I looked up the 723, and the maximum suggested resistors are 10K. The
National data sheet doesn't give the input bias currents, but judging by
the time the 723 has been around, bias currents and offsets were not as
well controlled as in today's processes. Thus using 100K resisitors
could cause problems - only though if you're designing for production,
rather than as a one off. I guess you could use a 555 comparator if you
had a good stable supply voltage to give the reference, (eg a 78L08 or
something) and if you used a dual 555, the other half could be used for
the filament delay. This would give the advantage that the exponential
rise in voltage on the capacitor charge when first turned on would give
the filament delay, while the discharge time would be proportional to
filament off time. Thus the delay after an interruption would be
proportional to the length of the interruption - with about the same
exponential curve of the cathode time constant. This again could be
useful if you were an amplifier to meet the IEC specs on EMC - the
sections on power interrupts.
Rich says:
>Rich's design philosophy: There is ALWAYS a performance trade-off to
>making a circuit more stable.
> - For instance, the c.100k ohm input resistors undoubtedly decrease
the
>regulation ability of the 723, however, after the 723 was made
>RF-resistant, I performed a load regulation test. With the output set
>for 1100v, a 250mA load was placed on the screen supply. The output v
>under load was 1100v. (measurement taken with John Fluke 8022A DMM)
>Sure, a performance drop would no doubt have been detected if a 4 1/2
>digit DMM had been used instead of a 3 1/2 digit DMM--but for a AB1
>screen supply, would it matter, Peter?
The 100K input resistors will not affect the regulation, but will affect
the thermal stability of the trip point because of drifting offsets and
bias currents. The frequency of the RF field doesn't affect it until you
get to the point where the junctions in the 723 won't rectify. I did
some tests to determine whether the problem was RF picked up on the
leads to the device (conducted susceptability) or the effect of radiated
RF (radiated susceptability). The results came down in favour of a
radiated susceptability problem with 723s, which is manufacturer
dependent - as you would expect, because of the variations in the
process that different people use to make 723s. The other thing I have
gainst 723s is that the reference uses a zener, rather than a band gap,
which is inherently less process dependent, more stable, and according
to some, more inherently reliable.
> - For instance #2: The unavoidable trade-off with the use of high
>VHF-Rp/low VHF-Q parasitic suppressors is a roughly 1% decrease in
29MHz
>power output. Is such a reduction in output just below the VHF range
>worth the price? I have a collection of kaput gold-sputtered 8877s and
>grid-fil. shorted 3-500Zs which suggest to me that it is.
> - Rich -
Are you sure that it wasn't the 'Rocky Point effect', rather than
parasitics?
73
Peter G3RZP
--
FAQ on WWW: http://www.contesting.com/ampfaq.html
Submissions: amps@contesting.com
Administrative requests: amps-REQUEST@contesting.com
Problems: owner-amps@contesting.com
Search: http://www.contesting.com/km9p/search.htm
|