[Trailboss]

Looking for info on this coupling capacitor in the oscillator ckt.
Metal can Axial type
..5 µF
900 WVDC

Is this can an oil filled capacitor?

Would Polypro or Polystyrene type be a suitable replacement?

Any help appreciated.

Vin

[Patrick Turner]

On Oct 2, 6:28*am, "Trailboss" wrote:
Looking for info on this coupling capacitor in the oscillator ckt.
Metal can Axial type
.5 F
900 WVDC

Is this can an oil filled capacitor?

Would Polypro or Polystyrene type be a suitable replacement?

Any help appreciated.

Vin

You say 0.5F, which is 0.5 Farad, which would be a very large cap.
What amount of capacitance do you really have? Is it 0.5uF, or 0.5
nF???????????

Polypropylene or Polyester will be fine as a replacement as long as
the capacitance value is the same and the voltage rating is the same,
but only if the original cap's capacitance and V ratings were correct.

Patrick Turner.

[Trailboss]

Patrick,

..5µF
point five micro farad,
the micro (using Alt 230 didn't come across for some reason) µ

thanks Patrick,

Vin


"Patrick Turner" wrote in message
...

On Oct 2, 6:28 am, "Trailboss" wrote:
Looking for info on this coupling capacitor in the oscillator ckt.
Metal can Axial type
.5 F
900 WVDC

Is this can an oil filled capacitor?

Would Polypro or Polystyrene type be a suitable replacement?

Any help appreciated.

Vin

You say 0.5F, which is 0.5 Farad, which would be a very large cap.
What amount of capacitance do you really have? Is it 0.5uF, or 0.5
nF???????????

Polypropylene or Polyester will be fine as a replacement as long as
the capacitance value is the same and the voltage rating is the same,
but only if the original cap's capacitance and V ratings were correct.

Patrick Turner.

[Patrick Turner]

On Oct 2, 3:15*pm, "Trailboss" wrote:
Patrick,

.5 F
point five micro farad,
the micro (using Alt 230 didn't come across for some reason)

thanks Patrick,

If you use 0.47uF ployprop pot polyester, rated for 1,000V it should
be OK. If the oscillator F depends on the C being exactly 0.5uF, then
strap 0.027uF plus 0.0033uF across the 0.47uF to increase C to exactly
0.5uF. But any chosen 0.47uF may not be exactly 0.047uF, so the actual
value of additional parallel C to trim the value depends on the
calibration of your oscillator, and the C should be adjusted up OR
down until you get the correct frequency according to your frequency
meter and then the dial of the oscillator adjusted to suit the real
frequency, so that when you set the dial to say 100Hz, you get F =
100Hz +/- 3 Hz, and not oustide such tolerances.

Old oscillators are like old men, they get all wandery - blundery in
their thortatious processes, rather like young men, and only men of
middle age and at the height of their potential powers can muddle thru
with more accuracy than those yunger or older. Never stop thinking or
asking questions. Wives will hate you, but the wise will love ya.

Patrick Turner.

[Trailboss]

Patrick,

Words of wisdom, well said.

Would you have any idea why HP would use a 900 volt cap in this ckt?

Plate Voltage of the 6AU6 is 110vdc coupled by the .5 micro F to the grid of
6AQ5.

TIA
Vin


"Patrick Turner" wrote in message
...

On Oct 2, 3:15 pm, "Trailboss" wrote:
Patrick,

.5 F
point five micro farad,
the micro (using Alt 230 didn't come across for some reason)

thanks Patrick,

If you use 0.47uF ployprop pot polyester, rated for 1,000V it should
be OK. If the oscillator F depends on the C being exactly 0.5uF, then
strap 0.027uF plus 0.0033uF across the 0.47uF to increase C to exactly
0.5uF. But any chosen 0.47uF may not be exactly 0.047uF, so the actual
value of additional parallel C to trim the value depends on the
calibration of your oscillator, and the C should be adjusted up OR
down until you get the correct frequency according to your frequency
meter and then the dial of the oscillator adjusted to suit the real
frequency, so that when you set the dial to say 100Hz, you get F =
100Hz +/- 3 Hz, and not oustide such tolerances.

Old oscillators are like old men, they get all wandery - blundery in
their thortatious processes, rather like young men, and only men of
middle age and at the height of their potential powers can muddle thru
with more accuracy than those yunger or older. Never stop thinking or
asking questions. Wives will hate you, but the wise will love ya.

