Forums :: Resources :: Features :: Photo Gallery :: Vintage Radio Shows :: Archives
Support This Site: Contributors :: Advertise


It is currently Dec Tue 10, 2019 12:53 am


All times are UTC [ DST ]





Post New Topic Post Reply  [ 48 posts ]  Go to page Previous  1, 2, 3  Next
Author Message
 Post subject: Re: CRT circuit question
PostPosted: May Fri 11, 2018 12:16 am 
Member

Joined: Jan Thu 01, 1970 1:00 am
Posts: 2334
Location: Saskatoon
Tube Radio wrote:
That said I still don't get why I am seeing the same retrace lines on the 2246 oscope when it is connected to the deflection plate driver transistor collectors even with the deflection plates disconnected when I can easily get a retrace line free display on the scope when connecting to the clock board output or the OP-AMP output.

That question has been answered several times, and I'll answer it again here. The effect of the deflection plate capacitance is one contributor to the problem, but not the only contributor. The fact that you get no retrace lines when connecting the Tek scope to the output of op amp but do get retrace lines when connecting to the collector of the transistor is telling you that the transistor deflection amp has a serious slew rate problem. You need to redesign the transistor deflection amp.

Edit: I went looking for a deflection amp as used in a decent quality scope to use as an example, but found this page instead which may be more useful.
http://www.labguysworld.com/ES_CRT_DEFLECTION.htm
Note that the circuit is similar, but uses compensating caps across the emitter resistors to increase the bandwidth. That would be an easy fix. You'd have to experiment with different values.


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: May Fri 11, 2018 4:16 am 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20392
Location: Warner Robins, GA
BobWeaver wrote:
That question has been answered several times, and I'll answer it again here. The effect of the deflection plate capacitance is one contributor to the problem, but not the only contributor. The fact that you get no retrace lines when connecting the Tek scope to the output of op amp but do get retrace lines when connecting to the collector of the transistor is telling you that the transistor deflection amp has a serious slew rate problem. You need to redesign the transistor deflection amp.

Edit: I went looking for a deflection amp as used in a decent quality scope to use as an example, but found this page instead which may be more useful.
http://www.labguysworld.com/ES_CRT_DEFLECTION.htm
Note that the circuit is similar, but uses compensating caps across the emitter resistors to increase the bandwidth. That would be an easy fix. You'd have to experiment with different values.


If it had a serious slew rate problem then wouldn't the display be affected far as the clock not displaying correctly aside from the retrace lines such as it being nearly unreadable?

I could understand the retrace lines being present if it was just with the deflection plates connected, but the retrace lines are there without the deflection plates connected and connecting the deflection plates doesn't seem to change the retrace lines at all.

That said the calibration square built into the clock is nearly perfect with a very slight barely noticeable gap on the top left side.

As said previously I did try some compensation across the 1K resistors connecting the emitters together and it did seem to help reduce the gap, but I used one fixed cap. Will try the second cap I have like the first variable and see if that works.

I think the main problem is that I'm trying to make a circuit designed for a 3" CRT work with a 5" CRT.

The only reason for ditching the original scope circuit is because it wasn't all that good as it used several power resistors in the power supply to derive a few voltages and used one 10K 10 watt resistor just for the blanking transistor. Also the vertical and horizontal positions would drift slightly as the scope warmed up. Then over the last 2-3 months I had to fix several things one after the other starting with a high leakage between the CRT heater winding and the 6.3 Vac winding (must have been for graticule illumination in a different scope) I used to power the clock which burnt out one 470K 3 watt resistor in the CRT power supply. I fixed that by connecting the clock to the main low voltage B+ line of 35Vdc through a 500 ohm resistor. The last straw was when I had already purchased the CRT driver boards with the intent of using the spare scope I have to build it in then one night recently I went to my building to get something and came back to a dim display which upon turning up the intensity control was flickering. It was at that point I decided to go ahead with the CRT driver build in that scope.

The scope has a neon pilot light that ran off the 150Vdc B+ line and I needed a way to power it easily so I tried it in the focus control circuit between one end of the focus control and the 1 meg resistor going to ground. Lights just fine on around 74Vdc and helps put the focus control nearly dead center for proper focus. Plus it serves as an indicator that the CRT HV is present.

