Coda Technologies was started by a small group of people who had formerly worked
at Threshold. This amp and preamp share some of the physical attributes of
Threshold equipment, aesthetic beauty and wonderful build quality, but what
sets this equipment apart from the crowd is the Amplifier System 100. It is
in two sections: The front-end of the overall power amplifier, the 100v voltage
amplifier, is in a chassis the size of the preamp; the larger part of the amp,
the I00i current amplifier, contains the output stages and their power supply.
The 100v and 100i are connected via two conductor shielded cables, one per channel, terminated with XLR connectors. The fundamental reason, according to Coda Technologies, for separating these functions is to keep crosstalk from the high currents in an output stage from getting into the frontend circuitry and causing distortion. The use of XLR connectors for the interconnects suggests balanced operation, but that is not the case here. One conductor is the unbalanced signal to the 100i, and the other conductor carries the d.c. control voltage that switches the related channel of the 100i between standby and full operation. Although it is not specifically mentioned in the owner's manual, I believe the intention is for the 100v to be placed with your system preamplifier and for the 100i to be placed wherever it is convenient.
There is no mention in the manual about any limit to the interconnect length between the two pieces, and it would seem that the 100v would be capable of driving quite long lines without any problems.
Another thing that distinguishes the System 100 from other power amplifiers is the number and type of output devices used.
There are 58--count 'em, 58--output transistors per channel in this beast! They are high-speed, TO-220 plastic-package devices that are usually used for driver and small-signal applications.
The front panel of the 100v features a rotary function switch with positions for "Standby," "Balanced/Bias On," and "Unbalanced/Bias On." A red LED indicator to the left of this switch comes on in either of the "Bias On" positions. Another red LED at the right half of the panel indicates when power is on. There is no power on/off switch on the 100v, as it is intended to be powered all the time for best sound. Rear panel connectors include an IEC a.c. line cord socket, a pair of XLR jacks for hookup to the 100i, a pair of XLR jacks for balanced signal input, and a pair of phono jacks for unbalanced input.
While there are no controls on the front panel of the 100i, LEDs indicate the presence of a.c. power and when each channel's bias is off (the standby state) or on. As with the 100v, the 100i is intended to be powered all the time, with the pair's operation controlled by the standby/operate knob on the 100v. In the standby state, the power consumption is very low, on the order of 15 to 20 watts for the System 100. In the operating state, the a.c. line current is a healthy 3.5 amperes, as the idling current in the output stages is rather high. On the 100i's rear panel are an IEC a.c. input connector with integral fuse-holder, the power switch, two pairs of dual binding posts for speakers, and a pair of XLR connectors for the input signal from the 100v.
The FET Preamplifier 01 does not have tone controls, and a phono stage is optional. Both balanced and unbalanced outputs are included. The unbalanced outputs use a pair of phono connectors, while the balanced outputs use a pair of XLR connectors. The optional phono stage has two gain settings, for MM and MC cartridges. Phono input resistance and capacitance are each selectable via internal DIP switches.
Front-panel controls include "Input Selector," "Record Selector," "Mode" ("Mono/ Stereo/Reverse"), "Balance," and "Output Level." The "Record Selector" arrangement in the FET 01 allows possible feedback oscillation in a connected tape recorder when the deck is in record mode. Coda Technologies does caution about this explicitly in the owner's manual, however. A red LED power indicator is above the Coda logo, near the right edge of the panel. On the rear panel are an IEC a.c. line-cord connector, a pair of XLR balanced output connectors, nine pairs of phono connectors, and a binding post for ground.
The amp and preamp chassis are constructed of machined aluminum plates that are bolted together to form the whole. The FET 01 and the 100v each contain one large p.c. board that carries, either directly or via smaller daughterboards, the majority of the parts. The 100i has a large toroidal power transformer, mounted towards the front, and filter capacitors at the rear. A p.c. board, mounted on top of these filter capacitors, serves as a control circuit, audio power supply, and signal-distribution board. Two more large p.c. boards, one per channel, are mounted to the heat-sinks; these boards interconnect all the power transistors with the supply rails and the output buses. Needless to say, build and parts quality used in this equipment is of the highest order.
