Driver to drive?

[legg]

On Sat, 2 May 2009 17:15:50 -0700 (PDT), Mark-T
<MarkTanner50@gmail.com> wrote:

DId anyone here see the problem presented in
the Science section of NY Times last week?
Quite startling, to see something so sophisticated
in a 'general readership' publication.

Is it solvable without a calculus of variations approach?

At least it makes more 'sense' than the duck and the fox.

There's no reason for the duck to leave the pond.

The only problem I see with the elaborate solution is the assumption
that the obviously ill 'killer' rabbit will react in any way to the
presence of suits around the pond's periphery. After all, a rabbit
with any sense wouldn't be in the middle of a pond in the first place.

RL

 

[Eeyore]

Bob Larter wrote:

Eeyore wrote:
John Larkin wrote:
[...]
Here's the box:

ftp://jjlarkin.lmi.net/99A260A1.JPG

ftp://jjlarkin.lmi.net/99A260A3.JPG

ftp://jjlarkin.lmi.net/99S260A.JPG

Seems to work now.

Ah ! Point to point wiring. The toob nuts would be proud of you.
;~)

Well, that's one way to reduce leakage!

Certainly is.

Graham

 

[riverman]

On May 3, 10:17 am, Sylvia Else <syl...@not.at.this.address> wrote:

Mark-T wrote:
DId anyone here see the problem presented in
the Science section of NY Times last week?
Quite startling, to see something so sophisticated
in a 'general readership' publication.

Is it solvable without a calculus of variations approach?

--
Mark

Yes.

Spoiler below 13

0

1

2

3

4

5

6

7

9

10

11

12

13

Since the agent can run four times as fast as the rabbit can swim, no
matter what the agent does, the rabbit can swim away from the centre,
and keep the agent on the far side as long as the rabbit is no more than
1/4 of the radius from the centre. So the rabbit can reach a point 3/4
of a radius from the edge while the agent is still on the opposite side.

Only if the agent stands still until the last possible moment. I think
the agent will move continuously, and as soon as the rabbit starts
moving in any direction, the agent will move toward the point that is
the closest to the rabbit at that moment. Let's call the agent 'Xeno'.

 From that point, if the rabbit swims directly to the edge, he has to
swim 3/4 of a radius. In that time, the agent can run at most 4 times as
far, or 3 * radius. But the distance to the rabbit's exit point is Pi *
radius, or a bit further, so the rabbit is free.

For the second part, instead of 4, write m, and let R be the radius.

The rabbit can reach a point R/m from the centre while keeping the agent
opposite. From that point, the rabbit needs to swim R * (m - 1) / m to
reach the edge. In the available time, the agent can run R * (m - 1). So
the break even point is where R * (m - 1) = R * Pi, or where m = Pi + 1.

Sylvia.

 

[Nobody]

On Mon, 04 May 2009 13:10:03 +0100, Nick wrote:

No - until the rabbit is 1/4 of the radius away from the centre, he
can swim fast enough, in a suitable direction, to keep the agent
exactly opposite. So it doesn't matter if the agent moves.

Which shows, I think (to return to the very original question about
whether this needs horrible maths to solve) that there is a simple proof
that pretty-well anyone can follow that shows the rabbit can escape in
the specific case (it goes to where it can just swim faster than the
agent, swims round in a circle until it gets to 180 degrees away from
the agent, and can then make it to the shore faster than the agent can
run round), but that the optimum strategy, and hence the answer to the
unasked question about which size of ponds or relative speeds allow the
rabbit to escape do require the horrid maths.

Simple algebra is sufficient to determine that the agent's speed needs
to be at least pi+1 times the rabbit's speed.

If the radius is r, the agent's speed is a and the rabbit's speed is b:

#1. The rabbit can remain diametrically opposed to the agent up
to a distance of r*b/a from the centre, i.e. r-r*b/a = r*(1-b/a) from the
shore.

#2. From there, the time it takes for the rabbit to reach the closest
point on the shore is r*(1-b/a)/b = r*(1/b-1/a).

#3. The time it takes the agent to reach that point is pi*r/a.

#4. They will draw when the times in #2 and #3 are equal, i.e.

r*(1/b-1/a) = pi*r/a
=> 1/b-1/a = pi/a
=> a/b-1 = pi
=> a/b = pi+1

If the ratio is greater than pi+1, the rabbit can't win using this
strategy. But proving that this strategy is optimal (i.e. it can't win
using another strategy) may be rather more involved.