Here's the bandwidth data on the circuit.

Quote:
The -3dB bandwidth of the design is dominated by two poles. The TL082 op-amp has a fixed gain of 6 and has an unity gain frequency (gain-bandwidth product) of 3MHz, so the -3dB frequency of the op-amp is 3MHz/6 = ~500KHz. The ZTX458 emitter degenerated CS amplifier with the given resistor values has a -3dB frequency at ~300KHz. The board has been characterized to have 1/2 the low frequency gain at ~200KHz. Note that there exists a non-dominant frequency response pole that comprises of the 56K resistors in the deflection amplifiers combined with the CRT deflection plate capacitance that is typically 10-25pF so this third pole is typically >1MHz.


From

http://www.catahoulatech.com/en/pcb-kit ... driver-p-1

So based on what tests I've done I know a 500KHz bandwidth is enough to perfectly display the clock.

That said I am not sure what the bandwidth of the deflection transistor stage is given I use different collector and emitter resistors.

I did notice that if I set the position pots to where 0Vdc was on the base then shorted the base to ground the image expanded slightly which means the transistor which the position pot controls increased in gain slightly.

I do know if the position pots are turned either way (horizontal position pot in this instance) beyond a certain point the displayed image on the 2246 will start to be a line on one side and the farther the position pot is turned the more of the image goes into the line.

Think I'm going to ditch the current position pot arrangement and duplicate the dual gang position pot setup like I used in the DuMont 292 scope. That way as I vary the position pots nothing at all in the deflection amp stage is being changed. Perhaps that will help with the bandwidth.

And

Looking at the DuMont 292 scope the dual gang 4.7 meg position pots will not work because that scope AC couples its deflection plates. So unless I want to AC couple the deflection plates which may present its own set of problems I'm going to have to work with the circuit as is.

Now is it possible to rewire the deflection transistor circuit to the original 68K collector and 4.7K emitter resistors then use a second transistor stage wired up as an emitter follower which will provide the necessary low impedance or would I lose too much deflection voltage and not have enough to properly deflect the clock on the 5" CRT and/or would the position pots not move the beam enough to properly position the display?

I did try altering the 1K resistor in one deflection amp circuit using a 5K pot and didn't get very good results using anything other than 1K.

Now here's a thought.

Wonder what would happen if I replaced the 1K resistor with two 500 ohm resistors then ditch the 2.1K emitter resistors using a 1.05K resistor between where the two 500 ohm resistors connect to each other and the -5 volt line?



Posting this again for easy reference.

http://www.bunkerofdoom.com/lit/belland ... lscope.jpg

If you'll look at the horizontal deflection amp you will see the same sort of position control that is across a + and - 9 Vdc supply which changes the bias on the base of a transistor. The main difference is between the base and ground there's a 10uF capacitor which I suppose bypasses any AC signal to ground.

Also it uses two 220 ohm resistors with a transistor at the junction connected to the -9Vdc supply. The transistor affects how the deflection amp displays the signal, but it does use the same technique to develop the push pull signal and the beam position as this driver circuit uses and it directly drives the bases of the deflection transistors.

So it looks like maybe I can use the deflection transistor stage on the board to provide the drive voltage lowering the emitter resistors and collector resistors to the values shown on the scope schematic then connect the collector resistors to the +5Vdc line instead of the HV and then duplicate the deflection plate driver circuit to drive the deflection plates.

My only concern is the + and - 5Vdc not providing enough voltage swing for the new deflection transistor stage to fully deflect the beam over the 5" CRT.

First I will try my best to make the existing circuit work then if it does not I will mod the deflection amp stage to add emitter followers and if that doesn't work I will mod it to be similar to the original scope's deflection circuit.

Also I will try shorting the 100 ohm resistor in series with the 2K resistor to see what effect that has on the circuit.


Last edited by Tube Radio on May Fri 11, 2018 4:47 am, edited 2 times in total.

Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: May Fri 11, 2018 4:40 am 
Member

Joined: Jan Thu 01, 1970 1:00 am
Posts: 2334
Location: Saskatoon
Tube Radio wrote:
If it had a serious slew rate problem then wouldn't the display be affected far as the clock not displaying correctly aside from the retrace lines such as it being nearly unreadable?

No. The slew rate problem wouldn't affect the parts that are supposed to be visible, because the beam is moving slowly when it's drawing those parts, the output of the deflection amp is changing slowly relative to the retrace part, and is within the slew rate limitations.

As I mentioned in my last post, you could probably mitigate the problem with a compensating cap across the 1k common emitter resistor. I'd start with 100pF and increase the value in steps of 100pF to see if that helps. If that doesn't help, then I'd reconfigure the emitter circuit as per the circuit that I linked to in my previous post.


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: May Fri 11, 2018 4:46 am 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20392
Location: Warner Robins, GA
BobWeaver wrote:
Tube Radio wrote:
If it had a serious slew rate problem then wouldn't the display be affected far as the clock not displaying correctly aside from the retrace lines such as it being nearly unreadable?

No. The slew rate problem wouldn't affect the parts that are supposed to be visible, because the beam is moving slowly when it's drawing those parts, the output of the deflection amp is changing slowly relative to the retrace part, and is within the slew rate limitations.

As I mentioned in my last post, you could probably mitigate the problem with a compensating cap across the 1k common emitter resistor. I'd start with 100pF and increase the value in steps of 100pF to see if that helps. If that doesn't help, then I'd reconfigure the emitter circuit as per the circuit that I linked to in my previous post.


Oh ok I think I get it now. Than's for explaining it in more detail.

Yes I'm going to try the compensating caps tomorrow.

I've got two 250-1000pF variable caps which were part of the original vertical circuitry of this scope and I had two of the scopes which gives me two caps.

Here's the schematic of how the deflection amps will be.

Attachment:
Deflection 2.png
Deflection 2.png [ 7.09 KiB | Viewed 1624 times ]




EDIT:

On a side note.

Sacrificing the original scope circuitry for the new circuitry has provided me with the proper slide switches I needed for my McIntosh C-24 preamp. :D :D :D



Thanks for posting that link, Bob.

This picture is pretty much what my display looks like except the geometry is a little better on mine

http://www.labguysworld.com/ES_AMP_007.jpg

Wish I had seen this page first as I could have saved some money and had a much simpler circuit for driving the CRT deflection plates.

That said I may look into altering the gain of the OP-AMP as there is entirely too much gain as is.

The driver board was designed to take into account the DC voltage the scope clock has on its outputs though as I tried AC coupling the horizontal and it wouldn't fully deflect horizontally on the left side without distortion which might just be due to the uncompensated deflection amp.


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: May Fri 11, 2018 2:47 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20392
Location: Warner Robins, GA
Well I've got very good news.

I tried both variable caps one on the horizontal and one on the vertical then tweaked them. I found that I could adjust them to where I'd get a faint line extending from the lower left horizontal of the calibration square going left and the vertical would extend vertically I think from the top left side.

So I adjusted the trimmers until the lines no longer extended past the square and I was hansomely rewarded with a proper display that is at least as good as what I got with the original scope deflection circuitry.

I also redid some grounds using larger wire for the clock board ground which also now serves as the ground wire for the 6.3vac winding that powers the deflection board LV section and moving the 800 volt winding ground to where the HV CRT circuitry is grounded.

I am not that sure it helped much with the ground issues, but I did install the case cover over the CRT side to block the light from the super bright flashing green LED on the clock board and I noticed the display no longer did what it was doing as bad as it was.



EDIT:

Got the driver board and board with the two variable caps on it mounted to the scope chassis using double sided tape and everything works great.

I tried the stock retrace blanking circuit and it didn't do good at all.

In fact since day 1 I could never get the retrace blanking to work right as it would always start to blank the first part of the first number of the date display and some of the retrace lines near the hands, but wouldn't do anything for the retrace lines around the edge of the clock face.

