开云体育

All,

Below is the link to the Oren Elliott website and the page with the
roller inductors. It doesn't say what the form material is, just that
it's "newly redesigned".





Best,

Will

--- In ham_amplifiers@..., "pentalab" <jim.thomson@...>
wrote:

--- In ham_amplifiers@..., R L Measures <r@...> wrote:

On Oct 22, 2006, at 7:39 AM, pentalab wrote:

#### any idiot knows you can't use Delrin rod as a form for a
roller inductor..

Oren Elliot mfg the roller inductor.

### well I guess Oren Elliot is an idiot then... so is MFJ. If
they had half a brain... they woulda known Delrin is a no-no. You
can tell it's Delrin from 8' away.... that stuff has a distinct
look about it.

Not necessarily. MFJ could have blundered on the specs
when ordering, or ignored some rated specs. I do know
that Oren Elliott changed the form though, to which type
I can't remember. You might take a look at OE's website,
seems to me they might say what it is. In other words,
they could have both blundered. This was Oren Elliots first
time of manufacturing roller inductors. They use steatite
on the air caps. Who knows, MFJ may have told Oren Elliott
that Delrin would be what to use? MFJ was the first customer
on these, and most likely was who told them they would buy
them if they made them.

Even at this, the OE products are still built better in my
opinion than some made by others as far as the other
components. Some of the cheapest stuff I've seen is by Palstar
and some stuff that MFJ made themselves such as their air caps.
The frames and brushes are pretty flimsy on them. Palstars
roller inductors are made about the same as their air caps.
Since OE changed the form material, they're the top of the line
I think.

If you dug into this, I'll about bet MFJ is who mentioned
using Delrin and hooked OE up with a plastic supplier. I just
know OE makes some pretty good air caps with a very good design.



Best,

Will

--- In ham_amplifiers@..., R L Measures <r@...> wrote:

On Oct 22, 2006, at 4:47 PM, craxd wrote:

See below;

--- In ham_amplifiers@..., R L Measures <r@...> wrote:

On Oct 22, 2006, at 7:39 AM, pentalab wrote:

--- In ham_amplifiers@..., R L Measures <r@...>

wrote:

his Ameritron/ Heath/ Dentron /Amp supply/ MFJ 'engineering'

MFJ has a rather well deserved reputation as the junkiest
commercially built Hamstuff on this planet. I have heard

numerous

Hams refer to it as Mother F___ing Junk. Also, as I understand

it,

an MFJ "3000w" antenna tuner was the only one ever tested in

the

ARRL

Lab to ignite at 900W.

#### any idiot knows you can't use Delrin rod as a form for a
roller inductor..

Oren Elliot mfg the roller inductor.

Oren Elliott changed the design of the coil form using another
material after this happened. I remember e-mailing Steve Elliott
about this and he mentioned them changing it. When they designed

the

coil, they didn't check it at RF frequencies.

All they had to do is check the published D-factor for Delrin.

ABS

would have worked well around RF, but it's not as nice to turn

down

on a lathe.

I think they more or less took the suggestion from a plastics peddler
saying that Delrin would for sure work in this application. I'll have
to look up the difference in the D factor between Delrin and say
Teflon, etc. I have the Modern Plastics Encyclopedia here, I've just
not looked through it in a while.

Evidently, they didn't
have anything there to test at high power with at the time, if I
remember the conversation correctly. Anyhow, they're safe for use

now.

What material is used now?

I can't remember if he even told me what they did change it
to. Seems to me like they may mention it on their website.
Also, there's several different designs you can get on these
similar to the air variables like the brushes etc. Some cost
a little more than others according to what you order. They
also figure the buyer does the calculations to determine which
size to buy, or at least know enough to. In caps, I just told
them what model, shaft type, and air gap I wanted to place an
order. It's similar for ordering their roller inductors.

I'm pretty sure MFJ bought these direct as I used to buy direct
from them. You don't need to buy large orders either, 1-2 pieces
can be ordered.

The problem here sounds to me like MFJ continued selling the

tuners

with the delrin form inductors after knowing them to be bad. It

also

says the MFJ must have NEVER tested the tuner under full power

before

they started building them! Any good manufacturer would have

delayed

the sales and shipment until new replacement inductors could be
installed. It sounds to me like they sent them out hoping some

would

make it. The ones that didn't, simply send a new inductor and get

out

of the labor. That is unless it failed under warranty and they

had to

do the labor. It still sounds an awful lot like a bunch of penny
penching happened.

They saved pennies but they wasted dollars. That a mfg would
advertise something as "3000W" without testing it at even 900W

tells

me to avoid purchasing their products.

That's the way I seen it too!

