From the datasheets which I have in my yellow Eimac binder, collected
over the years, there appear to be some discrepancies which could
have been typographical mistakes or design changes. It looks like
there is been some changes over the years with what they are calling
feedthru capacitance in GG operation, where the older datasheets
always called it Cgp. This usually was the value measured in a GK
fixture, with the numbers switched for Cgp and Cout. The most recent
show it as Cpk, which is the acutual capacitance from output to input
for a grounded grid mode. In all cases, the tubes were measured using
EIA standard RS191 techniques in carefully shielded fixtures that
don't add appreciably to the values. Eimac did measure some of their
larger tubes without the shielded fixtures, which are reported as
Nominal capacitance, measured without shielded fixtures. I am not
sure which measurements are more useful for circuit designers using
their tubes.
The other cold capacitance parameter which changed is for the various
data revisions for the 3CX1200A7, the 3CX1200D7/YU-121, and for the
3CX1200Z7/YU-181. In some cases the data was measured for GK, while
the tube was touted to be excellent for GG operation. It was
important to realize that before designing. Of course, the great
advantage of GG is that the shielding of the control grid will reduce
the feedthru capacitance from output to input (from P-K) to a lower
value, and the output capacitance appears higher as it is dominated
by Cgp. The DC values of feethru capacitance don't include the
additional factor of amplication, from the Miller effect, which could
raise them slightly.
Here are the examples I found---------
The original 3CX1200A7 tentative datahsheet that I got from the Eimac
saleman back in 1984 states:
Cin=20 pf
Cout=0.2
Cgp=12
This is assumed to be in grounded grid operation, as the tube was
being touted as a replacement for a pair of 3-500Zs at that time. A
later datasheet dated 3/15/89 states:
grounded cathode (GK)-
Cin=20
Cout=0.6
Cgp=10.3
grounded grid (GG)-
Cin=20
Cout=10.3
Cgp=0.6
Cgp is specified as 0.9 in the same data sheet under range for equp
design specs
The data which is presently on the CPI/Eimac website for this tube is
only specified for GK:
Cin=20
Cout=0.2
feedthru=12
--------------------------
The 3CX1200D7/YU-121 datasheet dated 3/15/89 states:
Cin=17.6
Cout=0.4
Cgp=9.7
all estimated values at that time
The data which is presently on the CPI/Eimac website for this tube is:
Cin=17.6
Cout=0.57
feedthru=9.7
-------------------------
The 3CX1200Z7/YU-181 is strictly for GG operation, with a bolted grid
flange. Datasheet (3/91) states:
Cin=17.6
Cout=0.08
Cgp=9.7
estimated values
The data which is presently on the CPI/Eimac website for this tube is:
Cin=17.6
Cout=9.7
feedthru=0.8
------------------------
The 3CX1500A7/8877 by comparison, datasheet (3/91) states:
GK-
Cin=38.5
Cout=0.1
Cgp=10
GG-
Cin=38.5
Cout=10
Cgp=0.1
A later datasheet for the same tube (undated) states:
GK-
Cin=38.5
Cout=0.1
Cgp=10
GG-
Cin=38.5
Cout=10
Cpk=0.1
The data on the website is substantially the same except for 10 being 10.2 pf
-----------------------
The 3CX1500D7 was later recommended as a replacement for a pair of
3-500Z, and is specified as:
GG-
Cin=18.6
Cout=7.2
Cpk=0.4
The data on the website is the same.
-----------------------
For comparison, the old 3-500Z datasheet () states:
GK-
Cin=8.3
Cout=0.07
Cgp=4.7
GG-
Cin=8.3
Cout=4.7
Cgp=0.07
Cgp is specified at 0.18 in the same datasheet under range of equip
design specs
----------------------
And finally, the 3-1000Z datasheet (8/15/74) states:
GK-
Cin=17
Cout=0.2
Cgp=7.5
GG-
Cin=17
Cout=7.5
Cgp=0.2
---------------------
To summarize, the printed values of C(feedthu) in pf are tabularized for the
various tubes in GG configuration. There is quite a range described:
3-500Z 3-1000Z 3CX1200A7 3CX1200D7 3CX1200Z7 3CX1500A7 3CX1500D7
0.07 0.2 0.2-0.9 0.4-0.57 0.08-0.8
0.1 0.4
I just wanted to list all of this because of the confusion which it
brings when hams talk about the different ceramic metal triodes for
grounded grid amplifiers. The 3CX1500A7/8877 is clearly a low feedthu
tube, and can be run up to very high frequencies with ease,
especially if the 8938 version is used (without pins on the socket).
