Cable

[bartie]

I am looking for a cable to transport a 3kV, 500kHz signal over a few meters. When using standard cables (or high voltage rated cables), the signal tends to decrease to 1 or 2 kV after 1 or 2 meters.

Any suggestions of cables I could use?

Thanks!
Bart

 

[VEBill]

I assume that your apparent voltage drop is not cause by I*R ;-) !! Your numbers are very suspicious.

Look in the general direction of large diameter RF coaxial cable. Some of the larger ones are capable of handling 3kV and, as a bonus, will carry 0.5 MHz with essentially no loss.

 

[jimkirk]

What's your source output impedance? Could be capacitive loading.

 

[davidbeach]

VE1BLL, I would not be surprised to find that typical power cable has a very high inductive reluctance at 500kHz.

 

[itsmoked]

Wonder about RFI...

 

[VEBill]

1 or 2 m is pretty darn short compared to 500kHz (unless the cables are ferrite-loaded...). Therefore I suspect that there might be something wrong with his measurements.

Perhaps we need more details...

 

[davidbeach]

VE1BLL, you may be right A quick calculation shows #6, 5kV cable as having an inductive reactance of around 1 ohm/foot, not accounting for skin effect. For 6 feet of that (close enough to 2 meters), it would take a current of 166.7A to drop 1kV. Maybe the source just can't carry the load imposed.

 

[VEBill]

166.7A times 3kv = 0.5MW, into a 2m long 'antenna'.

Smelly. Very very smelly.

 

[bartie]

Thanks for all this feedback...

I think that essentially the issue is the frequency. When I transport a similar signal at a lower frequency, everything works fine.

Jimkirk, source output is of the order of 1nF.

Davidbeach, how did you calculate the 166A? Was this at 50/60Hz?

VE1BLL, measurements were done using an oscilloscope with a 1000/1 differential probe (input impedance 20pF).

 

[jimkirk]

Any series resistance with that 1nF? A few Kohms driving several tens of pF lumped capacitance will have considerable drop at half a MHz.

 

[davidbeach]

1kV/6ohms=166.7A.

 

[rbulsara]

How much current is in the signal?
Find out the inductance of the cable at a given frequency.

The inductance of the cable at 500khZ will be 8333 times that at 60hz. 1 ohm inductive reactance at 60hz becomes 8333 ohms at 500khz. (Z=2*PI*f*L)

so a current of 0.00012A (0.12mA) will cause drop of 1kV (This is IF 1 ohm/ft is a correct assumption)

 

[davidbeach]

The 1ohm/ft was at 500kHz, I guess I didn't make that clear.

 

[rbulsara]

OK Davidbeach:

 

[VEBill]

Maybe the mysterious voltage drop is cause by arc-over within the 1 or 2m cable. Is the cable rated for 3kV? Does it account for the peak voltage of the (500kHz) AC?

Is the 500kHz still a nice clean sine wave at the far end?

Does the probe + scope indicate 3kV at the input to the cable?

Is the cable open circuit at the far end?

And most importantly of all, what the heck are you doing with 3kV of 500kHz ????

(PS: Is 500kHz still a distress frequency? It used to be.)

 

[bartie]

Cable I have been using so far was hi voltage rated. However, I think that cable flashover voltages are specced at 50/60 Hz or DC, so I have no idea what happens at higher frequencies. An RCL measurement showed no spectacular results, but this was done at 5V...

Something more about the measurement: 3kV is measured at the output of the source, when no cable is applied. 1 kV is measured at the far end of the cable (so with the cable connected to the source). I have a feeling it is the C that is causing this trouble and not the L (no current is flowing), but of course any fresh ideas on this are welcome.

BTW, the signal is used in the igniter of a discharge lamp.

Bart

 

[VEBill]

One more thing to measure would be the source voltage (at the input to the cable) when the cable is connected. This will inform us if the apparent voltage drop is in the cable or in the source.

($2 on 'in the source' please...)

 

[Higgler]

Maybe try other lengths of cable if you have any.
It might be a simple test to show your cable is ugly.

Also, check your cable with an ohm meter, maybe you'll be surprised.

kch

 

[IRstuff]

What's the drive capability of your driver?

Your signal has a period of 2 us. Assuming that your desired rise/fall time is 0.2 us, the dv/dt = 15 kV/us, which is quite sporty, even for a high-performance op-amp. Since you're getting as much as 2/3 rd reduction, your driver is undersized by at least a factor of 3. Conversely, you cannot run faster than about 100 kHz with this driver and this load.

With a 1 nF load, the displacement current is 15A, further increasing the sportiness.

TTFN

 

[shagooo]

Try using RG style coaxial cable, RG-220. We use it at 100kHz and run up to 15kV on it. I don't think the losses change much at 500kHz.
You cannot use standard wire to transmit the 500kHz (skin effect)