Patrick Turner.

[Engineer[_2_]]

On Oct 2, 9:33*am, "Trailboss" wrote:
Patrick,

Words of wisdom, well said.

Would you have any idea why HP would use a 900 volt cap in this ckt?

Plate Voltage of the 6AU6 is 110vdc coupled by the .5 micro F to the grid of
6AQ5.

TIA
Vin

(snip)

No idea. I'd just pick one of the 630 VDC caps to hand... orange drop
or the yellow ones. I've never had any problem with these as plate to
grid coupling caps. BTW, don't use ceramics in any audio coupling
capacity (no pun intended!)
Cheers,
Roger

[Engineer[_2_]]

On Oct 2, 10:49*am, Engineer wrote:
On Oct 2, 9:33*am, "Trailboss" wrote: Patrick,

Words of wisdom, well said.

Would you have any idea why HP would use a 900 volt cap in this ckt?

Plate Voltage of the 6AU6 is 110vdc coupled by the .5 micro F to the grid of
6AQ5.

TIA
Vin

(snip)

No idea. *I'd just pick one of the 630 VDC caps to hand... orange drop
or the yellow ones. *I've never had any problem with these as plate to
grid coupling caps. *BTW, don't use ceramics in any audio coupling
capacity (no pun intended!)
Cheers,
Roger

Sorry, forgot to confirm... use 0.47uF or two 0.22uF caps in parallel,
or other such parallel combination - close enough. My guess is that
this one does not affect the oscillator frequency.
Cheers,
Roger

[Trailboss]

Thanks Roger



"Engineer" wrote in message
...

On Oct 2, 10:49 am, Engineer wrote:
On Oct 2, 9:33 am, "Trailboss" wrote: Patrick,

Words of wisdom, well said.

Would you have any idea why HP would use a 900 volt cap in this ckt?

Plate Voltage of the 6AU6 is 110vdc coupled by the .5 micro F to the
grid of
6AQ5.

TIA
Vin

(snip)

No idea. I'd just pick one of the 630 VDC caps to hand... orange drop
or the yellow ones. I've never had any problem with these as plate to
grid coupling caps. BTW, don't use ceramics in any audio coupling
capacity (no pun intended!)
Cheers,
Roger

Sorry, forgot to confirm... use 0.47uF or two 0.22uF caps in parallel,
or other such parallel combination - close enough. My guess is that
this one does not affect the oscillator frequency.
Cheers,
Roger

[John L Stewart]

Originally Posted by Trailboss

Thanks Roger



"Engineer" wrote in message
...

On Oct 2, 10:49 am, Engineer wrote:
On Oct 2, 9:33 am, "Trailboss" wrote: Patrick,

Words of wisdom, well said.

Would you have any idea why HP would use a 900 volt cap in this ckt?

Plate Voltage of the 6AU6 is 110vdc coupled by the .5 micro F to the
grid of
6AQ5.

TIA
Vin

(snip)

No idea. I'd just pick one of the 630 VDC caps to hand... orange drop
or the yellow ones. I've never had any problem with these as plate to
grid coupling caps. BTW, don't use ceramics in any audio coupling
capacity (no pun intended!)
Cheers,
Roger

Sorry, forgot to confirm... use 0.47uF or two 0.22uF caps in parallel,
or other such parallel combination - close enough. My guess is that
this one does not affect the oscillator frequency.
Cheers,
Roger

I've looked at the cct, that cap C6 has no bearing on the frequency generated. Anything close to 0.5 microfarad, 600 volt will work OK.

Frequency of this oscillater is determined by the R's & C's in the first stage Wein Bridge. Amplitude is held constant by Bill Hewlett's cunning use of an incandecent bulb as part of the FB cct.

Thats the patent that help found the Hewlett-Packard company.

Not Packard-Hewlett, was the other way round at the flip of a coin!

Cheers, John

[Alex Pogossov]

"Trailboss" wrote in message
...
Patrick,

Words of wisdom, well said.

Would you have any idea why HP would use a 900 volt cap in this ckt?

There is a line of thinking that if the rated voltage is higher, then the
leakage at lower voltage will be lower.
Possibly, it might be generally true for new paper caps, for old paper caps
leakage grows over the years irrespective of the rated voltage.