...
R L Measures, AG6K, 805.386.3734
r@..., , rlm@..., www.somis.org

--- In ham_amplifiers@..., R L Measures <r@...> wrote:

On Oct 22, 2006, at 5:20 PM, craxd wrote:

Jim,

It's a wonder Tom didn't add a "by the way" saying, scopes and dip
meters are very inaccurate and should never be used. : ) At least

he

answered you by e-mail, he never would me.

He did not tell me this via e-mail. He said dipmeters were

worthless

during a discussion of various means of grounding grids after a
dipmeter measurement showed a result that contradicted his dicta
about grounding grids.

No, I meant I wouldn't doubt that he would have said
this to Jim.

If a scope is calibrated properly, or a dip meter the same, they
can't lie. They can only show you the truth.

That IS the problem here.

If a dip meter shows a
resonance, there is a resonance at its tuned frequency. Forget
reading the freq off the dial, simply couple it to a known

accurate

freq counter, it will show the truth if you want precision. A
calibrated scope can only show a waveform that it produces from

what

it sees at the input jack.

Correct. He didn't like the fact that an oscilloscope indicated a
worst-case potential in a SB-220 that was only about 1/3 of what

he

predicted it would be. This potential is on the verge of arcing

the

Tune-C, so even if the voltage tried to rise much higher, the cap
would zap and limit the voltage like a zener diode. The reason he
wanted the potential to be higher was to explain away what I said

was

parasitic arcing by the SB-220's occasional 110MHz oscillation.

The peak or peak to peak voltage one reads is exact, or
within the scopes calibration limits. It's according to
who calibrated it and the precision of the cal equipment
that was used, but at least < = +/- 3%.

The only way to make it lie is to not
calibrate it or set a control properly. If using a Tek or other
quality scope, your viewing something pretty accurate. I still

think

a scope is one of the better ways of measuring output power while
monitoring the signal for over-modulation and cleanliness. This
method has been used for years.

A calibrated oscilloscope is one way power meters can be

calibrated,

and it's much faster than using a bomb-calorimeter. However, an
oscilloscope is as useless as tits on a boar hog for measuring
cleanliness of a SSB signal.

Well it's according to what you term clean. Harmonics no, unless you
build or use a front end for this like the poor mans spectrum
analyzer, or one of the commercial add-on units. Over modulation,
hum, noise, regeneration, and parasitics can all be seen using a
standard scope. In the Sylvania scope book they give the following;

Modulation percentage either using the wave pattern or the trap
pattern.

Modulator output low, modulator mis-matched to RF stage.

Regen in RF stage, plus low AF output or modulator mis-match.

Low grid bias in RF stage, low excitatation, or both.

RF stage modulated incompletely neutralized.

Low excitation, low bias, or both in high-mu modulated RF stage.

Driven parasitics in modulated RF stage.

Hum modulation.

Noise modulation.

Modulation percentage or H2 - H1 / H2. H1 is the unmodulated height.

Each of the above having a distinct pattern on the scope. All can
contribute to one being un-clean.

...

R L Measures, AG6K, 805.386.3734
r@..., , rlm@..., www.somis.org

Best,

Will

--- In ham_amplifiers@..., R L Measures <r@...> wrote:

On Oct 24, 2006, at 1:35 AM, pentalab wrote:

### This sucks... and I don't like it one bit.

RICH SEZ... So use Matsushita/Panasonic MOFs instead of wire-wounds.

### Rich have u actually tried opening off one resistor [in a

one resistor per cap set up] and fired up the supply with a
small variac ??

RICH SEZ... Are you crazy? I was about 20-feet from where that

sucker exploded in the cal lab.

### Or should I not think about it... and hope my 24 x new

2500 > UF @ 450 V lytics don't got off like firecracker's
someday ??

RICH SEZ... To me, it sounded more like a jumbo cherry bomb.

Panasonic MOF resistors are pretty reliable.

### Flash ! with 4 x resistor's per cap... and one cut

loose. [assuming caps are no more than 75% of their v rating] The
one cap with the bad R will be almost maxed out.

RICH SEZ... That would work.

### 4 x resistor's in parallel = 25 K = lotsa heat per
cap. Probbaly 6-8 x resistor's.... each say 600-800K... all
in parallel... PER cap, would be the ultimate solution.....

then if anything opened up.... no chance of all ur caps blowing

up. The heat would be zip... per resistor.

ok problem solved.... now i can sleep

RICH SEZ.... But what if the sky falls?

later........ Jim VE7RF

########### Here's my problem Rich.... Those 2500 Uf @ 450 V CD
brand 10 A CCS ripple current rated caps I got NEW, by the case
load are listed at an unreal price tag.... like $100.00 EACH...
in single lot quantities. I have 72 of em... and will be using
em in groups of 24. I got em dirt cheap... they all tested
good, when checked on the bench. I would be just pissed if they
started exploding cuz of even the remotest chance of one resistor
lead opening.

