Help with measuring small signals on high voltage rails

[RoundsToThree]

Hi there,
This is my first post on the forum so I hope I'm not doing anything wrong and sorry in advance for being a bit verbose below, just didn't want to skip over anything that may be important.
I have been researching different ways to figure out how to properly interface with a old photomultiplier tube I found online to be able to attempt Gamma Spectrometery. I found that they need roughly around 1 to 1.6kV to operate but at low currents and it appears that when a photon originating from a gamma ray that excited a scintillator at the front of the tube enters the tube, it is amplified through 9 stages of dynodes and hits what may be a photodiode allowing a small current to flow through. I believe this current is then put through a resistor of high resistance and then to the ground rail where a voltage drop is measured across it, however I have found some conflicting schematics.
The tube: http://lampes-et-tubes.info/pm/FEU-35a.pdf
A design I found with a similar tube: https://www.instructables.com/Multi-Channel-Analyzer-for-Gamma-Spectroscopy-With/
I was planning on generating these high voltages using either a boost converter I found on ebay that is adjustable to 1.2kV or as the instructables suggested, a CCFL tester.

From the instructables article, It seems as though the signal can last 100us to ms and is of amplitude 50mV but may be floating on a high voltage rail. I think the article actually uses a negative high voltage so that the signal is not floating on hv but ground referenced with negative amplitude. The part that's been stumping me is when an issue comes up in the build/design stage of the circuit and I need to diagnose it, I don't really want to cook my scope (SDS1104X-E). After digging around and speaking with Dave a bit, I think I have a much better understanding of using a differential scope or a high voltage probe to address this issue but there are still some concerns bugging me a little. In terms of using a HV probe, I found that the Hantek T3100s were pretty affordable and had some relatively ok reviews on this forum, I figured two of them may work for a differential measurement but the signal would probably be drowned out by using very little of the hv dynamic range. So looking at the differential probes, I came across the HVP70 here, some no name brand 20v ones on ebay and Micsigs catalog of their hv differential probes which looked promising. The Micsigs have probes with 500x and 2000x attenuation for measuring up to 1.3kV and 5.6kV respectively which seem great but I figure that I'm not going to even see the 50mV signal which may be well into the uV by that point if present at all. My fear of just using a standard low voltage differential across the circuit is that some of these probes have a common mode rating of 700v or 1kV which may be exceeded even if only the differential part is small, am I correct in believing this?
I did plan on running my circuit during testing off some 12v SLA batteries to easily get a negative voltage if needed and I hear that this may actually render my concerns above null as the ground referenced can be tied to any point of the battery circuit if its floating but I still am concerned that even with a HV probe in this case, I may cook something.
If a differential probe is the way to go, is there a way that I can find one that i can use for both relatively low voltages (10s or 100sv) and this application at the same time to avoid purchasing one in the future?
Since differential probes will probably set me back a few hundred, is it practical to just purchase a cheap handheld scope meter instead which isn't earth referenced?
Also on a similar note, I was thinking, maybe this wont work at all, but could i use a high turn ratio transformer in series with the anode resistor to generate a higher voltage than 50mV thats isolated every time a pulse comes through?

Also for reference, this was another article I read through while researching that does discuss the same tube but with a different circuit:
https://www.fusor.eu/gammaspec.html
 
Tldr; what's the best way to measure a low voltage signal on a hv rail and will a hv differential probe work?

Thanks for your patience with me, any help is greatly appreciated!

[antenna]

I am way over my head on this one, but why can't you simply decouple with a HV capacitor to block the DC component followed by a pair of back to back diodes to ensure the output voltage doesn't spike in excess of 700mV (to protect the measurement device) when switching the HV power on and off? 

[Marco]

A PMT can have a simple resistor to ground at the anode for readout.

Even if you want to have a postive high voltage at the anode, capacitive coupling will work fine.

[Terry Bites]

Please don't link to unsecured pdfs

The anode is usually tied to ground though it's load resistor. The cathode is tied to a negative EHT source.
This means that the output pulse is ground referenced. That's the norm.
Get info from tube manufacturer's sites not from hobby ones.

PMT circuits can kill you so its really important to understand your HV circuit and the risks present.
The supply needs to be regulated or the PMT gain will drift. You want low noise from your psu as well.

[ledtester]

The way I see it if the HV power supply is powered by a floating supply (such as batteries) then you can make the ground of the system the same as earth ground and the output will also be referenced to earth ground. If the output is low voltage then your scope should have no problem with it.

The author connected the output to a scope as well as a PC via a USB sound dongle. You could email him and ask what his setup was - his email address can be found here:

https://stoppi-homemade-physics.de/impressum/

[coppercone2]

you need a power supply probe for your oscilloscope
https://www.tek.com/en/products/oscilloscopes/oscilloscope-probes/power-rail-probes

but some version thats higher voltage.

maybe hook it up to a HV probe ?

[RoundsToThree]

Thanks for your replies! Also i'll make sure to not post links to unsecure pdfs in the future.
I took onboard the advice and simulated the tube with a hv capacitor to decouple the signal and some diodes for hv protection (They would drown out the signal if i had them on the hv decoupled side so i'm not sure if they still are working here). This looks perfect as I'm getting a 300mV signal on the hv line although my probe line is still at HV so I think I'm missing something.
I did notice though that if i switch the supply to -1.6kV, I can move the ground to the middle so all readings are ground referenced but I'm not yet 100% sure if I want to start with a negative hv.


I tried a few more decoupling circuits which in hindsight make more sense as they actually decouple from the hv but the voltage was far too low to measure effectively (uV), especially since the end goal is to integrate the signal. So using some op-amps I managed to get a higher potential signal (300mV) but im worried my amplification ratio is a bit high, am I better off with more amplification stages or will an op amp not effectively work in this environment. I'll admit that I've never used an op-amp in practice before, only in theoretical scenarios so forgive me if this doesn't make any sense what I have drawn up.



If I were to test this circuit, would I be better off with a HV probe, differential probe (hv or not), or do I not need any new probes?

With the safety of this test, I do plan on isolating the dut and standing back with probes already connected before turning it on just to make sure. I have a few other ideas too to improve safety but I think its more likely I pop something than myself during the tests so I do want to make sure I have the ability to probe safely.
Thanks again for your help!