For modern plastic film caps the leakage is negligible. You do not need
excessive voltage rating for nothing. 630V is too much and makes the cap
unnecessarily large and expensive. Probably 250V rating is OK, depending on
the maximum voltages in the circuit (when all the tubes are pulled out,
except for the rectifier).


Plate Voltage of the 6AU6 is 110vdc coupled by the .5 micro F to the grid
of 6AQ5.

TIA
Vin

[Lord Valve]

Trailboss wrote:

Would you have any idea why HP would use a 900 volt cap in this ckt?

They got 'em cheap - on sale that week.

LV

[Trailboss]

HP and Tek definitely built stuff to last using higher rated components, but
I think Will is also correct in that they used what was available and cheap
at the time.

"Lord Valve" wrote in message ...

Trailboss wrote:

Would you have any idea why HP would use a 900 volt cap in this ckt?

They got 'em cheap - on sale that week.

LV

[Arny Krueger[_4_]]

"Trailboss" wrote in message
...
Looking for info on this coupling capacitor in the oscillator ckt.
Metal can Axial type
.5 µF
900 WVDC

Is this can an oil filled capacitor?

Would Polypro or Polystyrene type be a suitable replacement?

Any help appreciated.

I find no such part in the schematic in the HP 201 schematic located at

http://www.hparchive.com/Manuals/HP-201C-Manual.pdf .

What am I missing?

[Arny Krueger[_4_]]

"Arny Krueger" wrote in message
...

"Trailboss" wrote in message
...
Looking for info on this coupling capacitor in the oscillator ckt.
Metal can Axial type
.5 µF
900 WVDC

Is this can an oil filled capacitor?

Would Polypro or Polystyrene type be a suitable replacement?

Any help appreciated.

Generally, the capacitor types you propose are superior replacements for oil
filled capacitors in this kind of application.

I find no such part in the schematic in the HP 201 schematic located at

http://www.hparchive.com/Manuals/HP-201C-Manual.pdf .

What am I missing?

Whoops, there it is on the input to the 6AQ5 - C6.

Interesting that the online manual is missing section 6 and several
appendices.

[Arny Krueger[_4_]]

"Trailboss" wrote in message
...
HP and Tek definitely built stuff to last using higher rated components,
but I think Will is also correct in that they used what was available and
cheap at the time.

"Lord Valve" wrote in message ...

Trailboss wrote:

Would you have any idea why HP would use a 900 volt cap in this ckt?

They got 'em cheap - on sale that week.

Probably more like they used what they thought was the best at the time,
which was well before the ready availability of poly caps.

[Trailboss]

Hi Arny,

Yeah I was going to look over at boat anchor for another copy to see if they
are there.

Vin

"Arny Krueger" wrote in message
news

"Arny Krueger" wrote in message
...

"Trailboss" wrote in message
...
Looking for info on this coupling capacitor in the oscillator ckt.
Metal can Axial type
.5 µF
900 WVDC

Is this can an oil filled capacitor?

Would Polypro or Polystyrene type be a suitable replacement?

Any help appreciated.

Generally, the capacitor types you propose are superior replacements for oil
filled capacitors in this kind of application.

I find no such part in the schematic in the HP 201 schematic located at

http://www.hparchive.com/Manuals/HP-201C-Manual.pdf .

What am I missing?

Whoops, there it is on the input to the 6AQ5 - C6.

Interesting that the online manual is missing section 6 and several
appendices.

[Trailboss]

That 10 watt bulb providing FB for the oscillator, where can I source one, I
am not sure what voltage it should be, any help appreciated.

Vin

"Arny Krueger" wrote in message
news

"Arny Krueger" wrote in message
...

"Trailboss" wrote in message
...
Looking for info on this coupling capacitor in the oscillator ckt.
Metal can Axial type
.5 µF
900 WVDC

Is this can an oil filled capacitor?

Would Polypro or Polystyrene type be a suitable replacement?

Any help appreciated.

Generally, the capacitor types you propose are superior replacements for oil
filled capacitors in this kind of application.

I find no such part in the schematic in the HP 201 schematic located at

http://www.hparchive.com/Manuals/HP-201C-Manual.pdf .

What am I missing?

Whoops, there it is on the input to the 6AQ5 - C6.

Interesting that the online manual is missing section 6 and several
appendices.