 

[Michael A. Terrell]

Bob Larter wrote:

Eeyore wrote:

John Larkin wrote:
[...]
Here's the box:

ftp://jjlarkin.lmi.net/99A260A1.JPG

ftp://jjlarkin.lmi.net/99A260A3.JPG

ftp://jjlarkin.lmi.net/99S260A.JPG

Seems to work now.

Ah ! Point to point wiring. The toob nuts would be proud of you.
;~)

Well, that's one way to reduce leakage!

All Eeyore knows is cheap audio gear.


--
You can't have a sense of humor, if you have no sense!

 

[John Larkin]

On Mon, 04 May 2009 17:45:11 +1000, Bob Larter <bobbylarter@gmail.com>
wrote:

John Larkin wrote:
On Sun, 3 May 2009 17:32:18 -0700 (PDT), Greegor <Greegor47@gmail.com
wrote:

On May 3, 7:23 pm, John Larkin
jjlar...@highNOTlandTHIStechnologyPART.com> wrote:
These crappy RatShack terminal posts are actually conductive!
http://www.radioshack.com/product/index.jsp?productId=2103639

A pair of them leak 12 pA to the chassis at +1 volt. If I ground
myself and hold the plastic screw part of one, it goes up to 20.

So I'll have to replace them with some Pomonas or something. What a
nuisance.

Pity; they do look nice.

John
0.000000012 Amp? LOL

I just finished machining a couple of big slots in the aluminum box,
where the critical binding posts go, and added a polycarb plate that
the terminal strips mount on. Now the leakage is showing about 100 fA,
basically my measurement limit. It was ugly, machining the
already-"finished" electronics. I tried milling it dry, but that
didn't work, so I wound up getting chips and cutting fluid everywhere.
Had to clean all that up.

It's amazing what a few drops of Tapmatic will do.

Here's the box:

ftp://jjlarkin.lmi.net/99A260A1.JPG

What's with the huge bolts sticking out of the front? It'd look a lot
nicer if you'd put the heads on the outside.

That's to hold down the bent-aluminum measurement cover. At pA levels,
just moving around in the vicinity of the box will slam the
middle-node voltage around, and, unshielded, there's lots of 60 Hz
there too. These sorts of measurements don't work in free air.

John

 

[Michael A. Terrell]

Bob Larter wrote:

jcdrisc@melbpc.org.au wrote:
As a regular reader of the electronics postings I am heartily sick of
all the
advertising for clothing, shoes and vanity items. I recall in the old
days these would be
deleted by a moderator. I can only assume Google's advertising
policies are a bit out of control. We have a situation where the
majority of the group is polluted by this garbage.
I would hope something is done.

Not a chance. Google lets spammers sign up for as many free accounts as
they want, so if you get one nuked, they come straight back under
another name. This is why many people filter out all Google Groups posts.
I've tried campaigning to have Google treated the way other
spammer-friendly NSPs are treated, but Google spends too much money for
any of their news-peers to drop them.

Supernews was dropping anything posted to five or more newsgroups
from google groups. Then Giganews bought them and dropped all the
filtering they did.

--
You can't have a sense of humor, if you have no sense!

 

[John Larkin]

On Mon, 04 May 2009 17:47:29 +1000, Bob Larter <bobbylarter@gmail.com>
wrote:

Eeyore wrote:

John Larkin wrote:
[...]
Here's the box:

ftp://jjlarkin.lmi.net/99A260A1.JPG

ftp://jjlarkin.lmi.net/99A260A3.JPG

ftp://jjlarkin.lmi.net/99S260A.JPG

Seems to work now.

Ah ! Point to point wiring. The toob nuts would be proud of you.
;~)

Well, that's one way to reduce leakage!

This circuit has only one really leakage-sensitive node, and it's not
on the terminal strip. Its insulation is mostly air. The only leakage
is on the body of the opamp (National claims 10 fA typ) and the
polycarb thing I added to the front panel.

Unfortunately, on the LMC6001 the ni input is pin 3, and pin 4 is -5V.

John

 

[John Larkin]

On 4 May 2009 09:13:37 GMT, Jasen Betts <jasen@xnet.co.nz> wrote:

On 2009-05-04, James Rollins <james.rollines@gmail.com> wrote:

You may as well use a single series fet+resistor, pwm controlled, and
one output cap. Or just use a linear regulator. All are about equally
efficient.