=========
SPECS
Preamplifier:
Frequency Response: Line section, 0 Hz to 200 kHz, +0,-3 dB; phono stage, RIAA ±0.2 dB, with subsonic roll-off at 14 Hz.
THD (20 Hz to 20 kHz): Line section, less than 0.01% at 6 V peak into more than 600 ohms; phono stage, less than 0.01% at 3 V peak.
S/N (re: 1 V): Line section, greater than 100 dBA; phono stage, greater than 85 dBA. Phono Stage Gain at 1 kHz: MM, 37 dB; MC, 57 dB.
Maximum Output: 26 V, peak to peak.
Output Impedance (Resistive): Unbalanced, 75 ohms; balanced, 150 ohms.
Dimensions: 19 in. W X 1 1/4 in. H X 8 in. D (48.3 cm X 4.4 cm X 20.3 cm).
Weight: 8 lbs. (3.6 kg).
Price: With phono stage, $2,750; without phono, $2,450.
Amplifier Power Output (20 Hz to 20 kHz): Stereo, 100 watts per channel, Class A, into 8 ohms; bridged, 400 watts (100 watts Class A) into 8 ohms; paralleled, 100 watts, Class A, into 8 ohms or 200 watts into 4 ohms.
Frequency Response: 0 Hz to 100 kHz, +0,-3 dB.
Distortion (10 Hz to 20 kHz): Stereo, less than 0.1% at 100 watts into 1 to 8 ohms; bridged, less than 0.1% at 400 watts into 2 to 8 ohms; paralleled, less than 0.1% at 100 watts into 0.5 to 8 ohms.
Gain: Stereo, 26 dB; bridged, 32 dB; paralleled, 26 dB.
Maximum Peak Current: Stereo and bridged modes, more than 100 amperes; paralleled, more than 200 amperes.
Slew Rate: Stereo, 50 V/uS; bridged, 100 V/µS; paralleled, 50 V/µS.
Noise (re: Rated Output):-100 dB.
Input Impedance: Unbalanced (stereo and paralleled modes), 100 kilohms; balanced (stereo and bridged modes), 2 kilohms.
Output Impedance (20 Hz to 20 kHz): Stereo, 0.03 ohm; bridged,
0.06 ohm; paralleled, 0.015 ohm.
Power Requirements: 450 watts.
Dimensions: 100v voltage amplifier, 19 in. W X 8 in. H X 1 /4 in. D (48.3 cm X 20.3 cm x 4.4 cm); 100i current amplifier, 19 in. W X 7 in. H X 19 in. D (48.3 cm x 17.8 cm x 48.3 cm).
Weight: 70 lbs. (31.8 kg).
Price: $6,500; additional 100i current amplifier, $4,350.
Company Address: 9941 Horn Rd., Suite A, Sacramento, Cal. 95827.
=========
Circuit Description
The FET 01's phono circuitry consists of two gain blocks. The first block appears to be flat, with the high-frequency RIAA roll off taking place in the coupling between gain blocks. The RIAA bass boost is accomplished in the feedback loop of the second gain block. The line section's gain function is implemented with one gain block and produces the positive phase output. Another gain block, configured as a unity-gain inverter, is fed from the output of the main line-amp output and generates the negative phase for the balanced output. A simple J-FET source follower with a J-FET current source functions as a tape output buffer and connects whatever is selected by the "Record Selector" to the tape out jacks.
The selected source for listening goes through the balance control, into the volume control, and then into the line amplifier input.
The basic circuit topologies used in the preamp gain blocks and the 100v power amplifier front-end are variations on the following theme: An N-channel J-FET differential amplifier as the first stage, followed by a P-channel MOS-FET or PNP bipolar differential second stage with NPN bipolar turnarounds. In all cases except the phase inverters for the line output, this second stage is cascoded with a PNP bipolar common-base stage. Both first and second stages have bipolar current sources. Complementary emitter followers are used in the FET 01 line output amplifiers and in the output of the 100v. The output of the first gain block in the phono stage is a single-ended NPN emitter follower; the second phono block has no output-follower circuit but is coupled directly from the collectors of the appropriate phase of the second stage's output. In the phono input stage, the input amplifier uses two J-FET devices in parallel for each half of the differential configuration, to lower input noise.