Now the original retrace blanking circuit consisted of a 10K 10 watt resistor a MPSU10 transistor (same ones used for the deflection and intensity control) and a 68 volt zener between the transistor collector and ground.

Perhaps I need to try that circuit or is there a better retrace blanking circuit for a 5" CRT that is guaranteed to work?

That link I was given here talks about a retrace blank circuit, but then doesn't provide a schematic for it unfortunately.


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: May Mon 14, 2018 6:19 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20392
Location: Warner Robins, GA
So even though I got the geometry perfect meaning there's enough bandwidth for the clock to display properly I still see retrace lines on the Tektronix 2246 100MHz scope when I monitor the deflection transistor outputs using 10X probes.

Is there any actual way to increase the bandwidth more if that is even the problem?

Now I do expect some retrace lines due to the CRT I used, but I don't think I should be seeing retrace lines when monitoring the deflection outputs with the deflection plates disconnected.

My thought is that those retrace lines are what's making it difficult to get the retrace blanking perfected.


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: May Mon 14, 2018 11:47 pm 
Member

Joined: Jan Thu 01, 1970 1:00 am
Posts: 2334
Location: Saskatoon
Are you monitoring the transistor output while it's connected to your display CRT and displaying the clock pattern? If so, does connecting the scope probes make any change in the display on the display CRT?

What I'm getting at is that the compensating capacitors that you added have a similar function to the compensation of a scope probe. You can under-compensate and you can over-compensate. The correct amount of compensation depends on what you have connected to the output transistors. Basically, if the clock pattern looks good on the display CRT, then the circuit is correctly compensated.

Rule of thumb: when you get something to work, it's time to stop fixing it. If you continue fixing it after it's working, then it will soon stop working again.


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: May Tue 15, 2018 2:52 am 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20392
Location: Warner Robins, GA
Best as I can remember the display on the Tektronix 2246 looked the same whether or not the CRT deflection plates were connected at least as far as the retrace lines are concerned.

I do have it now to where the retrace lines are barely visible on the 5UP1 CRT which came about quite by accident.

I was trying to reduce the sensitivity of the OP-AMP which didn't work because any change would muck with the deflection plate driver stage to where I could no longer compensate for it with the variable caps.

Same with adding a 100K resistor in series with the signal input end of the 100K pot.

I had a couple of the original 5K pots from the scope so I tried one and then tried a 14K resistor I had handy in series with the input end of each pot. Had to remove the fixed caps across the variable caps to provide the proper compensation and was rewarded with a much better display with retrace lines that are extremely hard to see such that I don't think I even need a retrace blanking circuit.

The best screen for setting the horizontal compensation is the standard clock face and I adjust until the 1 is perfectly straight at the bottom.

The best screen for the vertical compensation is the binary clock as it has a lot of horizontal lines and I just adjust until they are perfectly straight at least that's how it should work. As I adjust one way on the upper half of the CRT the lines turn up at the left end and on the bottom half of the screen the lines turn down on the left end. The opposite happens when I adjust the other way.

The only problem is this.

I cannot get the lines perfectly straight so that means I need individual compensation for each transistor in the vertical and I might do the horizontal as well just to keep things the same and provide full adjustment.

Here's an example of the binary clock screen. The horizontal lines I look at are the tops of the dots.

http://www.dutchtronix.com/images/ScopeClock-Binary.JPG

Here's an example of the regular clock face. The 1 I look at is 1 o'clock

http://www.dutchtronix.com/images/ScopeClock-date.JPG

I of course first adjust with the cal screen so that the line does not extend horizontally past the left side of the square and so that the line does not extend vertically past the bottom left of the square. The other two screens are the fine adjust.

What I might try doing is connecting a variable cap between the base of the transistor the position control connects to and ground and see what that does.

If that fixes it then good. If not I'll try something else.


If you'll look at this picture my display looks at least this good if not slightly better, although mine isn't quite as good with the focus, but is still quite sharp.

http://www.dutchtronix.com/images/ScopeClock-time.JPG

The best display I ever had is when I had the sparkfun version of the clock in my Tektronix 606A X/Y monitor. Nice crisp display and very sharp due to the focus circuit the display has.

