Starting a complete restoration of another MC2100. If you're interested in purchasing the amp please contact me directly. This project will include new paint, polished chassis, new McIntosh output power boards, new North Reading Preamp, new filtering caps, new hardware including machine screws, new output transistors, new multi-cap and other goodies.
The chassis is in excellent condition and the transformers are fine.
Some photos of the teardown.
The amp was purchased with many parts missing. I have many of the parts on hand to restore the amplifier to near new condition!
MC2100_teardown_5.jpg
DSC_6537.JPG
MC2100_teardown_6.jpg
MC2100_teardown_4.jpg
MC2100_teardown_1.jpg
MC2100_teardown_2.jpg
First thing to tackle is the small signal cables from the amplifier front-end potentiometers to the preamplifier board. McIntosh factory leads are good quality, tinned, single conductor with metalized mylar sheild and tinned drain wire. A pigtail was attached to the drain via a crimp which was then secured using shrink.
I prepped the leads per the factory approach with the exception of soldering the pigtail to the drain rather than using a crimp fitting.
drain_1.jpg
drain_2.jpg
drain_3.jpg
Fuseholder: Bussman HKP 15A 250V
41mpgL+Y9TL._AC_.jpg
Replace the crack prone, hard plastic factory chassis feet with rubber. Includes integral washer installed. Requires a longer sheet metal screw for installation into the chassis.
https://www.mcmaster.com/9540K914/
9540K914_Unthreaded BumperX.gif
The McIntosh preamplifier board will be replaced with a North Reading Engineering preamp.
First consider the factory preamp board, the MC2100 preamplifier used low noise BC238 NPN amplifiers at Q301, 303 (top) and Q304, 302, each pair forming a differential input pair (one for each channel). These are paint coded orange-green by the factory, likely sorted by Hfe (DC current gain).
OEM_diff_pair_1.jpg
OEM_diff_pair_2.jpg
OEM_NRE_preamp.jpg
Here's the static characteristic curves of the two devices shown in the first photograph. The two form a differential pair for the input signal. The plots are the collector current (Y-axis) as a function of collector-emitter voltage (X-axis) over a range of fixed base currents. As is shown, there's differences between the factory pair, they can be better matched.
Q301_302.jpg
The North Reading Engineering board uses a pair of BC550 small signal, low noise NPN BJTs to form the differential input pair. There's a pair for each channel of the preamplifier board. The static characteristic of a selected pair is shown and, as evidenced, are nearly identical. These devices are from the same production lot which makes the job of sorting and finding similar pairs a bit easier.
NRE_Q2_Q3.jpg
Here's the schematic for the North Reading preamplifier. The pink highlighted areas show the primary differences between it and the factory preamp.
NRE_preamplifier_schematic_1.jpg
First (top) is a current mirror which loads the collectors of the differential paired NPN transistors. A current mirror synthesized using individual transistors is a bit of a challenge and requires careful matching of the two PNPs used to fabricate the mirror. On an integrated circuit, this is no issue, but here it's a compromise. That said, after sorting thru 20 or so 2N5087s, all from the same production lot, I was able to find two pair that are quite close based on the static characteristic curves I measured.
To minimize thermal drift, mechanically uniting the two devices as shown below does improve things a wee bit. The white stuff is thermal grease. There're small metal clips for this sort of thing but the tie-wrap will do.
current_mirror_2.jpg
What the current mirror provides is a significant increase in differential gain which allows for some emitter degeneration at the input pair (not done in the factory unit either). The currents are quite small in this design, on the order of about 200uA, which manages to keep thermal effects to a minimum.
The lower highlighted region is a current source which replaces the factory long-tailed-pair resistor arrangement which suffers from variations in tail current as the common mode signals change. This effects the gain of the amplifier causing distortion. The tail current is determined by one diode drop across the 2kOhm resistor (~215uA).
Below is the factory MC2100 preamplifier board schematic which can be compared to above.