Power-supply circuitry in the FET 01 consists of a full-wave, capacitor-input, positive and negative d.c. supply followed by constant-current-fed zener diodes for regulated positive and negative reference voltages. After these zener diodes come emitter followers, of appropriate polarity, in five regulator pairs. These regulator pairs serve to feed each gain block with its own regulated supply.
The 100i is basically one giant Darlington complementary power emitter follower. Input from the 100v is appropriately level-shifted to the bases of a pair of complementary driver transistors. Output from the drivers is to the 58-transistor composite complementary emitter-follower output stage. No negative loop feedback is used in the 100i, only the current feedback inherent in the emitter-follower topology. A separate bias supply for each channel applies regulated voltage to the input bias-voltage dividers in each channel. These bias supplies are enabled by the d.c. control voltage from the 100v, to place the 100i in the operative state. In the standby state, the bias supply is shut off and the giant output stages sleep, with their supply rails up but current conduction cut off and signal input shorted by a relay.
Power circuitry in the 100i consists of a generous-sized toroidal transformer with individual-channel secondary windings, capacitor input filtered to the positive and negative supplies that power the two channels. The filters use four 50,000-µF, 50-V capacitors. Two more secondary windings power the bias supplies.
Measurements
Voltage gain and IHF sensitivities of the FET Preamplifier 01 are listed in Table I. Performance of the preamplifier in general was well matched between channels, and any of the measurements cited here are representative of both channels unless otherwise mentioned. Frequency response of the main positive phase of the 01's line amp is shown in Fig. 1 as a function of the volume control's setting. Results were substantially the same with either my instrument load or the IHF load. High-frequency bandwidth is reduced a bit with the volume set about 6 dB down from maximum but returns to full bandwidth at attenuations of about 20 dB or more.
Fig. 1--Frequency response of FET 01 line stage vs. volume setting.
Fig. 2--Square-wave response of preamp line stage; see text.
Fig. 3--Preamp line section THD + N vs. output level.
Fig. 4--RIAA equalization error; see text.
Square-wave response of the preamp's line section with instrument load is shown in Fig. 2. The top trace is for 20 kHz at an output level of about 10 V peak to peak. In the middle trace, the volume control was reduced to an output level of about 4 V peak to peak, and the reduced high-frequency response is evidenced by an approximate doubling of the rise and fall times (lengthened from about I to 2µS). The bottom trace is for 20 hz; the d.c. coupling in the line amp is evident in the absence of any tilt in the waveform's steady-state levels.
The preamp's distortion as a function of output level is plotted in Fig. 3 for three frequencies. The measurement bandwidth for the 20-Hz curve extends down below 10 Hz; consequently, there is more hum in these readings than in the 1- and 20-kHz curves, whose measurement bandwidth extends only down to 400 Hz. Results were about the same with the instrument or the IHF load. The turn-up in distortion at about 7.5 V is due to premature clipping on the positive peak of the output waveform.
The output amps of the FET 01 can drive a 600-ohm load with reasonable grace, putting out some 20 dBm (7.78 V) at about 0.25% distortion.
Output impedance of the line outputs was 60 ohms, and the input impedance varied from about 20 kilohms (with volume at max) to a higher value of more like 25 kilohms (with the volume control at 50% rotation).
Fig. 5--THD + N for 20 kHz into IHF load (A) and instrument load (B) and
for 1 kHz into instrument load (C) and IHF load (D).
Table I--Gain and sensitivity, FET 01 preamp.
Table II--Output noise of FET 01's line section for full-clockwise and counterclockwise
positions of volume control. IHF S/N was 85.5 dB for the left channel and
87.3 dB for the right.
Fig. 6--Phono overload vs. frequency at 3% THD + N.
Fig. 7--Frequency response of System 100 amplifier; see text.