### So... the real solution [aside from the usual 1 kv-6A safety
diode, rvs connected across eacg lytic] is to parallel 6 x 300 K
MOF resistors across EACH cap. [50 K for the paralled mess].
If one ever opened... the voltage would only increase 20% on
that one cap. Total diss is 53 W. Each of the 144 resistor's
would only dissipate .37 w. Heck, with 6 x 150 K
resistors, Total diss rises to 106 W.... or a measely .74 watt
for each of the 144 resistor's.

### For a 8000 V [no load] supply... with 24 x caps..... normal V
per cap is 333 V [74% of the 450 V max rating]. With 6 x 300K /6
x 150 K resistor's across each cap.... and say one of the 6 opened
up.... V across that one cap will increase to 400 V.......well
within the 450 V rating of the cap....... end of problem.... end of
story.

### later...... Jim VE7RF

--- In ham_amplifiers@..., Bill Turner <dezrat@...>
wrote:

ORIGINAL MESSAGE:

On Mon, 23 Oct 2006 00:20:30 -0000, "craxd" <craxd1@...>
wrote:

If a dip meter shows a
resonance, there is a resonance at its tuned frequency. Forget
reading the freq off the dial, simply couple it to a known

accurate

freq counter, it will show the truth if you want precision.

------------ REPLY FOLLOWS ------------

One needs to be careful with this. The typical single-tube or
single-transistor dip meter can be pulled considerably off the true
resonant frequency if coupling is very tight to a high-Q circuit.

Just

physically move the dip meter away far enough so the dip is barely
perceptible and the accuracy will improve greatly.

If you have used a dip meter much, you have no doubt had the
experience of tuning slowly across the dip, and as you continue

tuning

have the dip suddenly disappear as the pulling effect disappears.
Loose coupling will prevent this from happening.

### Highly agreed. Loose coupling is as good as it gets.... for
acuracy anyway. Mike Stahl, K6MYC warned me about this 25 yrs
ago.... he didn't trust em.. when tweaking yagi's.... I wouldn't
either.

### For a real laff.... I bought the "mating" grid dip osc
coils for my MFJ-259 analyzer. Two coils cover the entire
spectrum. I can't grid dip ANYTHING with em... even simple
stuff.. like a coil in parallel with a cap... on the test bench.
Then I find out nobody else can grid dip anything either! Then Rauch
sez they don't really work. They still don't. It was
suggested not to buy the optional coil(s).

### The inductance measurement is not to be trusted in a MFJ-
259-B either. Inductance measurements drop like a rock with
increasing freq. Even at 1.8 mhz.... coils will show no where
near what my B+K 875A/B does. Per Tom.... the reason is the
the stray C between turns on a coil will subtract from the coils
inductance.

### To prove that theory one way or another.... Using the
various PI spreadsheets available..[like GM3SEK's... which will
factor in the stray L of the parasitic suppressor, and all stray L
between tube and C1 cap... and also factor in the UH/XL of the RFC
[which will require an equal amount of XC... coming from the
C1 cap]. His sheet will also factor in all stray C... plus anode
to chassis of tube... in the socket.

### When the sheets say C1 should be XXX and L should be
ZZZ, and C2 should be LLL. When I use my B+K 875... and
actually measure C1 L C2... and then tie the interconnecting
straps together.... then use a resistor between anode to chassis
[to simulate the plate load Z]... and a MFJ-259 on the output....
guess what? Flat swr, or at the most... just a very tiny teak
on the C1/C2 cap to resonate.
Now the B+K operates at 1 khz... and has never let me down
yet.

### Known coils from various manufacturer's always measure dead
on with the B+K. [rare exception was this 14 uh multronics
coil used on the last project.. which measured 12.2 uh]

## If I had used the inductance measuring function of the 259-B...
at the freq under test for the simple PI [say 80/40/20m], the
resulting coil would be WAY smaller.... and of course, the C1 and
C2 values had to be INCREASED by a huge amount to compensate....
then the loaded Q is WAY up... and the resulting circulating
currents are way up.

### Moral of the story.... use a MFJ-259B to measure inductance
for anything... like PI nets, LC networks, you name it... and you
will be guaranteed 100%, to have complete failure. Get a real LCR
meter. If you can't measure L + C accurately..[that's 90% of
radio].... you are dead in the water.

Later...... Jim VE7RF

Bill, W6WRT

ORIGINAL MESSAGE:

On Mon, 23 Oct 2006 00:20:30 -0000, "craxd" <craxd1@...>
wrote:

If a dip meter shows a
resonance, there is a resonance at its tuned frequency. Forget
reading the freq off the dial, simply couple it to a known accurate
freq counter, it will show the truth if you want precision.