McIntosh_preamplifier_schematic_1.jpg
Capacitor C1 in the North Reading preamplifier is straightforward to compensate for instability associated with feedback. The effect of varying C1 from 10p to 100pF, in increments of 10p is shown in the simulation. The response is simulated from 1k to 500kHz. The preamplifier slew rate is lowered as C1 in increased, the price for stability.
NRE_bandwidth_1.jpg
The filter caps for the new preamplifier board arrived this weekend which allowed me to complete the preamplifier board and begin testing it.
NRE_pre_1.jpg
A few preliminary measurements to determine if the assembled board works and doesn't oscillate. 20kHz (top) and 10kHz square wave response, each channel. Input is 180mVRMS. Signals sourced from Standford Reasearch Ultralow Distortion generator set at 50Ohm output impedance.
10_20kHz_sq.jpg
20kHz (bottom) and 10kHz sinewave response.
10_20kHz_sine.jpg
200kHz sinewave. Output reduces to about 1.43VRMS at 200kHz from 2.01VRMS at 20kHz or a change of about -2.96dBV
200kHz_sine.jpg
Plot below is simulation results of open loop gain of each preamplifier. A significant 8dB increase is realized in the North Reading design. Experimentally, this is a bit difficult to measure (requires some fabricated circuits, low cost but time consuming). Final %THD numbers of the preamplifier in the amp, which are straightforward to perform, support the simulation trends. I'll simulate the closed loop gain and then shown the loop gain and explain why this is worth examining.
To simulate the open loop gain requires the use of a simulation tool called a Tian probe which is easy to execute in LTSPICE and I'll show that also.
open_loop_gain.jpg
In the LTSPICE environment, it's relatively easy to execute a Tian probe analysis by finding the file in the LTC library (LoopGain2.asc) and pulling both the probe (highlighted in the box below), the simulation directive:
.step param prb list -1 1 ; set prb=0 to turn off probe
and the plotting expression:
-1/(1-1/(2*(I(Vi)@1*V(x)@2-V(x)@1*I(Vi)@2)+V(x)@1+I(Vi)@2))
from the file.
The probe is inserted on the inverting side of the differential pair after the feedback resisitor circuit.
Tian_probe.jpg
Prepping the bottom chassis cover for paint...
is done (my way) using abrasive loaded Scotch-brite, specifically the brown pad. This is ideal for breaking the factory paint and prepping the surface for a fresh finish. Emery paper can take the paint off completely which is not desired. The best primer is the original paint.
Before any paint is applied, the chassis should have a matte appearance and be free of any visible scratches. It also must be impeccably clean and free of residual abrasive grit or paint soot. A hard rub down with clean towel or clean tee shirt will eliminate any soot from the surface. I do not use solvents to clean the surface, they're not needed.
chassis_prep_1.jpg
chassis_prep_3.jpg
chassis_prep_4.jpg
Burrs should be filed down before paint too. This will ensure the new feet sit firmly to the chassis.
chassis_prep_2.jpg
Two very light coats later. I'll place the chassis in a small heating box set at about 100F for three hours to harden the paint and drive off any remaining solvent.
chassis_1.jpg
chassis_3.jpg
chassis_4.jpg
chassis_2.jpg
The new rubber feet attached. These have a metal washer molded into them, they're also rubbery. The sheet metal screws are a bit longer, stainless steel.
chassis_8.jpg
chassis_9.jpg
chassis_6.jpg
Back at the chassis.
Buffed out most of the light scratches using polishing paste.
chassis_12.jpg
chassis_13.jpg
chassis_14.jpg
I replaced the Cinch terminal block and replaced the hardware with stainless fasteners. #6-32 x5/8"
chassis_11.jpg
Soldering the terminal blocks.
Form the leads around the hook.
speaker_con_1.jpg
And solder, like we were never here!
terminal_2.jpg
When removing leads from the edge connectors (Amphenol parts that are NLA) the "trick" to prepping the solder lugs for new connections is to first trim as much of the wire stranding and solder as possible from the lug using a sharp wire cutter.
The connector can be destroyed by pulling or twisting the stranding from the lug and if that happens
you're screwed.
card_1.jpg
The large filter cans can be replaced.
