are they tested by any chance "connected"

 

Hello,

As said, have a look at the datasheets.
RG 174 is 1.4kV DC, 1 kV RMS.
RG 58 is 4 kV DC, 2 kV RMS.
For RG 213 this is given:
Max. Operating Voltage : 3700 rms
Corona Extinction Voltage : 5000V rms
Voltage Withstand : 10000V rms
Spark Test : 5000V rms

Bertus

 

Hello,

As said, have a look at the datasheets.
RG 174 is 1.4kV DC, 1 kV RMS.
RG 58 is 4 kV DC, 2 kV RMS.
For RG 213 this is given:
Max. Operating Voltage : 3700 rms
Corona Extinction Voltage : 5000V rms
Voltage Withstand : 10000V rms
Spark Test : 5000V rms

Bertus

Are these not all coax cables , the op is talking about an un specified connector.

 

Interesting question
I've never looked in details at the connectors max voltage
only ever used the manufacturers specification,

Your saying the things test as per spec, but your calculations show it should be much lower, by factor of two ?

I dont know
Are they tested by any chance "connected"

Yes, really racking my brain on this one. The connectors test per spec when we sent them to an outside lab, which would be no breakdown @ 670 Vrms at 5MHz. Have in-process oscilloscope and EUT pictures in report so not a mis-reporting. They were tested "connected" back to back with a .012 rear pin captured between the rear contact sockets. We sent them in for testing assembled as such so we know.

As other posters have said, we can trust the spec, especially with test data to back it up, the question is why or how this is possible. Electrostatic field analysis shows a coaxial structure of this geometry should not be able to reach this voltage before breakdown. I'm open to hearing the calculations are wrong...

Field analysis shows it should breakdown at 467Vrms, or about 660V DC, less than tested and rated. We actually have a DC and 60hz AC on-site HIPOT tester and some connectors and I'm leaning towards destructive testing to see if they can hold that here, disregarding the 5MHz, to help satisfy curiosity.

 

An impedance mismatch will cause standing waves on the coax. This can cause (in the worst case scenario) up to twice the source voltage to appear at certain spots along the line.

 

The connector is spec'd for 670V RMS and confirmed by lab tests but what is the breakdown voltage?

 

The connector is spec'd for 670V RMS and confirmed by lab tests but what is the breakdown voltage?

I wouldn't know exactly unless additional destructive testing was performed. They only spec "voltage rating" at sea level and 70,000 ft. Probably a safety factor involved, which is why we did not see breakdown when testing at 670Vrms at sea level. This is higher than calculated 660VDC breakdown for the rear geometry.

 

An impedance mismatch will cause standing waves on the coax. This can cause (in the worst case scenario) up to twice the source voltage to appear at certain spots along the line.

You are correct, but this phenomenon would make the actual tested rating lower than calculated, which is the opposite of what is happening. I am also skeptical of the magnitude of reflection at the 5MHz tested.

However, if we did had 100% reflection, the oscilloscope might read 670 Vrms when actual input was only half that (dont know much about scopes). Math says this should still breakdown, unless the critical sections are closer to nodes of the standing wave, avoiding the min/max altogether. We know a min/max location for standing wave will be the point of reflection for an open, which with a 60m wavelength is far to close to the critical section on such a small connector to believe it to be avoiding min/max of standing waves. Testing with HIPOT DC could shed some light on this particular point.

 

RE:I would like to get a better handle on maximum voltage.

--> Had to work many times with russian laser power supply what used silicon-rubber based flexible 12 mm coaxial to form the 10 nsec 50 kV voltage spikes. Consequently, from 5 such lasers in one year we got 4 defective PS, whose all had identical defect - carbonized parts of the cable... So I obtained American high voltage cable instead but even that served not much longer. Cable is wound around the permaloy ring core thus making 1:4 wave transformer.

About Your problem: just fill the plug after soldered with proper filler - light colored sorts of epoxydes, or silicones, or just fill in the high transparent sort of grease, teflon powder, polyethylene in hot melted state etc etc. I cant obey You shall get the 50 kV, but at least 3x folding is highly probable.