------------ REPLY FOLLOWS ------------

One needs to be careful with this. The typical single-tube or
single-transistor dip meter can be pulled considerably off the true
resonant frequency if coupling is very tight to a high-Q circuit. Just
physically move the dip meter away far enough so the dip is barely
perceptible and the accuracy will improve greatly.

If you have used a dip meter much, you have no doubt had the
experience of tuning slowly across the dip, and as you continue tuning
have the dip suddenly disappear as the pulling effect disappears.
Loose coupling will prevent this from happening.

Bill, W6WRT

Well, by some means Yahoo removed the text I wrote after my "Read this:" statement and before my sign. What I suggested to DJ7SW was to read the very bottom of the message, where there are instructions to unsubscribe from the list and group.

Best regards
Guillermo - LU8EYW.

GGLL escribi:

On Oct 24, 2006, at 5:25 AM, GGLL wrote:

R L Measures escribió:

On Oct 24, 2006, at 1:35 AM, pentalab wrote:

--- In ham_amplifiers@..., R L Measures <r@...> wrote:

On Oct 23, 2006, at 3:23 PM, GGLL wrote:

**** I have changed the subject ****

Please see below:

R L Measures escribió:

...
Hello Rich, how fast (or slow) you think voltage will rise
in the capacitor with its bleeder open?.

Probably about as fast as the capacitors with resistors discharge.

One could think that the losses of the capacitor could
be considered as "resistance", but of very high value, so
applying plain
voltage divider equations the higher potential will be there.

###### Exactly my thoughts. That's the only explanation that
works.

The fly in the pie is that the leakage R of the C is changing.

With voltage? or because of own unstable characteristics?.

Correctomundo. When the potential on the 200uF, 500wvdc electrolytic
with the open bleeder/equalizer R goes above 500v, leakage-I as well
as heating increases exponentially and the fireworks show is about to
start

### Rich have u actually tried opening off one resistor [in a one
resistor per cap set up] and fired up the supply with a small
variac ??

Are you crazy? I was about 20-feet from where that sucker exploded
in the cal lab.

Perhaps scaling down components (voltage specially, seriously
talking) and
being at a safe distance (100 mtrs or so) with remote measuring and
a robot
turning on the whole thing (not seriously talking)....

chortle

But what if the sky falls?

That's the only thing that worries Asterix and his chief.

guffaw

cheers, Guillermo

R L Measures, AG6K, 805.386.3734
r@..., , rlm@..., www.somis.org

On Oct 24, 2006, at 4:57 AM, Tony King - W4ZT wrote:

R L Measures wrote:
<snip>

### Lemmee ask u this Rich..... IF u were gonna build a HB 2 x 3-
500Z linear from scratch... would u use the 3 x 200 pf caps [ or
any othwer value] or not ??????????????????

I would use: 1500 - 2000 pF total per socket consisting of three
different values per each, a grid fusing element, a FWD/FWB
switchable PS, and I would use a glitch R, followed by an 8160.

<snip>
Rich, he asked about a pair of 3-500Z's... but, since you mentioned the
8160, do you mean you would NOT directly ground the grid of the 8160 and
put a fusing element between B- and ground?

The 3-500Zs would be the buffer amplifier, and the 8160 would be the final amplifier. With the latter, as you realized, since the grid is grounded by its socket, a grid fusing device would have to go between the grid-I meter shunt and gnd. Overall power gain would be c. 200x, which is about the same as an 8281 tetrode in AB1.

cheers, Tony.

73, Tony W4ZT

R L Measures, AG6K, 805.386.3734
r@..., , rlm@..., www.somis.org

Read this:

Best regards
Guillermo - LU8EYW.

dj7sw escribi:

Unsubscribe

R L Measures escribi:

On Oct 24, 2006, at 1:35 AM, pentalab wrote:

--- In ham_amplifiers@..., R L Measures <r@...> wrote:

On Oct 23, 2006, at 3:23 PM, GGLL wrote:

**** I have changed the subject ****

Please see below:

R L Measures escribi:

...
Hello Rich, how fast (or slow) you think voltage will rise
in the capacitor with its bleeder open?.

Probably about as fast as the capacitors with resistors discharge.

One could think that the losses of the capacitor could
be considered as "resistance", but of very high value, so applying plain
voltage divider equations the higher potential will be there.

###### Exactly my thoughts. That's the only explanation that
works.

The fly in the pie is that the leakage R of the C is changing.

With voltage? or because of own unstable characteristics?.

### Rich have u actually tried opening off one resistor [in a one
resistor per cap set up] and fired up the supply with a small
variac ??