In addition it is a short tube with low inductance connections. It
also costs a lot. The old 3-500Z was a great tube, based on the
printed data. Two would have 0.14 pf of Cgp. The 1200A7 and D7 had a
lot more feedthru, making the Z7 preferable if that was the parameter
you wanted to keep low. Some of these parameters would obviously be
more or less important in other applications, such as ISM industrial
oscillator/amplifiers. One thing to keep in mind is that the cold
capacitances do not give you the lead inductances for these tubes.
The effect of series L in the connections (esp for screen, grid and
cathode) makes a big difference in stability, gain-bandwidth product,
and efficiency that tubes will exhibit in circuits. So, knowing the
feedback or feedthru capacitance is only part of the knowledge
required to built good amplifiers. While this value is not specified
on CPI/Eimac data sheets, it is implied indirectly with the Max rated
frequency. For example, the 4CX5000A (110 Mhz max) is demonstrably
worse than a 4CX3500A (220 MHz) at 50-100 MHz in the same socket and
cavity circuit. If special tweaks are made, such as resonating the
inductances at the operating frequency, the older tube can be made
to work welll - at the expense of frequency agility.
Some companies, such as Burle Industries (RCA), give additional
values on their datasheets for their VHF and UHF tubes. These are the
strap resonance measurements, where a short cavity is put on the tube
and the first 1/4 wave resonance is measured for the input and
output. From this one can deduce what sort of L and C parameters they
might use in a circuit model. RCA also would print the exact cavity
line lengths and diameters for the 1/4 and 3/4 transmission line
modes for some of their Cermelox tetrodes. These sort of measurements
were not standardized by EIA to my knowledge, but they still give the
HF and VHF designer a lot more insight into the internal design of
the tubes. As the operating frequency and power is raised, new
parasitic modes in tubes appear, such as circumferential resonances
due to the waveguide dimensions around the tube, between the grids
and anode. This just adds more fun to the designers kit of practices
and real world experiences, stuff not seen with traditional computer
models.
I guess I should not digress further on this, but you can see that
feedback capacitance isn't the only important parameter when
designing PAs with tubes.
73
John
K5PRO
>Date: Mon, 29 May 2006 09:51:36 +0000
>From: Chris Pedder <chris@g3vbl.co.uk>
>Subject: Re: [Amps] 4-1000 amp
>To: R L Measures <r@somis.org>, amps@contesting.com
>Message-ID: <7.0.1.0.2.20060529094317.039b7c50@g3vbl.co.uk>
>Content-Type: text/plain; charset="us-ascii"; format=flowed
>
>At 09:29 29/05/2006, R.L. Measures wrote:
>
>> The 3-1000Z has a grid that does a marginal job of shielding the
>>input from the output -- which means it has more feedback-C than need
>>be. The 3cx1200A7 is the external anode version of the 3-1000Z and
>>it apparently uses the same grid and filament. Both tubes have a
>>reputation for squirreliness -- especially with higher anode PS
>>voltages. Eventually, Eimac discontinued making the 3-1000Z and the
>>3cx1200A7 evolved into the 3cx1200Z7, which has less feedback-C --
>>and better stability. Unfortunately, the ...A7 and the ...Z7 are not
>>interchangable.
>
>I cannot help but wonder why you keep pedalling this one. I have
>been, yet again, to the Eimac/CPI data sheets which read as follows:
>
>
>grounded grid capacitance in pf
>
> in out feedback
>3CX1200A7 20 10.3 0.6
>3CX1200Z7 17.6 9.7 0.8
>
>This does not accord with your comments. Are the Eimac figures incorrect?
>
>Chris
>
>
>Yes. On other sheets the Z7 has 0.08 pF, not 0.8pF. Does it make
>sense that the improved version would have more feedback?
>R L MEASURES, AG6K. 805-386-3734
>r@somis.org
>
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
|