Only thing I don't like is the sensitive intensity control which only adjusts the intensity over the last 1/4 CW rotation.

it is a 1 meg control with a 470K resistor between it and ground.

Suppose I could try putting a 1 meg across gthe 1 mwg pot then using a 500K resistor in series with the 470K resistor which will keep the load the same thusfore keeping the voltage the same, but might allow more of a range of control.



EDIT:

Got the vertical issue sorted.

I used the original vertical pots from the two identical scopes as vertical and horizontal controls. They are a triple pot with two 2K sections ganged and adjusted with a screwdriver by removing the knob which is the coarse adjust. The fine adjust is a 100 ohm pot connected between both 2K pots. Using those pots provides very good position control and also cleaned up the image some.

Found out that I had another of those variable caps so I tried connecting it between the base of the transistor and the center terminal of the vertical position pot and I put a 10K resistor in parallel with it.

I adjusted the one vertical cap until I got the horizontal lines reasonably straight then I tweaked the other vertical cap and went back and forth between the two caps until I got the horizontal lines as straight as possible.

The display is nearly perfect.

The retrace lines are so faint that I won't need a blanking circuit.



Here's the original intensity circuit.

Attachment:
Intensity control.png
Intensity control.png [ 1.92 KiB | Viewed 1575 times ]


Now here's the idea I had on how to make the control work over more of its range.

The idea is to keep the total resistance the same so the B- voltage stays the same.

Attachment:
Intensity control 2.png
Intensity control 2.png [ 2.17 KiB | Viewed 1575 times ]




EDIT 2:

The necessary control range is -60 volts

I can do it one of two ways.

1. Place a 100 volt zener across the 1 meg intensity pot and figure out what resistance I need to add to the 470K resistor to keep the B- voltage the same.

2. Use a 100K pot and a 1.375K 2 watt resistor which in my case will be made up of a single 300K resistor a 75K resistor and two 1 meg resistors in series which are in parallel with two 1 meg resistors for a total of 1 meg 2 watts which will put -94Vdc across the 100K pot.


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: May Wed 16, 2018 5:28 am 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20392
Location: Warner Robins, GA
Here's some pictures. The camera nor the 800 X 600 size do the display justice as it looks better in person and the focus is a bit better than the camera shows.

Here's the display at full brightness. The original clear plastic piece that went in front of the graticule I left in place to provide protection for the CRT face.

Attachment:
Full bright.jpg
Full bright.jpg [ 106.3 KiB | Viewed 1573 times ]


Normal brightness

Attachment:
Normal bright.jpg
Normal bright.jpg [ 112.22 KiB | Viewed 1573 times ]


Binary display. Note the horizontal lines that form the top of the dots how straight they are.

Attachment:
Binary display.jpg
Binary display.jpg [ 95.04 KiB | Viewed 1573 times ]


Easier to see the lines in a dark room

Attachment:
Binary display in dark room.jpg
Binary display in dark room.jpg [ 44.16 KiB | Viewed 1573 times ]


Display in dark room

Attachment:
Display in dark room.jpg
Display in dark room.jpg [ 48.11 KiB | Viewed 1573 times ]


DS32KHZ TCXO used to replace the on board RTC oscillator XTAL for better accuracy and is fed from same point as RTC so it works off the battery when the clock is turned off.

Attachment:
DS32KHZ TCXO.jpg
DS32KHZ TCXO.jpg [ 122.72 KiB | Viewed 1573 times ]


Clock board. RED LED on the cabinet is wired in series with the flashing green LED so I don't have to open the case to see the status.

Attachment:
Clock board.jpg
Clock board.jpg [ 154.63 KiB | Viewed 1573 times ]


Whole CRT circuit.

Attachment:
CRT driver circuit.jpg
CRT driver circuit.jpg [ 168.26 KiB | Viewed 1573 times ]


CRT driver board and compensation caps. Vertical compensation top. Horizontal compensation bottom.

Attachment:
CRT driver board + compensation caps.jpg
CRT driver board + compensation caps.jpg [ 157.94 KiB | Viewed 1573 times ]


CRT controls + power supply and vertical compensation cap. Gray caps are epoxied to the chassis.

