Of transistor amplifiers and CRTs...

Simply design as if hfe was much lower. That way, the influence of bias current will be minimal.

Graham

It looks like the big problem I've got is that the voltage input spans a range of 0V to 4V relative to ground, while the transistor ("Q") emitter is going to be about 500V below ground. Putting a capacitor between Q.base and Vin and adding a resistor between Q.base and Q.emitter should get rid of the "offset", but if the signal goes DC (which is quite likely) then the beam will turn off. An opamp would be one way to deal with the offset, but I expect opamps that can handle 600V are few and far between (not to mention expensive).

Later, Phil.

I don't quite understand what you are asking for: What voltages do you need on the CRT electrode when the beam is on, and what voltage do you need when the beam is off? If you could post a drawing of the circuit on a website somewhere then that would be helpful too.

Chris

OK, the voltage at the transistor emitter will be about -500V relative to ground. The collector will be at about -400V.

The input ranges from 0V to 4V relative to ground.

This is fine - in theory. Except the transistor's base is referenced to the emitter. There's 500V or so between base and emitter, so it's locked on hard. What I need is for the transistor to sink between 0 and 200uA as the input varies between 0 and 4V.

Later.

--
Phil.                              | Acorn RiscPC600 SA220 64MB+6GB 100baseT
philpem@philpem.me.uk              | Athlon64 3200+ A8VDeluxe R2 512MB+100GB
http://www.philpem.me.uk/          | Panasonic CF-25 Mk.2 Toughbook
No software patents!           /  
... After we pull the pin, Mr. Grenade is NOT our friend!

Philip Pemberton wrote: ...

...

This is a classic problem with driving electrostatic deflection CRTs. Some CRTs even had a separate blanking electrode that could be at anode potential to avoid the issue.

What bandwidth do you require?

AC coupling with DC restoration is a common approach but does not work at DC. You may be able to arrange that you periodically blank the beam even when you are displaying very slow moving traces - you probably won't notice it. Analog scopes with on screen displays periodically steal the beam to update the character display.

A variation on this approach is to combine it with a chopped version of the modulating signal that is coupled by a small capacitor and rectified to provide the low frequency component of the signal.

Another method is to use an optocoupler although this has speed limitations and requires a 50V or so power supply referenced to the cathode of the CRT. Obtaining linear response may also be difficult.

Yet another method is to use a high-voltage PNP transistor with it's current modulated to control the blanking signal - this also requires a power supply referenced to the cathode.

You can see a couple of examples at these links:

kevin

A few hundred kHz at most.

Looks like that's the best plan. Even if I make a scope add-on for the display, I guess I'm still going to need to shut down the beam during a retrace.

So now I need to work out the biasing... Time to borrow a copy of AoE, methinks.

I haven't got any PNPs though - I ordered NPNs for the deflection circuit, but forgot to order the PNPs for the blanking/brightness circuit.

Thanks.

--
Phil.                              | Acorn RiscPC600 SA220 64MB+6GB 100baseT
philpem@philpem.me.uk              | Athlon64 3200+ A8VDeluxe R2 512MB+100GB
http://www.philpem.me.uk/          | Panasonic CF-25 Mk.2 Toughbook
No software patents!           /  
... Biography: One of the terrors of death.

I still don't understand - it sounds like you have already chosen a circuit and that it cannot work because transistors can't work with 500V between base and emitter. What I was asking about is the overall function that your circuit is supposed to perform, in case I could suggest a different way of doing it which would be easier to bias. I am also still not really sure what your circuit looks like so if you could post a diagram on a website or in ABSE then I would probably be able to give you a more useful answer.

Chris

Yup, the OP still hasnt made clear what V and swing are needed on what electrodes. ac coupling and clamping is a common approach.

There is one more possible way to do it which gets dc voltages displayed ok. compare the probe's V_in with a hf sawtooth wave, so you get a digitised output, then use this to switch the V on the electrode. This will give full dc response. You can either smooth the hf before feeding to the electrode, or feed chopped as is.

No, I havent done this with CRTs, so someone else might spot a problem.

NT

Right, I've dragged it through SPICE and got some voltage figures. There IS a schematic online at - that just covers the resistor chain though.

Basically, when the Z amplifier gets a +4V input relative to 0V, I want the equivalent of a 170k resistor placed between "G1" and "X". That provides a

200uA (ish) path, and pulls down the cathode, thus increasing the brightness of the trace. Brightness is determined by the voltage difference between the cathode (which is wired to the 33k/680k junction to the right of "X") and G1. The lower G1 is relative to the cathode, the dimmer the trace gets.

OK, so for "normal" brightness with the 100K "Intensity" control at mid-position and no resistor over G1 and X: V(x) = -509.00V V(g1) = -554.40V V(ca) = -479.04V Diff1 = 45.40V Diff2 = 29.96V And with the 170k across X and G1: V(x) = -517.92V V(g1) = -553.67V V(ca) = -487.31V Diff1 = 35.75V Diff2 = 30.61V

V(ca): Cathode voltage relative to 0V V(x): Volts at "x" relative to 0V V(g1): Volts at "G1" relative to 0V Diff1: Difference between V(x) and V(g1) Diff2: Difference between V(x) and V(ca)

This doesn't take into account the effect the CRT has on the voltages - it's just a simulation of the resistor chain.

A voltage difference of about 32V is enough to get a full-brightness trace without causing the focus to go wildly out. About 55V is enough to shut down the beam.

