Searching \ for '[EE]: Fun with nested potential dividers' in subject line. ()
Help us get a faster server
FAQ page: massmind.org/techref/method/math.htm?key=divide
Search entire site for: 'Fun with nested potential dividers'.

Exact match. Not showing close matches.
'[EE]: Fun with nested potential dividers'
2005\10\20@183458 by

Hi,
I'm trying to tweak the resistor values in an intensity/focus chain for a
CRT. The circuit ATM looks like this ('scuse the ascii art):

47k     20k          82k     100k      940k (2x470k)
-600V <<<---/\/\/---/\/\/---*----/\/\/---/\/\/------/\/\/-----> 0V
^             |              ^
|             |              |
|             |              |
|             |              |
Vg1 (inten)   Vk (cathode)   Vg3 (focus)

Now, this looks to me like two potential dividers, which combine to make one
potential divider and two "sub-dividers" if that makes any sense. If I change
the total resistance on one side, the other side will be affected and vice
versa. So how do I pick resistor values for this thing?

Here's how I simplified/understood it:

67k           1122k
-600V <<<---/\/\/----*----/\/\/-----> 0V
|
|
Vk

At the moment, this circuit's outputs are (assuming my maths is correct):
Vg1: -600V to -574V
Vk:  -564V fixed
Vg3: -522V to -492V

What I need is:
Vg1: -600V to -550V
Vk:  -500V fixed
Vg3: -500V to -380V

So how should I go about modifying this circuit? So far this is the best way
I've come up with:

1- Work out the values for the simplified circuit, so that Vk = -500V for
R1+R2=1Meg2. New R1=x, new R2=y.
2- Split the circuit down into two dividers.
3- Calculate the component values required for the Vg1 divider, with R1+R2=x
4- Calculate the component values required for the Vg3 divider, with R1+R2=y

Does this sound like it'll work?

Thanks,
--
Phil.                              | Acorn RiscPC600 SA220 64MB+6GB 100baseT
philpemphilpem.me.uk              | Athlon64 3200+ A8VDeluxe R2 512MB+100GB
http://www.philpem.me.uk/          | Panasonic CF-25 Mk.2 Toughbook
Beware of programmers who carry screwdrivers.
Not so sure if this is the preferred way to do thing but anyway
I will put forward the proposition to use a circuit simulator.

Throw the schematics into PSPICE (the student version will do) and
do a parametric sweep.

For this particular circuit, hand calculation is still possible.
However if we put some more resistors in series and in parallel,
the PSpice or other tools will be easier.

If this were a low voltage application, it would have been possible to
use some potentiometers to do the simulation.

Regards,
Xiaofan

{Original Message removed}

Nothing says he has to use 600V to play with pots. Just scale
everything by factors of 10. 100x lower voltage, 10x lower R
with 10x lower target voltages to measure with DVM.

Don't forget to account for the currents drawn by the
CRT. Notably Cathode current as a function of intensity.

The usual way to hand solve this is to compute the current,
600V/(47k+20k+82k+100k+940k) then multiply I by each R
to find their individual drops, and then sum them down the chain.

Robert

Chen Xiao Fan wrote:

{Quote hidden}

> {Original Message removed}

>Throw the schematics into PSPICE (the student version
>will do) and do a parametric sweep.

Use the Linear technology LTSpice/Switchercad III package. It is a fully
fledged spice implementation that is fast and free, with no limitations.
There is also a yahoo support group which is independent of Linear
Technology, and seems to be populated by some pretty clued up people
(including one or two from this list I notice). Bits from the yahoo group
have been incorporated into the help file of LTSpice. (Linear technology
would like to drop the Switchercad III part of the name, but it is so well
known that they have to keep it going was the impression I got).

I went to a presentation that Linear Technology did here in the UK last
week, where the guy who wrote and maintains LTSpice did a presentation, and
came away real impressed with its capabilities.You can download a copy of
the power point presentation and all the support files from
http://ltspice.linear.com/software/handout.zip It is just under 8MB in size.

Alan B. Pearce said:
> Oh setting out to play with the DG7-32 are we?

Yes indeedy... Bought one to build a scope clock, then noticed the voltage levels for the focus chain I built were so far wrong the spot on the tube was unfocussable and completely lacked any form of brightness control. Thankfully the "unfocussable" aspect stopped the phosphor getting fried when the brightness hit maximum.

> of articles back in the 1960s/70s. They used a cathode resistor to the VG1
> voltage, so the cathode current dropped enough voltage across the resistor
> to generate the required voltage difference between the two electrodes, just
> like a normal valve circuit biasing arrangement where the grid is taken to
> ground through a resistor, and the cathode also has a resistor that causes
> it to be +ve wrt ground. This gave what they described as "automatic
> brightness control". Unfortunately the magazines are half way around the
> world from me, so I cannot just go and look it up.

