Teardown of the Rane MA3 Multichannel Commercial Amplifier (Discontinued)

[Ground Noise Industries]

Rane MA3 Multichannel Commercial Amplifier

MA3 Multichannel Amplifier

The MA3 is a 3-channel amplifier designed to operate reliably in commercial environments, for use in paging, foreground and background music distribution.

www.ranecommercial.com

Discontinued Rane Products

Manuals, data sheets and schematics are still available to download for Rane's discontinued models.

www.ranecommercial.com

• 3 Independent Amplifiers
• LM3886 (11-Lead TO-220 Package). https://www.ti.com/product/LM3886
• 60W per Channel Continuous Average Power into 4Ω, 20-20k Hz
• 40W per Channel Continuous Average Power into 8Ω, 20-20k Hz
• Load Sensitive Dynamic Limiters and Headroom Meters
• SPiKe® Protection Circuitry
• High Capacity Linear Power Supply
• Sealed Heat-Tunnel Forced-Air Cooling
• 80 Hz High Pass Filter Selection per Channel
• UL / cUL Certified

Let's begin with unboxing and details like the meters and the I/O.
Then we can open it up and see what makes it tick.

Opening the box. Well packed and supported. FedEX guy wasn't
really happy because it is 26 lbs. and was double boxed.

First View. Love that smell of new electronics.

Headroom Meters. More on these later.

Balanced Inputs via euroblock connectors.

Outputs. Notice the box...old tape. The box smells clean but old
like Christmas ornament boxes stored in the attic. This amp
was sold as new and it is new. But it was manufactured in
September of 2015. . .NOS. . .but no tubes here. Fully analogue,
solid state, no dsp amplifier.

 

[Ground Noise Industries]

The power input. This amp can be reconfigured to run from
240vac mains. Read the manual and schematics for instructions
on how to do the switch to 240vac. You can also easily engage
the built in 80 Hz high pass filter by moving the jumper inside.
Originally for use with 70.7V step up transform for 70.7V
distribution sound systems.

The forced air cooling fan on the rear panel. Airflow is from the
rear and exits at the front.

Inside power input board. Black wires go to front mounted power
switch. The blue, white, yellow grey go to the main transformer
that has three center tapped secondaries.

Here is an overview. The heatsink and fan are on the left but
not shown here.

Toroidal transformer. I will be opening it up again soon and get the
specs and maker info.

 

[Ground Noise Industries]

A closer look at the lm3886. The previous versions of the amp
had through the hole components. This last generation switched
to smd for most of the components. Takes a much steadier hand
than mine to work on one of these. I assume this is a two layer
circuit board to sneak the signals around without jumpers.

And another angle. Power supply caps in lower left. Lower center
is the +/- 15VDC regulator on silver colored heatsinks. Just above
the regulators is the vca ic and buffer oamps. The schematics
identify each part. The right side is the three inputs with the vca
pots. The red coils at the top and black cylinders near the bottom
right are output inductors.

Schematic of output inductors. Take a look at the link in the first
post to see all schematics. You can even look back at the older
ones to see how Rane made revisions. There is a very long trace
between the two output inductors -- across the length of the main board.

Constant current fan circuit. More about this shortly. The heatsink
must have air cooling because it is undersized for the thermal output
produced by three lm3886 chips.

Voltage regulators to make nice +/- voltages for them opamps
and vca ic.

 

[Ground Noise Industries]

Power supply FULL WAVE BRIDGE diode ic x3.

Power supply. 4700 micro farad, 50V, 85C Nichicon caps.

A closer look at the voltage regulators for the opamps.

The led headroom circuit board. I don't like those types of ribbon
cables. Not much in the way to modify.

Meter board leds tucked in there.

 

[Ground Noise Industries]

A shot on the input layout. Note the three black jumpers like you
would find on an old ide hard drive to configure master or slave.
They are to the right of the pairs of white caps (c12 and c14 for
channel 2). They are for setting the 80Hz high pass filter.

You can see J3 and J4 better in this image. This particular amp
was manufactured in September of 2015. Several of the parts are
now discontinued. Maybe that is one reason this model has been
discontinued. It is like a relic but most wonderful in simplicity.

The fan and heat sink. the chassis is one piece from front to base
and to rear panel. It is slightly rolled up on the sides. Then there is
a single thinner top/side cover made of thinner metal. This chassis
does not twist, it is solid. Interesting to note, when I had the top off,
the output had some strange sounds. I assume the top cover ensures
complete shielding and/or interacts with the toroidal transformer to
tame its electro-magnetic field. When to top cover was reinstalled,
the noise went away.

Another chassis shot to see how the metal was formed. We can see
from this angle the front exit for the fan on the left side.

Interesting fuse (Littelfuse series 382). Not sure why this was chosen
over a slow blow. Maybe the transformer just pulls too much at start
and the slow blow fuse would offer no protection for faults while running.