Are you crazy? I was about 20-feet from where that sucker exploded in the cal lab.

Perhaps scaling down components (voltage specially, seriously talking) and being at a safe distance (100 mtrs or so) with remote measuring and a robot turning on the whole thing (not seriously talking)....

But what if the sky falls?

That's the only thing that worries Asterix and his chief.

Best regadrs
Guillermo - LU8EYW.

R L Measures wrote:
<snip>

### Lemmee ask u this Rich..... IF u were gonna build a HB 2 x 3-
500Z linear from scratch... would u use the 3 x 200 pf caps [ or
any othwer value] or not ??????????????????

I would use: 1500 - 2000 pF total per socket consisting of three different values per each, a grid fusing element, a FWD/FWB switchable PS, and I would use a glitch R, followed by an 8160.

<snip>
Rich, he asked about a pair of 3-500Z's... but, since you mentioned the 8160, do you mean you would NOT directly ground the grid of the 8160 and put a fusing element between B- and ground?

73, Tony W4ZT

On Oct 24, 2006, at 1:35 AM, pentalab wrote:

--- In ham_amplifiers@..., R L Measures <r@...> wrote:

On Oct 23, 2006, at 3:23 PM, GGLL wrote:

**** I have changed the subject ****

Please see below:

R L Measures escribió:

...
Hello Rich, how fast (or slow) you think voltage will rise
in the capacitor with its bleeder open?.

Probably about as fast as the capacitors with resistors discharge.

One could think that the losses of the capacitor could
be considered as "resistance", but of very high value, so
applying plain
voltage divider equations the higher potential will be there.

###### Exactly my thoughts. That's the only explanation that
works.

The fly in the pie is that the leakage R of the C is changing.

Even though u would still be drawing some bleeder
current right at the moment u opened one off.... the caps would
have v on em.... plus a 100 k across it. The v would drop on
all the caps.... and the open one would have sky high V across
it... kablamo........ and putting 2 x 100k reistors across each
cap ain't gonna help much.

Maybe we wouldn't have to wait as long for the fireworks show with
more bleeder current?

## If one resistor in the pair opened up... that V would
double the rest.... and still blow up. it would be bleeder
crrent x twice as much R for the one cap... v would double...
and no more.... and bye bye cap. Never thought of this b4.
These resistor's would have to be terminated very carefully... no
screw ups.

### I'd suggest on any retrofit with rich's resistors... or
bringing up caps that have been sitting for years..... to use a
variac.... and bring em up from 0 to full bore... in 500 v
increments... every 15 mins. 0-500-1000-1500 etc. Of
course.... b4 u do any of that... just bring it up with a variac
to say no more than 450 V to start with.... then get in there
with a dvm... and make sure all the caps have equal voltage.

I would be inclined to replace tired electrolytics with new ones
rather than trying to resuscitate old ones with an inferior design.

After that... just keep raising it up every 15 mins. On screq
term caps.... I terminate the resistors in st-kon connectors...
[use round ones, NOT spade type] I then solder em as well.... then
terminate the sta-kons with the machine screw...and use an inside
tooth lockwasher . On caps with pins.... just wrap and solder..
and leave plenty of resistor lead sticking out..... which is the
heatsink on em.

The eventual potential across the 3 capacitors with resistors is
zero, so theoretically the potential across the capacitor with no
resistor would rise to 1520v -- assuming it could withstand this much
voltage.

If all four would fail, the voltages will be more "equalized"
than if only one blows, is this ok?.

My guess is that without the 8, 50k-ohm equalization resistors the
HV PS would probably function okay.

## u mean 8 x 100 K reistor's.

The 922 uses 50k, 10w wire-wounds.

### Correct. the HV will divide, proportionally to UF per cap.
Now if the caps were all matched perfectly... u probably would not
need eq resistors.... just a bleeder from B+ to B-

agreed, but adding diodes across the caps sounds like a pretty good
idea.

### The chances of ALL the eq resistor's opening up is NIL....
ain't gonna ever happen. I can see ONE end of ONE resitor
opening up... cuz of an int connection, etc.

especially with wire-wounds.

### I don't like this at all.... it's an accident waiting to
happen. I just tried it[on paper] with 3 x 100K resistors per
cap.... works... but STILL no good if one cap has one of
it's 3 x resistor's open up...... even running the caps at
75% of their V rating won't work in this case.... u still end
up with a solid 500 V across the cap with only 2 of it's 3
x resistor's intact.

### This sucks... and I don't like it one bit.

So use Matsushita/Panasonic MOFs instead of wire-wounds.

### Rich have u actually tried opening off one resistor [in a one
resistor per cap set up] and fired up the supply with a small
variac ??

