ARC VT100 repair and test.
- From: Patrick Turner <info@xxxxxxxxxxxxxxxxxx>
- Date: Wed, 28 Jun 2006 01:41:33 GMT
A week ago I said i would report after testing the ARC VT100 amp I
repaired.
A report on what experiences i had and my recomendations are included :-
There were a few problems with the amp.
Each VT100 channel has 5 solid state constant current sources each using
a TO92 package depletion mosfet for cathode current supplies; this does
not make the amp a hybrid because all the signal gain action is handled
solely by the 8 triodes in the input/driver stages.
One channel had two shorted mosfets which were probably because of the
location of the mosfets on small heatsinks close to the tubes where the
heatsink would heat up the mosfet rather than cool it. Placing 2 new
mosfets on the side of the board opposite to the tubes in the left
channel had them running cool as a cucumber so I doubt heat stress will
ruin them again.
The left channel then seemed to to work fine when the mosfet current was
adjusted in the two
new mosfets which were generics from Mouser.
The amp had been sent to me with a brand new set of Ei 6DJ8 plugged in.
While testing the right channel, I found that I could not adjust each
pair
of anodes on the input pair so that I could get +60V for each side of
the
cross coupled input LTP.
After replacing the two Ei 6DJ8 one at a time I found one Ei to have one
triode section that tended to become saturated. A 6DJ8 was running away
and unable
to react to its cathode bias.
But once the crook tube was replaced the right channel ran fine, The
mosfets were all fine.
But while checking the right channel two of the left channel ouput tubes
started to
glow red hot and I turned the amp off fast.
I removed the output tubes to check the applied grid bias to each tube
which comes from the cathode follower buffers on each side of the PP
circuit.
Sure enough, after 5 minutes one grid voltage began to climb above the
-45V
normal voltage and within a minute was +20V. That explained the two hot
tubes.
But I thought perhaps a coupling cap to the CF grids was leaky, but
after
removing the CF tube, another Ei 6DJ8, there was no rise in the applied
grid voltage to the buffer stages.
After replacing the suspected faulty tube twice I was sure yet another
Ei
6DJ8 had become faulty with a section that tended to become saturated
with current.
A replacement from my stocks worked fine. The 6DJ8 is a high gm twin
triode.
Its designed operation condition in the VT100 means none are under any
great stress.
So I guess that in two of the Ei 6DJ8 some cathode oxide material had
found its way
onto the grid, and when heated up the grid was emitting electrons as
well as the cathode.
Only one triode of the two within each dodgy twin triode tube was
affected.
Once the dodgy Ei tubes were replaced both channels produced identical
results for power, gain and Rout.
I advise anyone with Ei 6DJ8 with the gold lettering on the tube to be
beware;
Maybe a new batch of russian or NOS would be better.
6CG7 would also be usable but require that the two NFB loops have the
resistors changed to
increase the feedback by about 5dB. The VT100 is rather insensitive and
increasing loop NFB
when using the 6CG7 which has 5dB less gain than 6DJ8 will require a
preamp with
5dB more gain...
I set the 4 adjustable CCS in the two input amps so that equal current
is in each and there is near as possible equal anode voltages close to
the +60V nominated.
It wasn't possible to adjust each side of the input LTPs for exactly
+60V.
This means the direct coupled following LTP gain block has serious DC
imbalance which I found could not be
eliminated, only minimised. At normal levels this causes no problem but
ARC's way of doing things is inferior to
most other amps which do not rely so heavily on matched driver tubes and
mosfets and potentiometers.
At clipping levels there is slightly asymetrical clipping.
If or when the direct coupled cathode follower drivers to the output
stage
fail the output tubes have NO PROTECTION and will sit there red hot
while trying vainly to radiate
100 watts each and the "hot tube event" I witnessed did not cause any
fuse to blow or any
over current detector to work or resistor to fuse open.
The only place for a fuse or fusible resistance link is in the cathode
circuit of each tube because
if you have one in the anode circuit and it fuses open then the screen
which has +470V on it tries to
absorb far too many electrons and it just melts down and ensures that
the
tube will definately die.
