Photomultipliers in Practice

Forum for discussion of narrow-bandwidth mechanical television

Moderators: Dave Moll, Andrew Davie, Steve Anderson

Postby Panrock » Fri Jan 11, 2008 6:05 am

Ah ha! Will do. I'm surprised it's that sensitive. (EDIT - just done it. It's dead steady at 1015v under all conditions.) There's a 'gain' pot in the middle of the resistor chain too. The best setting seems to be at about ¾ travel. Over that and you tend to get 'bleaching' on whites. I still haven't really learned how to drive it though.

They are very sensitive. This afternoon I was getting usable pictures at f11 (though a bit noisy) in semi-twilight conditions. Yes, I had to close the blinds in my workshop and turn off all the lights! 8)

If you can figure out a way to make an LM338K ultra-low in impedance for reducing the chunks being taken out of the motor supply, let me know. However, maybe it will already be upto the job. I won't know till I try...

Steve
Panrock
Green padded cells are quite homely.
 
Posts: 870
Joined: Mon Feb 05, 2007 8:25 am
Location: Sedgeberrow, England

Postby Steve Anderson » Fri Jan 11, 2008 6:31 am

Panrock wrote:There's a 'gain' pot in the middle of the resistor chain too. The best setting seems to be at about ¾ travel. Over that and you tend to get 'bleaching' on whites. I still haven't really learned how to drive it though. Steve


Reading through the Hamamatsu data I posted here a few days ago they don't seem to be that keen on this method of gain control, see their page 99 and read between the lines. I would dispense with this pot and keep the resistive divider equal on all stages. The implication being that gain control is better achieved in the semiconductor or optical stages.

Panrock wrote:If you can figure out a way to make an LM338K ultra-low in impedance for reducing the chunks being taken out of the motor supply, let me know. Steve


At 12V out the LM338 has an output Z of around 0.012 ohms, more will be introduced by your wiring! Also put some suppression accross the motor terminals and if really persistant consider some series inductors. Yet another downside to using DC brushed motors.

Steve A.
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
 
Posts: 5360
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand

Postby Steve Anderson » Fri Jan 11, 2008 6:49 am

Panrock wrote:(EDIT - just done it. It's dead steady at 1015v under all conditions.) Steve


Hmm, well that can't be the source of the sensitivity to adjustment then. Let me think on this further...I assume that it's not the same as the excess gain you initially had with the Luxeon driver? i.e. Is the gain too high? However, when you do swap over to 80-lines it might be useful.

It might be useful to see what a still camera makes of the same varying scenes, and how its f-stop varies (or exposure time). Does it correlate to the adustments you need to make to keep a constant peak white? One f-stop is the same as going from 50% pot setting to 100%.

Steve A.
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
 
Posts: 5360
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand

Postby Panrock » Fri Jan 11, 2008 6:57 am

Let's get this in proportion... I am delighted with the performance of this rig as a whole (except currently for the synch) compared with how it was. I've just had a look at my favourite 'Daz' packet and was stunned by the photo-like detail now visible. It's like looking through a telescope of greater aperture.

I am going to have an intensive session on the synch tomorrow, and probably will order (yet another) transformer and regulator to improve operation in the monitor. It's a shame the hash on the camera pulses isn't always there - makes it hard to trace. Presumably it's caused by that wonderful brushed motor... :P

Steve
Panrock
Green padded cells are quite homely.
 
Posts: 870
Joined: Mon Feb 05, 2007 8:25 am
Location: Sedgeberrow, England

Postby Steve Anderson » Fri Jan 11, 2008 7:15 am

Panrock wrote:It's a shame the hash on the camera pulses isn't always there - makes it hard to trace. Steve


Could be....it just shows up in the pulses? Not white or dark pulses (shash) within the picture?

I have no idea how you're handling sync, perhaps you could fill me in.

Anyway, I'm off to bed, it's 1:45am here..

tomorrow..

Steve A.
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
 
Posts: 5360
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand

Postby Panrock » Sat Jan 12, 2008 7:59 am

Talk about dedication... thanks Steve, but it has been worth it.

