How can we measure low-current (microamps) at high voltage (3kV) ?? 2

[Khelz]

Hey Guys,

I am currently working on a biochemical application in which I send 3 kV into a syringe needle which is separated by an air gap from a grounded counter-electrode. Here is a simple schematic of what our setup looks like.

Link

Now, what we want to do is to monitor the current behavior on the HV side. In order to do so adequately, the measuring method needs to cover a bandwidth of 0-50 kHz.

So basically, the request is:

-measure microamps
-support 3.5 kV
-cover 50 kHz of bandwidth (fast response time)

So far, the only technology that I find was maybe suitable is photomupliers (LED which communicates with a photodiode when a current passes in the circuit --> the current is then measured on a low voltage side since the photodiode is protected from the HV) but I am really not familiar with it. According to my research, it seems that this particular challenge (fast response low current measurement on a HV line) is not much covered in the literature so I was hoping that you guys could help me figure out how to achieve this.

Thanks
Khelz

 

[VEBill]

EDA Board links don;t work for most of us: "You are not logged in or you do not have permission to access this page."

 

[VEBill]

Can you measure the current (or the voltage on a series resistor) on the return leg where the voltage should be much closer to ground?

You'd probably want to have an arrangement of resistors so that, in the event of an arc or short circuit, the peak voltage on the oscilloscope or sampler is low enough to avoid damage to the inputs.

There are also some safety issue to deal with, so another series resistor at the source (to enforce the micro amp limit) might be a good safety feature.

 

[Khelz]

Thanks for your reply VE1BLL. In our case, we are interested in the current on the high voltage leg. You see we are already measuring the current on the counter-electrode side, and what happens is the highly charged droplets that are emanated from the syringe needle do not all reach the counter-electrode. So we want to compare both legs together, plus we are also interested in the temporal side of the events as well.

As for the safety issues, this is definitely a good point. As in the source side, the high voltage supply has a really low power consumption so current is already internally limited to 3 mA.

Do you know any technology beside photomultiplier that would be able to measure on the high voltage side?

 

[IRstuff]

High voltage is only a problem if you're referenced against ground. If you're referenced against the 3.5 kV, then it's not a big deal. That's why PNPs and P-channel FETS exist.

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529

 

[Khelz]

Hi IRstuff,

I understand the concept but I do not really see how I could use it concretely. In my case, we want to digitally record the temporal behavior of the current with an acquisition device, so I am expecting to see some kind of current-to-voltage converter which would then send the information to a scope for example. How could I do that without damaging the equipment by sending 3 kV in it?

 

[LiteYear]

Should be feasible with a couple of resistive dividers and an optocoupler. At least with those parameters you can easily generate a good reference signal - say a 1kOhm precision resistor in series to give you a decent voltage to measure, if your circuit will support it (I can't see your schematic but I assume there's already a limiting resistor of about this size in the circuit). Then use an optocoupler like the TIL300 to give you isolation from the 3kV. The output is then a voltage with range of your choice, referenced as you please.

In theory this relatively straight forward. To get it just right would no doubt require some serious head-scratching. I wouldn't recommend this path unless you have someone with experience laying out measurement circuits nearby.

If you want something you can do without electronics knowledge, consider the same series precision resistor idea but use differential voltage probes for an oscilloscope to measure the voltage. Might not be as neat or flexible, but doesn't require any circuit board design.

 

[Khelz]

Hey LiteYear,

This idea with the optocoupler seems a good idea. In our setup, precision is definitely of high importance so I guess going with the voltage probes would be option B. Quick question though, when do you say it would need some serious head-scratching, how would that be? Tuning or something?

 

[Khelz]

Also, if you could say why using the voltage probes would be less neat and flexible that would help me justifying my choice. My guess is that the voltage divider in the probes slows down the response but I am not sure.

 

[ijl]

I once followed a (remotely) similar project. Their solution was a "floating" measurement unit, that was connected only to the HV. The current was measured and digitized in the unit. The result was then sent to the users using a RF link.

 

[OperaHouse]

Opto isolators are not linear so that is where the tweaking and hair pulling comes in. Maybe this doesn't need to be real time anyway. I would think about floating one of those little scope boxes that hook up to a computer and making a high speed isolated USB.