Are you crazy? I was about 20-feet from where that sucker exploded
in the cal lab.

### other than a bleeder current sense setup.... or measuring the
v aacross each cap.... I don't see a fix for this... cept maybe
one big oil cap... or 1-XXX oil caps in PARALLEL.

Using MOFs makes more $ense.

### Or should I not think about it... and hope my 24 x new 2500
UF @ 450 V lytics don't got off like firecracker's someday ??

To me, it sounded more like a jumbo cherry bomb.

That's 24 resistor's with a total of 48 x leads. This is for
the 7900 V HV supply, currently under construction.

## I'd almost be inclined to leave ALL the resistor's OFF... don't
use em. Bring it up with a variac... and at 1/4 V.... measure
the V across each cap. I have 72 caps... so I'm sure I can
match em in 3 x groups of 24.

Panasonic MOF resistors are pretty reliable.

### Flash ! with 4 x resistor's per cap... and one cut loose.
[assuming caps are no more than 75% of their v rating] The one
cap with the bad R will be almost maxed out.

That would work.

### 4 x resistor's in parallel = 25 K = lotsa heat per
cap. Probbaly 6-8 x resistor's.... each say 600-800K... all
in parallel... PER cap, would be the ultimate solution..... then
if anything opened up.... no chance of all ur caps blowing
up. The heat would be zip... per resistor.

ok problem solved.... now i can sleep

But what if the sky falls?

later........ Jim VE7RF

Best regards
Guillermo - LU8EYW.



Yahoo! Groups Links

R L Measures, AG6K, 805.386.3734
r@..., , rlm@..., www.somis.org

Yahoo! Groups Links

R L Measures, AG6K, 805.386.3734
r@..., , rlm@..., www.somis.org

--- In ham_amplifiers@..., R L Measures <r@...> wrote:

On Oct 23, 2006, at 3:23 PM, GGLL wrote:

**** I have changed the subject ****

Please see below:

R L Measures escribió:

Example: In a series string of 4 electrolytic filter caps,

with an

open bleeder on one cap, the potential on the other 3 caps

would

slowly bleed down to zero volts, and the voltage on the cap

with the

open bleeder would rise 4x to the full potential coming from the
rectifiers. For example, a TL-922: It uses 4, 200uF, 500V caps

in

each half of a FWD. Normally there is c. 380v on each filter

cap.

If one bleeder/equalizer R opened, the potential across the cap

with

the open bleeder would rise to 4x normal or 1520v - but of

course a

500v electrolytic would probably fail before the potential

rose to

600v.

Hello Rich, how fast (or slow) you think voltage will rise

in the

capacitor
with its bleeder open?.

Probably about as fast as the capacitors with resistors discharge.

One could think that the losses of the capacitor could
be considered as "resistance", but of very high value, so

applying

plain
voltage divider equations the higher potential will be there.

###### Exactly my thoughts. That's the only explanation that
works. Even though u would still be drawing some bleeder
current right at the moment u opened one off.... the caps would
have v on em.... plus a 100 k across it. The v would drop on
all the caps.... and the open one would have sky high V across
it... kablamo........ and putting 2 x 100k reistors across each
cap ain't gonna help much.

## If one resistor in the pair opened up... that V would
double the rest.... and still blow up. it would be bleeder
crrent x twice as much R for the one cap... v would double...
and no more.... and bye bye cap. Never thought of this b4.
These resistor's would have to be terminated very carefully... no
screw ups.

### I'd suggest on any retrofit with rich's resistors... or
bringing up caps that have been sitting for years..... to use a
variac.... and bring em up from 0 to full bore... in 500 v
increments... every 15 mins. 0-500-1000-1500 etc. Of
course.... b4 u do any of that... just bring it up with a variac
to say no more than 450 V to start with.... then get in there
with a dvm... and make sure all the caps have equal voltage.
After that... just keep raising it up every 15 mins. On screq
term caps.... I terminate the resistors in st-kon connectors...
[use round ones, NOT spade type] I then solder em as well.... then
terminate the sta-kons with the machine screw...and use an inside
tooth lockwasher . On caps with pins.... just wrap and solder..
and leave plenty of resistor lead sticking out..... which is the
heatsink on em.

The eventual potential across the 3 capacitors with resistors is
zero, so theoretically the potential across the capacitor with no
resistor would rise to 1520v -- assuming it could withstand this

much

voltage.

If all four would fail, the voltages will be more "equalized"

than if only one blows, is this ok?.

My guess is that without the 8, 50k-ohm equalization resistors the

HV PS would probably function okay.