[Patrick Turner]

On Oct 3, 12:33*am, "Trailboss" wrote:
Patrick,

Words of wisdom, well said.

Would you have any idea why HP would use a 900 volt cap in this ckt?

Plate Voltage of the 6AU6 is 110vdc coupled by the .5 micro F to the grid of
6AQ5.

TIA
Vin

Ah, HP oscillator. 400V rated rated polyprop or polyester would be
fine then; the cap is just a coupling cap and value isn't critical if
it is not used to determine any F, as it would be if there was a wien
bridge network for an oscillator.

HP might have had a cheap supply of 900V rated caps that year.

But usually companies never give anyone anything extra, like better
caps than then the mimimum which is needed. But I've often see 600V
ror 1,000V rated caps from anode to 0V in old radio audio amps to stop
them amplifying anything above about 8kHz.

Patrick Turner.

[Patrick Turner]

On Oct 4, 8:33*am, "Trailboss" wrote:
That 10 watt bulb providing FB for the oscillator, where can I source one, I
am not sure what voltage it should be, any help appreciated.


The light bulbs used to regulate the NFB or PFB may be quite difficult
to find now. The wien bridge NFB path network usually has a fixed
resistor in series with lamp.
R is from output of oscillator then join of R and lamp goes to cathode
of an input tube and lamp is then grounded.

When operating, the WARM resistance of the lamp = 1/2 the resistance
of the R, so the OP tube must be able to drive the NFB network, ie,
load should be maybe 3kohms, and the lamp is 1k0, when warm, but it
may only be 400 ohms when cold, and that's a fairly high resistance
lamp. Both positive FB and negative FB applied to the two inputs of
the gain amp will measure about 1/3 of the output, with PFB just
slightly more than NFB.

Sometimes you can substitute some low power lamps for 12.6Vac. I once
used 10 x "grain of wheat" tiniest 12V lamps I could buy, and the
added warm R of the lamps was about 1k. Even if the lamp is rated for
full brightness at 12V, you don't need anywhere near that much voltage
to just warm the filamant and have the lamp operating as a variable
resistance that responds well to temperature change caused by slight
voltage variations.

I've made lots of WB oscillators. Instead of a lamp, one might use a j-
fet or transistor and exploit the non linear properties, try googling
all ya can on WB oscillators.

My simplest best WB uses NE5534 opamp and does 2Hz to 200kHz, and at
1kHz THD = 0.0045%, at about 5Vrms, with rails at about +/- 12Vdc.
Most tubed WB amps have 0.5% THD, not so clean, because open loop gain
is low and huge 2H 3H THD reduction is not possible. Try to understand
what you build.

Patrick Turner.

[Arny Krueger[_4_]]

"Trailboss" wrote in message
...
That 10 watt bulb providing FB for the oscillator, where can I source one,
I am not sure what voltage it should be, any help appreciated.

Usually, this is a 120 volt bulb.

If you get it wrong, it will either not control the output level of the
oscillator, or LF distoriton will be too high.

[John L Stewart]

Originally Posted by Arny Krueger[_4_]

"Trailboss" wrote in message
...
That 10 watt bulb providing FB for the oscillator, where can I source one,
I am not sure what voltage it should be, any help appreciated.

Usually, this is a 120 volt bulb.

If you get it wrong, it will either not control the output level of the
oscillator, or LF distoriton will be too high.

The lamps used in my HP 200CD are shown as 250 volt, 10 watts (pn 2140-0007) in that service manual. The HP 201C was in production at about the same time so it makes sense to use the same lamp if possible.

The attachment is from a parts catalogue published in 1960, so the time frame looks good. I think RT11 is a 10S6/10-250V, bottom of the list. But don't expect to get one for that price today.

[Arny Krueger[_4_]]

"John L Stewart" wrote in message
...

'Arny Krueger[_4_ Wrote:
;939854']"Trailboss" wrote in message
...-
That 10 watt bulb providing FB for the oscillator, where can I source
one,
I am not sure what voltage it should be, any help appreciated.-

Usually, this is a 120 volt bulb.

If you get it wrong, it will either not control the output level of the

oscillator, or LF distortion will be too high.

The lamps used in my HP 200CD are shown as 250 volt, 10 watts (pn
2140-0007) in that service manual. The HP 201C was in production at
about the same time so it makes sense to use the same lamp if possible.