Fuse details.

Signal input has this device that was seen earlier on each input. I
looked at the schematics and electrical characteristics for this device.
I'm still not sure how this is something needed it this circuit.

The output network. Last stop before reentering the real world. Here
we see the medium sized traces from the lm3886 output feeding the
previous resistor/inductor paralleled and in series with the output. I
think it helps prevent high frequency oscilations. My old Yamaha M-45
had something similar on its output.

 

[Ground Noise Industries]

Let's start at the beginning. . .plus a bunch of billions of years. . .

When the Rane MA3 is powered on via the front mounted switch, there is no transient noise (+1 point). Unfortunately, when powered off, there can be a rather 'yucky' transient noise and a quick loudness increase in the music prior to the transient noise(-2 points). The transient noise is louder if the volume was up high when powered off and quieter if volume is minimized, but ever present. I think the vca turns off prior to the muting circuit of the lm3886 ic so there is a brief increase of the loudness of what was playing and then the off transient noise--that is why a minus 2 score. This is my first of two issues with this unit.

The main power supply produces three unregulated center tapped supplies, one for each lm3886. They float around 66.4vdc to 66.8vdc (33.2vdc-ct-33.2vdc). The input mains were 121.1vac the day I tested and the secondaries prior to rectification were at 50.3vac. This amp has the ability--via swapping the primaries from parallel to series--to operate off of 240vac mains and at 50Hz. The regulated +/- 15vdc supply and fan circuit are power off of the channel 1 main power supply. All three channels sound the same so it would take more advanced instruments than my ears to detect an issue with this arrangement.

Some quick numbers

The headroom meters are at -12dB(green), -6dB(green), -3dB(amber) and 0dB(red).
With a 61Hz sine wave as source across an 8 ohm and 4 ohm resistor.
Testing was with one channel driven since I only have one test resistor on a heatsink.
Balanced input signal was used. The output voltages are below.
The leds light up at different voltages depending on the load of 8 or 4 ohms, it isn't perfect

At about 4.50 to 4.70 Vrms the -12dB led fully lights up.
At about 8.98 to 9.38 Vrms the -6dB led fully lights up.
At about 12.68 to 13.25 Vrms the -3dB led fully lights up.
At about 17.91 to 18.72 Vrms the 0dB led fully lights up.

Above 0dB the limiter slowly clamps down and the maximum voltage achieved was around 22Vrms. Not sure how long you could push the amp at that level. It is forced air cooled so maybe the length of "Welcome to the Jungle". The amplifier meets design specs of 40 watts at 8 ohms and 60 watts at 4 ohms. At 4 ohms, the amp easily exceeds its listed 'ratings' but probably at increased distortion and would result in the lm3886 protection circuit jumping in, which is something to be avoided apparently.

Using the rear mounted 'level' pots to adjust the vca, the apparent 'gain' can be modified. At max(10) the 'gain' is between 26.6dB to 27.07dB, as expected. Around (7) the 'gain' is effectively 20dB. The lm3886 is probably fixed at 27dB gain but the vca prior in the circuit is doing its thing. Around (5) the 'gain' is effectively 16dB. Below (5) the gain drops off rapidly and goes negative. The specs state that the maximum input voltage is +20dBu (7.746Vrms). At just 20dB gain, I calculate 77.46Vrms or 750 watts into 8 ohms. That just isn't going to happen. You will have to change the level pot below (5) to utilize that +20dBu source voltage. Setting matching 'gain' between the three amps is easily done as long as you don't set the 'gain' less than 16dB.

The fan is controlled by a constant current circuit(you can see the schematic posted earlier or at rane site links). The real issue is the noise. This is the second issue with the amplifier(-1) Yes, it is a 'commercial' amplifier expected to be put in the a/v rack in some remote closet. Noise isn't an issue when you hide it away. My current idea is to replace it with a quiet fan. It is a 24V, 3100 rpm, 8cmx2.5cm 41.3 cfm. Seems there is room for improvement. In home use where it won't be driven to the limiters all day, the cooling rate could be reduced. Under normal use, the fan is dropping about 18.92vdc and 59mA. The input to the circuit is at full ps voltage of +32vdc.

The amplifier sounds good to me and working out the turn off noise and fan noise will be some fun. It is well built and rather heavy coming in at 26 lbs.

 

[Ground Noise Industries]

Some points of interest.

Trace size matters! I would have expected a larger output
trace considering the larger ps traces. It is possible that the
trace is deeper, thus same or similar area. I would like to pull
the whole board to view the backplane traces. Then most of
us will add our 12 awg speaker cables with the bottle neck
being the traces at the amp end and the crossover pinch points
added to the transmission problem of signals. This is a complex
transmission system!