Placing fuses in each cathode circuit means that the fuse value has to
be
low to react to say 2 times the idle current. This is a problem because
the ac peak current
can be 4 times the idle current.
Nuisance fuse blows could occur, perhaps in just one of the 4 tubes, and
nobody would know for awhile until another fuse blows, or the single
tube left on one side of the PP pair
decides the expire.
I never rely on fuses for bias failure and fit all my amps with 10 ohm 5
watt cathode resistors
and dc detectors which filter out the ac signal content from the top of
the 10 ohms and the dc signals gor
to an SCR which trips the amp off if any tube suffers more than 100mA dc
for longer than 4 seconds.
The VT100 speaker secondary of the OPT is used to provide feedback
voltage which is applied
to the output tube cathodes.
It makes fitting a protection circuit slightly more difficult, but it
could
be done using a special board with 8 detectors and a bunch of diodes to
the same point then the most
overloaded tube will trigger the protect SCR and relay to turn the amp
off.
An auxilliary transformer must be fitted to power the protect circuit
and to run the relay
but if no aux tranny the HT winding can at least be switched off with a
relay run from
DC generated off the heater circuits.
Active protection is standard on all my amps, and worth every cent.
I never have direct coupled cathode follower buffers like ARC's useless
set up. Normally
such buffers are used to power the amp into class AB2 with grid current
when the grids are forced to go positive.
But the VT100 has largish value series grid R between the buffer
cathodes and
6550 grids so there is virtually no AB2 power possible. There is thus
certainly no need for
the buffers, and the more reliable method to drive output tubes is via a
simple cap coupling from a low impedance driver gain stage using vastly
more rugged EL84 in triode for the driver
LTP.
The load on the output stage with 8 ohms connected to the 0-8 connection
reflect 1k6 a-a for the four tubes so the 110watts is possible.
The amount od pure class A is small with the rated load connected to the
matching taps.
I tried 8 ohms connected across the 0-4 ohms outlet points and got 80
watts
with about twice the amount of class A power and much less THD for the
same power when the rated load is matched
with the nominal outlets for those loads.
The VT100 amp should have the 0-8 re-signed 0-16, and the 0-4 re-signed
to 0-8.
The sound levels available by using 8 ohms plugged into the 4 ohm outlet
would only be 2dB less, ie, negligibly lower.
The output resistance is 1.5 ohms from the 0-8 outlets, so DF = only 6.
By using 8 ohms on the 4 ohm outlet, DF is effectively doubled, and the
THD
is about halved.
With 8 ohms on the 8 ohm tap the THD was predictably high at a dB below
clipping, 0.6%,
but at a few watts was the normal everyday amount of THD so typical of a
UL amp.
The output stage has a 0.5 ohm resistor for measuring bias and there
should be
0.13Vdc across this R for correct bias but I would only have 0.10Vdc so
that the least heat possible is
generated in the tubes, yet still allow quite enough class A power.
There is a heatsink for a solid state regulator which is very close to
one output tube.
It is amazing the devices on the heatsink have not secumbed to heat
stress
because as you should know, SS device reliablity falls by 10% for each
10degrees C above room temperature,
so hot SS is always unreliable to some extent but when
the temp rise is 100C, the SS has 100% likelyhood of failure.
I suspect that regulator is hanging by a thread.....
Keeping the output tube bias lower than normal is the major heat saving
measure that
should be taken.
Placing a fan in the amp can't do any harm except disturb the peace on
quiet evenings.
Having the fan blow first onto the regulator that gets hot before
wafting around everything else is a good
idea. And indeed waft around is all you'd get,
because it would be difficult to direct air to all the output tubes.
I am in no hurry to copy the circuitry used in the VT100.
Patrick Turner.
.
- Follow-Ups:
- Re: ARC VT100 repair and test.
- From: Ian Iveson
- Re: ARC VT100 repair and test.
- From: Bret Ludwig
- Re: ARC VT100 repair and test.
- Prev by Date: Re: Why We Should Ignore Bill Ramsay
- Next by Date: Re: For You Die-Hard DIY Types Out There
- Previous by thread: Tube Testers
- Next by thread: Re: ARC VT100 repair and test.
- Index(es):
Relevant Pages
|
