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


It is currently Jul Tue 16, 2019 7:44 pm


All times are UTC [ DST ]





Post New Topic Post Reply  [ 48 posts ]  Go to page 1, 2, 3  Next
Author Message
 Post subject: CRT circuit question
PostPosted: May Wed 09, 2018 11:31 am 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20037
Location: Warner Robins, GA
I am using two of these circuits to drive the deflection plates of a 5UP1 CRT.

Attachment:
Deflection.png
Deflection.png [ 6.26 KiB | Viewed 2070 times ]


If I for instance feed the circuit with a vector signal such as an oscilloscope clock I get retrace lines on the CRT.

I expected some light retrace lines given my experience with another scope using the same CRT displaying a vector signal, but I get a good bit more than that.

What's even stranger is this.

I can take the same vector signal and display it on a Tektronix 2246 scope and it has no retrace lines whatsoever

BUT

If I feed the scope from the deflection transistor outputs using the same 10X probes I used to feed the vector signal in separately I get the same retrace lines as I do on the CRT even with the CRT deflection plates disconnected.

I can take the signal from the inputs and outputs of the TL-082 OP-AMP and no retrace lines whatsoever.

So it appears like the deflection transistor stage is adding the retrace lines, although I have no clue why.

Unless it has to do with the emitter resistores going to the -5 volt supply, although the negative 5 volt supply has a 100uF 16 volt cap to ground to provide filtering.


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: May Wed 09, 2018 11:56 am 
Member
User avatar

Joined: Jul Mon 26, 2010 8:30 pm
Posts: 26014
Location: Annapolis, MD
I am confused...
If you are simply using two sources to drive the X and Y plates, you will get all of the signal from both. To have any kind of blanking, another circuit needs to look at one or both of the signals to decide when to shut off the beam.

In a scope, the blanking signal is generated in the sweep / triggering circuits. The blanking action is independent of whatever signal is going through the vertical amplifier.

_________________
-Mark
"Measure voltage, but THINK current." --anon.


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: May Wed 09, 2018 12:06 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20037
Location: Warner Robins, GA
Agreed, but if the vector signal moves the beam fast enough between the areas on the scren that are meant to be seen, retrace lines will be either somewhat visible to very visible on scopes with low bandwidth or barely visible to not visible at all on scopes with high bandwidth.

Like the Tektronix 2246 scope I used as a test display.

I can display an oscilloscope clock without using any sort of blanking and it displays perfectly with either no or barely visible retrace lines.

The retrace lines I get when I display the clock taking the signal from the deflection transistor collectors using the same 2246 scope as a test display are about as bright as the lines that are intended to be drawn.

Also I have another solid state scope using the 5UP1 and without blanking, the scope clock has barely visible retrace lines so it has to be something with this circuit.

Granted that circuit was meant for driving 3" CRTs though and for a 5" CRT I had to lower the collector and emitter resistor values so as not to affect the geometry any.


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: May Wed 09, 2018 12:47 pm 
Member
User avatar

Joined: Jul Mon 26, 2010 8:30 pm
Posts: 26014
Location: Annapolis, MD
So...you are referring to a blanking effect that is associated with how fast the beam is moving. For a given beam current, the amount of light emitted by a specific point in the screen is directly proportional to the time the beam spends in that location.

The contrast between "scan" and "retrace" will be directly proportional to the times for each. Assuming that you drive the circuit with a perfect square wave, then the retrace brightness is dependent on the the slew rate of the driver circuit. That, in turn, is determined by output impedance and the capacitance of the deflection plates.

_________________
-Mark
"Measure voltage, but THINK current." --anon.


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: May Wed 09, 2018 1:23 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20037
Location: Warner Robins, GA
Yes that is the blanking effect I am referring to.

That said if the slew rate of the deflection transistor stage wasn't fast enough wouldn't I also get a clock display with geometry issues?