The attachment is from a parts catalogue published in 1960, so the time
frame looks good. I think RT11 is a 10S6/10-250V, bottom of the list.
But don't expect to get one for that price today.


+-------------------------------------------------------------------+
|Filename: Candelabra Screw Base Lamps.jpg |
|Download: http://www.audiobanter.com/attachment.php?attachmentid=249|
+-------------------------------------------------------------------+


The area around the light bulb in the 200 schematic at

http://www.hparchive.com/Manuals/HP-...anual-1951.pdf

looks very similar to the coresponding area of the 201. Therefore, there is
as you suggest, a good chance that the light bulbs are the same - 10 watt
250 volt.

I forgot one of the other indications that you have the wrong light bulb -
the waveform has a sine wave envelope imposed on it.

I just did a search on 10 watt 250 volt bulbs and found that they are
available for such a low price that you might order not only the 250 volt
bulb but also others with similar wattage and voltage including 120 volt
types - the shipping fee would otherwise be much greater than the cost of
the actual bulb. http://www.1000bulbs.com/product/77/...0S06CC250.html
as an example.

[Trailboss]

Hi John,

Just pulled a schematic off of BAMA for an HP 200i. Should have done this a
long time ago, it answered a multitude of my inane dumb questions, there is
so much more tech info included in this manual. Including the manufacture
reference of the incandescent bulb used for FB. It is a GE 10S6 250v.

Also that coupling cap from the oscilator to the gain stage, HP part made by
Sangamo Electric point five mic 400 vdcw 10%.
Looks like Will was right on, must have been a fire sale on 900 vdcw caps
that week,

thanks to all,

Vin







...-
That 10 watt bulb providing FB for the oscillator, where can I source
one,
I am not sure what voltage it should be, any help appreciated.-

Usually, this is a 120 volt bulb.

If you get it wrong, it will either not control the output level of the

oscillator, or LF distoriton will be too high.

The lamps used in my HP 200CD are shown as 250 volt, 10 watts (pn
2140-0007) in that service manual. The HP 201C was in production at
about the same time so it makes sense to use the same lamp if possible.

The attachment is from a parts catalogue published in 1960, so the time
frame looks good. I think RT11 is a 10S6/10-250V, bottom of the list.
But don't expect to get one for that price today.


+-------------------------------------------------------------------+
|Filename: Candelabra Screw Base Lamps.jpg |
|Download: http://www.audiobanter.com/attachment.php?attachmentid=249|
+-------------------------------------------------------------------+



--
John L Stewart

[Trailboss]

Hi John,

Thanks for the info on the 10S6 lamp.


The attachment is from a parts catalogue published i51n 1960, so the time
frame looks good. I think RT11 is a 10S6/10-250V, bottom of the list.
But don't expect to get one for that price today.


I found 'em for $1.11 ea or in quantity 74 cent!!! LOL
Not much more than the 51 cent back in the fifties,,,,


http://www.1000bulbs.com/product/77/...0S06CC250.html

Thanks John,

Vin


"John L Stewart" wrote in message
...


'Arny Krueger[_4_ Wrote:
;939854']"Trailboss" wrote in message
...-
That 10 watt bulb providing FB for the oscillator, where can I source
one,
I am not sure what voltage it should be, any help appreciated.-

Usually, this is a 120 volt bulb.

If you get it wrong, it will either not control the output level of the

oscillator, or LF distoriton will be too high.

The lamps used in my HP 200CD are shown as 250 volt, 10 watts (pn
2140-0007) in that service manual. The HP 201C was in production at
about the same time so it makes sense to use the same lamp if possible.

The attachment is from a parts catalogue published in 1960, so the time
frame looks good. I think RT11 is a 10S6/10-250V, bottom of the list.
But don't expect to get one for that price today.


+-------------------------------------------------------------------+
|Filename: Candelabra Screw Base Lamps.jpg |
|Download: http://www.audiobanter.com/attachment.php?attachmentid=249|
+-------------------------------------------------------------------+



--
John L Stewart

[Patrick Turner]

On Oct 5, 12:35*pm, "Trailboss" wrote:
Hi John,

Thanks for the info on the 10S6 lamp.

The attachment is from a parts catalogue published i51n 1960, so the time
frame looks good. I think RT11 is a 10S6/10-250V, bottom of the list.
But don't expect to get one for that price today.