John


If you don't disconnect the load there will be a voltage spike and
also a voltage dividing effect. i.e., the load is connected to the
voltage source directly and depending on the ratio of the two
resistors the load will generally see a higher voltage than you might
want. Hence the reason why the load is disconnected. If the voltage
supply resistance is too small then it requires faster rise/fall times
to quickly stop charging the capacitor as it will end up charging up
much faster.

your circuit avoids neither of those problems.

The same problem still exists with your idea though. The fet's gate
will need to be held at a voltage that is within the source/drain of
usually around +-20V max. But if the fet is sorta "floating" I am
unsure how to accomplish such a task. At least in a way that keeps
with the simplicity of the design.

I've thought about using a simple linear regulator idea but
unfortunately similar problems as well as other problems exist. Mainly
in this case the regulation seems much poorer and it is less
efficient. Although I'm not too interested in efficiency as I am
regulation. Trying to find cost effective HV bjt's is a bit of a
problem as compared to fets.

a linear series regulator is no less efficient than a series resistor for
dropping voltage, but it handles load fluctuations better.

you should do an energy audit on your design.

One of my products has a power switcher (variable output, 25 to 200
volts) using a series switching fet and a BIG power resistor feeding a
BIG output cap. A microprocessor/ADC checks the output voltage once
every millisecond and turns the fet on or off. Works great, and keeps
the room warm. It's about as efficient as a linear reg, but simpler,
and the heat's mostly in the resistor, not in the fet.

Sort of a delta-sigma regulator.

John

 

[Bob Larter]

John Larkin wrote:

On Mon, 04 May 2009 17:47:29 +1000, Bob Larter <bobbylarter@gmail.com
wrote:

Eeyore wrote:
John Larkin wrote:
[...]
Here's the box:

ftp://jjlarkin.lmi.net/99A260A1.JPG

ftp://jjlarkin.lmi.net/99A260A3.JPG

ftp://jjlarkin.lmi.net/99S260A.JPG

Seems to work now.
Ah ! Point to point wiring. The toob nuts would be proud of you.
;~)
Well, that's one way to reduce leakage!

This circuit has only one really leakage-sensitive node, and it's not
on the terminal strip. Its insulation is mostly air. The only leakage
is on the body of the opamp (National claims 10 fA typ) and the
polycarb thing I added to the front panel.

Unfortunately, on the LMC6001 the ni input is pin 3, and pin 4 is -5V.

With the kinds of currents you're measuring, I'd be worrying about skin
oil from your fingers on the chip package, terminals, etc. It might be
worth your while to brush everything down with isopropyl alcohol.

--
W
. | ,. w , "Some people are alive only because
\|/ \|/ it is illegal to kill them." Perna condita delenda est
---^----^---------------------------------------------------------------

 

[Bob Larter]

John Larkin wrote:

On Mon, 04 May 2009 17:45:11 +1000, Bob Larter <bobbylarter@gmail.com
wrote:

John Larkin wrote:
On Sun, 3 May 2009 17:32:18 -0700 (PDT), Greegor <Greegor47@gmail.com
wrote:

On May 3, 7:23 pm, John Larkin
jjlar...@highNOTlandTHIStechnologyPART.com> wrote:
These crappy RatShack terminal posts are actually conductive!
http://www.radioshack.com/product/index.jsp?productId=2103639

A pair of them leak 12 pA to the chassis at +1 volt. If I ground
myself and hold the plastic screw part of one, it goes up to 20.

So I'll have to replace them with some Pomonas or something. What a
nuisance.

Pity; they do look nice.

John
0.000000012 Amp? LOL
I just finished machining a couple of big slots in the aluminum box,
where the critical binding posts go, and added a polycarb plate that
the terminal strips mount on. Now the leakage is showing about 100 fA,
basically my measurement limit. It was ugly, machining the
already-"finished" electronics. I tried milling it dry, but that
didn't work, so I wound up getting chips and cutting fluid everywhere.
Had to clean all that up.

It's amazing what a few drops of Tapmatic will do.

Here's the box:

ftp://jjlarkin.lmi.net/99A260A1.JPG
What's with the huge bolts sticking out of the front? It'd look a lot
nicer if you'd put the heads on the outside.