Table III--Output noise of phono section with inputs short-circuited, with
100-ohm load on MC input, and with l-kilohm load on MM input.
With IHF MM artificial source, A-weighted noise for the left and right channels was 0.87 and 0.75 dB, respectively. IHF S/N for the MC section was 66.3 dB for the left channel and 68.5 dB for the right; for the MM section, it was 75.4 and 76.5 dB for the left and right channels, respectively.
Table IV--Output noise, System 100 amplifier.
The A-weighted IHF S/N ratio was 89.9 dB for the left channel and 91.3 dB for the right.
Output noise levels as a function of measurement bandwidth, channel, and volume-control position are listed in Table II. Usually, output amplifier noise in a preamp is lowest when the volume is turned down, becomes worse at some point about 6 dB down from full volume, and goes back down nearly to its minimum value when the volume control is fully up.
(The preceding assumes that the inputs are terminated normally, with a source impedance of I kilohm or less.) The FET 01's output noise, however, is highest at full volume, due to some line-harmonic noise from the filter-capacitor charge currents getting into the line amp when the volume is up and the selected input is terminated. With the inputs open-circuited, there was somewhat less noise at full volume. The noise level here, some 100 µV or more, might be audible with a quiet source connected and volume set near maximum in a system using a power amplifier with high gain (30 to 40 dB) and high-efficiency speakers.
Interchannel crosstalk of the FET 01 was found to be down by more than 90 dB at frequencies up to 400 Hz, increasing at 6 dB per octave up to-60 dB at 16 kHz this with the volume control at maximum. With the volume control set for 6 dB of attenuation, crosstalk was better than 90 dB down at frequencies up to 200 Hz, increasing at 6 dB per octave up to-60 dB at 10 kHz.
Volume-control tracking was within 0.5 dB down to-55 dB, increasing to a 1-dB error at-66 dB and further increasing to about a 2-dB error at-95 dB. This is quite good performance.
Phono performance was mostly investigated in the high-gain (MC) mode, and differences between this and the low-gain (MM) mode are noted where significant. The RIAA equalization error, measured at the tape output jacks, is plotted in Fig. 4 for two different settings of the input selector. The curves are flatter when the selector is set to a line input and show a low-frequency roll-off when the selector is set to "Phono." This latter condition loads the phono output's coupling capacitor with the impedance of the volume and balance-control circuitry and forms a first-order high-pass filter with a cutoff frequency of some 14 Hz, by intent, to reduce effects of turntable rumble.
The reason the frequency response is flatter in the low end when the input selector is not set to "Phono" is that the input impedance of the tape output buffer is much higher than that of the volume and balance-control circuitry, resulting in a much lower low-frequency cutoff. If you want the best low-frequency response from records, don't monitor directly by setting the input selector to "Phono"; set the tape deck to record mode, and listen through the tape playback monitor. In the case of the flatter curve, equalization accuracy is very high, but in practice you don't get this curve when listening through the "Phono" input setting. I don't care for this high-pass filter being a default condition without a defeat.
Phono distortion at two frequencies as a function of output level and loading is shown in Fig. 5. The right channel (shown), starts its distortion rise at about 6 V, noticeably sooner than the left channel did (whose distortion rise started at about 8 V). In the MM mode, distortion was virtually identical to Fig. 5 above an output of 5 V, with lower readings due to lower noise below 5 V. The noise was lower because the phono input was being fed with 10 times greater signal levels, due to the overall lower MM gain. Figure 6 shows phono overload versus frequency and load, both for attainable output levels at 3% THD + N and for the input levels corresponding to them; the curves for IHF and instrument loads are almost identical. Again, data is shown for the right channel, which has a reduced output capability below 100 Hz, whereas the left channel was more flat below 100 Hz. Overload capability is generally quite good, about 10 mV at 1 kHz in the MC mode and about 100 mV in the MM mode. As an example, imagine an MC cartridge with a 1-mV output at standard cutting level at I kHz; this phono preamp would allow a 20-dB increase over standard cutting level through most of the audio range, an event of low likelihood. Obviously, an MC cartridge of lower output would have even more margin. I wouldn't recommend using moving-coil outputs of greater than 2 mV for the MC mode. With those and MC or MM cartridges of higher output, you should switch to MM mode.