Drat. Anyone else got copies of the aforementioned articles?

> This series of articles is how I knew about the DG7-32 and its voltages when
> you enquired the other day.

Heh. Guess I'm not the only person that likes playing with old technology then.
Apparently the DG7-32 was used in the Mullard "Serviceman's Oscilloscope" and a few Cossor, Philips, etc. scopes in the 1950s/1960s. Oh, and the Marantz 10B - apparently 10B owners are buying up DG7-32s as spares and causing prices to go through the roof... I paid £19 for my DG7-32 and a socket, which seemed fair enough, so I'm not grumbling.

What is annoying is the fact that I miscalculated the size of the transformer I needed to boost the 12V supply to 230V for the HT multiplier and rectifiers. I bought one of those cheap 100mA mains transformers and it bottomed out as soon as I hooked up the CRT :(

Thanks,
--
Phil.
philpemdsl.pipex.com
http://www.philpem.me.uk/

-----------------------------------------
This message sent by OakMailer, a free/donation-ware script
provided by Oakbox Productions http://www.oakbox.com.

Anti-spam notice, this message originated from IP Address:
161.112.232.17
-----------------------------------------

2005\10\21@082158 by
>Drat. Anyone else got copies of the aforementioned articles?

I'll see if I can persuade my father to find them, and do digital photos of
the pages.

> Drat. Anyone else got copies of the aforementioned articles?
>
>> This series of articles is how I knew about the DG7-32 and its
>> voltages when
>> you enquired the other day.

If you can quote exact magazine, date & volume or whatever I MAY have
them.

RM

>> This series of articles is how I knew about the DG7-32 and its
>> voltages when
>> you enquired the other day.

>If you can quote exact magazine, date & volume or whatever
>I MAY have them.

I don't have the dates, except that it would be somewhere during the 60's,
but possibly into the 70's.

In message <291f01c5d63a\$00ee9490\$e7bdf682space.rl.ac.uk>
"Alan B. Pearce" <A.B.Pearcerl.ac.uk> wrote:

> I'll see if I can persuade my father to find them, and do digital photos of
> the pages.

That's great - thanks. I've got the cathode and focus grid supplies working
fine now, but for some reason the brightness control doesn't have much of a
span. I suspect it'll work a lot better once I build up the deflection
controls though, but it would have been nice to have the brightness variable
over a wider range...

I guess I still need to tweak the voltages a bit. The datasheets are great
for ideas on what voltages are needed, but utterly crap for "This is what you
need to set Vg1 to in order to get rid of the spot" type stuff.

It does, however, work, which is nice. I suspect I just need to fiddle with
the resistor chain a little more. Maybe move the 100k to the other side of
the Intensity pot, put a 47k on each side or maybe eliminate the 100k and use
a 220k pot instead. Here are the values I'm using now:

100k    100k                220k     880k (680k+100k+100k)
-600V <<<---/\/\/---/\/\/---*------------/\/\/---/\/\/----------------> 0V
^             |              ^
|             |              |
|             |              |
|             |              |
Vg1 (inten)   Vk (cathode)   Vg3 (focus)

I also tried adding a 47k pot between +250 and GND with the wiper driving the
CRT anode (to allow for astigmatism control), but that loaded down the HT too
much. Guess I need to get some more potentiometers, or add a parallel
resistor to tweak the ones that I've got...

Thanks,
--
Phil.                              | Acorn RiscPC600 SA220 64MB+6GB 100baseT
philpemphilpem.me.uk              | Athlon64 3200+ A8VDeluxe R2 512MB+100GB
http://www.philpem.me.uk/          | Panasonic CF-25 Mk.2 Toughbook
Tagline? What tagline, Officer? I didn't see no tagline.

On the piclist, talking about biasing tubes... :)

Gotta love it!

This is a pretty typical approach. My guess is that the original
calculations were done with approximate values, then hand trimmed.
The calculation intensive approach would have been a real PITA with just
paper and pencil.

Don't forget to drop in some caps so that the beam current variations
don't visibly wobble the brightness and focus.

On Thu, 20 Oct 2005, Philip Pemberton wrote:

>             47k     20k          82k     100k      940k (2x470k)
> -600V <<<---/\/\/---/\/\/---*----/\/\/---/\/\/------/\/\/-----> 0V
>              ^             |              ^
>              |             |              |
>              |             |              |
>              |             |              |
>          Vg1 (inten)   Vk (cathode)   Vg3 (focus)

Make an assumption about the maximum tube current, say 100uA, and choose
a current through the divider of 10x this, e.g. 1mA. That would be
600kOhms total. Then calculate the stage resistors (easy, 1kOhm/volt).
Use several resistors such that none see more than 50V across them (for
0.25W usual resistors). Alternately use oversized 1W resistors which can
take the voltage.

Peter

More... (looser matching)
- Last day of these posts
- In 2005 , 2006 only
- Today
- New search...