That's to hold down the bent-aluminum measurement cover. At pA levels,
just moving around in the vicinity of the box will slam the
middle-node voltage around, and, unshielded, there's lots of 60 Hz
there too. These sorts of measurements don't work in free air.

Oh, I see.

--
W
. | ,. w , "Some people are alive only because
\|/ \|/ it is illegal to kill them." Perna condita delenda est
---^----^---------------------------------------------------------------

 

[Bob Larter]

Michael A. Terrell wrote:

Bob Larter wrote:
jcdrisc@melbpc.org.au wrote:
As a regular reader of the electronics postings I am heartily sick of
all the
advertising for clothing, shoes and vanity items. I recall in the old
days these would be
deleted by a moderator. I can only assume Google's advertising
policies are a bit out of control. We have a situation where the
majority of the group is polluted by this garbage.
I would hope something is done.
Not a chance. Google lets spammers sign up for as many free accounts as
they want, so if you get one nuked, they come straight back under
another name. This is why many people filter out all Google Groups posts.
I've tried campaigning to have Google treated the way other
spammer-friendly NSPs are treated, but Google spends too much money for
any of their news-peers to drop them.


Supernews was dropping anything posted to five or more newsgroups
from google groups. Then Giganews bought them and dropped all the
filtering they did.

A friend of mine is the ex-Supernews news-admin. He's the reason that I
know about Googles peering arrangements.

--
W
. | ,. w , "Some people are alive only because
\|/ \|/ it is illegal to kill them." Perna condita delenda est
---^----^---------------------------------------------------------------

 

[Eeyore]

Bob Larter wrote:

John Larkin wrote:
Bob Larter <bobbylarter@gmail.com> wrote:
Eeyore wrote:
John Larkin wrote:
[...]
Here's the box:

ftp://jjlarkin.lmi.net/99A260A1.JPG

ftp://jjlarkin.lmi.net/99A260A3.JPG

ftp://jjlarkin.lmi.net/99S260A.JPG

Seems to work now.
Ah ! Point to point wiring. The toob nuts would be proud of you.
;~)
Well, that's one way to reduce leakage!

This circuit has only one really leakage-sensitive node, and it's not
on the terminal strip. Its insulation is mostly air. The only leakage
is on the body of the opamp (National claims 10 fA typ) and the
polycarb thing I added to the front panel.

Unfortunately, on the LMC6001 the ni input is pin 3, and pin 4 is -5V.

With the kinds of currents you're measuring, I'd be worrying about skin
oil from your fingers on the chip package, terminals, etc. It might be
worth your while to brush everything down with isopropyl alcohol.

This is where a vapour cleaning tank could come in useful.

Graham

 

[Spehro Pefhany]

On Mon, 04 May 2009 06:48:58 -0700, John Larkin
<jjlarkin@highNOTlandTHIStechnologyPART.com> wrote:

On Mon, 04 May 2009 17:47:29 +1000, Bob Larter <bobbylarter@gmail.com
wrote:

Eeyore wrote:

John Larkin wrote:
[...]
Here's the box:

ftp://jjlarkin.lmi.net/99A260A1.JPG

ftp://jjlarkin.lmi.net/99A260A3.JPG

ftp://jjlarkin.lmi.net/99S260A.JPG

Seems to work now.

Ah ! Point to point wiring. The toob nuts would be proud of you.
;~)

Well, that's one way to reduce leakage!

This circuit has only one really leakage-sensitive node, and it's not
on the terminal strip. Its insulation is mostly air. The only leakage
is on the body of the opamp (National claims 10 fA typ) and the
polycarb thing I added to the front panel.

Unfortunately, on the LMC6001 the ni input is pin 3, and pin 4 is -5V.

John

Does it have to be -5V? Maybe you could bootstrap it off the output.
;-)

 

[Anon]

On Sat, 2 May 2009 17:15:50 -0700 (PDT), Mark-T
<MarkTanner50@gmail.com> wrote:

DId anyone here see the problem presented in
the Science section of NY Times last week?
Quite startling, to see something so sophisticated
in a 'general readership' publication.

Is it solvable without a calculus of variations approach?

Some very basic calculus is helpful. A solution is given at
<http://www.mathrec.org/old/2003jul/solutions.html>

 

[marcodbeast]

Eeyore wrote:

Archimedes' Lever wrote:

Jim Thompson wrote:

Success!