Other than retrace lines the clock display seems to have perfect geometry.

The deflection plate capacitance is why I had to lower the collector and emitter resistor values to what they are now as the original values of 68K and 4.7K were causing geometry issues and the displayed image was not right.

If I look at the scope I have that uses the 5UP1 its deflection plate drve transistor resistors are 10K collector and 390 ohms emitter.

If I do that with this circuit I no longer have enough output to fill the CRT screen with the image and would either need to increase the B+ and B- to the OP-AMP or add another amplification stage.

That said I might just give those values a try and see what I get while measuring if the OP-AMP output or the deflection transistors distort first.

If it would help I can measure at what frequency the output from the deflection circuit starts to decrease in amplitude at the output of the OP-AMP and also the deflection transistor stage.


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: May Wed 09, 2018 1:59 pm 
Member

Joined: Jan Mon 28, 2013 9:35 pm
Posts: 1444
Location: Santa Rosa, CA
Part of the problem is with the CRT you are using. The 5UP1 has relatively
large deflection plates to lower the volatge needed to move the beam.
This means higher capacitance which reduces usable bandwidth.
CRTs used in higher bandwidth scopes use CRTs with small plates
and their amps are compensated to match.

Ran into this myself when building a CRT clock. The high capacitance of the
deflection plates is distorting the nice, sharp vector waveforms.
Steve

_________________
'cell phones and the internet are tools, not a lifestyle'


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: May Wed 09, 2018 3:05 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20037
Location: Warner Robins, GA
I knew to expect some retrace lines given the experience with the other scope using this CRT, but the retrace lines weren't nearly as bad as they are here as with the other scope I could make the retrace lines barely visible with a green filter, but cannot using this circuit.

Also the displayed image is not distorted in any way and has near perfect geometry which was achieved after adjusting the deflection transistor collector and emitter values.

Also that doesn't explain why I can probe the outputs of the deflection amp using proeprly compensated 10X probes even with the deflection plates not conencted and see the same exact retrace lines on the Tektronix 2246 scope as I do on the 5UP1, whereas connecting the clock board output straight to the scope's 10X probes or conencting the 10X probes after the OP-AMP produces no retrace lines at all or they are so faint that I cannot see them due to the much higher bandwidth of that scope.

I did discover that the OP-AMP does indeed go into distortion before the deflection transistors do so it could be possible that I try the 10K collector resistors and the 390 ohm emitter resistors and either increase the B+ and B- to the OP-AMP or add an amplification stage.



EDIT:

Here's a link to the schematic for the scope that uses this CRT

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

Perhaps there's some compensation circuitry for the CRT?


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: May Wed 09, 2018 9:39 pm 
Member
User avatar

Joined: Aug Wed 26, 2015 9:41 pm
Posts: 1188
Location: Baton Rouge, Louisiana
You may need to utilize the Z axis. perhaps provide a blanking amplifier.

From the Dutchtronix FAQ:
Intensity control is present with both positive (Z) and negative (ZINV) drive levels.

Image quality varies from scope to scope of course. X-Y monitors, which are like oscilloscopes but without the time-base section, work best. All pictures showing the Dutchtronix AVR Oscilloscope clock with V4.0 firmware were taken on a Tektronix 608 X-Y monitor using the Intensity Control output ('Z') available on the board.

Q: What is the purpose of the ‘Z’ pin on the AVR clock?

A: The Z (Intensity) output on the AVR clock dims the beam at the appropriate moments. This output will decrease its voltage towards 5V to decrease the scope intensity. Please connect this signal to your oscilloscope intensity control input, if present. It is very effective in suppressing highlighted dots on the screen.

Q: What is the purpose of the ‘ZINV’ pin on the AVR clock?

A: The ZINV (Inverted Intensity) output on the AVR clock dims the beam at the appropriate moments. This output will increase its voltage towards 5V to decrease the scope intensity (Tektronix 400 scopes convention). Please connect this signal to your oscilloscope intensity control input, if present. It is very effective in suppressing highlighted dots on the screen.