Today I have separated the sync and video power completely from the motor (by using a different transformer) and put a 9v zener + emitter follower into the sync board power. The result is that the previous interaction between picture content and sync lock is definitely cured. However, there is still a little hunting and I think the next step will be to regulate the DC motor supply. It's already 'smoothed' by 10,000 + 22,000uF with an R in-between but clearly this isn't enough. (EDIT I've now found Pete Smith's circuit works best with a critical 11 or 12v on the motor supply - to the extent that hunting is virtually eliminated).

There really is far too much modulated light available from the luxeons at 30-lines and I find I'm usually running the system at a fraction of its capacity because when I give it some welly, the flicker on the brilliant picture becomes objectionable. I haven't the slightest doubt there will be enough light for 80-lines. The luxeons themselves are running cool, helped along by the proximity of the fan (for the motor) near the heatsink.

It's a great hobby. :) It's hard to convey why looking at a test pattern or a Daz packet across the workshop in flickering, brilliant colours give so much satisfaction - but it does!

I won't be able to get the motor supply regulator until Tuesday. More then.

Steve O
Panrock
Green padded cells are quite homely.
 
Posts: 870
Joined: Mon Feb 05, 2007 8:25 am
Location: Sedgeberrow, England

Postby Steve Anderson » Sat Jan 12, 2008 5:07 pm

Panrock wrote:Talk about dedication... thanks Steve, but it has been worth it. Steve O


Don't be fooled! Being up at 1:45am wasn't for NBTV, sadly, but my 'real' work. I just check the forum a couple of times a day and respond when appropriate.

It seems like you're making progress on motor control. There's not a lot I can help on without the mechanics in front of me, it all appears to be a trial-and-error process.

The only thought I have had (and this will depend on the actual motor characteristics), is there is no electric braking. The MOSFET drives the motor with volts, if an overspeed occurs it switches off or at least reduces drive volts.

Some form of push-pull drive similar to those in the output stages of audio power amps might help in keeping the motor in line. But I'm only speculating here, it's an area I haven't really been involved in.

Steve A.
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
 
Posts: 5360
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand

Postby Panrock » Sun Jan 13, 2008 5:51 am

There is some light hash on the camera syncs which is causing false triggering on the monitor. I noticed this is only present when the red channel in the camera is turned well up. The other channels don't cause this. The red picture is normal.

I've screened the camera sync feed and pulse circuitry but it's still there. Could this could be the red channel breaking into RF oscillation (but not affecting the picture) ?

EDIT UPDATE: I've now cheated by putting 1000pF straight across the camera sync output. This soaks up the hash and barely touches the waveform. 8)

Steve
Panrock
Green padded cells are quite homely.
 
Posts: 870
Joined: Mon Feb 05, 2007 8:25 am
Location: Sedgeberrow, England

Postby Panrock » Sun Jan 13, 2008 11:17 pm

Steve Anderson wrote:
Panrock wrote:There's a 'gain' pot in the middle of the resistor chain too. The best setting seems to be at about ¾ travel. Over that and you tend to get 'bleaching' on whites. I still haven't really learned how to drive it though. Steve


Reading through the Hamamatsu data I posted here a few days ago they don't seem to be that keen on this method of gain control, see their page 99 and read between the lines. I would dispense with this pot and keep the resistive divider equal on all stages. The implication being that gain control is better achieved in the semiconductor or optical stages.


I've now - I think - mastered the use of this pot. It seems definitely necessary. Basically, if you set it at maximum, you get a noisy picture. Reducing it (and resetting the other gains in the chain) results in a much better, noise-free picture but less sensitivity. Set for low noise, the camera will give good pictures in dim room light, whereas before it would operate in semi-twilight. This means when I lose 3 stops at 80 lines it will probably need dull daylight.

I've tried to capture the results with my camera, but for some reason am now getting 'shutter' effects or over-exposure and am unable to do the pictures justice. Anyway, here's a view of the monitor from the rear as it is at present 'in bits'. Hope I'm not boring everybody.

The sync is pretty good now but I think I shall probably fit the regulated motor supply next week anyway. Then I'll fit the first 80-line disc to the monitor to judge raster quality before deciding whether to proceed further.

Steve
Attachments
Imgp1226.jpg
Imgp1226.jpg (66.41 KiB) Viewed 13028 times
Panrock
Green padded cells are quite homely.
 
