Forums

Magnetic field sensitivity of PMT

Started by Joerg December 21, 2015
Folks,

Does anyone have experience with operating small switchmode transformers 
close to a PMT?

What I am considering in order to avoid a split into multiple boards is 
to run a Royer with a CCFL transformer on a board that plugs into the 
PMT. It would also contain the HV ladder, some amps, linear regulators 
and other stuff. In the trade this whole enchilada referred to as a 
"socket". I'll try my best to keep harmonics low but running it somewhat 
sinusoidal. But in terms of shielding there won't be much, towards the 
tube a full copper plane on the board and on the component side away 
from the tube a Laird shield can. I could also plop a shallow Laird 
shield on the bottom side if absolutely needed. The transformer would be 
just an inch or so from the tube contacts on the opposite side of the 
board. The switching transistors and filter inductors would also live 
under that shield.

Commercial "sockets" usually have some sort of switcher in there. I 
don't have one so can't look inside but probably they spent a lot of 
time optimizing the shielding and positioning of the transformer or 
inductor.

Considering that the oscillator runs in the 50-100kHz range the shield 
effectiveness could be low, around 20-30dB. PMTs are sensitive to 
magnetic fields but not sure how critical this is.

-- 
Regards, Joerg

http://www.analogconsultants.com/
On Monday, December 21, 2015 at 12:23:33 PM UTC-5, Joerg wrote:
> Folks, > > Does anyone have experience with operating small switchmode transformers > close to a PMT? > > What I am considering in order to avoid a split into multiple boards is > to run a Royer with a CCFL transformer on a board that plugs into the > PMT. It would also contain the HV ladder, some amps, linear regulators > and other stuff. In the trade this whole enchilada referred to as a > "socket". I'll try my best to keep harmonics low but running it somewhat > sinusoidal. But in terms of shielding there won't be much, towards the > tube a full copper plane on the board and on the component side away > from the tube a Laird shield can. I could also plop a shallow Laird > shield on the bottom side if absolutely needed. The transformer would be > just an inch or so from the tube contacts on the opposite side of the > board. The switching transistors and filter inductors would also live > under that shield. > > Commercial "sockets" usually have some sort of switcher in there. I > don't have one so can't look inside but probably they spent a lot of > time optimizing the shielding and positioning of the transformer or > inductor. > > Considering that the oscillator runs in the 50-100kHz range the shield > effectiveness could be low, around 20-30dB. PMTs are sensitive to > magnetic fields but not sure how critical this is. > > -- > Regards, Joerg > > http://www.analogconsultants.com/
Hi Joerg, I know only a little. But I think the mu-metal shields on PMT's are there so that if you change the orientation of the pmt in the earth's field then the gain won't change. (I've never tried to measure the effect, but I've used PMT's with mu-metal shields.) So at a guess if you keep the B-field well below that of the Earth's maybe you'll be fine. Toroidal inductors? I think you give commercial socket makers too much credit. We have some hamamatsu sockets that put varying amounts of (I think) 40 kHz crud from the socket C-W multiplier onto the anode. We use it in photon counting mode and the crud is just an annoyance, not a real problem. What mode will you be using? George H.
On a sunny day (Mon, 21 Dec 2015 09:23:46 -0800) it happened Joerg
<news@analogconsultants.com> wrote in <ddqqseFqk29U1@mid.individual.net>:

>Folks, > >Does anyone have experience with operating small switchmode transformers >close to a PMT? > >What I am considering in order to avoid a split into multiple boards is >to run a Royer with a CCFL transformer on a board that plugs into the >PMT. It would also contain the HV ladder, some amps, linear regulators >and other stuff. In the trade this whole enchilada referred to as a >"socket". I'll try my best to keep harmonics low but running it somewhat >sinusoidal. But in terms of shielding there won't be much, towards the >tube a full copper plane on the board and on the component side away >from the tube a Laird shield can. I could also plop a shallow Laird >shield on the bottom side if absolutely needed. The transformer would be >just an inch or so from the tube contacts on the opposite side of the >board. The switching transistors and filter inductors would also live >under that shield. > >Commercial "sockets" usually have some sort of switcher in there. I >don't have one so can't look inside but probably they spent a lot of >time optimizing the shielding and positioning of the transformer or >inductor. > >Considering that the oscillator runs in the 50-100kHz range the shield >effectiveness could be low, around 20-30dB. PMTs are sensitive to >magnetic fields but not sure how critical this is.
Depends on the PMT, how much gain, but I think an inch is very little. Depends on the transformer too. My setup: http://panteltje.com/pub/PMT_HV_generator_size_img_3174.jpg PMT socket on the right Can do up to 2 kV, but now runs a simple Russian FEU35 at much lower voltage, say 500 to 600 V. http://panteltje.com/pub/Russian_PMT_FEU35_front_img_3189.jpg At that point I have no problems.. But I have a strong signal from a scintillator crystal. AND it is not a switchmode, but a SINE wave generator. smft
On 2015-12-21 09:55, George Herold wrote:
> On Monday, December 21, 2015 at 12:23:33 PM UTC-5, Joerg wrote: >> Folks, >> >> Does anyone have experience with operating small switchmode transformers >> close to a PMT? >> >> What I am considering in order to avoid a split into multiple boards is >> to run a Royer with a CCFL transformer on a board that plugs into the >> PMT. It would also contain the HV ladder, some amps, linear regulators >> and other stuff. In the trade this whole enchilada referred to as a >> "socket". I'll try my best to keep harmonics low but running it somewhat >> sinusoidal. But in terms of shielding there won't be much, towards the >> tube a full copper plane on the board and on the component side away >> from the tube a Laird shield can. I could also plop a shallow Laird >> shield on the bottom side if absolutely needed. The transformer would be >> just an inch or so from the tube contacts on the opposite side of the >> board. The switching transistors and filter inductors would also live >> under that shield. >> >> Commercial "sockets" usually have some sort of switcher in there. I >> don't have one so can't look inside but probably they spent a lot of >> time optimizing the shielding and positioning of the transformer or >> inductor. >> >> Considering that the oscillator runs in the 50-100kHz range the shield >> effectiveness could be low, around 20-30dB. PMTs are sensitive to >> magnetic fields but not sure how critical this is. >> >> -- >> Regards, Joerg >> >> http://www.analogconsultants.com/ > > Hi Joerg, I know only a little. But I think the mu-metal shields on PMT's > are there so that if you change the orientation of the pmt in the earth's field > then the gain won't change. (I've never tried to measure the effect, but I've > used PMT's with mu-metal shields.) So at a guess if you keep the B-field > well below that of the Earth's maybe you'll be fine.
Probably even less because an AC field creates a signal in the tube.
> Toroidal inductors? >
Unfortunately not in the cards, this is all on a very small budget. I have to live with what's available and right now only CCFL transformers are left. For a while. Soon they'll also go lalaland because monitors are migrating to LED backlighting.
> I think you give commercial socket makers too much credit. > We have some hamamatsu sockets that put varying amounts of > (I think) 40 kHz crud from the socket C-W multiplier onto the > anode. We use it in photon counting mode and the crud is just an > annoyance, not a real problem.
That's what I also assumed, I always wondered how that could ever work. Well, in this case it didn't which is why they want a new kind of circuit that performs better.
> ... What mode will you be using? >
I don't have much information about the use, just that it all has to be very stable, clean and low noise. Unfortunately that also means large. -- Regards, Joerg http://www.analogconsultants.com/
On 2015-12-21 10:00, Jan Panteltje wrote:
> On a sunny day (Mon, 21 Dec 2015 09:23:46 -0800) it happened Joerg > <news@analogconsultants.com> wrote in <ddqqseFqk29U1@mid.individual.net>: > >> Folks, >> >> Does anyone have experience with operating small switchmode transformers >> close to a PMT? >> >> What I am considering in order to avoid a split into multiple boards is >> to run a Royer with a CCFL transformer on a board that plugs into the >> PMT. It would also contain the HV ladder, some amps, linear regulators >> and other stuff. In the trade this whole enchilada referred to as a >> "socket". I'll try my best to keep harmonics low but running it somewhat >> sinusoidal. But in terms of shielding there won't be much, towards the >> tube a full copper plane on the board and on the component side away >>from the tube a Laird shield can. I could also plop a shallow Laird >> shield on the bottom side if absolutely needed. The transformer would be >> just an inch or so from the tube contacts on the opposite side of the >> board. The switching transistors and filter inductors would also live >> under that shield. >> >> Commercial "sockets" usually have some sort of switcher in there. I >> don't have one so can't look inside but probably they spent a lot of >> time optimizing the shielding and positioning of the transformer or >> inductor. >> >> Considering that the oscillator runs in the 50-100kHz range the shield >> effectiveness could be low, around 20-30dB. PMTs are sensitive to >> magnetic fields but not sure how critical this is. > > Depends on the PMT, how much gain, but I think an inch is very little. > Depends on the transformer too. > My setup: > http://panteltje.com/pub/PMT_HV_generator_size_img_3174.jpg
A transformer fastened with rope, cool. That's a new one to me. I've used lacing cord but not packaging cord (yet).
> PMT socket on the right
The problem in my case is that I don't have those 11 centimeters.
> Can do up to 2 kV, but now runs a simple Russian FEU35 at much lower voltage, say 500 to 600 V. > http://panteltje.com/pub/Russian_PMT_FEU35_front_img_3189.jpg >
I hope Vladimir's guys don't know about this ...
> At that point I have no problems.. But I have a strong signal from a scintillator crystal. > > AND it is not a switchmode, but a SINE wave generator. > smft >
Mine will be very close to a sine wave so no worries from harmonics. -- Regards, Joerg http://www.analogconsultants.com/
On Monday, December 21, 2015 at 2:06:36 PM UTC-5, Joerg wrote:
> On 2015-12-21 09:55, George Herold wrote: > > On Monday, December 21, 2015 at 12:23:33 PM UTC-5, Joerg wrote: > >> Folks, > >> > >> Does anyone have experience with operating small switchmode transformers > >> close to a PMT? > >> > >> What I am considering in order to avoid a split into multiple boards is > >> to run a Royer with a CCFL transformer on a board that plugs into the > >> PMT. It would also contain the HV ladder, some amps, linear regulators > >> and other stuff. In the trade this whole enchilada referred to as a > >> "socket". I'll try my best to keep harmonics low but running it somewhat > >> sinusoidal. But in terms of shielding there won't be much, towards the > >> tube a full copper plane on the board and on the component side away > >> from the tube a Laird shield can. I could also plop a shallow Laird > >> shield on the bottom side if absolutely needed. The transformer would be > >> just an inch or so from the tube contacts on the opposite side of the > >> board. The switching transistors and filter inductors would also live > >> under that shield. > >> > >> Commercial "sockets" usually have some sort of switcher in there. I > >> don't have one so can't look inside but probably they spent a lot of > >> time optimizing the shielding and positioning of the transformer or > >> inductor. > >> > >> Considering that the oscillator runs in the 50-100kHz range the shield > >> effectiveness could be low, around 20-30dB. PMTs are sensitive to > >> magnetic fields but not sure how critical this is. > >> > >> -- > >> Regards, Joerg > >> > >> http://www.analogconsultants.com/ > > > > Hi Joerg, I know only a little. But I think the mu-metal shields on PMT's > > are there so that if you change the orientation of the pmt in the earth's field > > then the gain won't change. (I've never tried to measure the effect, but I've > > used PMT's with mu-metal shields.) So at a guess if you keep the B-field > > well below that of the Earth's maybe you'll be fine. > > > Probably even less because an AC field creates a signal in the tube. > > > > Toroidal inductors? > > > > Unfortunately not in the cards, this is all on a very small budget. I > have to live with what's available and right now only CCFL transformers > are left. For a while. Soon they'll also go lalaland because monitors > are migrating to LED backlighting. > > > > I think you give commercial socket makers too much credit. > > We have some hamamatsu sockets that put varying amounts of > > (I think) 40 kHz crud from the socket C-W multiplier onto the > > anode. We use it in photon counting mode and the crud is just an > > annoyance, not a real problem. > > > That's what I also assumed, I always wondered how that could ever work. > Well, in this case it didn't which is why they want a new kind of > circuit that performs better.
So are you using HV and a resistor chain or a cockroft-walton type thing? Are you sure it's the B-field in the pmt that is causing the problems? Phil H. had a thread a week or two ago about PMT drives. There they were running the pmt at fairly high currents, and I think he mostly made the voltages sources driving the last few dynodes stiffer. I think he had a link to a good pmt paper. George H.
> > > > ... What mode will you be using? > > > > I don't have much information about the use, just that it all has to be > very stable, clean and low noise. Unfortunately that also means large. > > -- > Regards, Joerg > > http://www.analogconsultants.com/
On a sunny day (Mon, 21 Dec 2015 11:11:14 -0800) it happened Joerg
<news@analogconsultants.com> wrote in <ddr15vFs9d0U1@mid.individual.net>:

>A transformer fastened with rope, cool. That's a new one to me. I've >used lacing cord but not packaging cord (yet).
That is actually a rubber band, and it is just fitting component placing. IIRC it is now fixed with a tie-ribbon.
>> PMT socket on the right > > >The problem in my case is that I don't have those 11 centimeters.
The problem is that you always have a problem, why not sell it ;-)
>> Can do up to 2 kV, but now runs a simple Russian FEU35 at much lower voltage, say 500 to 600 V. >> http://panteltje.com/pub/Russian_PMT_FEU35_front_img_3189.jpg >> > >I hope Vladimir's guys don't know about this ...
I am more worried about demonrats.
>Mine will be very close to a sine wave so no worries from harmonics.
You may like this movie: http://panteltje.com/pub/PMT_FEU35_interface_mvi_3210.avi
On 2015-12-21 12:01, George Herold wrote:
> On Monday, December 21, 2015 at 2:06:36 PM UTC-5, Joerg wrote: >> On 2015-12-21 09:55, George Herold wrote:
[...]
>> >>> I think you give commercial socket makers too much credit. >>> We have some hamamatsu sockets that put varying amounts of >>> (I think) 40 kHz crud from the socket C-W multiplier onto the >>> anode. We use it in photon counting mode and the crud is just an >>> annoyance, not a real problem. >> >> >> That's what I also assumed, I always wondered how that could ever work. >> Well, in this case it didn't which is why they want a new kind of >> circuit that performs better. > > So are you using HV and a resistor chain or a cockroft-walton type thing? > Are you sure it's the B-field in the pmt that is causing the problems? > Phil H. had a thread a week or two ago about PMT drives. > There they were running the pmt at fairly high currents, > and I think he mostly made the voltages sources driving the last few dynodes > stiffer. >
That's all stable and very stiff. Phil is the guru on this stuff.
> I think he had a link to a good pmt paper. >
I must have missed the whole thread, was busy with an x-ray design. Right now I am only concerned about how much of the remaining magnetic field could cause trouble. Since I don't have a CCFL transformer here I can't really measure it. So I just wanted to know whether it's serious or way down below the noise. One issue with split boards is that if you have the HV gen on a separate board, some measurement hooks on the tube socket and then for some reason the ground return becomes loose ... bzzzt. [...] -- Regards, Joerg http://www.analogconsultants.com/
On 2015-12-21 12:09, Jan Panteltje wrote:
> On a sunny day (Mon, 21 Dec 2015 11:11:14 -0800) it happened Joerg > <news@analogconsultants.com> wrote in <ddr15vFs9d0U1@mid.individual.net>: > >> A transformer fastened with rope, cool. That's a new one to me. I've >> used lacing cord but not packaging cord (yet). > > That is actually a rubber band, > and it is just fitting component placing. > IIRC it is now fixed with a tie-ribbon. > > > >>> PMT socket on the right >> >> >> The problem in my case is that I don't have those 11 centimeters. > > The problem is that you always have a problem, why not sell it ;-) >
You can't sell a problem :-) Anyhow, there is always a solution, I am just looking for the simplest one.
> >>> Can do up to 2 kV, but now runs a simple Russian FEU35 at much lower voltage, say 500 to 600 V. >>> http://panteltje.com/pub/Russian_PMT_FEU35_front_img_3189.jpg >>> >> >> I hope Vladimir's guys don't know about this ... > > I am more worried about demonrats. >
Me, too.
> >> Mine will be very close to a sine wave so no worries from harmonics. > > You may like this movie: > http://panteltje.com/pub/PMT_FEU35_interface_mvi_3210.avi >
It is blocked on my PC, can't see it. Do you have it on Youtube also? -- Regards, Joerg http://www.analogconsultants.com/
On a sunny day (Mon, 21 Dec 2015 12:33:07 -0800) it happened Joerg
<news@analogconsultants.com> wrote in <ddr5vfFthkaU1@mid.individual.net>:

>>>> Can do up to 2 kV, but now runs a simple Russian FEU35 at much lower voltage, say 500 to 600 V. >>>> http://panteltje.com/pub/Russian_PMT_FEU35_front_img_3189.jpg >>>> >>> >>> I hope Vladimir's guys don't know about this ... >> >> I am more worried about demonrats. >> > >Me, too.
>> You may like this movie: >> http://panteltje.com/pub/PMT_FEU35_interface_mvi_3210.avi >> > >It is blocked on my PC can't see it.
I knew :-)