Attachment:
CRT supply + controls.jpg
CRT supply + controls.jpg [ 159.14 KiB | Viewed 1573 times ]


Coming across 1.7KV on a .1uF cap hurts :shock:


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: May Mon 21, 2018 3:06 am 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20392
Location: Warner Robins, GA
Well I thought I had everything good until I took it home and noticed I could not fill up the whole CRT face anymore without horizontal distortion.

Turns out my line voltage is 121Vac and it must be a little lower at work.

So given I have compensation caps I may look into either increasing the value of the collector and emitter resistors or finding a way to regulate the voltage.

I learned a valuable lesson Friday.

Pay close attention to where your HV probe is if you cannot connect it by anything other than a clip lead which leaves the probe and exposed.

Had the probe sitting on another piece of test equipment and with the clock on I went to install another small neon light on the chassis. Next thing I know I'm seeing a brief flash of light and feeling the same feeling in my tongue one does when a 9 volt is put on the tongue. Turns out my head or hair must have brushed against the probe end with one hand on the scope chassis.



EDIT 2:

Ran into a problem today.

I increased the collector resistors to 50K and the emitter resistors to 4.7K.

That worked ok, but found a 5K in parallel with each 4.7K worked better.

I then with a variable resistor found I could get more horizontal output by reducing the value of one of the 4.7K resistors.

Did that and while testing it worked fine.

Did the vertical and that's where the problem started.

The TL-082 got hot and there was some weird pattern on the CRT so I removed the OP-AMP and I get a small line pointing towards the right top of the CRT and viewing the deflection plate signal I see a sawtooth waveform on both vertical and horizontal plates that is not affected by anything, although the position controls work.

Not sure what the issue is as I don't think anything is shorted together that shouldnt be.

What i did was tweak the horizontal for the maximum output while still being able to properly compensate it.

That way I can have more outout for other 5" CRTs such as the 5BP4 which does work with the current circuit and has the sharpest display but will not fill the full screen without distortion.

The information is being sent to the circuit designer so he can update his site with info on how to make it work with the 5" CRTs. Perhaps that will spark more interest in his circuit.

What I really want to do is find the largest electrostatically deflected CRT and make a scope clock out of it.


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: May Wed 23, 2018 2:15 am 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20392
Location: Warner Robins, GA
An update.

I was able to get the deflection transistor stage to produce around 430Vpp deflection plate to deflection plate in both vertical and horizontal.

That drives the 5DEP1 very nicely and I could easily increase the B- to the CRT and have plenty of drive voltage to fill up the CRT, but running on around -1600Vdc it is more than bright enough.

The problem.

I had to alter the compensation by adding more 390pF caps in parallel to the trimmer caps used for the compensation adjustments.

Image is just slightly worse than it was previously and is something I can live with, but the perfectionist in me has to get it right.

I can now use the 5BP4 and have enough drive to fully cover the screen and then some.

I haven't tried adjusting the compensation for that one yet.

With the 5BP4 I can get better brightness at around 1.7KV which is what I get when I short the 470K resistor in the B- supply, but I cannot fully expand the image to cover the whole CRT face (covers maybe 4 to 4 1/2 inches) without distortion.

Also still have to try the 5ADP1 and adjust the compensation for it.

Next thing I may try is to connect the emitters in the vertical together and do the same for the horizontal which will make the deflection amp better and provide higher bandwidth and in so doing it may be that I will no longer need compensation circuitry or very minimal compensation.


I did notice something.

While troubleshooting a sawtooth waveform that shouldn't be there I found it was coming from the -5Vdc supply. A 4700uF cap for the main filter and a 2200uF cap after the regulator took care of it.

I then saw that the initial problem of what looked like another dimmer image swimming around the main image which I attributed to a ground issue seems to be completely gone now.


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: May Wed 23, 2018 4:44 am 
Member

Joined: Jan Thu 01, 1970 1:00 am
Posts: 2334
Location: Saskatoon
Tube Radio wrote:
Next thing I may try is to connect the emitters in the vertical together and do the same for the horizontal which will make the deflection amp better and provide higher bandwidth and in so doing it may be that I will no longer need compensation circuitry or very minimal compensation.


