High voltage oscilloscope probes, make or buy?

[Mark]

Soon I will need to look at wave shapes and make some measurements which involve 1.2/50us impulses at 10kV.   Can anyone recommend an oscilloscope probe for this task?  Is it easy to roll my own?  I'm not sure if I have the time or patience to tweak a home-made probe, but if it's a case of following a schematic and having it work first time, then I'm willing to try. 

Price limit for a bought probe is about 500UK Pounds. 

[swingbyte]

Hi what you need is a tek p6015 HV probe.  If you are lucky you can get one for reasonable cost second hand. Look at the Tesla coil websites for how to make your own probes but if you want high bandwidth the tek is the only option.

[MarkL]

I would not try to make your own.  Why risk yourself and your equipment.

I agree with swingbyte that the Tek P6015A would be the best choice, and possibly the P6013A which is an older model.  The P6015A is still being sold by Tektronix.

Manuals and data for the two are here:

  http://www.ko4bb.com/manuals/index.php?dir=Tektronix/Tektronix_-_Probes

In particular, look at the charts that derate the maximum voltage based on the frequency and duration of measurement.

There's a couple of both types on ebay right now within your price range.

[electr_peter]

10kV is a bit too much to do DIY probe if you have no experience, risk is far too high.

There are two reasonable options - differential HV probe and HV probe with divider. Pintek has a range of both types (http://www.pintek.com.tw/). Other manufacturers also have similar offerings.
Differential probes are more expensive by default, passive HV probes should be cheaper.

[Mechatrommer]

quite agree, unlike normal mains, KV range can jump off to your nearest skin if not careful. and characterising the homebrew probe is just as difficult unless you have high quality KV signal generator.

[tautech]

50 MHz 1000:1 15kV passive probe

http://www.siglentamerica.com/prodcut-fjxx.aspx?fjid=407&id=25&tid=1&T=2

US$505 + taxes if any

[DanielS]

A basic HV probe is little more than a voltage divider network and some compensation capacitors. You can build one fairly easily and cheaply with just a little bit of care:


It might not look pretty but it will get the job done.

[Rupunzell]

Frequency response is going to be pretty lumpy due to the lump constant resistor/capacitor divider in a metal case. Beyond this, this design does not account for probe cable reflections or impedance mis-matches.

Easier, safer and much more accurate results by simply purchasing a Tektronix P6013, P6015 or P6015A


Bernice

A basic HV probe is little more than a voltage divider network and some compensation capacitors. You can build one fairly easily and cheaply with just a little bit of care:


It might not look pretty but it will get the job done.

[DanielS]

Beyond this, this design does not account for probe cable reflections or impedance mis-matches.

When was the last time probe cable reflections and mis-matches have ever been an issue? Passive probes use lossy coax to prevent any significant signal reflections from appearing despite the gross mismatch between the unknown source impedance, the probe, the probe's cable and the scope's input.

[Mark]

OK so I decided I'd take a look on ebay and managed to get a Tek P6015.  There is a tube of fluorocarbon shown in the photo, but it will still go to 13kV without it, which is enough for me.  It hasn't arrived yet. 
Thanks for your advice and opinions everyone, I'll let you know how it goes with the P6015. 

[Rupunzell]

Lossy coax alone does not cure the probe problem as it still has reflections/ringing and requires damping in the probe termination and peaking to compensate for the loss in HF response.

This is the probe termination box of a Tektronix P6105 which is spec'd at 100Mhz. It is a pretty basic Kobbe-Polits scope probe using a lead-lag network. Not the best scope probe as it does produce reflections at the probe tip and has less waveform fidelity compared to Tek and other higher quality probes.

Tek P6105, Term box side:



Tek P6105, Term box bottom:



This is Tektronix P6106 probe (200Mhz)  termination box, it has two inductors and a LOT more Trimmer and Cap adjustments. Why is this?