_________________
Mark


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: May Thu 10, 2018 1:22 am 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20037
Location: Warner Robins, GA
The CRT driver circuit does have a blanking circuit, but ot doesn't get rid of the retrace lines with the circuit as is.

Somehow the deflection amp transistors are making the retrace lines clearly visible and I don't think the blamking circuit was designed to handle that.

Covering all the bases the position controls are connected to the circuit using shielded wire.

Could that cause the problem?


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: May Thu 10, 2018 4:27 am 
Member

Joined: Jan Thu 01, 1970 1:00 am
Posts: 2284
Location: Saskatoon
The way the blanking circuit works is that the blanking signal is asserted to turn off the beam. Then the deflection circuit moves the electron beam to the new position. So there's no visible trace when this happens. Then, when the electron beam is correctly positioned, the blanking signal is de-asserted, and the trace becomes visible again. These signals are generated by the scope clock microprocessor, and it's possible that if it doesn't have sufficient delay between changing the state of the blanking signal and moving the beam, your external driver circuit may not react fast enough, and you may see part of a trace that should not be visible and you may also miss part of a trace which should be visible.

Your driver circuit output is taken from the transistor collectors, so that your output impedance is essentially equal to the collector resistors: 22k which is quite high if you want a fast slew rate. You may have to add a low impedance buffer to the output in order to get the required slew rate.


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: May Thu 10, 2018 5:20 am 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20037
Location: Warner Robins, GA
BobWeaver wrote:
Your driver circuit output is taken from the transistor collectors, so that your output impedance is essentially equal to the collector resistors: 22k which is quite high if you want a fast slew rate. You may have to add a low impedance buffer to the output in order to get the required slew rate.


The board was designed for 3" CRTs with the Dutchtronix clock and with the original values of 68K and 4.7K I believe it would have displayed the clock properly.

The issue I have is that of retrace lines being present on an external scope that is of sufficient bandwidth to where no blanking circuit is really needed for a proper display with no retrace lines.

I get a display with retrace lines only from the deflection transistor collectors and it doesn't matter if the deflection plates are connected or not. I am completely puzzled by that because there should be no retrace lines and if the slew rate wasn't good enough the retrace lines would be the least of my worries as the display wouldn't look right, but aside from the retrace lines the display looks just as good as it did with the original scope circuitry driving the CRT.

I've tried experimenting with the blanking, but no matter the signal level the circuit feeds to the CRT grid 1 it still will not properly blank the CRT which i think somehow has to do with the retrace lines that are somehow being produced by the deflection transistor stage.

Has to be that stage as the output from the OP-AMP is clean and produces a display so good that I can connect the vertical or horizontal to the OP-AMP output and the other to the OP-AMP input and get the same good display on the 2246 scope of course adjusting the scope for the signal difference of one channel.

I know that I am missing something here, but have no clue what.

It would help if I could take pictures, but I am working on the clock at work when I have some free time due to better test equipment and soldering tools than what i have at home and cameras aren't allowed.

The best I can do is this.

Picture this display only with retrace lines that are more visible between each line forming the round clock face.

http://s863.photobucket.com/user/jmcinv ... ort=3&o=10

A bit about that picture. Think I had the intensity control up a bit high to specifically show the retrace lines as they were normally only slightly visible with a normal intensity setting and barely visible with the green filter in front of the CRT.

This is a picture of how the clock should look.

http://s863.photobucket.com/user/jmcinv ... ort=3&o=11

Think I did use retrace blanking in that version of the scope clock, but the clock displayed on the 2246 scope looks the exact same without retrace blanking when the signal is taken from the clock board, OP-AMP input or OP-AMP output, but not from the deflection transistor collectors.

That said I'm not worried about the retrace blanking just yet as I first have to get the other problem solved.