I found 'em for $1.11 ea or in quantity 74 cent!!! *LOL
Not much more than the 51 cent back in the fifties,,,,

http://www.1000bulbs.com/product/77/...0S06CC250.html

Thanks John,

Vin

"John L Stewart" *wrote in ...

'Arny Krueger[_4_ Wrote:

;939854']"Trailboss" wrote in message
...-
That 10 watt bulb providing FB for the oscillator, where can I source
one,
I am not sure what voltage it should be, any help appreciated.-

Usually, this is a 120 volt bulb.

If you get it wrong, it will either not control the output level of the

oscillator, or LF distoriton will be too high.

The lamps used in my HP 200CD are shown as 250 volt, 10 watts (pn
2140-0007) in that service manual. The HP 201C was in production at
about the same time so it makes sense to use the same lamp if possible.

The attachment is from a parts catalogue published in 1960, so the time
frame looks good. I think RT11 is a 10S6/10-250V, bottom of the list.
But don't expect to get one for that price today.

A 10Watt x 250V lamp draws 40mA, and its 'on' resistance when fully
bright is 6k3 ohms.

I recall 240V x 10W globes and 15W globes were easily available here,
often for a fridge light that turned on when the door was opened.

But when its cold the R may be 1k0, and with say 5Vrms applied by the
NFB network the working warmer R will only be slightly more than cold
R.
The value of working R will be 1/2 the value of the other R used in
the NFB network. There has to be some considerable change in R lest
ambient T ranges tends to affect globe R too much.
I recall measuring the R of many lamp globes with a variac and series
10 ohms current sensing R while searching for an ideal type of globe
to use. I do recall a certain 25V type was good, but most I measured
were not, unless I put a number of globes in series to get the R up so
the NFB didn't load the tube driver too much

If you have a 6AQ5 or 6V6 output tube, and Idle Ia was say 35mA, then
with RL of NFB = 3k0, then to get low THD the Ia change wants to be
less than +/-10mA peak which is 21Vrms into 3k0.

Japanese Kikusui oscillators of the same vintage as the HP had
thermistors mounted in what looks like a lamp. This meant that there
was no air to cool the thermistor so it was sensitive to feeble
current changes produced on an OP tube like 6AQ5. Such thermistors may
not have been made for the last 40 years.
It was supposed to work better than a tungsten filamant lamp. Most
thermistors one might buy now to use instead of a lamp have operating
R way too low like most lamps.
In theory, if a pair of mosfets are used as a direct coupled
complentary follower pair as a buffer after the tube, then the NFB
load R can be much lower, and the PFB network also can have lower R so
F will go higher. But trying to get a WB oscillator to go to 2Mhz
with 2 tubes is a big ask, and with only one set of dial numbers,
so that with six ranges 2-20, 20-200, 200-2k, 2k-20k 20k-200k, 200k-2M
is possible, and you only have to switch the R and the tuning gang
does the rest.

I have a spare 4 gang tuner from a HP333A THD meter, and total C for
each gang = 500pF, so I should be able to make a very nice WB
oscillator, time permitting.

I'm thinking of an LC oscillator for RF with level controlled by a
lamp but I'm not sure how well the idea .may work.

I've too much work to try to make more oscillators than I have
already.

Patrick Turner.

[John L Stewart]

For the curious, here are some actual numbers relating E & I for a couple of these bulbs. I made these measurements about 10 years ago as part of an article demonstrating the CC Source as a useful tool in the workshop, in particular for non-linear devices. You can read about it in the July 2001 issue of AudioXpress magazine.

For the 6S6/250V as used in the HP 200CD
3.32 ma results in 2.31 volts or 0.696 K
6.24 ma results in 6.73 volts or 1.079 K
9.92 ma results in 15.73 volts or 1.586 K
11.66 ma results in 20.9 volts or 1.792 K

For the 6S6/24V
3.37 ma results in 3.37 mv or 9.555 R
10.2 ma results in 102.4 mv or 10.039 R
19.3 ma results in 242 mv or 12.539 R
30.2 ma results in 603 mv or 19.967 R
94.8 ma results in 4.94 volts or 52.11 R
178 ma results in 14.26 volts or 80.112 R

Looks to me like the 24 volt lamp is not a good choice for a vacuum tube oscillator.

