On Jun 2, 12:00*am, John Byrns wrote:
In article ,
*Bret Ludwig wrote:
*Saul Marantz (yes, I know) was determined to make the finest quality
hi-fi equipment and spared little in the quest to do so. His build
cost was double and then some of what his nearest competitor,
McIntosh, spent per unit.
I've never wanted anything made by Marantz, could always have been a
lot better.
*Have you ever seen a marantz 2, 5 or 8B amplifier or a 9 or 9C
preamp???????????
I owned a Marantz 8B in the day, I would have thought it would have received
high marks from Patrick, at least as high he would give any commercial product. *
It even uses the sort of circuit he espouses, I'm not sure if the OPT would be
up to his standards though.
I never heard of the Marantz 9 or 9C?
--
Regards,
John Byrns
Surf my web pages at, *http://fmamradios.com/
I had a look at all Marantz models 5 to 9.
They do look well on paper with a 6BH6 ( like a 6AU6 ) in triode for
V1, then a 6CG7 direct coupled LTP like Mullard does, with EL34
outputs in UL. Not too many problems with all that but the anode
loads for 6CG7 are way too low, and consist of 5k pot for AC
balancing.with 15k to each anode when it really should be 47k The
obvious mod to make it all work better is to have CR coupling from V1
SET 6BH6 to one 6CG7 grid biased for 0V, then have the other grid
grounded, then have a negative rail of say -200Vdc to -400Vdc for the
common Rk, and then get the B+ to be higher and then you have 6CG7 Ea
and Ia higher with RL higher and maybe the THD of the stage is much
reduced. A CCS LTP tail taken to say -50V would be best though because
the pot for AC balance can be omitted if the RL for each anode is
equal which is easy to do these days.
The Marantz OPT takes most of its LF NFB from the 4 ohm tap on the
sec. But there is an additional NFB winding with its live end connectd
to the 16 ohm tap via 10 ohms, and the NFB winding feeds HF NFB to the
V1 cathode via 1.2nF. I'd guess the NFB winding does not experience
the same phase shift caused by LL and the load at HF, so Marantz
manages to apply NFB over a wide BW without trying to fix the phase
lag at HF. But without having a Marantz here to examine, I cannot say
exactly how well the arrangement works, or understand exactly how
Marantz wound its OPTs. How can I endorse something I have never
examined properly? Impossible. But if Marantz had used a 5P x 4S
interleaving pattern with 6 secondary sections each for 16, 8 &4 ohms
and all paralleled, there would have been no need for a NFB winding
because such interleaving with paralleled secs give wide BW for this
sort of amplifier as Williamson and McIntosh have so so clearly
established over the last 60 years. Tapped secondaries usually have
higher winding resistance than where 4 sec sections are strapped in
different patterns for the load match variations. Most manufactuers
have used appallingly thin wire in OPT secs and rarely ever have 4 Sec
sections; two if you are lucky, and often the two will each be a match
for 4 ohms, so that the two series S sections give 16 ohms, a tap
along one section gives 8 ohms, and only one section is then used for
the 4 ohm match and it is with 4 ohms that there is the highest LL and
worst BW and highest instability and highest winding losses. As i
said, for good performance for a 50W OPT, you'll need 5P x 4S, and
each Sec section should be a match for 16 and so with 4 ohms you still
have 5P x 4S interleaving, even though 1/2 the sec turns are unused.
The strapped method of changing Z match on OPTs is often done poorly.
For example, in Quad-II there are only two link changes needed for 8
or 16 ohms, but no official match for 4 ohms. Walker must have
pondered how to set up the OPT and its strapping method long and hard
to satisfy a Nation Full of Dumbos who are confused if forced to count
beyond 3. Quad-II would be a MUCH better amp if it were possible to
strap all OPT secs to get a wasteless arrangement of secs so that
every length of secondary wire had the same current density.
Then the Quad-II would have allowed a P-S load match of 4k0 : 4 ohms
which would better suit most modern LS, and of course using 8 ohms on
the 4 ohm outlet gives 8k0 : 8 ohms and nearly all class A which suits
the old junk circuit best. See why I say old junk is OLD JUNK?
Its because I can see that the old gizas who designed all that old
junk *could* have made masterpeices with an extra 2 hour's thought and
4 hour's more labour. But alas, the designers were addicted to
mediocrity.
One more thought about LTPs.
I find LTP with CCS cathode tail and high RL values and lotsa Ia with
a 6CG7 or a pair of EL84, EL86 etc, are good. Consider the operation
of any LTP with a 6SN7/6CG7 low µ triode pair and equal + and -
30Vrms outputs, and gain of each triode = 15. With one grounded grid,
you need -2Vrms input to the live grid you get -1V at the common
cathodes. So the Va across the input tube is 31Vrms and Va across the
grounded grid tube it is 29Vrms, and so there is slight imbalance of
Ea and therefore of the 2H and the stage will produce 2H even if both
triodes are really identical. But balance is fair, and the 2H is
usually negligible. Maybe it cancels with 2H from the SET input stage,
or with some slight 2H in the OP stage, and maybe it don't, but PP
amps all seem to produce *some* 2H.
But the main THD from an LTP is the 3H, which adds to that produced in
the OP stage unless the OP stage 3H has oppositely phased 3H which
rarely ever occurs, so one should be keen to reduce 3H anywhere it is
produced, while also minimizing 2H anywhere.
To get the best Va balance across each triode in an LTP the pair
should have balanced drive to each grid, as it was done with
Williamson's balance drive amp. Now Williamson used SET input + SET
concertina, but he could have used the same 2 triodes as an LTP input
stage which would have largle eliminated any 2H from teh SET input
stage. Alas he didn't have a decent transistor to bung in as the tail
for the input stage CCS. But where the input LTP only has to produce
about a volt from each anode to drive the following balanced amp it
doesn't matter about the Va imbalance and 2H because this 2H will bea
microscopic amount. If the balanced amp as Willy made it has its
common cathode R increased to about 4k7, then the stage self balances
far better and gives less THD. In recent years I have used the LTP
with CCS forthe first stage with input to one grid and GNFB to the
other grid and then had a pair of EL84 in triode in a balanced amp
with a common Rk of at least 4k7 to -100V, and the dynamics and
clarity of music became astounding. Each EL84 in triode with Ia = 10mA
does about the same job with same gain as 4 triode sections of
6SN7/6CG7, or is like 4 x 6J5 in parallel. But the Ra of the EL84 is
only 2k2 at ia = 10mA so any Miller C in the OP stage is handled
better. In Ming-Da I have here with PP 845, there are a pair of 300B
in LTP for the driver amp. It sounds fantabulous. Old dead bean
counters are spinning in their graves at what Chinese sometimes do to
make their amps spectacular looking. Unfortunately, the Chinese often
turn music time into a spectacular event with thunder and flames, and
the repairman has to find where the Cracker was hidden, ready to go
off when conditions would soon allow. But once you remove Crackers
from Ming Da, wow, PP 845 are really something. How can I just
consider British and American trends of the past without thought of
the present, and the the Chinese? Not easily.
Patrick Turner.