Once I get it to where I get a perfect display on the 2246 from the deflection amp collectors then the retrace blanking should be easy.


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: May Thu 10, 2018 8:08 am 
Member

Joined: Jan Thu 01, 1970 1:00 am
Posts: 2284
Location: Saskatoon
Yeah, those photos look pretty much as I expected.

The TL082A op amp has a slew rate of about 13 V/µs. So, if you need to switch the output from -5 to +5, that's a 10 V swing and with the spec'd slew rate it will take 10/13 = 0.77 µs, assuming that your input signal switches instantly. Now, if you want the transistor deflection driver to swing full scale in the same amount of time, it needs to switch from 0 to 180 volts in 0.77 µs, which gives a required slew rate of 180/0.77 = 234 V/µs. Much higher than that of the op amp. If you want to swing that much voltage in that short a time, you have to start looking at all of the parasitic elements in your circuit. How much inductance is in the power leads to the transistor collectors for example? How much stray capacitance at the collectors even if you ignore the CRT deflection plates?


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: May Thu 10, 2018 11:14 am 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20037
Location: Warner Robins, GA
So even though the display geometry looks perfect aside from the retrace lines which are greater in intensity than what should be there including being there on a scope that should show no retrace lines I could have an issue with stray inductance and/or capacitance?

Only thing I can think of is the shielded wire running to the position pots, but in both deflecton amps I have the base of the transistor where the center wiper of the control connects bypassed to ground with a .1uF cap which made the calibration square the clock has look better as the left side without the caps in circuit had a very small gap between the left vertical line and the top and bottom horizontal lines before the caps were added.

I do know the OP-AMP has a bit too much gain as the size pots only have to be turned 1/4 of the way to fill the CRT.

That said I can adjust the size pot on the clock board to compensate which would also affect the deflection circuit given the dc coupling of the clock board and deflection circuit.


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: May Thu 10, 2018 11:46 am 
Member
User avatar

Joined: Jul Mon 26, 2010 8:30 pm
Posts: 26014
Location: Annapolis, MD
I'm under-sampling this thread, so apologies if I missed something.....

If I understand correctly, what you are building does not have blanking like what a scope has. Rather, you are depending on the timing of the drive waveform. If that's correct, then I don't understand the significance of how a scope displays the signals. Your display will have a retrace brightness determined by the relative timing of the drive signal. Let's assume the behavior is linear---if so, the retrace time might have to be at least a factor of 100 less than the scan time for the retrace to be not visible.

You can of course use your scope to see exactly what timing you are getting with your circuit.

I intuitively agree that driving the plates with a source impedance of 22Kohms might be way too high.

_________________
-Mark
"Measure voltage, but THINK current." --anon.


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: May Thu 10, 2018 2:39 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20037
Location: Warner Robins, GA
The circuit does have a Z axis and I decided not to mess with getting that right until I solve the other problem.

It could be that 22K is too high, but I don't think that would explain why I am getting retrace lines on the Tektronix 2246 100MHz scope used as a test display when I connect it using 10X probes to the deflection plate drive transistors even with the deflection plates disconnected when I get no retrace lines at all when the clock is directly fed to the scope or fed from the OP-AMP outputs to the scope.

It's something about that transistor stage which is causing retrace lines to appear on the Tektronix scope.

Unforetunately I don't think I can go to 10K collector resistors unless I use a different transistor or further reduce B+

The original scope circuit I believe ran 100 volts on the collectors of its horizontal deflection output transistors with a 10K 7 watt collector resistor and a 390 ohm emitter resistor for each transistor.

Now given how the position control works which varies the voltage on a transistor base between -5 and +5 volts I should be able to disconnect the bases of both transistors and see if that clears up the retrace lines. If so then I know the shielded cable is the cause. If not then it is something else.



EDIT:

Not intending to sound like I'm stubborn at all.