 

[Ground Noise Industries]

The Toroidal Transformer in the Rane MA3 3-Channel Commercial Amplifier

1. Dual Primary for 120vac or 240vac, 50Hz or 60Hz operation
2. Three Independent Secondaries
3. 50.3vac CT (+/- 25.15vac) at idle
4. 66.6vdc CT (+/- 33.3vdc) at idle

CHAIN MATE INDUSTRIAL CO., LTD
09408
E206119
Viking B-2
1445
CHAIN MATE INDUSTRIAL CO., LTD. is dedicated in "Transformer for Electronics, Power Transformer, Input/Output Transformer, toroidal Transformer, Epoxy Cast CT, Current Transformer, Choke and Coil, Low Frequency Transformer" with operations in Taiwan.

The Toroidal Beast

 

[Scrivs]

Thank you for the teardown, photos, and clear and concise analysis. Easily one of my favorite things about ASR.

 

[Ground Noise Industries]

Thank you for the teardown, photos, and clear and concise analysis. Easily one of my favorite things about ASR.

You are welcome. Keep it simple but informational. I like both ASR and DIY Audio. Learned a bunch and enjoyed it all the way. It is good to have found a home. I'm jealous, though. . .enjoy your evening on the island!

 

[Ground Noise Industries]

The Fan Noise


Here is the stock fan circuit. You can also see it in the schematics posted
at the Rane site. I posted the links in the first post. Since the unit is mostly
surface mount devices, the fan speed adjustment will take place after the
fan jumper. Apparently this is a constant current circuit; my goal was to
keep the load resistance similar to the original to keep the 2SC4793 happy
because it is not on a heat sink.

With a little math and 'trial and error' I came up with the drop resistor
and parallel resistance. This drops the voltage across the fan from
18.68 vdc down to 12.79 vdc while maintaining a similar resistance
(probably impedance since the fan is very inductor like). This particular
fan, The NMB 3110SB-05B, MinebeaMitsumi, 24 vdc, 41.3 cfm fan.
It is rated from 12v to 27.6v so there is plenty of room to adjust. The
noise reduction is very helpful. The temperature of the 2SC4793 is
running somewhat higher than 110 F (42C). If I can get my fingers down
there, I will give that transistor some help with a small heatsink.

I also did some quick temperature measurement with the thermometer
that came with my multimeter. It gives me a ballpark number and I assume
that all temps are probably higher by few degrees down at the die level.

+/- 15 vdc voltage regulators at idle
U14 (7815 +) 134F (58C)
U15 (7915 -) 116F (49C)

LM3886 at idle was 72F (22C) with fan on modified circuit)

 

[raindance]

I used to build bridge amplifiers using LM3886 chips. The application was fairly loud background music for bars. I discovered why it was important to use sufficient heatsink compound really quickly - the chip has a rather small piece of real estate to attach to the heatsink. After the magic smoke came out a few times I got this figured out.

The fan cooled tunnel style heatsink used by Rane is a good idea as these guys really get hot.

 

[Ground Noise Industries]

I used to build bridge amplifiers using LM3886 chips. The application was fairly loud background music for bars. I discovered why it was important to use sufficient heatsink compound really quickly - the chip has a rather small piece of real estate to attach to the heatsink. After the magic smoke came out a few times I got this figured out.

The fan cooled tunnel style heatsink used by Rane is a good idea as these guys really get hot.

Thanks for the heads up; the undersized heat sink with three chips and forced air cooling was probably the only way to keep them under thermal control, the amp at 2 rack units high and the weight under 30 lbs. They can act a little like nitroglycerin--poof--without proper heat dissipation; I do plan to install a switch for low and high fan speed. Such a unique chip amp. There have been so many different designs with them that we now have a great template for future improvements and modifications. The LM3886 is like a 'custom' Lego block that can be the kernel of a new build.

Earlier versions of the amp had a rear mounted speed switch; this last generation amp took that option away--now it is fool-proof? Stick it in a closet and forget it. You can view earlier version schematics at the links provided in the first post of this thread(or see one in this post). I will be monitoring the heat closely. I might need to pick your brain on some other ideas I have in mind.

The amp is being used in a home environment with a pair of Klipsch KG4. They are around 93 to 94 dB SPL @ 1 watt. Listening area is about 7 feet. I can't sit on the couch with the -12dB headroom LEDs flashing (around 4 or 5 volts across 4 ohms). Knowing Rane equipment and over engineering (I used the MA6 amps back in the 90s, I also have the useful VP-12--I should have bought 2), I'm betting they designed this amp to be run with the limiters maxed 24/7 and never overheat. I like my music but not loud these days. I would rather listen at low levels all day thana few loud songs and fatigue my ears for the rest of the day. I always wear ear plugs to the movie theater(started with THX, ugh)--I even wear them to Disneyland(too expensive now). I may not hear much above 12kHz anymore but everything else sounds fine at low to moderate levels.

1998 MA3 with fan control

Both the 1998 and 2003 versions had the fan speed control.