## u mean 8 x 100 K reistor's.

### Correct. the HV will divide, proportionally to UF per cap.
Now if the caps were all matched perfectly... u probably would not
need eq resistors.... just a bleeder from B+ to B-

### The chances of ALL the eq resistor's opening up is NIL....
ain't gonna ever happen. I can see ONE end of ONE resitor
opening up... cuz of an int connection, etc.

### I don't like this at all.... it's an accident waiting to
happen. I just tried it[on paper] with 3 x 100K resistors per
cap.... works... but STILL no good if one cap has one of
it's 3 x resistor's open up...... even running the caps at
75% of their V rating won't work in this case.... u still end
up with a solid 500 V across the cap with only 2 of it's 3
x resistor's intact.

### This sucks... and I don't like it one bit.

### Rich have u actually tried opening off one resistor [in a one
resistor per cap set up] and fired up the supply with a small
variac ??

### other than a bleeder current sense setup.... or measuring the
v aacross each cap.... I don't see a fix for this... cept maybe
one big oil cap... or 1-XXX oil caps in PARALLEL.

### Or should I not think about it... and hope my 24 x new 2500
UF @ 450 V lytics don't got off like firecracker's someday ??
That's 24 resistor's with a total of 48 x leads. This is for
the 7900 V HV supply, currently under construction.

## I'd almost be inclined to leave ALL the resistor's OFF... don't
use em. Bring it up with a variac... and at 1/4 V.... measure
the V across each cap. I have 72 caps... so I'm sure I can
match em in 3 x groups of 24.

### Flash ! with 4 x resistor's per cap... and one cut loose.
[assuming caps are no more than 75% of their v rating] The one
cap with the bad R will be almost maxed out.

### 4 x resistor's in parallel = 25 K = lotsa heat per
cap. Probbaly 6-8 x resistor's.... each say 600-800K... all
in parallel... PER cap, would be the ultimate solution..... then
if anything opened up.... no chance of all ur caps blowing
up. The heat would be zip... per resistor.

ok problem solved.... now i can sleep

later........ Jim VE7RF

Best regards
Guillermo - LU8EYW.



Yahoo! Groups Links

R L Measures, AG6K, 805.386.3734
r@..., , rlm@..., www.somis.org

On Oct 23, 2006, at 3:23 PM, GGLL wrote:

**** I have changed the subject ****

Please see below:

R L Measures escribió:

Example: In a series string of 4 electrolytic filter caps, with an
open bleeder on one cap, the potential on the other 3 caps would
slowly bleed down to zero volts, and the voltage on the cap with the
open bleeder would rise 4x to the full potential coming from the
rectifiers. For example, a TL-922: It uses 4, 200uF, 500V caps in
each half of a FWD. Normally there is c. 380v on each filter cap.
If one bleeder/equalizer R opened, the potential across the cap with
the open bleeder would rise to 4x normal or 1520v - but of course a
500v electrolytic would probably fail before the potential rose to
600v.

Hello Rich, how fast (or slow) you think voltage will rise in the
capacitor
with its bleeder open?.

Probably about as fast as the capacitors with resistors discharge.

One could think that the losses of the capacitor could
be considered as "resistance", but of very high value, so applying
plain
voltage divider equations the higher potential will be there.

The eventual potential across the 3 capacitors with resistors is
zero, so theoretically the potential across the capacitor with no
resistor would rise to 1520v -- assuming it could withstand this much
voltage.

If all four would fail, the voltages will be more "equalized" than
if only one
blows, is this ok?.

My guess is that without the 8, 50k-ohm equalization resistors the HV
PS would probably function okay.

Best regards
Guillermo - LU8EYW.



Yahoo! Groups Links

R L Measures, AG6K, 805.386.3734
r@..., , rlm@..., www.somis.org

On Oct 23, 2006, at 4:22 PM, pentalab wrote:

--- In ham_amplifiers@..., R L Measures <r@...> wrote:

On Oct 23, 2006, at 12:07 AM, pentalab wrote:

Tom's problem with dipmeters was when one showed that grid-

resonance Decreases in frequency when a grid is grounded with
heavy duty Cu straps instead of caps (SB-220, TL-922. L4B)

### IT does decrease.... by just 1 mhz if I remember. What I
don't understand is.... why does stability improve with grids

strapped directly to? chassis ??

It may and it may not. Moving one resonance a bit may appear to be
a sure cure, however, removing the perforated cover from a Henry
2K-4
will usually make it more stable but in my experiences parasitic
oscillations are inheriently on the ragged edge, so it's easy to be
deluded into thinking you made a slam-dunk by making one change
and seeing no fireworks during a few minute test.

##### well how about 3 doz L4B's, SB-220, SB-221, TL922,s...
that all had their grids directly grounded....... and have been
like that, running flawlessly for at least 4 -5 years now ???
They all have stock suppressor's in em too.