Do they need to be replaced? Who cares, just replace them.
filter_1.jpg
Filter caps installed.
And don't pay resellers on our favorite auction site for these capacitors. They're about $63/ea from Mouser.
filter_3.jpg
The pots have to go, they work but are tamper resistant, ie they will not accept a knob and must be adjusted using a flat blade screwdriver.
pots_1.jpg
They're CTS 500kOhm linear pots with a 470kOhm +/-10% resistor across lugs 1-3 to modify it to a 250kOhm linear. The solder job to connect the pots to the circuit looked a bit sketchy making me think these were added by an owner. There's nothing special about these and a panel mount, 250kOhm audio-taper with 3/8" bushing and 1/4" shaft will do the trick. Due to the type of pot used, the amp doesn't have input level adjust knobs which will change too.
A long-shaft bushing (similar to what's used in guitar pots) is a thought here as well. It would allow for setting the pot back a bit from the panel and make it easier to service.
pots_2.jpg
pots_3.jpg
AC simulation of 250kOhm pot with and without a taper resistor and how it alters the preamplifier voltage output from 5 to 95% cw rotation. When the vertical axis is a logarithmic scale (dB), a log-taper should show, approximately, equally spaced output levels for equal, incremental turns of the knob, in this case each increment is 1/20th of a full turn. The predictions on the right are with a taper resistor across the wiper to gnd. The taper resistor is 15% of 250KOhm (~37.5kOhm).
pots_4.jpg
Will install two, 250k, linear, Clarostat/Honeywell pots to replace the ones on the unit. These are pricey at ~$17/ea if you can find them. They're sealed construction, so nothing gets into the slider assembly. And made in USA!
Part Number: RV4NAYSD254A
RV4NAYSD254A.jpg
Datasheet here:
35dfa41e9654ac5abb8722a713c095c2.pdf
The Clarostats installed.
pots_5.jpg
Completed the right channel output stage. Replaced all six transistors with ON SEMI MJ15003 NPN (quasi-complimentary output stage).
right_channel.jpg
Datasheet
MJ15003-D.PDF
North Reading preamplifier installed.
NRE_preamp_installed.jpg
New power output boards and output transitors installed.
MC2100_7.jpg
Vintage knobs,
knobs.jpg
All buttoned-up,
MC2100_1.jpg
MC2100_.jpg
MC2100_4.jpg
1kHz %THD v. Watts
MC2100_%THD_Watts_48kHz.jpg
1kHz %THD v. Volts
MC2100_%THD_Volts_48kHz.jpg
%THD measured using Krohn-Hite 6900B at 20VRMS out, one channel,
Voltage
K-H_20VRMS.jpg
%THD at 20VRMS
K-H_20VRMS_%THD.jpg
%THD measured at 5VRMS out, one channel,
Voltage,
K-H_5VRMS.jpg
%THD at 5VRMS
K-H_5VRMS_%THD.jpg
The distortion number are similar to those measured in the CLIO FFT analyzer. Note the Krohn-Hite unit has significantly higher measurement bandwidth (5Hz-1MHz) than the CLIO unit (10Hz-192kHz).
On older McIntosh amps the sheet metal fasteners that hold the chassis bottom cover might no longer grab the chassis when tightened. This can be remedied by first filing down the upset metal on the chassis (not the cover) being careful not to gouge the area around the hole and then replacing the #6 x 1/4" with a #8 x 1/2". Both the #6 and #8 have the same size hex head, the flange being a tad larger (a good thing).
fasteners_1.jpg
90286A146_18-8 Stainless Steel Flanged Hex Head Screws for Sheet MetalX.gif
90286A194_18-8 Stainless Steel Flanged Hex Head Screws for Sheet MetalX.gif
After a few days of using the unit I noticed an intermittent connection at one of the RCA panel jacks. Replaced both with Switchcraft RCA panel jacks.
RCA_3.jpg
RCA_4.jpg
RCA_2.jpg
RCA_1.jpg
It will be packed and shipped today to its new owner in Kentucky.
complete_2.jpg