Cheers to all, John

[Arny Krueger[_4_]]

"Patrick Turner" wrote in message
...

Japanese Kikusui oscillators of the same vintage as the HP had
thermistors mounted in what looks like a lamp. This meant that there
was no air to cool the thermistor so it was sensitive to feeble
current changes produced on an OP tube like 6AQ5. Such thermistors may
not have been made for the last 40 years.

About 20 years ago I built a number of 5534-based wein bridge audio
generators that used thermistors that I ordered right out the Mouser
catalog, if memory served.

I also built auidio signal generators that used light bulbs, optoisolators,
FETs and multiplier chips to stabilize them. I also experimented with active
servo loops with various time constants. I also experimented with other op
amps, particularly those that were suited for operation at higher supply
voltages so that higher output voltages were possible. My goal was 10 vrms
into a 600 ohm load.

It was supposed to work better than a tungsten filamant lamp. Most
thermistors one might buy now to use instead of a lamp have operating
R way too low like most lamps.

In any of these schemes, one often finds that the stabilizer is too
effective all by itself and will cause less low frequency nonlinear
distoriton if padded with a series resistor or its equivalent. The boundary
conditions are amplitude modulation, low frequency nonlinear distoriton, and
poor flatness over the audio range.

I used Heathkit IG-18 and 5218 kits as breadboards for my experiments
because they contain a very nice switch-controlled Wein bridge.

[Stewart28]

You say 0.5f, which is 0.5 Farad, which is a huge ceiling.
How much capacity you really? This is a 0.5uF or 0.5
nF ???????????

Polypropylene or polyester is a great substitute for all the time
capacity value is the same and the voltage is the same,
but only if the capacity of the original CAP and V assessments were correct.

[Patrick Turner]

On Oct 7, 1:03*am, John L Stewart John.L.Stewart.
wrote:
For the curious, here are some actual numbers relating E & I for a
couple of these bulbs. I made these measurements about 10 years ago as
part of an article demonstrating the CC Source as a useful tool in the
workshop, in particular for non-linear devices. You can read about it in
the July 2001 issue of AudioXpress magazine.

For the 6S6/250V as used in the HP 200CD
3.32 ma results in 2.31 volts or 0.696 K
6.24 ma results in 6.73 volts or 1.079 K
9.92 ma results in 15.73 volts or 1.586 K
11.66 ma results in 20.9 volts or 1.792 K

These results underline what I was saying about the NFB network
loading resistance needed with a 6AQ5. The NFB network resistance
needs to be a load of thousands of ohms, not hundreds of ohms.
There is plenty enough variation in abobe lamp R to give good
regulation of the oscillator output when Vac across the lamp is
between 3Vrms and 6Vrms when the anode output signal is then between
9V and 18V.

By analysis of the HP200CD oscillator circuit and its NFB network R
values, the R value of the lamp while operating can be calculated.

I recall using 10 x "grain of wheat" lamps used in model train sets
for a suitable NFB R and with DC of the input tube flowing through it,
a slightly complex set up to get the Idc to heat the damn lamps up a
bit to put get their R up a bit and to into a region where the R rise
for V rise was most significant.

For the 6S6/24V
3.37 ma results in 3.37 mv or 9.555 R
10.2 ma results in 102.4 mv or 10.039 R
19.3 ma results in 242 mv or 12.539 R
30.2 ma results in 603 mv or 19.967 R
94.8 ma results in 4.94 volts or 52.11 R
178 ma results in 14.26 volts or 80.112 R

Looks to me like the 24 volt lamp is not a good choice for a vacuum tube
oscillator.

Agreed. Maybe OK if had a BJT or mosfet output stage, and then you may
as well have an all SS normal type amp with high internal gain like an
opamp, and direct coupling et all, and with say +/- 50V rails, amybe
working in class A. OK because the load is not wanted to be less than
600 ohms. For 50V rails, swing max = 30Vpk, or 21Vrms, and that's 35mA
rms into 600 ohms, 0.75Watts.