See...

http://analog-innovations.com/SED/TROLLFEEDER.jpg

Had to resort to NewsProxy until Agent gets a "References:" filter.

Will now just add, to Agent...

999 Delete: Subject:TROLLFEEDER

Then I won't even see the "TROLLFEEDER" tag.

Someone tell NymNuts, it _is_ universal, presently covering NymNuts,
Eeyore and Slowman follow-ups.

So, as far as I'm concerned, these "folks" don't exist anymore ;-)

...Jim Thompson

JimBob Brainlees Fart's head finally exploded.

It has been quite enjoyable watching him sink deeper and deeper
into his stupidity based seclusion desires.

Have fun being a net recluse. That has to be one of the most
retarded acts ever performed.

I have to agree with you.

The USA claims to be so in favour of 'free speech' yet it's the
Americans here who don't want to hear views that are contrary to
their own.

That's how they retain their cockeyed worldview. The right wing lie
aquariums (Fox, CNS, Newsmax, WND, etc.) their keepers built for them have
one universal feature - they're designed from the bottom up to make them
think that any other info sources, where they might hear actual facts, are
out to get them. There is literally no right wing k00khaus that doesn't
expend quite a bit of effort demonizing what they call the Mainstream Media,
can't have the dupes finding out they are in a fantasy world. =)

 

[John Larkin]

On Mon, 04 May 2009 10:42:25 -0400, Spehro Pefhany
<speffSNIP@interlogDOTyou.knowwhat> wrote:

On Mon, 04 May 2009 06:48:58 -0700, John Larkin
jjlarkin@highNOTlandTHIStechnologyPART.com> wrote:

On Mon, 04 May 2009 17:47:29 +1000, Bob Larter <bobbylarter@gmail.com
wrote:

Eeyore wrote:

John Larkin wrote:
[...]
Here's the box:

ftp://jjlarkin.lmi.net/99A260A1.JPG

ftp://jjlarkin.lmi.net/99A260A3.JPG

ftp://jjlarkin.lmi.net/99S260A.JPG

Seems to work now.

Ah ! Point to point wiring. The toob nuts would be proud of you.
;~)

Well, that's one way to reduce leakage!

This circuit has only one really leakage-sensitive node, and it's not
on the terminal strip. Its insulation is mostly air. The only leakage
is on the body of the opamp (National claims 10 fA typ) and the
polycarb thing I added to the front panel.

Unfortunately, on the LMC6001 the ni input is pin 3, and pin 4 is -5V.

John

Does it have to be -5V? Maybe you could bootstrap it off the output.
;-)

Too late; it's built. But yes, the V- could be made to track the
output, maybe a few tenths of a volt down. LMC6001 is a remarkably
poorly-specified amp, but Vcm seems to go to the - rail.


Yeah, this might work...


+5
|
|
|
Vcc


in------------IN+

OUT------------+------out
|
+----IN- |
| |
| |
| |
+-------------------------+
|
A
K big schottky diode
|
Vss----------------+
|
|
|
Isink to -5


(Or just an NPN follower, not so cute)

I could actively guard pin 3, on the chip itself and on the polycarb
insulator, with some difficulty. But I'm probably down below 100 fA
now, good enough for testing some diodes.

John

 

On May 4, 8:10 am, Nick <3-nos...@temporary-address.org.uk> wrote:

Sylvia Else <syl...@not.at.this.address> writes:
riverman wrote:
On May 3, 10:17 am, Sylvia Else <syl...@not.at.this.address> wrote:
Mark-T wrote:
DId anyone here see the problem presented in
the Science section of NY Times last week?
Quite startling, to see something so sophisticated
in a 'general readership' publication.
Is it solvable without a calculus of variations approach?
--
Mark
Yes.

Spoiler below 13

0

1

2

3

4

5

6

7

9

10

11

12

13

Since the agent can run four times as fast as the rabbit can swim, no
matter what the agent does, the rabbit can swim away from the centre,
and keep the agent on the far side as long as the rabbit is no more than
1/4 of the radius from the centre. So the rabbit can reach a point 3/4
of a radius from the edge while the agent is still on the opposite side.

Only if the agent stands still until the last possible moment. I think
the agent will move continuously, and as soon as the rabbit starts
moving in any direction, the agent will move toward the point that is
the closest to the rabbit at that moment. Let's call the agent 'Xeno'.