Tektronix P6106, termination box's other side:



Scope probes of this design work well, have low internal reflections with good pulse and frequency response.
Answer to this can be found starting at page 15 in this Tektronix probe book from 1969.
 http://www.davmar.org/TE/TekConcepts/TekProbeCircuits.pdf

A bit of probe development history:
http://www.vintagetek.org/oscilliscope-patents/

Low quality, poorly designed and produced probes are a MAJOR source of measurement error. This coupled with inadequate probe application and insensitivity to how the circuit behaves when probed ADDs to the measurement errors. Passive probes are the most error prone due to their loading at HF, probe reflections and grounding problems. Active probes can make a BIG difference when properly applied, Zo and numerous other probe variations can make the different between low circuit reaction to being probed with good waveform-waveshape fidelity and extreme measurement error. Analog folks can be extremely demanding on probe performance. Every lump, bump, ring and all manner of wave shape aberration can have meaning directly related to circuit behavior. 


This is the probe termination box from the P6015 HV probe, it is not a simple network and is LOTs of fun to set up. These are often found with the termination box completely out of whack which results in lumpy, bumpy pulse & frequency response. Getting one of these properly adjusted is a good network learning experience.

 

Bernice

Beyond this, this design does not account for probe cable reflections or impedance mis-matches.

When was the last time probe cable reflections and mis-matches have ever been an issue? Passive probes use lossy coax to prevent any significant signal reflections from appearing despite the gross mismatch between the unknown source impedance, the probe, the probe's cable and the scope's input.

[Mark]

Excellent info Bernice!  Thank you, I will bear that in mind if I attempt to adjust anything, I'm certainly glad I decided to buy and not make/bodge now. 

[HighVoltage]

If you bought one of the older P6015 probes that require to be filled with fluorocarbon, you can fill them instead with a high voltage stable silicone gel to have them permanently filled. This is the same as in the P6015A. Just make sure the probe works perfectly, before you fill it with silicone gel.

[Mechatrommer]

Beyond this, this design does not account for probe cable reflections or impedance mis-matches.

When was the last time probe cable reflections and mis-matches have ever been an issue?

when you try to diy it yourself. all this while we took this for granted, with off the shelf probe.

Passive probes use lossy coax to prevent any significant signal reflections from appearing..

i will be very glad if you can give me ebay link where to get lossy 50ohm coax. afaik its none, so reflection problem solving is you need to put a resistor at the very tip of your probe side (with some degree of "success"), not at the active circuit or dso side. otoh, yes building a voltage divider and its compensation is easy, the hard part is specifying it with real world signal measurement, in this case, you need HV generator with acceptable output impedance. if its easy to be get, then its easy to be done.

[korlatos]

If you bought one of the older P6015 probes that require to be filled with fluorocarbon, you can fill them instead with a high voltage stable silicone gel to have them permanently filled. This is the same as in the P6015A. Just make sure the probe works perfectly, before you fill it with silicone gel.

Does anybody have any recommendations for the type (and source of it in the USA) of a silicone gel suitable for filling P6015?

[TimFox]

Can you build the HV divider network directly into the circuit being tested, using appropriately rated components?
You can verify the operation at lower voltage, driving the divider from a good square wave or fast pulse, before hardwiring the top to the device under test.  You can also verify safe operation of the divider circuit at HV before connecting it to an expensive oscilloscope.

[SeanB]

Lucky for me I still can get R141B, which will probably work in these probes. Still have 5l in a disposacan under my bench. Getting expensive though, but as a cleaner it works wonders, removes pretty much anything organic.

[DanielS]

i will be very glad if you can give me ebay link where to get lossy 50ohm coax.

The cables off of defective probes are always an option. Another possibility is to design the divider network as an accessory for existing probes. Not ideal for precision measurements on fast edges or high frequencies but still good enough in many cases.