Noise levels in both MM and MC mode are listed in Table III. The results are satisfactorily low here, although nowhere near state of the art. Noise is greater in the MM mode due to the use of a high-value shunt feedback resistor in the first phono gain block and, possibly, to a greater contribution from the second phono gain block.
Crosstalk versus frequency in the MC mode was found to be better than 80 dB down from 20 Hz to about 13 kHz, where it crossed over the-80 dB level at a slope of +6 dB per octave.
The two pieces of the System 100 amp were interconnected by two 6-foot, two-conductor shielded cables supplied by Coda. Voltage gains and sensitivities for both channels were found to be virtually the same, at 26.5 dB and 133 mV. Again, as with the FET 01, performance of the two channels of the System 100 was very close and the results of the lesser channel are presented unless noted.
With the unit operating and the top cover off, I noted that the main power-supply rectifiers, mounted to the p.c. board atop the filter capacitors, do not have any heat sinking other than free-air radiation and convection. I think that they get too hot; I wasn't able to keep my fingers on them for more than an instant.
Frequency response for open-circuit, 8 ohm, and 4-ohm loading is plotted in Fig. 7. The results are so close that all three conditions just make the curve appear to be a little thicker with the resolution used in the graph. It would be fair to say that the System 100 is rather impervious to loading in regard to changes in frequency response.
Related is the square-wave response (Fig. 8). Here, the top trace is for a 10-kHz square wave and an 8-ohm load at an output level of 10 V peak to peak. Of interest is that the waveform remains exponential in shape all the way up to clipping.
Rise- and fall-times were 3.6 uS. In the middle trace, a 2-µF capacitor has been added across the 8-ohm load. The 100i does not have an RL output-buffering network, which usually causes much greater and slower ringing than shown here. The System 100 has unusually low reaction to the added 2µF, an excellent and desirable result. The bottom trace in the figure, for 40 Hz, illustrates the d.c., low-frequency response of the System 100.
Fig. 8--Amplifier square wave response; see text.
Fig. 9--Distortion vs. power output.
Fig. 10--Amplifier THD + N vs. frequency.
Fig. 11-Distortion spectrum of 1-kHz signal at 20 watts into 4 ohms.
Both THD + N at 1 kHz and SMPTE-IM distortion are plotted in Fig. 9 as functions of power output with 4and 8-ohm loading. Even though Coda says their Class-A biasing scheme is relatively insensitive to the effects of "running out of Class-A current" into lower impedance loads, the effect of this is visible in the plot of 4-ohm IM distortion as a null and a rise to greater distortion beyond. Overall, however, the System 100 has very low distortion, especially considering that there is no overall feedback loop encompassing the output stage. This design doesn't have much margin beyond its specified power output at stated distortion. As can be seen, the distortion is on the rise at the 100-watt power point. All these measurements were made from a 120-V a.c. line; lower line voltages would make the amplifier clip at lower power levels. Figure 10 shows THD + N as a function of frequency and power. A spectrum analysis of a 1-kHz signal at an output of 20 watts is shown in Fig. 11 for the left channel; the right channel had a bit more second harmonic but about the same amount of third. The amounts and orders of distortion are admirably low, but what appears to be some line-harmonic side bands are quite visible up to several kHz.
Output impedance was found to be about 0.03 ohm over most of the frequency range. This corresponds to a damping factor of some 267 referenced to 8 ohms, which is quite high for an output stage without global feedback.
Output noise as a function of measurement bandwidth, along with IHF S/N ratios, are listed in Table IV. The values obtained are exemplary.