That won't give you what you're expecting. Connecting the emitters together without the resistor will cause the gain to increase. You'll still need compensation, but you won't have any place to put it.


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: May Wed 23, 2018 11:41 am 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20392
Location: Warner Robins, GA
Depends.

I do know that the original scope circuit did tie emitters together like I'm thinking, but it used a constant current source from the emitters to ground with other stages after the initial push pull stage having some sort of compensation.

The circuit designer is going to experiment with something similar so perhaps a workable solution could be made that would increase the bandwidth and be better than the current circuit.

Can't hurt to try it myself though and see what I get.

I could always try a larger value for the 1K resistor or a smaller value and see what that does.

This is more of an experiment to see if I can get it working as perfectly as possible with the three 5" CRTs I have so the info can be passed along to the designer along with seeing what doesn't work. Also this circuit has been a blast to experiment with.



EDIT:

Decided I didn't need to mess around with tying the emitters together.

Got it working well on a 5DEP1, 5ADP1 (5CP1 should work as well) and a 5BP4.

Geometry is nearly perfect, but could use some improvement, although it looks good enough to me.

Perhaps I should take a cue from the DuMont 304H scope that uses a 5ADP1 with a choke in series with each deflection plate and try the same here.

I could do like the Heath scope did and incorporate some of its compensations, but that is impossible to do with just one stage of amplification after the OP-AMP stage.

Once I get the clock home I will take pictures with all three CRTs and post them here.

I haven't posted the modified full schematic as I would rather wait until it is on the designer's website first.

What I am wondering is if I could get better compensation by putting a resistor in series with each deflection plate and a variable cap across each resistor versus the scheme I have now.


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: May Fri 25, 2018 5:17 am 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20392
Location: Warner Robins, GA
More pictures with the circuitry in its final state.

I do not have pictures of the 5ADP1 because while testing I shorted the -5 volt regulator so I need to order one.

Because of the way the board is mounted I'll just clip the old regulator off and install the new one where the extra filter caps are mounted for the -5Vdc supply.

Attachment:
5BP4 lighted room.jpg
5BP4 lighted room.jpg [ 99.06 KiB | Viewed 1509 times ]


Attachment:
5BP4 darker room.jpg
5BP4 darker room.jpg [ 73.67 KiB | Viewed 1509 times ]


Attachment:
5DEP1 full bright lighted room.jpg
5DEP1 full bright lighted room.jpg [ 127.27 KiB | Viewed 1509 times ]


Attachment:
5DEP1 normal bright lighted room.jpg
5DEP1 normal bright lighted room.jpg [ 111.99 KiB | Viewed 1509 times ]


Attachment:
5DEP1 full bright darker room.jpg
5DEP1 full bright darker room.jpg [ 78.01 KiB | Viewed 1509 times ]


Attachment:
5DEP1 normal bright darker room.jpg
5DEP1 normal bright darker room.jpg [ 62.98 KiB | Viewed 1509 times ]


I actually like the 5BP4 the best due to the superior image clarity, but the rounded face and the lower intensity doesn't make it very practical for me as I wouldn't be able to use a color filter to change the color of the display plus the image isn't perfect because of the rounded face.


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: May Wed 30, 2018 3:56 am 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20392
Location: Warner Robins, GA
Here's a couple more pictures showing the geometry issues.

Attachment:
20180529_221619-600x800.jpg
20180529_221619-600x800.jpg [ 43.36 KiB | Viewed 1463 times ]


Attachment:
20180529_221507-600x800.jpg
20180529_221507-600x800.jpg [ 38.44 KiB | Viewed 1463 times ]


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: Jun Tue 05, 2018 5:36 am 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20392
Location: Warner Robins, GA
Looks like I done screwed the circuit up again.

:x :x :x

Was testing the 5DEP1 and I moved the CRT and heard a pop.

Now all I get is a dot on the 5DEP1 and it only moves horizontally by the position control.