Posts: 870
Joined: Mon Feb 05, 2007 8:25 am
Location: Sedgeberrow, England

Postby Steve Anderson » Mon Jan 14, 2008 5:50 pm

Panrock wrote:I noticed this is only present when the red channel in the camera is turned well up. The other channels don't cause this. The red picture is normal. Steve


Not sure why this should be happening, the only difference between the channels is the type of PMT. Might be worth swapping the PMTs around and see if the fault follows the red PMT...that's assuming base and pin-out is the same.

I don't have the circuits you're using for sync generation, but if you can post them here, that might help.

Steve A.
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
 
Posts: 5360
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand

Solid-State PMTs.

Postby Steve Anderson » Mon Jan 14, 2008 8:02 pm

Steve,

Just checked, it appears the R446 is the same base & pinning.

For those a little hesitant at working with 1000V of DC, I've come across these devices. SSPMs, they appear to be semiconductor versions of the high voltage vacuum PMTs, but without the high bias volts.

They do seem fussy though about applied volts and it would have to be well regulated and different for each colour sensitivity. I think these are quite a new development (read - expensive). The companies web site is mentioned within the datasheet.

Steve A.
Attachments
SSPM_0701BG_TO18.pdf
(62.42 KiB) Downloaded 865 times
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
 
Posts: 5360
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand

Postby Panrock » Tue Jan 15, 2008 5:30 am

Thanks. Fascinating about the little semiconductor PMTs - completely new to me!

I'm happy with that C absorbing the hash and so the sync is basically sorted. I could post the circuit but it's probably not worth it. It's basically the 'club' circuit, but with two 555 monostables between the pulse pickup photodiodes and the 4046, ensuring its inputs get nice clean pulses that match in length and amplitude (which I've always presumed is a good idea?).

Tomorrow I shall fit a 7½ amp regulator to the motor supply and set to 11v, which seems to be optimum.

Then I'll check the first 80-line disc for quality in the monitor, on a signal generator source.

Steve
Panrock
Green padded cells are quite homely.
 
Posts: 870
Joined: Mon Feb 05, 2007 8:25 am
Location: Sedgeberrow, England

Postby Steve Anderson » Tue Jan 15, 2008 4:34 pm

Panrock wrote:Then I'll check the first 80-line disc for quality in the monitor, on a signal generator source. Steve


I was thinking of knocking up a .wav file to suit your system which would be mono, but at least stable. As you are using seperate syncs I would have put those on the right channel and the (mono) video on the left.

But if you already have a source then there's no need.

It would have to be a 96kHz .wav file as 48kHz doesn't have the bandwidth you require. So you would need to check your PC can play them back.

Steve A.
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
 
Posts: 5360
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand

Postby Panrock » Tue Jan 15, 2008 8:18 pm

That would be great Steve, as it would also give the PWM drivers a thorough run-through.

There seems to be a problem photographing the picture that I didn't have before. However, to the eye, the pictures are much better (if a little bright!), sharper, more subtle, with lovely flat black level.

Positive-going syncs on the right channel would be fine.

My knowledge is limited - not sure what I need to check on my PC? Can I presume the sound card's stereo outputs would not introduce phase distortion...?

Steve
Panrock
Green padded cells are quite homely.
 
Posts: 870
Joined: Mon Feb 05, 2007 8:25 am
Location: Sedgeberrow, England

Postby Steve Anderson » Wed Jan 16, 2008 1:27 pm

Panrock wrote:There seems to be a problem photographing the picture that I didn't have before. However, to the eye, the pictures are much better (if a little bright!), sharper, more subtle, with lovely flat black level. Steve


There might be an issue with the cameras exposure system. Our eyes integrate the PWM so we see an average level. But perhaps the camera responds to this and gets somewhat confused. Not sure why a camera would need such a fast exposure response time, but it's a thought.

As for the 96kHz .wav file, the only thing you'll need to check is simply if your machine can play it, most likely it will unless is getting a little old. Give me a little while and I'll get on with it.

In the meantime attached is a sample 96kHz .wav file. It's mono (as in sound) 1kHz maximum amplitude 8-bit signal. If your PC plays this back OK there should be no problem with the stereo version. It's a sine-wave and it might be useful to check that it looks like one on a scope to ensure that no audio stages are overloading.

Steve A.
Attachments
96_1kHz.wav
(2.75 MiB) Downloaded 739 times
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
 
Posts: 5360
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand

PreviousNext

Return to Mechanical NBTV

Who is online

Users browsing this forum: No registered users and 26 guests

cron