But the amp used for any oscillator must have high open loop bandwidth
if you really want a good Wien Bridge oscillator which goes to 2Mhz.
Most simple CR coupled tube voltage amps with only pure R loading and
only 2 pentode/beam-tetrode stages battle to get to 1Mhz. I've tried
all this some 16 years ago when I built a balanced voltage amp which
used :-

Input = 6DJ8 paralleled for concertina phase inverter; it gets the
bandwidth 2MHz+
Driver to each 2 stage cascaded amp = E280F sq high gm pentode. 6k8
load. It just gives enough gain at 2MHz.
Outputs = 6CM5, -17V fixed bias, class A, Ia = 90mA, RL = 2k0 on
heatsink taken to +400V. Eg2 = 150V and regged. Vo max at 1Mhz =
100Vrms.
NFB = 9k plus Rk of V1, from V2 anode V1 cathode with output electro
couping cap, 47uF. Rout = 600 ohms.
The two amps give the two phases of 100Vrms from 2Hz to 1MHz, -3dB,
good enough to test most OPTs for their characteristics without load,
or woth load at low levels.

The only time I got a WB oscillator to get to a high F was with 6CM5
output plus a pentode input, so only 2 tubes as 3 will make things too
slow. I got up to 800kHz, but the scale for the HF had to be a
separate one.

It is easier to make the WB only have to go to 200kHz.

Then if one wants say 200kHz to 2MHz, then make an RF type oscillator
based around tunes circuits which can be easily tuned with a radio
tuning gang anywhere between 200kHz and 20Mhz. To make square waves at
HF a biased diode limiter may be used to feed say 13Vpk into a limiter
network which alows only 5Vpk output. Its not a perfect square but its
good enough for most apps, because the odd H present will be up to
many times Fo. The Kikusui oscillator which I completely re-designed
has this biased diode limiter set up to the peak square wave at any F
= peak sine wave level. Goes to 200kHz. Also has a mix of 4:1 for 50Hz
+ any HF signal, so IMD might be checked. But its THD is typically
high at about 0.5% like many tubed oscillators which just don't have
enough open loop gain and effective applied NFB to reduce the THD.

I do have a little wide band amp which does 16Hz to 30Mhz using 2 x
2N2222 in cascade, makes only a few V and gain = 20. Nice and quiet,
and good use with my mV meter. Does more FLAT bandwidth than most
opamps.

The other thing to remember is that the WB NFB works to reduce THD
just like any other series voltage NFB loop around any amp. But the
positive FB network also delivers the 2H and 3H back to the PFB input
port, so any 2H at the amp output is applied in to both PFB and NFB
ports at the amp input. The 2H and 3H are at readuced level and are
slightly out of phase, but this partial common mode of THD delivery to
both ports makes it impossible to reduce the 2H and 3H very much, and
anyone examining the response of a WB network will see why because the
Q is so low of such an RC circuit designed to give a peaked response
like a tuned circuit.
An opamp might have its maximum open loop gain of 100,000 at 100Hz,
many times anything achievable with a pair of tubes. So despite the
common mode application of THD by the PFB and NFB loops, the possible
THD reduction with say a NE5534 is quite pleasing, and at 1khz, THD
0.005% is quite possible. Just don't expect the same result at 10kHz
of 200kHz when open loop gain has become much lower.

If one used a whole bunch of switched RC tank circuits for the
oscillator, then you'd get purer tones generated. A number of
electronic organs used a 12AX7 plus slug tuned L and fixed Cs, one LC
for each note, and for a whole keyboard fill of notes there were a
lotta LC circuits. Such an instrument could be made to make a very
nice oscillator with switched F based on the F of the musical scale.
That's 12 F per octave.

In one WB oscillator with opamps I get by with a switched oscillator
with F at multiples of 2, 1.2, 1.6, 2.0. 2.5. 3.2, 4.0, 5.0. 6.0. 7.0.
8.5, 10.0. This means a 12 position wafer switch may be used, and I
had lots I took from defunct tube testers made in the 1930s and 40s.
Ranges are 2Hz to 200kHz. Dial is accurate for all switch pos.
The switches change the R value much more accurately and hence change
F more reliably than most logarithmic pots, and without amplitude
bounce because I have a limiter fitted to "catch" excessive Vo, before
it has a chance to overheat the NFB element.

The Kikusui oscillator has a dual tuning gang which I find is also far
more reliable than most pots. It has nicely variable F, but can't do
below 20Hz because the R value required becomes too large with only
such a small amount of C. One might use a pot and fixed C for LF, but
sooner or later the pot will become unreliable, you begin to swear at
the gear, and you must chuck the pot out!

Patrick Turner.