Vertical nothing at all.

Horizontal no deflection as I turn up the size pot.

Measuring voltage I get similar to what I'd normally get for the horizontal and the voltage changes on both deflection plates inversely as it should so that seems to indicate both horizontal transistors might be good, although it doesn't explain why no signal is being output at all which given the low impedance output of the OP-AMP should be able to put some signal into a 2.4K load.

Vertical no signal out of the OP-AMP as well.

Only thing I can figure is something caused the transistors to short briefly to the bases which may have fried the OP-AMP then the collectors opened up.

I feel like I'm so done with this project, but I gotta get it finished.

That means I'll have to use the transistors out of the parts I bought for the 3" clock I'm going to build and order more.

Also means I'll have to take the clock to work again so I can try and desolder the transistors from the board.

I'm going to try something from the RCA datasheet for the 5UP1 and use dual gang 500K pots for the position controls in series with 2.2 meg resistors with the signal being coupled via a .1uF cap.

I will then disconnect the original position controls and ground the end of the compensation caps that would have connected to the position pots.

Will see how well that works far as the displayed signal goes.

Now that might affect the compensation, but it shouldn't be anything I can't adjust.

That would also get me the proper AC coupling I needed so I can work on a different size for the menus and other square things plus it will keep the deflection amps biased properly given the bias won't change with the position pots anymore.


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: Jun Wed 06, 2018 11:51 am 
Member
User avatar

Joined: Aug Wed 26, 2015 9:41 pm
Posts: 1259
Location: Baton Rouge, Louisiana
You seem to be referring to this circuit.
I used a modified version of it and so far, it seems to work fairly well.
Attachment:
5up1_ckt.png
5up1_ckt.png [ 41.92 KiB | Viewed 1405 times ]

_________________
Mark


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: Jun Wed 06, 2018 7:33 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20392
Location: Warner Robins, GA
That's the circuit.

At one point a few years ago I was going to use that very circuit with a 5DEP1 (similar enough to a 5UP1) and install it in a scope chassis that originally used a 5BP1 CRT and have a display.

Never got beyond doing a few initial things though.

It will definitely be something worth trying.

With AC coupling of the deflection plates does a displayed image move about any?

Somewhere at home I think that I have a board with a few transistor sockets that will fit the TO-92 transistors so I may install those. That way I won't have to worry about desoldering the transistors should I ruin them again.

Looked on Mouser for some when I ordered the transistors and they are expensive.

Now provided this method will work correctly I will be able to do the mod of adding a resistor and switch in parallel with the size pot on the clock board.

I got in touch with the designer of the clock board asking if there was any firmware that would basically solve the problem of the square things being off the screen at the edges of the square when a round clock face is set to fill the whole usable part of a round CRT face. Haven't heard back yet.


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: Jun Thu 07, 2018 1:05 am 
Member
User avatar

Joined: Aug Wed 26, 2015 9:41 pm
Posts: 1259
Location: Baton Rouge, Louisiana
Quote:
With AC coupling of the deflection plates does a displayed image move about any?

I'm not sure, haven't tried it yet.
Quote:
the problem of the square things being off the screen at the edges of the square when a round clock face is set

There is a guy on the web that had a hardware problem with his crt.
He was able to place magnets on the neck of the crt to provide a good square,
then he adjusted the circle.
Referencing the schematic, what do R14, R15, R16, R17 do?
I don't have them in at the moment, my best guess is current limiters.
What you say?

_________________
Mark


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: Jun Thu 07, 2018 1:12 am 
Moderator

Joined: Jan Thu 01, 1970 1:00 am
Posts: 37561
Location: Livermore, CA
R14 etc prevents loading of the signal going to deflection plates. If centering controls were toward either end signals to deflection plate would be shorted out.

_________________
Norm


Top
 Profile  
 
Post New Topic Post Reply  [ 48 posts ]  Go to page Previous  1, 2, 3  Next

All times are UTC [ DST ]


Who is online

Users browsing this forum: No registered users and 7 guests



Search for:
Jump to:  




























Privacy Policy :: Powered by phpBB