[HighVoltage]

If you bought one of the older P6015 probes that require to be filled with fluorocarbon, you can fill them instead with a high voltage stable silicone gel to have them permanently filled. This is the same as in the P6015A. Just make sure the probe works perfectly, before you fill it with silicone gel.

Does anybody have any recommendations for the type (and source of it in the USA) of a silicone gel suitable for filling P6015?

Contact Ellsworth for Dow Corning type 3-4680
http://www.ellsworth.com/products/by-market/consumer-products/encapsulants/silicone/dow-corning-3-4680-silicone-gel-transparent-blue-210-ml-kit/
This material has a dielectric strength of 16 kV/mm

I have several of these P6015A probes and one P6015 that I filled with a similar German made silicone gel material
The P6015 with the silicone gel was tested by me to 35 kV pulse signals and I detected no problems
Just make sure you fill it free of any air bubbles.
Hold the probe in an angle and let the silicone gel run down the  housing, essentially filling it slowly from the bottom.
 

[korlatos]

Thank you, HighVoltage - that's great information. I saw some on-line references to people using mineral oil for filling P6015. Do you think that mineral oil would work as well as silicone gel? Were you able to successfully compensate your P6015 after filling it with silicone gel?

[tautech]

I saw some on-line references to people using mineral oil for filling P6015. Do you think that mineral oil would work as well as silicone gel?

It shouldn't be ANY mineral oil but a product with a high dielectric strength like Transformer oil.
Datasheets will guide you.

I'd go with HV's suggestion.

[HighVoltage]

Thank you, HighVoltage - that's great information. I saw some on-line references to people using mineral oil for filling P6015. Do you think that mineral oil would work as well as silicone gel? Were you able to successfully compensate your P6015 after filling it with silicone gel?

As tautech said, you can use a high quality transformer oil.
But it is nasty stuff and you don't want to handle it really.

Yes, I was able to compensate the P6015 correctly with the silicone gel as a insulator.

[PaulAm]

Excellent information, thanks for posting that.

I have a P6015 that I picked up the other year and I was wondering where I was going to find some r114.  That stuff is very difficult and very expensive to find and it has a tendency to leak out of any container, even if you keep it in a freezer to lower the vapor pressure.

[korlatos]

Thank you, HighVoltage - that's great information. I saw some on-line references to people using mineral oil for filling P6015. Do you think that mineral oil would work as well as silicone gel? Were you able to successfully compensate your P6015 after filling it with silicone gel?

As tautech said, you can use a high quality transformer oil.
But it is nasty stuff and you don't want to handle it really.

Yes, I was able to compensate the P6015 correctly with the silicone gel as a insulator.

That's great to know. Thanks again!

[Rupunzell]

One way to helping to get the air bubbles out for a filled space with silicone or similar compounds is to place the container, in this case the probe body into a larger sealed container and pump the air out creating a vacuum inside the sealed container with the probe body. The difference in pressure will help promote the air bubbles to move out of the filler compound leaving the probe body with much less air trapped inside the filler.

As for home brew probes -vs- high quality commercial probes, it comes down to waveform measurement accuracy and need. These Tektronix HV probes have been used in particle accelerator targets for high energy physics experiments. Since the probe comes apart and has a BNC connector on the main probe body housing, it can be easy adapted to being used for a host of non-standard applications. Measurement demand placed on probes in these applications are not the same as probes being used for hobbyist experiments and related. Demands placed on the resulting data are also very different between hobbyist and physics researchers or individuals doing serious development and research.

What is not often appreciated today venturing into the world of electronics is just how good and how far some of the test and measurement devices pushed the limits of physics and their understanding of nature. The Tektronix P6015 is just one example of Tektronix probes designed and made in the 1960's that are difficult to equal or better to this day. Other Tektronix probe examples are P6046, P6042 and others. Many of these Tektronix products are still in service to this day decades after they were produced.


Bernice

Just make sure you fill it free of any air bubbles.
Hold the probe in an angle and let the silicone gel run down the  housing, essentially filling it slowly from the bottom.