No - until the rabbit is 1/4 of the radius away from the centre, he
can swim fast enough, in a suitable direction, to keep the agent
exactly opposite. So it doesn't matter if the agent moves.

Which shows, I think (to return to the very original question about
whether this needs horrible maths to solve) that there is a simple proof
that pretty-well anyone can follow that shows the rabbit can escape in
the specific case (it goes to where it can just swim faster than the
agent, swims round in a circle until it gets to 180 degrees away from
the agent, and can then make it to the shore faster than the agent can
run round), but that the optimum strategy, and hence the answer to the
unasked question about which size of ponds or relative speeds allow the
rabbit to escape do require the horrid maths.

I'd hate to tackle it analytically. But numerically, it seems
perfectly tractable.

Assume that the rabbit isn't going for style points (swimming the
backstroke in circles to taunt the agent) and is going for the maximum
"lead" at the point where he exits the water.

Assume that the rabbit has reached his 1/4 r "jumping off" point. The
rabbit will initially head straight for shore. But only long enough
for the agent to pick a chase direction.

Once the agent has picked a direction, the rabbit chooses a new
straight line course. This course can be chosen by differentiation --
how much additional distance must the agent run versus how much
additional distance must the rabbit swim as a function of an
incremental change in course angle.

If agent running distance is a and rabbit swimming distance is h and
the rabbits selected swimming angle is theta then the hare is trying
maximize

a/4 - h

so he needs to solve:

d(a)/4 - d(h) / d(theta) = 0


The rabbit may continue making this calculation as it approaches the
shore. The result will remain unchanged as long as the agent does not
turn around. The selected angle only depends on the angle made
between the
shoreline and the rabbit's course line at the point where they
intersect. This means that the rabbit will swim a straight line
course.

If one is trying to zero in on the optimal value for the agent's
speed, one can can adjust the agent/rabbit speed ratio and until the
critical value is found.

 

[flipper]

On Mon, 4 May 2009 09:56:19 -0500, "marcodbeast" <its@casual.com>
wrote:

Eeyore wrote:
Archimedes' Lever wrote:

Jim Thompson wrote:

Success!

See...

http://analog-innovations.com/SED/TROLLFEEDER.jpg

Had to resort to NewsProxy until Agent gets a "References:" filter.

Will now just add, to Agent...

999 Delete: Subject:TROLLFEEDER

Then I won't even see the "TROLLFEEDER" tag.

Someone tell NymNuts, it _is_ universal, presently covering NymNuts,
Eeyore and Slowman follow-ups.

So, as far as I'm concerned, these "folks" don't exist anymore ;-)

...Jim Thompson

JimBob Brainlees Fart's head finally exploded.

It has been quite enjoyable watching him sink deeper and deeper
into his stupidity based seclusion desires.

Have fun being a net recluse. That has to be one of the most
retarded acts ever performed.

I have to agree with you.

The USA claims to be so in favour of 'free speech' yet it's the
Americans here who don't want to hear views that are contrary to
their own.

That's how they retain their cockeyed worldview. The right wing lie
aquariums (Fox, CNS, Newsmax, WND, etc.) their keepers built for them have
one universal feature - they're designed from the bottom up to make them
think that any other info sources, where they might hear actual facts, are
out to get them. There is literally no right wing k00khaus that doesn't
expend quite a bit of effort demonizing what they call the Mainstream Media,
can't have the dupes finding out they are in a fantasy world. =)

Congratulations on the near perfect emulation of a dog barking at his
own reflection.

 

[MooseFET]

On May 3, 4:09 pm, James Rollins <james.rolli...@gmail.com> wrote:

In my never ending quest for a cheap and simple high voltage
regulation I have come up with a simple dual capacitor mode. This is
similar to a Buck circuit without the inductor.

http://i41.tinypic.com/68zl03.jpg

That circuit is about as efficient as a linear regulator.

[....]

My requirements is a programmable constant voltage supply from around
0 voltages to 1kV driving a load of around 100kohms to 1Mohms. The
regulation should be less than 1% or better.

I have done a design that did that. How much ripple can you stand?

If you make the basic flyback switcher and drive a transformer with
it, getting 300V on the secondary isn't hard. A voltage tripler with
the two diodes in the flyback direction and a choke in series with the
diode in the forwards direction gives 1000V.

The reason to stay down at 300V is that I found a transformer that
could do that.