Earlier today , I finished the last set of R-4B mods. Those of you
following my progress, this marks the ten week point of incremental changes.
The last (and final, I swear!) set of modifications:
1) Added the Sherwood PD-4 active product detector; and
2) Replaced the Sherwood AMP-4 audio amp with something truly special: a
modified form of the SM0VPO Class AB design. This amp is rated at 4W into
4-ohms, so it can accommodate both headphones and an MS-4 speaker. The
circuit board was made by FAR Circuits, based on the artwork available on
the SM0VPO website. IMHO, this is the Holy Grail of "single supply" audio
amps.
Adding the PD-4 was easy enough now that the Sherwood PS-4 board is
delivering regulated and decoupled +12V throughout the R-4B. Findings? Not
a whole lot. AGC attack is much smoother but I notice no other changes. An
ancillary benefit of the active design is that In/Out isolation is better
than that of the diode product detector and as such, there's probably less
opportunity for audio to feed back into the AGC circuits.
I was not expecting much change between the Sherwood AMP-4 and the discrete
SM0VPO design. However, the new amp board elevates the R-4B audio to a new
level of audio performance. I searched for low-noise transistors for use in
the high gain stage and kept the TIP-41 types in the output stage as a truly
low-Z drive source for the headphones and MS-4 speaker. The output coupling
cap is a low ESR type with a whopping 2700 uF of output C. The only way
this amp can be improved is to utilize a bipolar supply to attain DC
coupling between stages. I cannot justify that level of extra effort just
for DC-coupling "bragging rights." Just adding the Sherwood PS-4 to the
Drake R-4B was complicated enough (see prior message threads for details).
http://72.52.250.47/images/AMP-1.jpg
The discrete audio amp has zero residual noise, no hiss, no hum, no buzz.
Unlike the monolithic LM3xx series, it's totally silent. I raised the amp's
ground buss potential above the R-4B's chassis and used a large conductor
size between the new amp board's ground buss and a common ground point on
the Sherwood PS-4. Likewise, the audio input shield ties back to the same
ground point. I did this to avoid buzz-inducing currents on the chassis
from the AC tube filaments and the CT high-voltage secondary of the power
transformer. The common point is in the corner tip of the R-4B chassis -- a
point where AC currents are minimum.
http://72.52.250.47/images/AMP-2.jpg
I had one issue develop: As I unplugged the headphones, I noticed
intermittent instability. For clues, I went to the National Semiconductor
datasheets for the familiar LM383, LM380,and LM386 output amps. Each uses a
small resistor in series with a cap (e.g., 2.7 ohm in series with 0.1 uF) at
the output to better stabilize the amp under changing loads (e.g.,
un/plugging headphones). That worked very well with the discrete design and
completely cured the instability without affecting performance.
When I have time in the next couple weeks, I will measure the THD and IMD of
the new amp. I have a Tektronix TM series test set with
ultra-low-distortion oscillators and automated distortion analyzer.
According to SM0VPO, typical THD at 2W is under 0.01%.
So, perhaps now is a good time to summarize all my R-4B mods:
1) Eliminated all power-hungry voltage dividers and substituted with a
Sherwood PS-4 board;
2) Replaced power transformer with an R-4C NOS unit thanks to the generosity
of a list member. This R-4B runs super cool to the touch;
3) Replaced the multi-section power supply cap with a multi-section type
from Tom at hayseedhamfest.com;
4) Removed the entire R-4B audio section and replaced with a modified SM0VPO
Class AB audio amp board;
5) Changed H/P jack to TRS so newer stereo phones can be used without
hunting for an adapter;
6) Replaced all 12.6V tubes to their 6.3V equivalents (e.g., 12AX7 is now
6AX7). This, because the R-4C transformer only uses a 6.3V filament
secondary since the R-4C uses all 6.3V tubes. From the factory, the R-4B
has a mix of 12.6V and 6.3V tubes -- and a CT 12V filament secondary;
7) Using an outboard inrush surge limiter on the AC line cord: Two CL-90
thermistors in series. In addition to inrush limiting to the filaments, the
filaments now see 6.3V instead of 6.8V from my high AC mains line voltage
(runs about 125VAC);
8) Replaced all paper caps with "yellow-jacket" metalized Polypropylene
types;
9) Replaced the diode product detector with the Sherwood PD-4;
10) Replaced both dial lamps with blue LEDs from N9OO. Love the new look
against the orange neon dial indicator;
11) Was going to replace a tired NE-2 bulb. For whatever reason, reversing
the neon lamp leads has given it new life;
12) Changed AC line cord to 3-wire for safety;
13) Changed AGC "Slow" time constant;
14) Changed several audio coupling and bypass caps to extend the audio
response. Low-end response for AM and ESSB reception is now close to DC
when zero-beating the internal 25 kHz oscillator.
There are probably a few more changes, but these are the major points.
Paul, W9AC
_______________________________________________
Drakelist mailing list
[email protected]
http://mailman.zerobeat.net/mailman/listinfo/drakelist