** ______

I'd still grnd the
grids directly.... just cuz the drive requirements drop by an

easy 20-25 watts.

That the 2 x 600pF of grid bypass on the pair of 3-500Zs in a 922,

L4- B or SB-220 dissipate 20 - 25w is hardly likely.

### agreed. The point here is the drive power requirement DOES
drop by 20-25 watts. I measured several of em with bird
wattmeter's.... that had been recently calibrated. During the b4
and after test.... it's always the same on a L4B.... exactly 22
watts LESS drive... WITH the grids directly grnded.

### MY tech buddy's with the SB-221 and 220 and the fellows
with the TL-922 reported between 20-25 watts LESS drive
required. The wattmeter's don't lie. And that test has
been done countless times by other's too. The TX imd improves
a lot too. That NFB is non existent with the 3 x 200 pf caps
in place. Run the xcvr at 20-25 watts LESS output.... and the
xcvr's IMD will clean up too...... total imd is way down. Try it
urself.

### Lemmee ask u this Rich..... IF u were gonna build a HB 2 x 3-
500Z linear from scratch... would u use the 3 x 200 pf caps [ or
any othwer value] or not ??????????????????

I would use: 1500 - 2000 pF total per socket consisting of three
different values per each, a grid fusing element, a FWD/FWB
switchable PS, and I would use a glitch R, followed by an 8160.

later... Jim VE7RF

Yahoo! Groups Links

R L Measures, AG6K, 805.386.3734
r@..., , rlm@..., www.somis.org

--- In ham_amplifiers@..., R L Measures <r@...> wrote:

On Oct 23, 2006, at 12:07 AM, pentalab wrote:

Tom's problem with dipmeters was when one showed that grid-

resonance Decreases in frequency when a grid is grounded with
heavy duty Cu straps instead of caps (SB-220, TL-922. L4B)

### IT does decrease.... by just 1 mhz if I remember. What I
don't understand is.... why does stability improve with grids

strapped directly to chassis ??

It may and it may not. Moving one resonance a bit may appear to be

a sure cure, however, removing the perforated cover from a Henry 2K-
4

will usually make it more stable but in my experiences parasitic
oscillations are inheriently on the ragged edge, so it's easy to

be

deluded into thinking you made a slam-dunk by making one change

and seeing no fireworks during a few minute test.

##### well how about 3 doz L4B's, SB-220, SB-221, TL922,s...
that all had their grids directly grounded....... and have been
like that, running flawlessly for at least 4 -5 years now ???
They all have stock suppressor's in em too.

I'd still grnd the
grids directly.... just cuz the drive requirements drop by an

easy 20-25 watts.

That the 2 x 600pF of grid bypass on the pair of 3-500Zs in a 922,

L4- B or SB-220 dissipate 20 - 25w is hardly likely.

### agreed. The point here is the drive power requirement DOES
drop by 20-25 watts. I measured several of em with bird
wattmeter's.... that had been recently calibrated. During the b4
and after test.... it's always the same on a L4B.... exactly 22
watts LESS drive... WITH the grids directly grnded.

### MY tech buddy's with the SB-221 and 220 and the fellows
with the TL-922 reported between 20-25 watts LESS drive
required. The wattmeter's don't lie. And that test has
been done countless times by other's too. The TX imd improves
a lot too. That NFB is non existent with the 3 x 200 pf caps
in place. Run the xcvr at 20-25 watts LESS output.... and the
xcvr's IMD will clean up too...... total imd is way down. Try it
urself.

### Lemmee ask u this Rich..... IF u were gonna build a HB 2 x 3-
500Z linear from scratch... would u use the 3 x 200 pf caps [ or
any othwer value] or not ??????????????????

later... Jim VE7RF

**** I have changed the subject ****

Please see below:

R L Measures escribi:

Example: In a series string of 4 electrolytic filter caps, with an open bleeder on one cap, the potential on the other 3 caps would slowly bleed down to zero volts, and the voltage on the cap with the open bleeder would rise 4x to the full potential coming from the rectifiers. For example, a TL-922: It uses 4, 200uF, 500V caps in each half of a FWD. Normally there is c. 380v on each filter cap. If one bleeder/equalizer R opened, the potential across the cap with the open bleeder would rise to 4x normal or 1520v - but of course a 500v electrolytic would probably fail before the potential rose to 600v.

Hello Rich, how fast (or slow) you think voltage will rise in the capacitor with its bleeder open?. One could think that the losses of the capacitor could be considered as "resistance", but of very high value, so applying plain voltage divider equations the higher potential will be there.
If all four would fail, the voltages will be more "equalized" than if only one blows, is this ok?.

Best regards
Guillermo - LU8EYW.