Dynamic power, measured using the IHF tone burst, was found to be 115 watts and 220 watts into 8 and 4 ohms, respectively. Since the unit is not rated explicitly for 4-ohm loading in normal stereo operation, the dynamic headroom of 0.61 dB applies to 8-ohm loading. Peak current attainable into a 1-ohm load was ± 39 amperes at the point of clipping. I don't have a lower impedance load to find out what this output stage can really do, but judging by its beef, I have no problem imagining 100-ampere peaks, as claimed. Steady-state output at the onset of clipping was 112 watts into 8 ohms and 198 watts into 4 ohms, yielding a clipping headroom of 0.5 dB. Power supply regulation is pretty good in this design, as the steady-state draw off the power supply is rather high and is in the flatter portion of the power-supply regulation curve.
As a matter of interest, the power supply's rail voltage and a.c. line current was ± 47.5 V and less than 100 mA in standby, ± 42.9 V and 3.5 amperes at idle, and ± 42.2 V and 4.0 amperes at 100 watts per channel into 8 ohms.
Use and Listening Tests
Equipment used in evaluating the System 100 amp and FET 01 preamp consisted of an Oracle turntable with a Well Tempered Arm and Spectral Audio MCR-1 Select cartridge, a Krell Digital MD-1 CD transport feeding PS Audio Ultralink and VTL D/A converters, a Nakamichi ST-7 FM tuner and 250 cassette recorder, and a Technics 1500 open-reel recorder. Other preamplifiers on hand during the review period were a Quicksilver Audio, a First Sound Reference II, and a Counterpoint SA-5000. Other power amplifiers used were Quicksilver Audio M-135 prototypes, the Arnoux Seven B (a prototype switching design), and a Crown Macro Reference. Speakers used were Win Research SM-10 monitors and an early experimental Genesis Technologies two-way design.
I first listened to the Coda equipment with the 100v located with the front-end equipment and stacked with the FET 01 preamp. A pair of AudioQuest Lapis 20 foot balanced cables were used to interconnect the 100v and 100i. My impressions at this time were that the sound was musically enjoyable but that tonal balance was just a little on the dark or muted side although highs were smooth and inoffensive.
My next listening was done without the FET 01, stacking the 100v on top of the 100i and treating the pair like any other power amplifier. The two were linked by a pair of 6-foot interconnects supplied by Coda. The preamp was linked to the amp by a pair of 20-foot Masterlink LP cables, from Music and Sound Imports, which I have been using happily for the better part of a year. At this time I thought the amplifier was very good; it gave an excellent sense of space. It also had excellent definition, bass quality, and dynamics and was free of irritation. The high end was a little subdued but very complimentary to CDs.
For my final evaluations of the sonic quality of this amp and preamp, I first listened to the preamp mostly with the Crown Macro Reference amp, which has really endeared itself to me for its sonic honesty and good sound. Results were generally pretty good, with any sins generally being ones of omission. Compared to the same signals heard through the First Sound passive preamp/controller and the Crown amp, the sound of line-level sources played through the Coda Technologies equipment seemed to have lost a bit of its "thereness" but was otherwise nicely musical and very listenable.
Using the Coda pair together yielded very listenable musical results. As a final test of the System 100, I used the First Sound preamp to feed the amp via 1-meter interconnects. I felt the resolution and presence to be superior to what I'd heard when feeding the amp from the FET 01 preamp via my 20-foot interconnects. This makes me think I like the amplifier a little better than I like the preamplifier. Don't get me wrong, though; this equipment sounds damn good, both individually and in combination.
In conclusion, and nit-picking aside, I really think the Coda Technologies equipment reviewed here is stunningly beautiful and well made. The System 100 and FET 01 should last "forever," with little or no trouble, and provide excellent sonics in the process. Go out and give this pair a listen.
-Bascom H. King
( Audio magazine, Mar. 1993)
Also see:
Conrad Johnson Premier Five Amplifier (Aug. 1986)
Conrad-Johnson Premier Seven-A Preamp and Evolution 2000 Amp (Equip. Profile, Jun. 1992)
David Berning EA-2101 Amp & TF-12 Preamp (Dec. 1991)
Counterpoint SA-220 Power Amp (Jul. 1990)
Cary Audio CAD-50 and CAD-505L Mono Amps (Jul. 1991)
= = = =