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: May Thu 10, 2018 6:06 pm 
Member
User avatar

Joined: Jul Mon 26, 2010 8:30 pm
Posts: 26014
Location: Annapolis, MD
If you disconnect the bases of the transistors, how would the circuit drive the deflection plates?

How about calculating the allowable capacitance?
First, what slew time is required? Maybe 1% of the signal?
Suppose you need RC = 0.1 microsecond. If R is 20K, that's an allowable C of 5 pF
1 microsecond....50 pF

What's the collector capacitance on those transistors?

_________________
-Mark
"Measure voltage, but THINK current." --anon.


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: May Thu 10, 2018 6:11 pm 
Member
User avatar

Joined: Jul Mon 26, 2010 8:30 pm
Posts: 26014
Location: Annapolis, MD
For that tube, the average deflection plate capacitance is just under 10pF

_________________
-Mark
"Measure voltage, but THINK current." --anon.


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: May Thu 10, 2018 6:18 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20037
Location: Warner Robins, GA
Not sure, but it is supposed to work well using the original 68K collector resistors, 4.7K emitter resistors and a 3" CRT such as a 3RP1 or other 3" CRTs.

That said I do know the output of both deflection transistors for vertical or horizontal aren't exactly equal which is due to the way the signal for the other transistor is derived.

So it is in effect using one deflection plate to alter the beam position along with both plates receiving the deflection signal.

Part of me wants to take two dual gang 4.7 meg pots and wire them up to alter the DC voltage on both deflection plates where as one plate DC voltage increases the other decreases which will provide a much better control over position and will also eliminate the issue of when moving the position control beyond a certain point it starts to distort the signal.

I modded a DuMont 292 scope using the 4.7 meg pots for the position controls as it just used a single gang pot for that and didn't work so well and that worked quite good.

I could then rework the circuit so that I have a true push pull circuit.



So with the deflection capacitance just under 10pF how would that affect the circuit?

EDIT:

Because I wasn't sure what to try next I used the variable cap used to compensate the vertical output in the scopes original circuit and connected it across the 1K resistor in the horizontal deflection transistor circuit and saw where it affected things.

So apparently I need compensation caps to get the display perfect.

I do have a second vertcal board so I can get an identical variable cap.

Also there's some slight swimming of the image, but if I reduce the clock boards gain pot to be lower and turn up the vertical and horizontal size pots of the driver board the image has an oscillation to it as though there's a grounding issue.

So somewhere I don't have a ground in the exact right place.


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: May Thu 10, 2018 8:48 pm 
Member
User avatar

Joined: Jul Mon 26, 2010 8:30 pm
Posts: 26014
Location: Annapolis, MD
there are several questions bubbling along in parallel.

I'm focused only on requirements:
Assuming that the visible retrace is related to slew rate, the challenge is to quantify the required value. If we assume that time and beam intensity are interchangeable, then you could do some experiments to see how much beam current reduction is needed to get acceptably low intensity, Then convert that to a dwell time ratio.
Example:
Suppose you can get good contrast with a beam current reduction of 50:1. Apply that factor to the time spent tracing out the image---if that's 100 microseconds, then we could guess that retrace needs to happen in 2 microseconds. From that, calculate allowable capacitance. (time constant = R*C ---gets us only in the ballpark--we need the time constant to be a fraction of the retrace time)

Unrelated: That tube is described as having "medium" persistence---could that be relevant?

_________________
-Mark
"Measure voltage, but THINK current." --anon.


Top
 Profile  
 
 Post subject: Re: CRT circuit question
PostPosted: May Thu 10, 2018 10:09 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20037
Location: Warner Robins, GA
Persistence does matter for sure.

When using the original scope circuit I got a better display with a little less visible retrace lines.

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.


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

All times are UTC [ DST ]


Who is online

Users browsing this forum: No registered users and 3 guests



Search for:
Jump to:  




























Privacy Policy :: Powered by phpBB