 

[itsmoked]

Not to mention opto-isolators are slow and variable with temperature. You will see reality far more with a resistor shunt and and a differential probe.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[OperaHouse]

My understanding is the droplet will get a slow (relatively) buildup charge and then the droplet will get a sudden discharge, then some particles will just get lost. It will need a computer somewhere to figure this all out because there will be a lot of math. This is not a situation where you will be reading meters. So might as well start with digital numbers. Analysis will likely be done on something like a one second sample. After the process is understood average values may be sufficient.

 

[Khelz]

Hi all,

Operahouse, you are right about the slow buildup and the sudden discharge, however I am not sure to understand what do you mean when you talk about a lot of math. The circuit we use right now implies an oscilloscope which samples the converted current values and store them into memory, which is then recorded and transferred to a computer for signal processing.

The goal is to do the same thing but on the HV section of the circuit.

A for itsmoked, I am not sure to follow you when you say optocouplers are slow. I don't know if you can have access to this but this is the first solution that was proposed to me. It implies optocouplers and photodiodes and the author state that the device is capable of reaching a bandwidth of 200 kHz.

Also, I am wondering why people would break their back to design solutions like this (there is even a new scientific paper from Aplin et al. [2008] that propose a new solution to measure nanoamps on a high voltage line) if you could have just placed a resistor shunt and a differential probe.

 

[ijl]

I am wondering why people would break their back to design solutions like this

These "people" are professional electrical engineers. Their main design principle is "safety first". That is why.

 

[OperaHouse]

I think two times 50,000 samples for one second is a lot of math. Simple math, but a lot. What your scope and whatever are doing right now. I think of it as a system to design. You look at it as where can I buy this piece. You would likt to hang a scope probe on a 3,500V line and measure a little signal. Gut feeling tells me that won't work out for the detail you want. I see puttig an A/D converter up there. And that will define the rest of your system.

 

[ScottyUK]

Geting power onto a 3500V DC line isn't a particularly big challenge given a bit of space and a bit of money. How about floating the whole measurement system including the scope at line voltage and using an optical fibre to transmit back to earth-referenced equipment over ethernet? You'd need to provide an insulated enclosure for the instrument, but again 3500V isn't all that high. I'm not sure of the breakdown voltage of the average fibre patch lead but I expect it is 'very high': high enough not to worry about anyway.

You could possibly use one of the battery-powered scopes if you don't need a long continuous measurement period, and avoid having to isolate the power supply, although you'd need a battery-powered fibre-copper converter too.

 

[Khelz]

Operahouse there definitely has to be an A/D converter in the system since we want to process the results afterward. However, the hard part is really to get the measurement on a lower voltage line. Let me also reassure you by definitely not looking at it from a where can I buy this thing perspective. I actually know for a fact that there is no commercial device that can do this thing all-in-one. Ijl, my point was that there is probably a downside to straight using a shunt resistor and voltage probe whether it's response time, accuracy, price, or any other.

Scotty Uk this seems like a plausible solution. Can you briefly explain how does the optical fiber transfer the signal from the high voltage line to the grounded voltage line without carrying the voltage? I am not too familiar with ethernet and optical fibers. All that I understand is that the optical fiber communicates with optical energy instead of electrical which probably explain the transfer from a high to low voltage line and ethernet is a wireless communication system (I think). However I don't really know how the conversion would be made from the oscilloscope to the optical fiber.

 

[ScottyUK]

Optical fibre is either glass or polymer (plastic), both of which are very good insulators. Data is transmitted using light. A meter of optical fibre will likely stand a few hundred kV before breakdown if it is externally clean. You could have something at one end of a fibre which is at a completely different potential to the other end, yet the two devices would never know because the only connection is via light and an insulator.

As for the first question, it depends on the scope. Most mid-range scopes and all high-end scopes have a comms port, whether it be serial, IEEE-488 / GPIP, or whatever. Recent ones often have an ethernet port, or USB, or both. Older ones tend to have GPIB and/or serial interfaces. You can usually control the scope and poll data from the instrument over the link. If you have a serial connection you can get a fibre link which carries RS-232 data from (e.g.) Hirschmann. If your scope has Ethernet capability then a fibre link is almost trivial.

What instrument are you using?

 

[Khelz]

Oh I understand a little bit more now. I am currently using a DPO3014 from tektronix. I just checked and it definitely has an ethernet cable slot plus a usb slot as well.