I've been using a transistor current source, as shown in AoE (emitter resistor controls the current, NPN transistor - from page 74) with the "ground" end of the emitter resistor wired to G1 and the collector wired to X. The catch is, because the emitter is at -500V relative to ground, feeding it a 0V input ends up tripping the PSU's output breaker. Feeding it either X or G1 as an input has the desired action - either increasing or decreasing the brightness.

Circuit as follows:

X >--------C E---[ Re ]----> G1 \ / ======= | ZTX458 (NPN Si HV) | ^ 0-4V Z axis input

So far, the best idea I've come up with is to add a resistor between the base and emitter and feed the input in through a capacitor. I'm not even sure if that'll work, though...

The current passing between G1 and X needs to vary between basically nothing (0V input) and the 200uA maximum (4V input). Ideally the 200uA figure needs to be adjustable over a range of 50uA or so.

Later.

--
Phil.                              | Acorn RiscPC600 SA220 64MB+6GB 100baseT
philpem@philpem.me.uk              | Athlon64 3200+ A8VDeluxe R2 512MB+100GB
http://www.philpem.me.uk/          | Panasonic CF-25 Mk.2 Toughbook
No software patents!           /  
... Our OS which art in CPU - RISC OS be thy name .....

That's a blank link. You were going to step way back and let us see the big picture, tell us what you're trying to do? It's the forest we need to see, not all the trees.

--
 Thanks,
    - Win

Hopefully MUCH later...get lost, sick TROLL.

You are falling victim to the TROLL technique- where they are deliberately vague and pretend to suffer from an always elementary confusion that cannot be altered by the well-intentioned response. The hapless retard working on the clock, the mystifying amplifier requirements, vague and partial circuit descriptions, the pretense of sincerity, the willingness to respond with even more prolific confusion and so forth. The best possible outcome is that he will act out his schizoid alter personality and electrocute himself.

I should really check links before I post them.. The working one is , which is the diagram of the resistor chain that generates the bias voltages for a DG7/32 CRT.

What I'm trying to do is take an analogue signal between 0V and 4V (where 0V is full-blank, 0.5V is "black", and 4V is "peak white") and use that to modulate the brightness of the CRT.

Generally, that's done by varying the difference between the cathode and G1 voltage. If Vg1 goes below Vcathode by about 160V, the beam is completely blanked. If it's set to Vcathode, you get a horrendously defocussed beam that (if focussed down properly) would be bright enough to burn the phosphor. Vg1 = Vcathode-70V gives maximum beam output without fouling up the focus.

So what I need to do is load down the output of the resistor chain by an amount that varies dependent on the control voltage (0V to 4V).

I've had a look at transistor current sources, which appear to be ideal for the task. Problem is, the control voltage is referenced to 0V, so if I wire it up to the resistor chain I end up with about 500V between base and emitter (or collector, depending on how it's wired). Needless to say, the transistors don't like that...

It seems the "industry standard" way of dealing with this is to put a huge resistor (about 10 megohms) in the path between the resistor chain and the CRT grid (G1), then use a 0.1uF (or 0.01uF) capacitor between Z-in and G1. I'm just trying to figure out how best to do this...

Thanks.

--
Phil.                              | Acorn RiscPC600 SA220 64MB+6GB 100baseT
philpem@philpem.me.uk              | Athlon64 3200+ A8VDeluxe R2 512MB+100GB
http://www.philpem.me.uk/          | Panasonic CF-25 Mk.2 Toughbook
No software patents!           /  
... Cookie pieces contain NO calories, breakage leaks calories.

Here is my suggestion:

-600V -478.9V -327V -127.4V -18.2V | Bright Ca ? ? ? 0V | 133k? | 680k | 220k | 120k | 20k | .-\/\-.-\/\/-..-\/\/\/-..-\/\/\/---\/\/\/---\/\/\/--- | | ^ | 220k | | | | .-\/\/\/--. | .--. ^ 680k |0V (Gnd) | |G3 .-\/\/--. | | |B (300V PNP) | |G1 | ----^--- Contrast | 100k | 560k | C / \ E 22k? .---/\/\/\-.--/\/\/\----.----/ \---/\/\/\-/\/\/\-.---logic(0-4V) 6.8k ^ | Logic 0: G1=555.2V | | .---. My suggestion is basically to keep the cathode bias voltage about constant (assuming beam current in the cathode is fairly small), and to modulate the G1 voltage. A 300V transistor should do, these can be obtained in TO-92 package with adequate power rating I think. The transistor is operated common-base. I can't be bothered simulating or calculating to find the exact resistor values you need for the correct brightness and contrast but you can do that. If the high resistor values make the circuit driving G1 too slow for your needs, you could reduce all of the resistor values but this will increase dissipation. You could also try putting a tiny cap across the 560k resistor but I would try without first. Some spark gaps might make the circuit a bit less prone to damage.

Chris

Looks promising, Chris - I'll have a play around once I get some HV PNPs.

Thanks.

--
Phil.                              | Acorn RiscPC600 SA220 64MB+6GB 100baseT
philpem@philpem.me.uk              | Athlon64 3200+ A8VDeluxe R2 512MB+100GB
http://www.philpem.me.uk/          | Panasonic CF-25 Mk.2 Toughbook
No software patents!           /  
... This Tagline is for sale.  Call 1-800-TAG-THIS!